Aaron B. Holley, MD

Beta-blockers are excellent drugs. They’re cheap and effective; feature prominently in hypertension guidelines; and remain a sine qua non for coronary artery disease, myocardial infarction, and heart failure treatment. They’ve been around forever, and we know they work. Good luck finding an adult medicine patient who isn’t on one.

Beta-blockers act by slowing resting heart rate (and blunting the heart rate response to exercise. The latter is a pernicious cause of activity intolerance that often goes unchecked. Even when the adverse effects of beta-blockers are appreciated, providers are loath to alter dosing, much less stop the drug. After all, beta-blockers are an integral part of guideline-directed medical therapy (GDMT), and GDMT saves lives.

Balancing Heart Rate and Stroke Volume Effects

The pulmonologist sees beta-blockers differently. To augment cardiac output and optimize oxygen uptake (VO₂) during exercise, we need the heart rate response. In fact, the heart rate response contributes more to cardiac output than augmenting stroke volume (SV) and more to VO₂ than the increase in arteriovenous (AV) oxygen difference. An inability to increase the heart rate commensurate with physiologic work is called chronotropic incompetence (CI). That’s what beta-blockers do ─ they cause CI.

Physiology dictates that CI will cause activity intolerance. That said, it’s hard to quantify the impact from beta-blockers at the individual patient level. Data suggest the heart rate effect is profound. A study in patients without heart failure found that 22% of participants on beta-blockers had CI, and the investigators used a conservative CI definition (≤ 62% of heart rate reserve used). A recent report published in JAMA Cardiology found that stopping beta-blockers in patients with heart failure allowed for an extra 30 beats/min at max exercise.

Wasserman and Whipp’s textbook, the last word on all things exercise, presents a sample subject who undergoes two separate cardiopulmonary exercise tests (CPETs). Before the first, he’s given a placebo, and before the second, he gets an intravenous beta-blocker. He’s a 23-year-old otherwise healthy male — the perfect test case for isolating beta-blocker impact without confounding by comorbid diseases, other medications, or deconditioning. His max heart rate dropped by 30 beats/min after the beta-blocker, identical to what we saw in the JAMA Cardiology study (with the heart rate increasing by 30 beats/min following withdrawal). Case closed. Stop the beta-blockers on your patients so they can meet their exercise goals and get healthy!

Such pithy enthusiasm discounts physiology’s complexities. When blunting our patient’s heart rate response with beta-blockers, we also increase diastolic filling time, which increases SV. For the 23-year-old in Wasserman and Whipp’s physiology textbook, the beta-blocker increased O₂ pulse (the product of SV and AV difference). Presumably, this is mediated by the increased SV. There was a net reduction in VO₂ peak, but it was nominal, suggesting that the drop in heart rate was largely offset by the increase in O₂ pulse. For the patients in the JAMA Cardiology study, the entire group had a small increase in VO2 peak with beta-blocker withdrawal, but the effect differed by left ventricular function. Across different studies, the beta-blocker effect on heart rate is consistent but the change in overall exercise capacity is not. 

Patient Variability in Beta-Blocker Response

In addition to left ventricular function, there are other factors likely to drive variability at the patient level. We’ve treated the response to beta-blockers as a class effect — an obvious oversimplification. The impact on exercise and the heart will vary by dose and drug (eg, atenolol vs metoprolol vs carvedilol, and so on). Beta-blockers can also affect the lungs, and we’re still debating how cautious to be in the presence of asthma or chronic obstructive pulmonary disease. 

In a world of infinite time, resources, and expertise, we’d CPET everyone before and after beta-blocker use. Our current reality requires the unthinkable: We’ll have to talk to each other and our patients. For example, heart failure guidelines recommend titrating drugs to match the dose from trials that proved efficacy. These doses are quite high. Simple discussion with the cardiologist and the patient may allow for an adjustment back down with careful monitoring and close attention to activity tolerance. With any luck, you’ll preserve the benefits from GDMT while optimizing your patient›s ability to meet their exercise goals.
 

Dr. Holley, professor in the department of medicine, Uniformed Services University, Bethesda, Maryland, and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center, Washington, disclosed ties with Metapharm, CHEST College, and WebMD.

A version of this article appeared on Medscape.com.

Beta-blockers are excellent drugs. They’re cheap and effective; feature prominently in hypertension guidelines; and remain a sine qua non for coronary artery disease, myocardial infarction, and heart failure treatment. They’ve been around forever, and we know they work. Good luck finding an adult medicine patient who isn’t on one.

Beta-blockers act by slowing resting heart rate (and blunting the heart rate response to exercise. The latter is a pernicious cause of activity intolerance that often goes unchecked. Even when the adverse effects of beta-blockers are appreciated, providers are loath to alter dosing, much less stop the drug. After all, beta-blockers are an integral part of guideline-directed medical therapy (GDMT), and GDMT saves lives.

Balancing Heart Rate and Stroke Volume Effects

The pulmonologist sees beta-blockers differently. To augment cardiac output and optimize oxygen uptake (VO₂) during exercise, we need the heart rate response. In fact, the heart rate response contributes more to cardiac output than augmenting stroke volume (SV) and more to VO₂ than the increase in arteriovenous (AV) oxygen difference. An inability to increase the heart rate commensurate with physiologic work is called chronotropic incompetence (CI). That’s what beta-blockers do ─ they cause CI.

Physiology dictates that CI will cause activity intolerance. That said, it’s hard to quantify the impact from beta-blockers at the individual patient level. Data suggest the heart rate effect is profound. A study in patients without heart failure found that 22% of participants on beta-blockers had CI, and the investigators used a conservative CI definition (≤ 62% of heart rate reserve used). A recent report published in JAMA Cardiology found that stopping beta-blockers in patients with heart failure allowed for an extra 30 beats/min at max exercise.

Wasserman and Whipp’s textbook, the last word on all things exercise, presents a sample subject who undergoes two separate cardiopulmonary exercise tests (CPETs). Before the first, he’s given a placebo, and before the second, he gets an intravenous beta-blocker. He’s a 23-year-old otherwise healthy male — the perfect test case for isolating beta-blocker impact without confounding by comorbid diseases, other medications, or deconditioning. His max heart rate dropped by 30 beats/min after the beta-blocker, identical to what we saw in the JAMA Cardiology study (with the heart rate increasing by 30 beats/min following withdrawal). Case closed. Stop the beta-blockers on your patients so they can meet their exercise goals and get healthy!

Such pithy enthusiasm discounts physiology’s complexities. When blunting our patient’s heart rate response with beta-blockers, we also increase diastolic filling time, which increases SV. For the 23-year-old in Wasserman and Whipp’s physiology textbook, the beta-blocker increased O₂ pulse (the product of SV and AV difference). Presumably, this is mediated by the increased SV. There was a net reduction in VO₂ peak, but it was nominal, suggesting that the drop in heart rate was largely offset by the increase in O₂ pulse. For the patients in the JAMA Cardiology study, the entire group had a small increase in VO2 peak with beta-blocker withdrawal, but the effect differed by left ventricular function. Across different studies, the beta-blocker effect on heart rate is consistent but the change in overall exercise capacity is not. 

Patient Variability in Beta-Blocker Response

In addition to left ventricular function, there are other factors likely to drive variability at the patient level. We’ve treated the response to beta-blockers as a class effect — an obvious oversimplification. The impact on exercise and the heart will vary by dose and drug (eg, atenolol vs metoprolol vs carvedilol, and so on). Beta-blockers can also affect the lungs, and we’re still debating how cautious to be in the presence of asthma or chronic obstructive pulmonary disease. 

In a world of infinite time, resources, and expertise, we’d CPET everyone before and after beta-blocker use. Our current reality requires the unthinkable: We’ll have to talk to each other and our patients. For example, heart failure guidelines recommend titrating drugs to match the dose from trials that proved efficacy. These doses are quite high. Simple discussion with the cardiologist and the patient may allow for an adjustment back down with careful monitoring and close attention to activity tolerance. With any luck, you’ll preserve the benefits from GDMT while optimizing your patient›s ability to meet their exercise goals.
 

Dr. Holley, professor in the department of medicine, Uniformed Services University, Bethesda, Maryland, and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center, Washington, disclosed ties with Metapharm, CHEST College, and WebMD.

A version of this article appeared on Medscape.com.

Beta-blockers are excellent drugs. They’re cheap and effective; feature prominently in hypertension guidelines; and remain a sine qua non for coronary artery disease, myocardial infarction, and heart failure treatment. They’ve been around forever, and we know they work. Good luck finding an adult medicine patient who isn’t on one.

Beta-blockers act by slowing resting heart rate (and blunting the heart rate response to exercise. The latter is a pernicious cause of activity intolerance that often goes unchecked. Even when the adverse effects of beta-blockers are appreciated, providers are loath to alter dosing, much less stop the drug. After all, beta-blockers are an integral part of guideline-directed medical therapy (GDMT), and GDMT saves lives.

Balancing Heart Rate and Stroke Volume Effects

The pulmonologist sees beta-blockers differently. To augment cardiac output and optimize oxygen uptake (VO₂) during exercise, we need the heart rate response. In fact, the heart rate response contributes more to cardiac output than augmenting stroke volume (SV) and more to VO₂ than the increase in arteriovenous (AV) oxygen difference. An inability to increase the heart rate commensurate with physiologic work is called chronotropic incompetence (CI). That’s what beta-blockers do ─ they cause CI.

Physiology dictates that CI will cause activity intolerance. That said, it’s hard to quantify the impact from beta-blockers at the individual patient level. Data suggest the heart rate effect is profound. A study in patients without heart failure found that 22% of participants on beta-blockers had CI, and the investigators used a conservative CI definition (≤ 62% of heart rate reserve used). A recent report published in JAMA Cardiology found that stopping beta-blockers in patients with heart failure allowed for an extra 30 beats/min at max exercise.

Wasserman and Whipp’s textbook, the last word on all things exercise, presents a sample subject who undergoes two separate cardiopulmonary exercise tests (CPETs). Before the first, he’s given a placebo, and before the second, he gets an intravenous beta-blocker. He’s a 23-year-old otherwise healthy male — the perfect test case for isolating beta-blocker impact without confounding by comorbid diseases, other medications, or deconditioning. His max heart rate dropped by 30 beats/min after the beta-blocker, identical to what we saw in the JAMA Cardiology study (with the heart rate increasing by 30 beats/min following withdrawal). Case closed. Stop the beta-blockers on your patients so they can meet their exercise goals and get healthy!

Such pithy enthusiasm discounts physiology’s complexities. When blunting our patient’s heart rate response with beta-blockers, we also increase diastolic filling time, which increases SV. For the 23-year-old in Wasserman and Whipp’s physiology textbook, the beta-blocker increased O₂ pulse (the product of SV and AV difference). Presumably, this is mediated by the increased SV. There was a net reduction in VO₂ peak, but it was nominal, suggesting that the drop in heart rate was largely offset by the increase in O₂ pulse. For the patients in the JAMA Cardiology study, the entire group had a small increase in VO2 peak with beta-blocker withdrawal, but the effect differed by left ventricular function. Across different studies, the beta-blocker effect on heart rate is consistent but the change in overall exercise capacity is not. 

Patient Variability in Beta-Blocker Response

In addition to left ventricular function, there are other factors likely to drive variability at the patient level. We’ve treated the response to beta-blockers as a class effect — an obvious oversimplification. The impact on exercise and the heart will vary by dose and drug (eg, atenolol vs metoprolol vs carvedilol, and so on). Beta-blockers can also affect the lungs, and we’re still debating how cautious to be in the presence of asthma or chronic obstructive pulmonary disease. 

In a world of infinite time, resources, and expertise, we’d CPET everyone before and after beta-blocker use. Our current reality requires the unthinkable: We’ll have to talk to each other and our patients. For example, heart failure guidelines recommend titrating drugs to match the dose from trials that proved efficacy. These doses are quite high. Simple discussion with the cardiologist and the patient may allow for an adjustment back down with careful monitoring and close attention to activity tolerance. With any luck, you’ll preserve the benefits from GDMT while optimizing your patient›s ability to meet their exercise goals.
 

Dr. Holley, professor in the department of medicine, Uniformed Services University, Bethesda, Maryland, and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center, Washington, disclosed ties with Metapharm, CHEST College, and WebMD.

A version of this article appeared on Medscape.com.

A few months ago, I posted a column on continuous positive airway pressure (CPAP) with the title, “CPAP Oversells and Underperforms.” To date, it has 299 likes and 90 comments, which are almost all negative. I’m glad to see that it’s generated interest, and I’d like to address some of the themes expressed in the posts.

Most comments were personal testimonies to the miracles of CPAP. These are important, and the point deserves emphasis. CPAP can provide significant improvements in daytime sleepiness and quality of life. I closed the original piece by acknowledging this important fact. Readers can be forgiven for missing it given that the title and text were otherwise disparaging of CPAP.

But several comments warrant a more in-depth discussion. The original piece focuses on CPAP and cardiovascular (CV) outcomes but made no mention of atrial fibrillation (AF) or ejection fraction (EF). The effects of CPAP on each are touted by cardiologists and PAP-pushers alike and are drivers of frequent referrals. It›s my fault for omitting them from the discussion.

AF is easy. The data is identical to all other things CPAP and CV. Based on biologic plausibility alone, the likelihood of a relationship between AF and obstructive sleep apnea (OSA) is similar to the odds that the Celtics raise an 18th banner come June. There’s hypoxia, intrathoracic pressure swings, sympathetic surges, and sleep state disruptions. It’s easy to get from there to arrhythmogenesis. There’s lots of observational noise, too, but no randomized proof that CPAP alters this relationship.

I found four randomized controlled trials (RCTs) that tested CPAP’s effect on AF. I’ll save you the suspense; they were all negative. One even found a signal for more adverse events in the CPAP group. These studies have several positive qualities: They enrolled patients with moderate to severe sleep apnea and high oxygen desaturation indices, adherence averaged more than 4 hours across all groups in all trials, and the methods for assessing the AF outcomes differed slightly. There’s also a lot not to like: The sample sizes were small, only one trial enrolled “sleepy” patients (as assessed by the Epworth Sleepiness Score), and follow-up was short.

To paraphrase Carl Sagan, “absence of evidence does not equal evidence of absence.” As a statistician would say, type II error cannot be excluded by these RCTs. In medicine, however, the burden of proof falls on demonstrating efficacy. If we treat before concluding that a therapy works, we risk wasting time, money, medical resources, and the most precious of patient commodities: the energy required for behavior change. In their response to letters to the editor, the authors of the third RCT summarize the CPAP, AF, and CV disease data far better than I ever could. They sound the same words of caution and come out against screening patients with AF for OSA. 

The story for CPAP’s effects on EF is similar though muddier. The American College of Cardiology (ACC)/American Heart Association (AHA) guidelines for heart failure cite a meta-analysis showing that CPAP improves left ventricular EF. In 2019, the American Academy of Sleep Medicine (AASM) CPAP guidelines included a systematic review and meta-analysis that found that CPAP has no effect on left ventricular EF in patients with or without heart failure.

There are a million reasons why two systematic reviews on the same topic might come to different conclusions. In this case, the included studies only partially overlap, and broadly speaking, it appears the authors made trade-offs. The review cited by the ACC/AHA had broader inclusion and significantly more patients and paid for it in heterogeneity (I2 in the 80%-90% range). The AASM analysis achieved 0% heterogeneity but limited inclusion to fewer than 100 patients. Across both, the improvement in EF was 2%- 5% at a minimally clinically important difference of 4%. Hardly convincing.

In summary, the road to negative trials and patient harm has always been paved with observational signal and biologic plausibility. Throw in some intellectual and academic bias, and you’ve created the perfect storm of therapeutic overconfidence. The cemetery for discarded medical therapies is crowded, but there’s room for CPAP, at least when it comes to using it to improve CV outcomes. 
 

Dr. Holley is a professor in the department of medicine, Uniformed Services University, Bethesda, Maryland, and a physician at Pulmonary/Sleep and Critical Care Medicine, MedStar Washington Hospital Center, Washington. He disclosed ties to Metapharm Inc., CHEST College, and WebMD.

A version of this article appeared on Medscape.com .

A few months ago, I posted a column on continuous positive airway pressure (CPAP) with the title, “CPAP Oversells and Underperforms.” To date, it has 299 likes and 90 comments, which are almost all negative. I’m glad to see that it’s generated interest, and I’d like to address some of the themes expressed in the posts.

Most comments were personal testimonies to the miracles of CPAP. These are important, and the point deserves emphasis. CPAP can provide significant improvements in daytime sleepiness and quality of life. I closed the original piece by acknowledging this important fact. Readers can be forgiven for missing it given that the title and text were otherwise disparaging of CPAP.

But several comments warrant a more in-depth discussion. The original piece focuses on CPAP and cardiovascular (CV) outcomes but made no mention of atrial fibrillation (AF) or ejection fraction (EF). The effects of CPAP on each are touted by cardiologists and PAP-pushers alike and are drivers of frequent referrals. It›s my fault for omitting them from the discussion.

AF is easy. The data is identical to all other things CPAP and CV. Based on biologic plausibility alone, the likelihood of a relationship between AF and obstructive sleep apnea (OSA) is similar to the odds that the Celtics raise an 18th banner come June. There’s hypoxia, intrathoracic pressure swings, sympathetic surges, and sleep state disruptions. It’s easy to get from there to arrhythmogenesis. There’s lots of observational noise, too, but no randomized proof that CPAP alters this relationship.

I found four randomized controlled trials (RCTs) that tested CPAP’s effect on AF. I’ll save you the suspense; they were all negative. One even found a signal for more adverse events in the CPAP group. These studies have several positive qualities: They enrolled patients with moderate to severe sleep apnea and high oxygen desaturation indices, adherence averaged more than 4 hours across all groups in all trials, and the methods for assessing the AF outcomes differed slightly. There’s also a lot not to like: The sample sizes were small, only one trial enrolled “sleepy” patients (as assessed by the Epworth Sleepiness Score), and follow-up was short.

To paraphrase Carl Sagan, “absence of evidence does not equal evidence of absence.” As a statistician would say, type II error cannot be excluded by these RCTs. In medicine, however, the burden of proof falls on demonstrating efficacy. If we treat before concluding that a therapy works, we risk wasting time, money, medical resources, and the most precious of patient commodities: the energy required for behavior change. In their response to letters to the editor, the authors of the third RCT summarize the CPAP, AF, and CV disease data far better than I ever could. They sound the same words of caution and come out against screening patients with AF for OSA. 

The story for CPAP’s effects on EF is similar though muddier. The American College of Cardiology (ACC)/American Heart Association (AHA) guidelines for heart failure cite a meta-analysis showing that CPAP improves left ventricular EF. In 2019, the American Academy of Sleep Medicine (AASM) CPAP guidelines included a systematic review and meta-analysis that found that CPAP has no effect on left ventricular EF in patients with or without heart failure.

There are a million reasons why two systematic reviews on the same topic might come to different conclusions. In this case, the included studies only partially overlap, and broadly speaking, it appears the authors made trade-offs. The review cited by the ACC/AHA had broader inclusion and significantly more patients and paid for it in heterogeneity (I2 in the 80%-90% range). The AASM analysis achieved 0% heterogeneity but limited inclusion to fewer than 100 patients. Across both, the improvement in EF was 2%- 5% at a minimally clinically important difference of 4%. Hardly convincing.

In summary, the road to negative trials and patient harm has always been paved with observational signal and biologic plausibility. Throw in some intellectual and academic bias, and you’ve created the perfect storm of therapeutic overconfidence. The cemetery for discarded medical therapies is crowded, but there’s room for CPAP, at least when it comes to using it to improve CV outcomes. 
 

Dr. Holley is a professor in the department of medicine, Uniformed Services University, Bethesda, Maryland, and a physician at Pulmonary/Sleep and Critical Care Medicine, MedStar Washington Hospital Center, Washington. He disclosed ties to Metapharm Inc., CHEST College, and WebMD.

A version of this article appeared on Medscape.com .

A few months ago, I posted a column on continuous positive airway pressure (CPAP) with the title, “CPAP Oversells and Underperforms.” To date, it has 299 likes and 90 comments, which are almost all negative. I’m glad to see that it’s generated interest, and I’d like to address some of the themes expressed in the posts.

Most comments were personal testimonies to the miracles of CPAP. These are important, and the point deserves emphasis. CPAP can provide significant improvements in daytime sleepiness and quality of life. I closed the original piece by acknowledging this important fact. Readers can be forgiven for missing it given that the title and text were otherwise disparaging of CPAP.

But several comments warrant a more in-depth discussion. The original piece focuses on CPAP and cardiovascular (CV) outcomes but made no mention of atrial fibrillation (AF) or ejection fraction (EF). The effects of CPAP on each are touted by cardiologists and PAP-pushers alike and are drivers of frequent referrals. It›s my fault for omitting them from the discussion.

AF is easy. The data is identical to all other things CPAP and CV. Based on biologic plausibility alone, the likelihood of a relationship between AF and obstructive sleep apnea (OSA) is similar to the odds that the Celtics raise an 18th banner come June. There’s hypoxia, intrathoracic pressure swings, sympathetic surges, and sleep state disruptions. It’s easy to get from there to arrhythmogenesis. There’s lots of observational noise, too, but no randomized proof that CPAP alters this relationship.

I found four randomized controlled trials (RCTs) that tested CPAP’s effect on AF. I’ll save you the suspense; they were all negative. One even found a signal for more adverse events in the CPAP group. These studies have several positive qualities: They enrolled patients with moderate to severe sleep apnea and high oxygen desaturation indices, adherence averaged more than 4 hours across all groups in all trials, and the methods for assessing the AF outcomes differed slightly. There’s also a lot not to like: The sample sizes were small, only one trial enrolled “sleepy” patients (as assessed by the Epworth Sleepiness Score), and follow-up was short.

To paraphrase Carl Sagan, “absence of evidence does not equal evidence of absence.” As a statistician would say, type II error cannot be excluded by these RCTs. In medicine, however, the burden of proof falls on demonstrating efficacy. If we treat before concluding that a therapy works, we risk wasting time, money, medical resources, and the most precious of patient commodities: the energy required for behavior change. In their response to letters to the editor, the authors of the third RCT summarize the CPAP, AF, and CV disease data far better than I ever could. They sound the same words of caution and come out against screening patients with AF for OSA. 

The story for CPAP’s effects on EF is similar though muddier. The American College of Cardiology (ACC)/American Heart Association (AHA) guidelines for heart failure cite a meta-analysis showing that CPAP improves left ventricular EF. In 2019, the American Academy of Sleep Medicine (AASM) CPAP guidelines included a systematic review and meta-analysis that found that CPAP has no effect on left ventricular EF in patients with or without heart failure.

There are a million reasons why two systematic reviews on the same topic might come to different conclusions. In this case, the included studies only partially overlap, and broadly speaking, it appears the authors made trade-offs. The review cited by the ACC/AHA had broader inclusion and significantly more patients and paid for it in heterogeneity (I2 in the 80%-90% range). The AASM analysis achieved 0% heterogeneity but limited inclusion to fewer than 100 patients. Across both, the improvement in EF was 2%- 5% at a minimally clinically important difference of 4%. Hardly convincing.

In summary, the road to negative trials and patient harm has always been paved with observational signal and biologic plausibility. Throw in some intellectual and academic bias, and you’ve created the perfect storm of therapeutic overconfidence. The cemetery for discarded medical therapies is crowded, but there’s room for CPAP, at least when it comes to using it to improve CV outcomes. 
 

Dr. Holley is a professor in the department of medicine, Uniformed Services University, Bethesda, Maryland, and a physician at Pulmonary/Sleep and Critical Care Medicine, MedStar Washington Hospital Center, Washington. He disclosed ties to Metapharm Inc., CHEST College, and WebMD.

A version of this article appeared on Medscape.com .

Continuous positive airway pressure (CPAP) is first-line therapy for sleep-related breathing disorders (SRBDs). Obstructive sleep apnea (OSA) is the major player in the SRBDs space, with a prevalence approaching 100% in adult men using current diagnostic criteria. Patients with OSA and comorbid cardiovascular disease (CVD) are diagnosed with OSA syndrome, and CPAP is prescribed. Primary care physicians and cardiologists are quick to refer patients with CVD to sleep docs to see whether CPAP can improve CVD-related outcomes.

What the Studies Show

There’s a problem though. CPAP doesn’t seem to improve CVD-related outcomes. In some cases, it’s even harmful. Let’s do a quick review. In 2005, the CANPAP study found CPAP didn’t improve a composite CVD outcome that included mortality. A post hoc analysis found that it actually increased mortality if central apneas weren’t eliminated. The post hoc analysis also found benefit when central apneas were eliminated, but for all-comers, CPAP didn’t improve outcomes. Strike one.

Enter adaptive servo-ventilation (ASV). If CANPAP showed that success depended on eliminating central apneas, why not use ASV for all patients with CVD and central apneas or Cheyne-Stokes respirations? ASV eliminates central apneas and Cheyne-Stokes. Well, that didn’t work either. The randomized, controlled SERVE-HF trial, published in 2015, showed that ASV increases all-cause and CVD-specific mortality. Oops. That’s two trials showing that CPAP and ASV can increase mortality in patients with heart failure. Strike two.

Alright. But that’s heart failure. What about hypertension or coronary artery disease (CAD)? Shouldn’t such patients be treated with CPAP to reduce CVD risk? After all, there’s all those surrogate outcomes data for CPAP — it improves vascular tone and lowers catecholamines and all that stuff. Doesn’t it lower blood pressure too? Surely CPAP benefits patients with CVD who don’t have heart failure, right?

Not really. The RICCADSA study, published in 2016, found that CPAP didn’t reduce a composite of CVD outcomes in patients with newly revascularized CAD. The SAVE trial published the same year had a similar design with similar results. CPAP did not improve CVD-related outcomes. Most recently, the ISAACC study was negative. That’s three negative randomized controlled trials in less than 5 years showing CPAP doesn’t affect CVD-related outcomes in high-risk populations with known disease. Strike three?

CPAP provides no benefit for CVD and possible harm when treating heart failure. Surely CPAP is useful for patients with hypertension. Let’s see. The American Academy of Sleep Medicine (AASM) conducted meta-analyses for the guideline it produced recommending CPAP for patients with comorbid hypertension. They note that 24-hour blood pressure measurements are best correlated with outcomes. CPAP did lead to significant 24-hour blood pressure reduction, but guess how large it was? For systolic blood pressure, it was 1.5 mm Hg; for diastolic pressure, it was 1.6 mm Hg. That’s it.

How did the AASM summarize and interpret the above data in their 2019 guidelines for prescribing CPAP? Although covered in their detailed review, both heart failure and CVD are left out of their primary recommendations . They do provide a conditional recommendation for prescribing CPAP to patients with comorbid hypertension that states, “The majority of well-informed patients would choose the intervention over no treatment.” Really? If you were told that CPAP provides less reduction in blood pressure than dietary changes and/or medications, would you choose to wear it or take a pill once a day? Remember, you have to take the pill anyway to get your blood pressure to target unless your pressure is only 1.5-1.6 mm Hg above normal. Where does one find patients who are anxious to wear a mask to bed for minimal benefit and a 20% copay? I’ve yet to meet one.

As always, the pressure pushers are undeterred by inconvenient evidence. A secondary analysis of adherent patients in RICCADSA resorts to the “bait and switch” that’s propped up CPAP enthusiasts for decades: Compare adherent patients versus those who are not (or those who refuse treatment) to prove benefit. The flaws to this approach are obvious. First, performing a post hoc analysis that reintroduces all of the confounding that plagues existing CPAP data negates the benefits of randomization, fancy statistics notwithstanding. Second, it belies the reality that in well-controlled, well-conducted randomized trials where patients get far more support than those in the community (and sometimes are preselected for adherence), a majority simply won’t use CPAP . Excluding the nonadherent or comparing them with the adherent is the epitome of selection bias.

The editorial accompanying the ISAACC study is a tour de force in CPAP apologies. The apnea-hypopnea index (AHI) isn’t the right metric — this one’s invoked often. Never mind that the very premise that OSA causes CVD is from observational data based on the AHI. If you abandon the AHI, don’t you lose your justification for prospective trials targeting CVD with CPAP?

Even better, in an argument fit for a Twitter ban, the author suggests that patients in ISAACC, SAVE, and RICCADSA couldn’t benefit because they already have CVD. The very concept, refuted by decades of secondary prevention research in cardiology, implies that CPAP should be used for primary prevention. Only a sleep researcher could spin a negative study into an expansion of CPAP indications. Others in the AASM have made similar proposals.

Final Thoughts

The sleep field lacks unblinded realists capable of choosing wisely. A little therapeutic underconfidence is warranted. Diseases and therapies will always have champions. Prudence and restraint? Not so much. The AASM could summarize the CPAP literature in a single recommendation: “If your patient is sleepy, CPAP might help them feel better if their disease is moderate or severe.” All other indications are soft.

A version of this article first appeared on Medscape.com.

Aaron B. Holley, MD, is a professor of medicine at Uniformed Services University in Bethesda, Maryland, and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington, DC. He covers a  wide range of topics in pulmonary, critical care, and sleep medicine .

Continuous positive airway pressure (CPAP) is first-line therapy for sleep-related breathing disorders (SRBDs). Obstructive sleep apnea (OSA) is the major player in the SRBDs space, with a prevalence approaching 100% in adult men using current diagnostic criteria. Patients with OSA and comorbid cardiovascular disease (CVD) are diagnosed with OSA syndrome, and CPAP is prescribed. Primary care physicians and cardiologists are quick to refer patients with CVD to sleep docs to see whether CPAP can improve CVD-related outcomes.

What the Studies Show

There’s a problem though. CPAP doesn’t seem to improve CVD-related outcomes. In some cases, it’s even harmful. Let’s do a quick review. In 2005, the CANPAP study found CPAP didn’t improve a composite CVD outcome that included mortality. A post hoc analysis found that it actually increased mortality if central apneas weren’t eliminated. The post hoc analysis also found benefit when central apneas were eliminated, but for all-comers, CPAP didn’t improve outcomes. Strike one.

Enter adaptive servo-ventilation (ASV). If CANPAP showed that success depended on eliminating central apneas, why not use ASV for all patients with CVD and central apneas or Cheyne-Stokes respirations? ASV eliminates central apneas and Cheyne-Stokes. Well, that didn’t work either. The randomized, controlled SERVE-HF trial, published in 2015, showed that ASV increases all-cause and CVD-specific mortality. Oops. That’s two trials showing that CPAP and ASV can increase mortality in patients with heart failure. Strike two.

Alright. But that’s heart failure. What about hypertension or coronary artery disease (CAD)? Shouldn’t such patients be treated with CPAP to reduce CVD risk? After all, there’s all those surrogate outcomes data for CPAP — it improves vascular tone and lowers catecholamines and all that stuff. Doesn’t it lower blood pressure too? Surely CPAP benefits patients with CVD who don’t have heart failure, right?

Not really. The RICCADSA study, published in 2016, found that CPAP didn’t reduce a composite of CVD outcomes in patients with newly revascularized CAD. The SAVE trial published the same year had a similar design with similar results. CPAP did not improve CVD-related outcomes. Most recently, the ISAACC study was negative. That’s three negative randomized controlled trials in less than 5 years showing CPAP doesn’t affect CVD-related outcomes in high-risk populations with known disease. Strike three?

CPAP provides no benefit for CVD and possible harm when treating heart failure. Surely CPAP is useful for patients with hypertension. Let’s see. The American Academy of Sleep Medicine (AASM) conducted meta-analyses for the guideline it produced recommending CPAP for patients with comorbid hypertension. They note that 24-hour blood pressure measurements are best correlated with outcomes. CPAP did lead to significant 24-hour blood pressure reduction, but guess how large it was? For systolic blood pressure, it was 1.5 mm Hg; for diastolic pressure, it was 1.6 mm Hg. That’s it.

How did the AASM summarize and interpret the above data in their 2019 guidelines for prescribing CPAP? Although covered in their detailed review, both heart failure and CVD are left out of their primary recommendations . They do provide a conditional recommendation for prescribing CPAP to patients with comorbid hypertension that states, “The majority of well-informed patients would choose the intervention over no treatment.” Really? If you were told that CPAP provides less reduction in blood pressure than dietary changes and/or medications, would you choose to wear it or take a pill once a day? Remember, you have to take the pill anyway to get your blood pressure to target unless your pressure is only 1.5-1.6 mm Hg above normal. Where does one find patients who are anxious to wear a mask to bed for minimal benefit and a 20% copay? I’ve yet to meet one.

As always, the pressure pushers are undeterred by inconvenient evidence. A secondary analysis of adherent patients in RICCADSA resorts to the “bait and switch” that’s propped up CPAP enthusiasts for decades: Compare adherent patients versus those who are not (or those who refuse treatment) to prove benefit. The flaws to this approach are obvious. First, performing a post hoc analysis that reintroduces all of the confounding that plagues existing CPAP data negates the benefits of randomization, fancy statistics notwithstanding. Second, it belies the reality that in well-controlled, well-conducted randomized trials where patients get far more support than those in the community (and sometimes are preselected for adherence), a majority simply won’t use CPAP . Excluding the nonadherent or comparing them with the adherent is the epitome of selection bias.

The editorial accompanying the ISAACC study is a tour de force in CPAP apologies. The apnea-hypopnea index (AHI) isn’t the right metric — this one’s invoked often. Never mind that the very premise that OSA causes CVD is from observational data based on the AHI. If you abandon the AHI, don’t you lose your justification for prospective trials targeting CVD with CPAP?

Even better, in an argument fit for a Twitter ban, the author suggests that patients in ISAACC, SAVE, and RICCADSA couldn’t benefit because they already have CVD. The very concept, refuted by decades of secondary prevention research in cardiology, implies that CPAP should be used for primary prevention. Only a sleep researcher could spin a negative study into an expansion of CPAP indications. Others in the AASM have made similar proposals.

Final Thoughts

The sleep field lacks unblinded realists capable of choosing wisely. A little therapeutic underconfidence is warranted. Diseases and therapies will always have champions. Prudence and restraint? Not so much. The AASM could summarize the CPAP literature in a single recommendation: “If your patient is sleepy, CPAP might help them feel better if their disease is moderate or severe.” All other indications are soft.

A version of this article first appeared on Medscape.com.

Aaron B. Holley, MD, is a professor of medicine at Uniformed Services University in Bethesda, Maryland, and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington, DC. He covers a  wide range of topics in pulmonary, critical care, and sleep medicine .

Continuous positive airway pressure (CPAP) is first-line therapy for sleep-related breathing disorders (SRBDs). Obstructive sleep apnea (OSA) is the major player in the SRBDs space, with a prevalence approaching 100% in adult men using current diagnostic criteria. Patients with OSA and comorbid cardiovascular disease (CVD) are diagnosed with OSA syndrome, and CPAP is prescribed. Primary care physicians and cardiologists are quick to refer patients with CVD to sleep docs to see whether CPAP can improve CVD-related outcomes.

What the Studies Show

There’s a problem though. CPAP doesn’t seem to improve CVD-related outcomes. In some cases, it’s even harmful. Let’s do a quick review. In 2005, the CANPAP study found CPAP didn’t improve a composite CVD outcome that included mortality. A post hoc analysis found that it actually increased mortality if central apneas weren’t eliminated. The post hoc analysis also found benefit when central apneas were eliminated, but for all-comers, CPAP didn’t improve outcomes. Strike one.

Enter adaptive servo-ventilation (ASV). If CANPAP showed that success depended on eliminating central apneas, why not use ASV for all patients with CVD and central apneas or Cheyne-Stokes respirations? ASV eliminates central apneas and Cheyne-Stokes. Well, that didn’t work either. The randomized, controlled SERVE-HF trial, published in 2015, showed that ASV increases all-cause and CVD-specific mortality. Oops. That’s two trials showing that CPAP and ASV can increase mortality in patients with heart failure. Strike two.

Alright. But that’s heart failure. What about hypertension or coronary artery disease (CAD)? Shouldn’t such patients be treated with CPAP to reduce CVD risk? After all, there’s all those surrogate outcomes data for CPAP — it improves vascular tone and lowers catecholamines and all that stuff. Doesn’t it lower blood pressure too? Surely CPAP benefits patients with CVD who don’t have heart failure, right?

Not really. The RICCADSA study, published in 2016, found that CPAP didn’t reduce a composite of CVD outcomes in patients with newly revascularized CAD. The SAVE trial published the same year had a similar design with similar results. CPAP did not improve CVD-related outcomes. Most recently, the ISAACC study was negative. That’s three negative randomized controlled trials in less than 5 years showing CPAP doesn’t affect CVD-related outcomes in high-risk populations with known disease. Strike three?

CPAP provides no benefit for CVD and possible harm when treating heart failure. Surely CPAP is useful for patients with hypertension. Let’s see. The American Academy of Sleep Medicine (AASM) conducted meta-analyses for the guideline it produced recommending CPAP for patients with comorbid hypertension. They note that 24-hour blood pressure measurements are best correlated with outcomes. CPAP did lead to significant 24-hour blood pressure reduction, but guess how large it was? For systolic blood pressure, it was 1.5 mm Hg; for diastolic pressure, it was 1.6 mm Hg. That’s it.

How did the AASM summarize and interpret the above data in their 2019 guidelines for prescribing CPAP? Although covered in their detailed review, both heart failure and CVD are left out of their primary recommendations . They do provide a conditional recommendation for prescribing CPAP to patients with comorbid hypertension that states, “The majority of well-informed patients would choose the intervention over no treatment.” Really? If you were told that CPAP provides less reduction in blood pressure than dietary changes and/or medications, would you choose to wear it or take a pill once a day? Remember, you have to take the pill anyway to get your blood pressure to target unless your pressure is only 1.5-1.6 mm Hg above normal. Where does one find patients who are anxious to wear a mask to bed for minimal benefit and a 20% copay? I’ve yet to meet one.

As always, the pressure pushers are undeterred by inconvenient evidence. A secondary analysis of adherent patients in RICCADSA resorts to the “bait and switch” that’s propped up CPAP enthusiasts for decades: Compare adherent patients versus those who are not (or those who refuse treatment) to prove benefit. The flaws to this approach are obvious. First, performing a post hoc analysis that reintroduces all of the confounding that plagues existing CPAP data negates the benefits of randomization, fancy statistics notwithstanding. Second, it belies the reality that in well-controlled, well-conducted randomized trials where patients get far more support than those in the community (and sometimes are preselected for adherence), a majority simply won’t use CPAP . Excluding the nonadherent or comparing them with the adherent is the epitome of selection bias.

The editorial accompanying the ISAACC study is a tour de force in CPAP apologies. The apnea-hypopnea index (AHI) isn’t the right metric — this one’s invoked often. Never mind that the very premise that OSA causes CVD is from observational data based on the AHI. If you abandon the AHI, don’t you lose your justification for prospective trials targeting CVD with CPAP?

Even better, in an argument fit for a Twitter ban, the author suggests that patients in ISAACC, SAVE, and RICCADSA couldn’t benefit because they already have CVD. The very concept, refuted by decades of secondary prevention research in cardiology, implies that CPAP should be used for primary prevention. Only a sleep researcher could spin a negative study into an expansion of CPAP indications. Others in the AASM have made similar proposals.

Final Thoughts

The sleep field lacks unblinded realists capable of choosing wisely. A little therapeutic underconfidence is warranted. Diseases and therapies will always have champions. Prudence and restraint? Not so much. The AASM could summarize the CPAP literature in a single recommendation: “If your patient is sleepy, CPAP might help them feel better if their disease is moderate or severe.” All other indications are soft.

A version of this article first appeared on Medscape.com.

Aaron B. Holley, MD, is a professor of medicine at Uniformed Services University in Bethesda, Maryland, and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington, DC. He covers a  wide range of topics in pulmonary, critical care, and sleep medicine .

Earlier this year, I stumbled across a podcast in a content update email from the Journal of the American Medical Association. The moderator was interviewing the first author of a study comparing hydrocortisone and fludrocortisone (hydro/fludro) to hydrocortisone alone for treatment of septic shock. In the introduction, the author commented on the discordance in practice among his peers at his hospital. It seemed that there was no consensus on whether fludrocortisone was necessary.

I thought this issue had been settled with publication of the COIITSS trial in 2010. This study randomly assigned 509 patients with septic shock to hydro/fludro versus hydrocortisone alone. There was a nonsignificant reduction in mortality with hydro/fludro and everyone I knew stopped adding fludrocortisone for septic shock. It wasn’t included in guidelines (and still isn›t). I figured the only docs still using it were also prescribing ivermectin and vitamin C – another treatment touted to work in an apocryphal podcast.

It wasn’t just COIITSS that killed fludrocortisone for me. Back in 2002, I was a loyal adherent. That year, a randomized controlled trial (RCT) published by “the lord of corticosteroids for critical illness” doctor, Djillali Annane, found benefit to hydro/fludro in septic shock . Everyone in that study had a cosyntropin stim test and only certain subgroups had better outcomes. As a medical resident paying obeisance to all things evidence-based medicine, I rigidly adopted their protocol for all septic patients. I also kept their insulin between 80 and 110 mg/dL, prescribed drotrecogin alfa, and made sure they were floating in crystalloid. But those are topics for another time.

Subsequent trials and meta-analyses cast doubt on the need for the stim test, and a consensus around hydrocortisone at moderate doses for patients with septic shock emerged. Because one part of the Annane protocol was already deemed unnecessary (the cosyntropin stim test), it was easy to dismiss fludrocortisone after COIITTS was published. Yes, I read Annane’s 2018 APROCCHSS trial, and I’m aware that it found that hydro/fludro reduced 90-day mortality. Like others, I rationalized this finding by framing it as a function of baseline mortality. The two Annane RCTs that found that hydro/fludro reduced mortality in enrolled patients who were considerably more likely to die than those enrolled in RCTs of hydrocortisone alone were negative. It was the target population mortality rate and not the addition of fludrocortisone that made the difference, right?
 

Rethinking hydro/fludro

The author interviewed for the recent JAMA podcast forced me to rethink my blithe dismissal of fludrocortisone. He contended that the COIITTS trial was underpowered and the two Annane RCTs that used fludrocortisone supply the evidence that shows corticosteroids reduce septic shock mortality. As discussed earlier, he found clinical equipoise among his colleagues. Last, he invoked pleiotropic mineralocorticoid effects, such as activation of innate immunity and clearance of alveolar fluid, to support the need to reexamine hydro/fludro.

In his study, he used Big Data to compare hospital records from 2016 to 2020. He analyzed a total of 88,275 patients with septic shock. Most were prescribed hydrocortisone alone (85,995 [97.4%] vs. only 2.6% hydro/fludro). After a number of statistical adjustments and sensitivity analyses, the authors concluded that the addition of fludrocortisone to hydrocortisone for patients with septic shock provides a 3.7% absolute risk reduction in mortality (or discharge to hospice) when compared with hydrocortisone alone. That’s a number needed to treat of 28 to prevent one death (or discharge to hospice).
 

Key takeaways

The study isn’t perfect. In their methods section they use terms like “ensemble machine learner (super learner)” and “immortal time bias.” The first is a fancy way of saying they did a form of propensity scoring, which in turn is a fancy way of saying they tried to control for confounding. The second is a way to adjust for time delays between drug administration. Both are attempts to compensate for the observational design, as is their argument for biologic plausibility. Here they’re on particularly thin ice when trying to prove causal inference. Biologic plausibility is never hard to find; after all, what compound doesn’t have pleiotropic effects? Furthermore, the analysis lacks any data to support their biologic plausibility hypothesis that fludrocortisone’s effect on mortality is mediated via activation of innate immunity and/or clearance of alveolar fluid.

The editorial accompanying this Big Data study endorsed adding fludrocortisone. We have very little that reduces ICU mortality so the low number needed to treat is enticing, especially in light of the low risk from adverse events, so I’m going to start using it. Do I think I’ll save one life for every 28 patients with septic shock to whom I give hydro/fludro instead of hydrocortisone alone? I sure don’t. No way an oral mineralocorticoid at that dose has that type of impact on top of hydrocortisone alone. I still believe that the Annane studies are positive because of the mortality rate in the population enrolled and not because fludrocortisone was added. It all comes full circle, though – 20 years after I abandoned hydro/fludro, I’m going back to it.

Aaron B. Holley, MD, is a professor of medicine at Uniformed Services University in Bethesda, Md., and a pulmonary/critical care and sleep medicine physician at MedStar Washington Hospital Center in Washington, D.C.

A version of this article first appeared on Medscape.com.

Earlier this year, I stumbled across a podcast in a content update email from the Journal of the American Medical Association. The moderator was interviewing the first author of a study comparing hydrocortisone and fludrocortisone (hydro/fludro) to hydrocortisone alone for treatment of septic shock. In the introduction, the author commented on the discordance in practice among his peers at his hospital. It seemed that there was no consensus on whether fludrocortisone was necessary.

I thought this issue had been settled with publication of the COIITSS trial in 2010. This study randomly assigned 509 patients with septic shock to hydro/fludro versus hydrocortisone alone. There was a nonsignificant reduction in mortality with hydro/fludro and everyone I knew stopped adding fludrocortisone for septic shock. It wasn’t included in guidelines (and still isn›t). I figured the only docs still using it were also prescribing ivermectin and vitamin C – another treatment touted to work in an apocryphal podcast.

It wasn’t just COIITSS that killed fludrocortisone for me. Back in 2002, I was a loyal adherent. That year, a randomized controlled trial (RCT) published by “the lord of corticosteroids for critical illness” doctor, Djillali Annane, found benefit to hydro/fludro in septic shock . Everyone in that study had a cosyntropin stim test and only certain subgroups had better outcomes. As a medical resident paying obeisance to all things evidence-based medicine, I rigidly adopted their protocol for all septic patients. I also kept their insulin between 80 and 110 mg/dL, prescribed drotrecogin alfa, and made sure they were floating in crystalloid. But those are topics for another time.

Subsequent trials and meta-analyses cast doubt on the need for the stim test, and a consensus around hydrocortisone at moderate doses for patients with septic shock emerged. Because one part of the Annane protocol was already deemed unnecessary (the cosyntropin stim test), it was easy to dismiss fludrocortisone after COIITTS was published. Yes, I read Annane’s 2018 APROCCHSS trial, and I’m aware that it found that hydro/fludro reduced 90-day mortality. Like others, I rationalized this finding by framing it as a function of baseline mortality. The two Annane RCTs that found that hydro/fludro reduced mortality in enrolled patients who were considerably more likely to die than those enrolled in RCTs of hydrocortisone alone were negative. It was the target population mortality rate and not the addition of fludrocortisone that made the difference, right?
 

Rethinking hydro/fludro

The author interviewed for the recent JAMA podcast forced me to rethink my blithe dismissal of fludrocortisone. He contended that the COIITTS trial was underpowered and the two Annane RCTs that used fludrocortisone supply the evidence that shows corticosteroids reduce septic shock mortality. As discussed earlier, he found clinical equipoise among his colleagues. Last, he invoked pleiotropic mineralocorticoid effects, such as activation of innate immunity and clearance of alveolar fluid, to support the need to reexamine hydro/fludro.

In his study, he used Big Data to compare hospital records from 2016 to 2020. He analyzed a total of 88,275 patients with septic shock. Most were prescribed hydrocortisone alone (85,995 [97.4%] vs. only 2.6% hydro/fludro). After a number of statistical adjustments and sensitivity analyses, the authors concluded that the addition of fludrocortisone to hydrocortisone for patients with septic shock provides a 3.7% absolute risk reduction in mortality (or discharge to hospice) when compared with hydrocortisone alone. That’s a number needed to treat of 28 to prevent one death (or discharge to hospice).
 

Key takeaways

The study isn’t perfect. In their methods section they use terms like “ensemble machine learner (super learner)” and “immortal time bias.” The first is a fancy way of saying they did a form of propensity scoring, which in turn is a fancy way of saying they tried to control for confounding. The second is a way to adjust for time delays between drug administration. Both are attempts to compensate for the observational design, as is their argument for biologic plausibility. Here they’re on particularly thin ice when trying to prove causal inference. Biologic plausibility is never hard to find; after all, what compound doesn’t have pleiotropic effects? Furthermore, the analysis lacks any data to support their biologic plausibility hypothesis that fludrocortisone’s effect on mortality is mediated via activation of innate immunity and/or clearance of alveolar fluid.

The editorial accompanying this Big Data study endorsed adding fludrocortisone. We have very little that reduces ICU mortality so the low number needed to treat is enticing, especially in light of the low risk from adverse events, so I’m going to start using it. Do I think I’ll save one life for every 28 patients with septic shock to whom I give hydro/fludro instead of hydrocortisone alone? I sure don’t. No way an oral mineralocorticoid at that dose has that type of impact on top of hydrocortisone alone. I still believe that the Annane studies are positive because of the mortality rate in the population enrolled and not because fludrocortisone was added. It all comes full circle, though – 20 years after I abandoned hydro/fludro, I’m going back to it.

Aaron B. Holley, MD, is a professor of medicine at Uniformed Services University in Bethesda, Md., and a pulmonary/critical care and sleep medicine physician at MedStar Washington Hospital Center in Washington, D.C.

A version of this article first appeared on Medscape.com.

Earlier this year, I stumbled across a podcast in a content update email from the Journal of the American Medical Association. The moderator was interviewing the first author of a study comparing hydrocortisone and fludrocortisone (hydro/fludro) to hydrocortisone alone for treatment of septic shock. In the introduction, the author commented on the discordance in practice among his peers at his hospital. It seemed that there was no consensus on whether fludrocortisone was necessary.

I thought this issue had been settled with publication of the COIITSS trial in 2010. This study randomly assigned 509 patients with septic shock to hydro/fludro versus hydrocortisone alone. There was a nonsignificant reduction in mortality with hydro/fludro and everyone I knew stopped adding fludrocortisone for septic shock. It wasn’t included in guidelines (and still isn›t). I figured the only docs still using it were also prescribing ivermectin and vitamin C – another treatment touted to work in an apocryphal podcast.

It wasn’t just COIITSS that killed fludrocortisone for me. Back in 2002, I was a loyal adherent. That year, a randomized controlled trial (RCT) published by “the lord of corticosteroids for critical illness” doctor, Djillali Annane, found benefit to hydro/fludro in septic shock . Everyone in that study had a cosyntropin stim test and only certain subgroups had better outcomes. As a medical resident paying obeisance to all things evidence-based medicine, I rigidly adopted their protocol for all septic patients. I also kept their insulin between 80 and 110 mg/dL, prescribed drotrecogin alfa, and made sure they were floating in crystalloid. But those are topics for another time.

Subsequent trials and meta-analyses cast doubt on the need for the stim test, and a consensus around hydrocortisone at moderate doses for patients with septic shock emerged. Because one part of the Annane protocol was already deemed unnecessary (the cosyntropin stim test), it was easy to dismiss fludrocortisone after COIITTS was published. Yes, I read Annane’s 2018 APROCCHSS trial, and I’m aware that it found that hydro/fludro reduced 90-day mortality. Like others, I rationalized this finding by framing it as a function of baseline mortality. The two Annane RCTs that found that hydro/fludro reduced mortality in enrolled patients who were considerably more likely to die than those enrolled in RCTs of hydrocortisone alone were negative. It was the target population mortality rate and not the addition of fludrocortisone that made the difference, right?
 

Rethinking hydro/fludro

The author interviewed for the recent JAMA podcast forced me to rethink my blithe dismissal of fludrocortisone. He contended that the COIITTS trial was underpowered and the two Annane RCTs that used fludrocortisone supply the evidence that shows corticosteroids reduce septic shock mortality. As discussed earlier, he found clinical equipoise among his colleagues. Last, he invoked pleiotropic mineralocorticoid effects, such as activation of innate immunity and clearance of alveolar fluid, to support the need to reexamine hydro/fludro.

In his study, he used Big Data to compare hospital records from 2016 to 2020. He analyzed a total of 88,275 patients with septic shock. Most were prescribed hydrocortisone alone (85,995 [97.4%] vs. only 2.6% hydro/fludro). After a number of statistical adjustments and sensitivity analyses, the authors concluded that the addition of fludrocortisone to hydrocortisone for patients with septic shock provides a 3.7% absolute risk reduction in mortality (or discharge to hospice) when compared with hydrocortisone alone. That’s a number needed to treat of 28 to prevent one death (or discharge to hospice).
 

Key takeaways

The study isn’t perfect. In their methods section they use terms like “ensemble machine learner (super learner)” and “immortal time bias.” The first is a fancy way of saying they did a form of propensity scoring, which in turn is a fancy way of saying they tried to control for confounding. The second is a way to adjust for time delays between drug administration. Both are attempts to compensate for the observational design, as is their argument for biologic plausibility. Here they’re on particularly thin ice when trying to prove causal inference. Biologic plausibility is never hard to find; after all, what compound doesn’t have pleiotropic effects? Furthermore, the analysis lacks any data to support their biologic plausibility hypothesis that fludrocortisone’s effect on mortality is mediated via activation of innate immunity and/or clearance of alveolar fluid.

The editorial accompanying this Big Data study endorsed adding fludrocortisone. We have very little that reduces ICU mortality so the low number needed to treat is enticing, especially in light of the low risk from adverse events, so I’m going to start using it. Do I think I’ll save one life for every 28 patients with septic shock to whom I give hydro/fludro instead of hydrocortisone alone? I sure don’t. No way an oral mineralocorticoid at that dose has that type of impact on top of hydrocortisone alone. I still believe that the Annane studies are positive because of the mortality rate in the population enrolled and not because fludrocortisone was added. It all comes full circle, though – 20 years after I abandoned hydro/fludro, I’m going back to it.

Aaron B. Holley, MD, is a professor of medicine at Uniformed Services University in Bethesda, Md., and a pulmonary/critical care and sleep medicine physician at MedStar Washington Hospital Center in Washington, D.C.

A version of this article first appeared on Medscape.com.

A 70-year-old woman is admitted to the intensive care unit with a pH of 7.1, an acute kidney injury (AKI), and ketonuria. She is volume depleted and her history is consistent with starvation ketosis. This LOL truly is in NAD (that’s little old lady in no acute distress, for those who haven’t read The House of God). She is clinically stable and seemingly unperturbed by the flurry of activity surrounding her admission.

Your resident is concerned by the severity of the acidosis and suggests starting an intravenous bicarbonate drip. The fellow is adamantly against it. He’s been taught that intravenous bicarbonate increases the serum pH but paradoxically causes intracellular acidosis. As the attending you elect to observe fellow autonomy – no bicarb is given. Because any debate on rounds is a “teachable moment,” you decide to review the evidence and physiology behind infusing bicarbonate.
 

What do the data reveal?

An excellent review published in CHEST in 2000 covers the physiologic effects of bicarbonate, specifically related to lactic acidosis, which our patient didn’t have. Aside from that difference, the review validates the fellow’s opinion. In short, the authors stated that a low pH may be a marker of a dangerous systemic condition, but it need not be corrected for its own sake. It is unlikely to provoke hemodynamic or respiratory compromise outside the setting of shock or hypercapnia. Intravenous bicarbonate can lead to intracellular acidosis, hypercapnia, hypocalcemia, and a reduction in oxygen delivery via the Bohr effect. The authors concluded that because the benefits are unproven and the negative effects are real, intravenous bicarbonate should not be used to correct a metabolic acidosis.

The CHEST review hardly settles the issue, though. A survey published a few years later found a majority of intensivists and nephrologists used intravenous bicarbonate to treat metabolic acidosis while the Surviving Sepsis Campaign Guidelines for the Management of Sepsis and Septic Shock published in 2017 recommended against bicarbonate for acidosis. It wasn’t until 2018 that we reached the holy grail: a randomized controlled trial.

The BICAR-ICU study randomly assigned patients with a pH of 7.20 or less, PCO₂ of 45 mm Hg or less, and sodium bicarbonate concentration of 20 mmol/L or less to receive no bicarbonate versus a sodium bicarbonate drip to maintain a pH greater than 7.30. There’s additional nuance to the trial design and even more detail in the results. To summarize, there was signal for an improvement in renal outcomes across all patients, and those with AKI saw a mortality benefit. Post–BICAR-ICU iterations of the Surviving Sepsis Campaign Guidelines have incorporated these findings by recommending intravenous bicarbonate for patients with sepsis who have AKI and a pH of 7.20 or less.

That’s not to say BICAR-ICU has settled the issue. Although it’s far and away the best we have, there were fewer than 400 total patients in their intention-to-treat analysis. It was open label, with lots of crossover. The primary outcome was negative for the entire population, with only a subgroup (albeit a prespecified one) showing benefit. Finally, the results weren’t stratified by etiology for the metabolic acidosis. There was also evidence of alkalosis and hypocalcemia in the treatment group.

Last but not least in terms of importance, in most cases when bicarbonate is being considered, wouldn’t some form of renal replacement therapy (RRT) be preferred? This point was raised by nephrologists and intensivists when we covered BICAR-ICU in a journal club at my former program. It’s also mentioned in an accompanying editorial. RRT timing is controversial, and a detailed discussion is outside the scope of this piece and beyond the limits of my current knowledge base. But I do know that the A in the A-E-I-O-U acute indications for dialysis pneumonic stands for acidosis.

Our patient had AKI, a pH of 7.20 or less, and a pCO₂ well under 45 mm Hg. Does BICAR-ICU support the resident’s inclination to start a drip? Sort of. The majority of patients enrolled in BICAR-ICU were in shock or were recovering from cardiac arrest, so it’s not clear the results can be generalized to our LOL with starvation ketosis. Extrapolating from studies of diabetic ketoacidosis (DKA) seems more appropriate, and here the data are poor but equivocal. Reviews are generally negative but don’t rule out the use of intravenous bicarbonate in certain patients with DKA.
 

Key takeaways

Our patient survived a 24-hour ICU stay with neither cardiopulmonary decompensation nor a need for RRT. Not sure how she did out of the ICU; presumably she was discharged soon after transfer. As is always the case with anecdotal medicine, the absence of a control prevents assessment of the counterfactual. Is it possible she may have done “better” with intravenous bicarbonate? Seems unlikely to me, though I doubt there would have been demonstrable adverse effects. Perhaps next time the fellow can observe resident autonomy?

Aaron B. Holley, MD, is a professor of medicine at Uniformed Services University of the Health Sciences, Bethesda, Md., and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center. He reported conflicts of interest with Metapharm, CHEST College, and WebMD.
 

A version of this article first appeared on Medscape.com.

A 70-year-old woman is admitted to the intensive care unit with a pH of 7.1, an acute kidney injury (AKI), and ketonuria. She is volume depleted and her history is consistent with starvation ketosis. This LOL truly is in NAD (that’s little old lady in no acute distress, for those who haven’t read The House of God). She is clinically stable and seemingly unperturbed by the flurry of activity surrounding her admission.

Your resident is concerned by the severity of the acidosis and suggests starting an intravenous bicarbonate drip. The fellow is adamantly against it. He’s been taught that intravenous bicarbonate increases the serum pH but paradoxically causes intracellular acidosis. As the attending you elect to observe fellow autonomy – no bicarb is given. Because any debate on rounds is a “teachable moment,” you decide to review the evidence and physiology behind infusing bicarbonate.
 

What do the data reveal?

An excellent review published in CHEST in 2000 covers the physiologic effects of bicarbonate, specifically related to lactic acidosis, which our patient didn’t have. Aside from that difference, the review validates the fellow’s opinion. In short, the authors stated that a low pH may be a marker of a dangerous systemic condition, but it need not be corrected for its own sake. It is unlikely to provoke hemodynamic or respiratory compromise outside the setting of shock or hypercapnia. Intravenous bicarbonate can lead to intracellular acidosis, hypercapnia, hypocalcemia, and a reduction in oxygen delivery via the Bohr effect. The authors concluded that because the benefits are unproven and the negative effects are real, intravenous bicarbonate should not be used to correct a metabolic acidosis.

The CHEST review hardly settles the issue, though. A survey published a few years later found a majority of intensivists and nephrologists used intravenous bicarbonate to treat metabolic acidosis while the Surviving Sepsis Campaign Guidelines for the Management of Sepsis and Septic Shock published in 2017 recommended against bicarbonate for acidosis. It wasn’t until 2018 that we reached the holy grail: a randomized controlled trial.

The BICAR-ICU study randomly assigned patients with a pH of 7.20 or less, PCO₂ of 45 mm Hg or less, and sodium bicarbonate concentration of 20 mmol/L or less to receive no bicarbonate versus a sodium bicarbonate drip to maintain a pH greater than 7.30. There’s additional nuance to the trial design and even more detail in the results. To summarize, there was signal for an improvement in renal outcomes across all patients, and those with AKI saw a mortality benefit. Post–BICAR-ICU iterations of the Surviving Sepsis Campaign Guidelines have incorporated these findings by recommending intravenous bicarbonate for patients with sepsis who have AKI and a pH of 7.20 or less.

That’s not to say BICAR-ICU has settled the issue. Although it’s far and away the best we have, there were fewer than 400 total patients in their intention-to-treat analysis. It was open label, with lots of crossover. The primary outcome was negative for the entire population, with only a subgroup (albeit a prespecified one) showing benefit. Finally, the results weren’t stratified by etiology for the metabolic acidosis. There was also evidence of alkalosis and hypocalcemia in the treatment group.

Last but not least in terms of importance, in most cases when bicarbonate is being considered, wouldn’t some form of renal replacement therapy (RRT) be preferred? This point was raised by nephrologists and intensivists when we covered BICAR-ICU in a journal club at my former program. It’s also mentioned in an accompanying editorial. RRT timing is controversial, and a detailed discussion is outside the scope of this piece and beyond the limits of my current knowledge base. But I do know that the A in the A-E-I-O-U acute indications for dialysis pneumonic stands for acidosis.

Our patient had AKI, a pH of 7.20 or less, and a pCO₂ well under 45 mm Hg. Does BICAR-ICU support the resident’s inclination to start a drip? Sort of. The majority of patients enrolled in BICAR-ICU were in shock or were recovering from cardiac arrest, so it’s not clear the results can be generalized to our LOL with starvation ketosis. Extrapolating from studies of diabetic ketoacidosis (DKA) seems more appropriate, and here the data are poor but equivocal. Reviews are generally negative but don’t rule out the use of intravenous bicarbonate in certain patients with DKA.
 

Key takeaways

Our patient survived a 24-hour ICU stay with neither cardiopulmonary decompensation nor a need for RRT. Not sure how she did out of the ICU; presumably she was discharged soon after transfer. As is always the case with anecdotal medicine, the absence of a control prevents assessment of the counterfactual. Is it possible she may have done “better” with intravenous bicarbonate? Seems unlikely to me, though I doubt there would have been demonstrable adverse effects. Perhaps next time the fellow can observe resident autonomy?

Aaron B. Holley, MD, is a professor of medicine at Uniformed Services University of the Health Sciences, Bethesda, Md., and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center. He reported conflicts of interest with Metapharm, CHEST College, and WebMD.
 

A version of this article first appeared on Medscape.com.

A 70-year-old woman is admitted to the intensive care unit with a pH of 7.1, an acute kidney injury (AKI), and ketonuria. She is volume depleted and her history is consistent with starvation ketosis. This LOL truly is in NAD (that’s little old lady in no acute distress, for those who haven’t read The House of God). She is clinically stable and seemingly unperturbed by the flurry of activity surrounding her admission.

Your resident is concerned by the severity of the acidosis and suggests starting an intravenous bicarbonate drip. The fellow is adamantly against it. He’s been taught that intravenous bicarbonate increases the serum pH but paradoxically causes intracellular acidosis. As the attending you elect to observe fellow autonomy – no bicarb is given. Because any debate on rounds is a “teachable moment,” you decide to review the evidence and physiology behind infusing bicarbonate.
 

What do the data reveal?

An excellent review published in CHEST in 2000 covers the physiologic effects of bicarbonate, specifically related to lactic acidosis, which our patient didn’t have. Aside from that difference, the review validates the fellow’s opinion. In short, the authors stated that a low pH may be a marker of a dangerous systemic condition, but it need not be corrected for its own sake. It is unlikely to provoke hemodynamic or respiratory compromise outside the setting of shock or hypercapnia. Intravenous bicarbonate can lead to intracellular acidosis, hypercapnia, hypocalcemia, and a reduction in oxygen delivery via the Bohr effect. The authors concluded that because the benefits are unproven and the negative effects are real, intravenous bicarbonate should not be used to correct a metabolic acidosis.

The CHEST review hardly settles the issue, though. A survey published a few years later found a majority of intensivists and nephrologists used intravenous bicarbonate to treat metabolic acidosis while the Surviving Sepsis Campaign Guidelines for the Management of Sepsis and Septic Shock published in 2017 recommended against bicarbonate for acidosis. It wasn’t until 2018 that we reached the holy grail: a randomized controlled trial.

The BICAR-ICU study randomly assigned patients with a pH of 7.20 or less, PCO₂ of 45 mm Hg or less, and sodium bicarbonate concentration of 20 mmol/L or less to receive no bicarbonate versus a sodium bicarbonate drip to maintain a pH greater than 7.30. There’s additional nuance to the trial design and even more detail in the results. To summarize, there was signal for an improvement in renal outcomes across all patients, and those with AKI saw a mortality benefit. Post–BICAR-ICU iterations of the Surviving Sepsis Campaign Guidelines have incorporated these findings by recommending intravenous bicarbonate for patients with sepsis who have AKI and a pH of 7.20 or less.

That’s not to say BICAR-ICU has settled the issue. Although it’s far and away the best we have, there were fewer than 400 total patients in their intention-to-treat analysis. It was open label, with lots of crossover. The primary outcome was negative for the entire population, with only a subgroup (albeit a prespecified one) showing benefit. Finally, the results weren’t stratified by etiology for the metabolic acidosis. There was also evidence of alkalosis and hypocalcemia in the treatment group.

Last but not least in terms of importance, in most cases when bicarbonate is being considered, wouldn’t some form of renal replacement therapy (RRT) be preferred? This point was raised by nephrologists and intensivists when we covered BICAR-ICU in a journal club at my former program. It’s also mentioned in an accompanying editorial. RRT timing is controversial, and a detailed discussion is outside the scope of this piece and beyond the limits of my current knowledge base. But I do know that the A in the A-E-I-O-U acute indications for dialysis pneumonic stands for acidosis.

Our patient had AKI, a pH of 7.20 or less, and a pCO₂ well under 45 mm Hg. Does BICAR-ICU support the resident’s inclination to start a drip? Sort of. The majority of patients enrolled in BICAR-ICU were in shock or were recovering from cardiac arrest, so it’s not clear the results can be generalized to our LOL with starvation ketosis. Extrapolating from studies of diabetic ketoacidosis (DKA) seems more appropriate, and here the data are poor but equivocal. Reviews are generally negative but don’t rule out the use of intravenous bicarbonate in certain patients with DKA.
 

Key takeaways

Our patient survived a 24-hour ICU stay with neither cardiopulmonary decompensation nor a need for RRT. Not sure how she did out of the ICU; presumably she was discharged soon after transfer. As is always the case with anecdotal medicine, the absence of a control prevents assessment of the counterfactual. Is it possible she may have done “better” with intravenous bicarbonate? Seems unlikely to me, though I doubt there would have been demonstrable adverse effects. Perhaps next time the fellow can observe resident autonomy?

Aaron B. Holley, MD, is a professor of medicine at Uniformed Services University of the Health Sciences, Bethesda, Md., and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center. He reported conflicts of interest with Metapharm, CHEST College, and WebMD.
 

A version of this article first appeared on Medscape.com.

The recent statement on interpretive strategies for lung testing uses the acronym PRISm for preserved ratio impaired spirometry. PRISm identifies patients with a normal forced expiratory volume in 1 second/forced vital capacity ratio but abnormal FEV₁ and/or FVC (usually both). Most medical students are taught to call this a “restrictive pattern,” and every first-year pulmonary fellow orders full lung volumes when they see it. If total lung capacity (TLC) is normal, PRISm becomes the nonspecific pattern. If TLC is low, then the patient has “true” restriction, and if it’s elevated, then hyperinflation may be present.

The traditional classification scheme for basic spirometry interpretation (normal, restricted, obstructed, or mixed) is simple and conceptually clear. Because this simplicity is achieved at the expense of precision, the “restrictive pattern” label is due for retirement. It turns out that many with this pattern won’t have an abnormal TLC, so the name is, in some ways, a misnomer and can be misleading. Enter PRISm, a more descriptive and inclusive term. The phrase also lends itself to a phonetic acronym that is fun to say, easy to remember, and likely to catch on with learners.

Information on occurrence and clinical behavior comes from large cohorts with basic spirometry, but without full lung volumes because PRISm no longer applies once TLC is determined. As may be expected, prevalence varies by the population studied. Estimates for general populations have been in the 7%-12% range; however, one study examining a database of patients with clinical spirometry referrals found a prevalence of 22.3%. Rates may be far higher in low- and middle-income countries. Identified risk factors include sex, tobacco use, and body mass index; the presence of PRISm is associated with respiratory symptoms and mortality. Thus, PRISm is common and it matters.

Along with PRISm, the nonspecific pattern is a new addition, if not a new concept, to the 2022 interpretative strategies statement. As with PRISm, the title is necessarily broad, though far less imaginative. Defined by reductions in FEV₁ and FVC and a normal TLC, the nonspecific pattern has classically been considered a marker of early airway disease. The idea is that early, heterogeneous closure of distal segments of the bronchial tree can reduce total volume during a forced expiration before affecting the FEV₁/FVC. The fact that the TLC is not a forced maneuver means there is proportionately less effect from more collapsible/susceptible smaller units. More recent data suggest that there are additional causes.

Because the nonspecific pattern requires full lung volumes, we have less population-level data than for PRISm. Estimated prevalence is approximately 9.5% in patients with complete test results. The two most common causes are obesity and airway obstruction, and the pattern is relatively stable over time. Notably, an increase in specific airway resistance or TLC minus alveolar volume difference predicts progression to frank obstruction on spirometry.

The physiologic changes that obesity inflicts on the lung have been well described. Patients with obesity breathe at lower lung volumes and are therefore susceptible to small airway closure at rest and during forced expiration. There is no doubt that the increased recognition of PRISm and the nonspecific pattern is in part related to the worldwide rise in obesity rates.
 

Key takeaways

In summary, PRISm and the nonspecific pattern are now part of the classification scheme we use for spirometry and full lung volumes, respectively. They should be included in interpretations given their diagnostic and predictive value. Airway disease and obesity are common causes and often coexist with either pattern. Many will not have a true, restrictive lung deficit, and a reductionist approach to interpretation is likely to lead to erroneous diagnoses. There were many important updates included in the 2022 iteration on lung testing interpretation that should not fly under the radar.

Dr. Holley is professor of medicine at Uniformed Services University in Bethesda, Md., and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington. He disclosed ties with CHEST College, Metapharm, and WebMD.

A version of this article appeared on Medscape.com.

The recent statement on interpretive strategies for lung testing uses the acronym PRISm for preserved ratio impaired spirometry. PRISm identifies patients with a normal forced expiratory volume in 1 second/forced vital capacity ratio but abnormal FEV₁ and/or FVC (usually both). Most medical students are taught to call this a “restrictive pattern,” and every first-year pulmonary fellow orders full lung volumes when they see it. If total lung capacity (TLC) is normal, PRISm becomes the nonspecific pattern. If TLC is low, then the patient has “true” restriction, and if it’s elevated, then hyperinflation may be present.

The traditional classification scheme for basic spirometry interpretation (normal, restricted, obstructed, or mixed) is simple and conceptually clear. Because this simplicity is achieved at the expense of precision, the “restrictive pattern” label is due for retirement. It turns out that many with this pattern won’t have an abnormal TLC, so the name is, in some ways, a misnomer and can be misleading. Enter PRISm, a more descriptive and inclusive term. The phrase also lends itself to a phonetic acronym that is fun to say, easy to remember, and likely to catch on with learners.

Information on occurrence and clinical behavior comes from large cohorts with basic spirometry, but without full lung volumes because PRISm no longer applies once TLC is determined. As may be expected, prevalence varies by the population studied. Estimates for general populations have been in the 7%-12% range; however, one study examining a database of patients with clinical spirometry referrals found a prevalence of 22.3%. Rates may be far higher in low- and middle-income countries. Identified risk factors include sex, tobacco use, and body mass index; the presence of PRISm is associated with respiratory symptoms and mortality. Thus, PRISm is common and it matters.

Along with PRISm, the nonspecific pattern is a new addition, if not a new concept, to the 2022 interpretative strategies statement. As with PRISm, the title is necessarily broad, though far less imaginative. Defined by reductions in FEV₁ and FVC and a normal TLC, the nonspecific pattern has classically been considered a marker of early airway disease. The idea is that early, heterogeneous closure of distal segments of the bronchial tree can reduce total volume during a forced expiration before affecting the FEV₁/FVC. The fact that the TLC is not a forced maneuver means there is proportionately less effect from more collapsible/susceptible smaller units. More recent data suggest that there are additional causes.

Because the nonspecific pattern requires full lung volumes, we have less population-level data than for PRISm. Estimated prevalence is approximately 9.5% in patients with complete test results. The two most common causes are obesity and airway obstruction, and the pattern is relatively stable over time. Notably, an increase in specific airway resistance or TLC minus alveolar volume difference predicts progression to frank obstruction on spirometry.

The physiologic changes that obesity inflicts on the lung have been well described. Patients with obesity breathe at lower lung volumes and are therefore susceptible to small airway closure at rest and during forced expiration. There is no doubt that the increased recognition of PRISm and the nonspecific pattern is in part related to the worldwide rise in obesity rates.
 

Key takeaways

In summary, PRISm and the nonspecific pattern are now part of the classification scheme we use for spirometry and full lung volumes, respectively. They should be included in interpretations given their diagnostic and predictive value. Airway disease and obesity are common causes and often coexist with either pattern. Many will not have a true, restrictive lung deficit, and a reductionist approach to interpretation is likely to lead to erroneous diagnoses. There were many important updates included in the 2022 iteration on lung testing interpretation that should not fly under the radar.

Dr. Holley is professor of medicine at Uniformed Services University in Bethesda, Md., and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington. He disclosed ties with CHEST College, Metapharm, and WebMD.

A version of this article appeared on Medscape.com.

The recent statement on interpretive strategies for lung testing uses the acronym PRISm for preserved ratio impaired spirometry. PRISm identifies patients with a normal forced expiratory volume in 1 second/forced vital capacity ratio but abnormal FEV₁ and/or FVC (usually both). Most medical students are taught to call this a “restrictive pattern,” and every first-year pulmonary fellow orders full lung volumes when they see it. If total lung capacity (TLC) is normal, PRISm becomes the nonspecific pattern. If TLC is low, then the patient has “true” restriction, and if it’s elevated, then hyperinflation may be present.

The traditional classification scheme for basic spirometry interpretation (normal, restricted, obstructed, or mixed) is simple and conceptually clear. Because this simplicity is achieved at the expense of precision, the “restrictive pattern” label is due for retirement. It turns out that many with this pattern won’t have an abnormal TLC, so the name is, in some ways, a misnomer and can be misleading. Enter PRISm, a more descriptive and inclusive term. The phrase also lends itself to a phonetic acronym that is fun to say, easy to remember, and likely to catch on with learners.

Information on occurrence and clinical behavior comes from large cohorts with basic spirometry, but without full lung volumes because PRISm no longer applies once TLC is determined. As may be expected, prevalence varies by the population studied. Estimates for general populations have been in the 7%-12% range; however, one study examining a database of patients with clinical spirometry referrals found a prevalence of 22.3%. Rates may be far higher in low- and middle-income countries. Identified risk factors include sex, tobacco use, and body mass index; the presence of PRISm is associated with respiratory symptoms and mortality. Thus, PRISm is common and it matters.

Along with PRISm, the nonspecific pattern is a new addition, if not a new concept, to the 2022 interpretative strategies statement. As with PRISm, the title is necessarily broad, though far less imaginative. Defined by reductions in FEV₁ and FVC and a normal TLC, the nonspecific pattern has classically been considered a marker of early airway disease. The idea is that early, heterogeneous closure of distal segments of the bronchial tree can reduce total volume during a forced expiration before affecting the FEV₁/FVC. The fact that the TLC is not a forced maneuver means there is proportionately less effect from more collapsible/susceptible smaller units. More recent data suggest that there are additional causes.

Because the nonspecific pattern requires full lung volumes, we have less population-level data than for PRISm. Estimated prevalence is approximately 9.5% in patients with complete test results. The two most common causes are obesity and airway obstruction, and the pattern is relatively stable over time. Notably, an increase in specific airway resistance or TLC minus alveolar volume difference predicts progression to frank obstruction on spirometry.

The physiologic changes that obesity inflicts on the lung have been well described. Patients with obesity breathe at lower lung volumes and are therefore susceptible to small airway closure at rest and during forced expiration. There is no doubt that the increased recognition of PRISm and the nonspecific pattern is in part related to the worldwide rise in obesity rates.
 

Key takeaways

In summary, PRISm and the nonspecific pattern are now part of the classification scheme we use for spirometry and full lung volumes, respectively. They should be included in interpretations given their diagnostic and predictive value. Airway disease and obesity are common causes and often coexist with either pattern. Many will not have a true, restrictive lung deficit, and a reductionist approach to interpretation is likely to lead to erroneous diagnoses. There were many important updates included in the 2022 iteration on lung testing interpretation that should not fly under the radar.

Dr. Holley is professor of medicine at Uniformed Services University in Bethesda, Md., and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington. He disclosed ties with CHEST College, Metapharm, and WebMD.

A version of this article appeared on Medscape.com.

It’s tough to keep up with the proliferation of monoclonal antibodies. Seems every day I’m confronted by a patient who’s using a new drug with a name ending in “mab.” That drug blocks a cellular receptor I haven’t heard of that’s involved in a cascade of interactions I haven’t thought about since medical school. The resulting disruption reduces disease burden, typically at great expense to the medical system, the patient, or both. We’ve truly entered the era of precision medicine. It’s not enough to understand disease; you also must know its heterogeneous expression so that you can prescribe the ‘mab that targets the biology responsible for variants in behavior. All diseases are, in fact, syndromes. This isn’t a bad thing, but it’s a challenge.

A series of ‘mabs have been approved for treating type 2 high (TH2) or eosinophilic asthma. We refer to this group of ‘mabs generically as biologics. The group includes omalizumab, mepolizumab, dupilumab, benralizumab, reslizumab, and tezepelumab. While mechanism of action varies slightly across drugs, the biologics all target a specific arm of the immune system. Efficacy is linearly related to serum eosinophil count and there’s little clinically or pharmacologically to distinguish one from another. Of course, no head-to-head comparisons of efficacy are available and there’s no financial incentive for them to be performed.
 

Latest research

A new randomized controlled trial (RCT) of dupilumab for chronic obstructive pulmonary disease (COPD) adds to the aforementioned biologic knowledge base. Turns out it works as long as the patients are carefully selected. Researchers enrolled GOLD D (or E depending on which iteration of the GOLD Statement you use) patients on triple inhaler therapy (inhaled corticosteroids [ICS]/long-acting beta-agonist [LABA]/long-acting muscarinic antagonist [LAMA]) with two moderate exacerbations or one exacerbation requiring hospitalization in the past year. Blood eosinophil counts were > 300 cells/mcL and chronic bronchitis was present clinically. The primary and multiple secondary outcomes were improved with dupilumab.

This is welcome news. I’ve treated countless patients with severe COPD who have repeated exacerbations despite my efforts to prevent them. These patients are on ICS/LABA/LAMA and azithromycin or roflumilast, and occasionally both. While every COPD guideline known to man forbids using chronic oral corticosteroids (OCS), I’ve prescribed them repeatedly because the benefits to keeping a recalcitrant, exacerbating patient out of the hospital seem to outweigh OCS risks. It would be nice to have a better option. Although we were taught that they were immutably distinct in medical school, every first-year pulmonary fellow knows that asthma and COPD share more similarities than differences, so it makes sense that proven asthma therapies would work for some patients with COPD.

However, the dupilumab study must be placed in context. Past studies haven’t been as positive. In 2017, two separate RCTs found that mepolizumab reduced the annual rate of moderate to severe exacerbations (primary outcome) in one trial but not the other. Interpretation gets more complicated when broken down by intention to treat (ITT) vs. modified ITT and when secondary outcomes are considered. Sparing you those details, this trial does not instill confidence, leading the Food and Drug Administration to refuse approval for mepolizumab for COPD. A second RCT of benralizumab for COPD was published in 2019. Much less cognitive load was required to interpret this one; it was negative. FDA approval was not requested.

Looking through the trial designs for the three RCTs of biologics for COPD, I couldn’t find major differences that could explain the discordant results. Sample size and enrollment criteria were similar. As stated, I don’t believe that the biologic data in asthma allow for predicting efficacy in one eosinophilic patient vs. another and I assume the same would be true for COPD. All three trials found that eosinophils were eliminated, so responses were biologically equivalent.
 

Key takeaways

If trial design and pharmacology don’t account for the disparate outcomes, how do we explain them? More important, how do we translate these trials into clinical practice? I looked for a review or editorial by a scientist-clinician smarter than I so I could steal their ideas and express them as pedantic euphemisms here. I found it curious that I was unable to find one. A recent publication in the American Journal of Respiratory and Critical Care Medicine suggests that the answer lies within the complex lattice of eosinophil subtypes, but I’m unqualified to judge the veracity of this “phenotype within a phenotype” theory.

For now, there will be no biologics prescribed for COPD – at least not by me. More trials in COPD are being done. We should have results on tezepelumab, that great savior that may cover noneosinophilic asthma phenotypes, within the next few years. Until then, we’re stuck defying guidelines with the anachronistic use of OCS for the COPD patient who exacerbates through ICS/LABA/LAMA, roflumilast, and azithromycin.

Dr. Holley is professor of medicine at Uniformed Services University in Bethesda, Md., and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington. He reported receiving income from CHEST College, Metapharm, and WebMD.

A version of this article first appeared on Medscape.com.

It’s tough to keep up with the proliferation of monoclonal antibodies. Seems every day I’m confronted by a patient who’s using a new drug with a name ending in “mab.” That drug blocks a cellular receptor I haven’t heard of that’s involved in a cascade of interactions I haven’t thought about since medical school. The resulting disruption reduces disease burden, typically at great expense to the medical system, the patient, or both. We’ve truly entered the era of precision medicine. It’s not enough to understand disease; you also must know its heterogeneous expression so that you can prescribe the ‘mab that targets the biology responsible for variants in behavior. All diseases are, in fact, syndromes. This isn’t a bad thing, but it’s a challenge.

A series of ‘mabs have been approved for treating type 2 high (TH2) or eosinophilic asthma. We refer to this group of ‘mabs generically as biologics. The group includes omalizumab, mepolizumab, dupilumab, benralizumab, reslizumab, and tezepelumab. While mechanism of action varies slightly across drugs, the biologics all target a specific arm of the immune system. Efficacy is linearly related to serum eosinophil count and there’s little clinically or pharmacologically to distinguish one from another. Of course, no head-to-head comparisons of efficacy are available and there’s no financial incentive for them to be performed.
 

Latest research

A new randomized controlled trial (RCT) of dupilumab for chronic obstructive pulmonary disease (COPD) adds to the aforementioned biologic knowledge base. Turns out it works as long as the patients are carefully selected. Researchers enrolled GOLD D (or E depending on which iteration of the GOLD Statement you use) patients on triple inhaler therapy (inhaled corticosteroids [ICS]/long-acting beta-agonist [LABA]/long-acting muscarinic antagonist [LAMA]) with two moderate exacerbations or one exacerbation requiring hospitalization in the past year. Blood eosinophil counts were > 300 cells/mcL and chronic bronchitis was present clinically. The primary and multiple secondary outcomes were improved with dupilumab.

This is welcome news. I’ve treated countless patients with severe COPD who have repeated exacerbations despite my efforts to prevent them. These patients are on ICS/LABA/LAMA and azithromycin or roflumilast, and occasionally both. While every COPD guideline known to man forbids using chronic oral corticosteroids (OCS), I’ve prescribed them repeatedly because the benefits to keeping a recalcitrant, exacerbating patient out of the hospital seem to outweigh OCS risks. It would be nice to have a better option. Although we were taught that they were immutably distinct in medical school, every first-year pulmonary fellow knows that asthma and COPD share more similarities than differences, so it makes sense that proven asthma therapies would work for some patients with COPD.

However, the dupilumab study must be placed in context. Past studies haven’t been as positive. In 2017, two separate RCTs found that mepolizumab reduced the annual rate of moderate to severe exacerbations (primary outcome) in one trial but not the other. Interpretation gets more complicated when broken down by intention to treat (ITT) vs. modified ITT and when secondary outcomes are considered. Sparing you those details, this trial does not instill confidence, leading the Food and Drug Administration to refuse approval for mepolizumab for COPD. A second RCT of benralizumab for COPD was published in 2019. Much less cognitive load was required to interpret this one; it was negative. FDA approval was not requested.

Looking through the trial designs for the three RCTs of biologics for COPD, I couldn’t find major differences that could explain the discordant results. Sample size and enrollment criteria were similar. As stated, I don’t believe that the biologic data in asthma allow for predicting efficacy in one eosinophilic patient vs. another and I assume the same would be true for COPD. All three trials found that eosinophils were eliminated, so responses were biologically equivalent.
 

Key takeaways

If trial design and pharmacology don’t account for the disparate outcomes, how do we explain them? More important, how do we translate these trials into clinical practice? I looked for a review or editorial by a scientist-clinician smarter than I so I could steal their ideas and express them as pedantic euphemisms here. I found it curious that I was unable to find one. A recent publication in the American Journal of Respiratory and Critical Care Medicine suggests that the answer lies within the complex lattice of eosinophil subtypes, but I’m unqualified to judge the veracity of this “phenotype within a phenotype” theory.

For now, there will be no biologics prescribed for COPD – at least not by me. More trials in COPD are being done. We should have results on tezepelumab, that great savior that may cover noneosinophilic asthma phenotypes, within the next few years. Until then, we’re stuck defying guidelines with the anachronistic use of OCS for the COPD patient who exacerbates through ICS/LABA/LAMA, roflumilast, and azithromycin.

Dr. Holley is professor of medicine at Uniformed Services University in Bethesda, Md., and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington. He reported receiving income from CHEST College, Metapharm, and WebMD.

A version of this article first appeared on Medscape.com.

It’s tough to keep up with the proliferation of monoclonal antibodies. Seems every day I’m confronted by a patient who’s using a new drug with a name ending in “mab.” That drug blocks a cellular receptor I haven’t heard of that’s involved in a cascade of interactions I haven’t thought about since medical school. The resulting disruption reduces disease burden, typically at great expense to the medical system, the patient, or both. We’ve truly entered the era of precision medicine. It’s not enough to understand disease; you also must know its heterogeneous expression so that you can prescribe the ‘mab that targets the biology responsible for variants in behavior. All diseases are, in fact, syndromes. This isn’t a bad thing, but it’s a challenge.

A series of ‘mabs have been approved for treating type 2 high (TH2) or eosinophilic asthma. We refer to this group of ‘mabs generically as biologics. The group includes omalizumab, mepolizumab, dupilumab, benralizumab, reslizumab, and tezepelumab. While mechanism of action varies slightly across drugs, the biologics all target a specific arm of the immune system. Efficacy is linearly related to serum eosinophil count and there’s little clinically or pharmacologically to distinguish one from another. Of course, no head-to-head comparisons of efficacy are available and there’s no financial incentive for them to be performed.
 

Latest research

A new randomized controlled trial (RCT) of dupilumab for chronic obstructive pulmonary disease (COPD) adds to the aforementioned biologic knowledge base. Turns out it works as long as the patients are carefully selected. Researchers enrolled GOLD D (or E depending on which iteration of the GOLD Statement you use) patients on triple inhaler therapy (inhaled corticosteroids [ICS]/long-acting beta-agonist [LABA]/long-acting muscarinic antagonist [LAMA]) with two moderate exacerbations or one exacerbation requiring hospitalization in the past year. Blood eosinophil counts were > 300 cells/mcL and chronic bronchitis was present clinically. The primary and multiple secondary outcomes were improved with dupilumab.

This is welcome news. I’ve treated countless patients with severe COPD who have repeated exacerbations despite my efforts to prevent them. These patients are on ICS/LABA/LAMA and azithromycin or roflumilast, and occasionally both. While every COPD guideline known to man forbids using chronic oral corticosteroids (OCS), I’ve prescribed them repeatedly because the benefits to keeping a recalcitrant, exacerbating patient out of the hospital seem to outweigh OCS risks. It would be nice to have a better option. Although we were taught that they were immutably distinct in medical school, every first-year pulmonary fellow knows that asthma and COPD share more similarities than differences, so it makes sense that proven asthma therapies would work for some patients with COPD.

However, the dupilumab study must be placed in context. Past studies haven’t been as positive. In 2017, two separate RCTs found that mepolizumab reduced the annual rate of moderate to severe exacerbations (primary outcome) in one trial but not the other. Interpretation gets more complicated when broken down by intention to treat (ITT) vs. modified ITT and when secondary outcomes are considered. Sparing you those details, this trial does not instill confidence, leading the Food and Drug Administration to refuse approval for mepolizumab for COPD. A second RCT of benralizumab for COPD was published in 2019. Much less cognitive load was required to interpret this one; it was negative. FDA approval was not requested.

Looking through the trial designs for the three RCTs of biologics for COPD, I couldn’t find major differences that could explain the discordant results. Sample size and enrollment criteria were similar. As stated, I don’t believe that the biologic data in asthma allow for predicting efficacy in one eosinophilic patient vs. another and I assume the same would be true for COPD. All three trials found that eosinophils were eliminated, so responses were biologically equivalent.
 

Key takeaways

If trial design and pharmacology don’t account for the disparate outcomes, how do we explain them? More important, how do we translate these trials into clinical practice? I looked for a review or editorial by a scientist-clinician smarter than I so I could steal their ideas and express them as pedantic euphemisms here. I found it curious that I was unable to find one. A recent publication in the American Journal of Respiratory and Critical Care Medicine suggests that the answer lies within the complex lattice of eosinophil subtypes, but I’m unqualified to judge the veracity of this “phenotype within a phenotype” theory.

For now, there will be no biologics prescribed for COPD – at least not by me. More trials in COPD are being done. We should have results on tezepelumab, that great savior that may cover noneosinophilic asthma phenotypes, within the next few years. Until then, we’re stuck defying guidelines with the anachronistic use of OCS for the COPD patient who exacerbates through ICS/LABA/LAMA, roflumilast, and azithromycin.

Dr. Holley is professor of medicine at Uniformed Services University in Bethesda, Md., and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington. He reported receiving income from CHEST College, Metapharm, and WebMD.

A version of this article first appeared on Medscape.com.

In late 2021, the Rome Proposal for diagnosing acute exacerbations of chronic obstructive pulmonary disease (AECOPD) and grading their severity was published. The 2023 Global Strategy for the Diagnosis, Management, and Prevention of Chronic Obstructive Lung Disease (GOLD) Report has adopted the Rome Proposal criteria. Given that an endorsement by GOLD is tantamount to acceptance by clinicians, researchers, and policymakers alike, I guess we’re all using them now.

Anyone who’s ever cared for patients with COPD knows that treatment and reduction of exacerbations is how we improve outcomes. AECOPD are associated with considerable morbidity, greater health care utilization and costs, and a long-term decline in lung function. While we hope our pharmacotherapies improve symptoms, we know they reduce AECOPD. If our pharmacotherapies have any impact on mortality, it’s probably via AECOPD prevention.

Methods for reducing AECOPD are not controversial, but the approach to AECOPD treatment is, particularly decisions about who gets an antibiotic and who doesn’t. Since antibiotic indications are tied to severity, using the Rome Proposal criteria may affect management in unpredictable ways. As such, it’s worth reviewing the data on antibiotics for AECOPD.
 

What do the data reveal?

To start, it’s important to note that GOLD doesn’t equate having an AECOPD with needing an antibiotic. I myself have conflated the diagnosis with the indication and thereby overprescribed. The bar for diagnosis is quite low. In previous GOLD summaries, any “change in respiratory symptoms” would warrant the AECOPD label. Although the Rome Proposal definition is more specific, it leaves room for liberal interpretation. It’s likely to have a greater effect on research than on clinical practice. My guess is that AECOPD prevalence doesn’t change.

Holley_Aaron_B_BETHESDA_web.jpg

The antibiotic hurdle is slightly higher than that for diagnosis but is equally open to interpretation. In part, that’s related to the inherent subjectivity of judging symptoms, sputum production, and changes in color, but it’s also because the data are so poor. The meta-analyses that have been used to establish the indications include fewer than 1000 patients spread across 10 to 11 trials. Thus, the individual trials are small, and the sample size remains nominal even after adding them together. The addition of antibiotics – and it doesn’t seem to matter which class, type, or duration – will decrease mortality and hospital length of stay. One study says these effects are limited to inpatients while the other does not. After reading GOLD 2013, GOLD 2023, and both the meta-analyses they used to support their recommendations, I’m still not sure who benefits. Do you have to be hospitalized? Is some sort of ventilatory support required? Does C-reactive protein help or not?

In accordance with the classic Anthonisen criteria, GOLD relies on sputum volume and color as evidence of a bacterial infection. Soon after GOLD 2023 was published, a meta-analysis found that sputum color isn’t particularly accurate for detecting bacterial infection. Because it doesn’t seem to matter which antibiotic class is used, I always thought we were using antibiotics for their magical, pleiotropic anti-inflammatory effects anyway. I didn’t think the presence of an actual bacterial infection was important. If I saw an infiltrate on chest x-ray, I’d change my diagnosis from AECOPD to community-acquired pneumonia (CAP) and switch to CAP coverage. I’ve been doing this so long that I swear it’s in a guideline somewhere, though admittedly I couldn’t find said guideline while reading for this piece.
 

Key takeaways

In summary, I believe that the guidance reflects the data, which is muddy. The Rome Proposal should be seen as just that – a framework for moving forward with AECOPD classification and antibiotic indications that will need to be refined over time as better data become available. In fact, they allow for a more objective, point-of-care assessment of severity that can be validated and tied to antibiotic benefits. The Rome criteria aren’t evidence-based; they’re a necessary first step toward creating the evidence.

In the meantime, if your AECOPD patients are hospitalized, they probably warrant an antibiotic. If they’re not, sputum changes may be a reasonable surrogate for a bacterial infection. Considerable uncertainty remains.

Aaron B. Holley, MD, is a professor of medicine at Uniformed Services University in Bethesda, Md., and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington. He reported conflicts of interest with Metapharm, CHEST College, and WebMD.

A version of this article first appeared on Medscape.com.

In late 2021, the Rome Proposal for diagnosing acute exacerbations of chronic obstructive pulmonary disease (AECOPD) and grading their severity was published. The 2023 Global Strategy for the Diagnosis, Management, and Prevention of Chronic Obstructive Lung Disease (GOLD) Report has adopted the Rome Proposal criteria. Given that an endorsement by GOLD is tantamount to acceptance by clinicians, researchers, and policymakers alike, I guess we’re all using them now.

Anyone who’s ever cared for patients with COPD knows that treatment and reduction of exacerbations is how we improve outcomes. AECOPD are associated with considerable morbidity, greater health care utilization and costs, and a long-term decline in lung function. While we hope our pharmacotherapies improve symptoms, we know they reduce AECOPD. If our pharmacotherapies have any impact on mortality, it’s probably via AECOPD prevention.

Methods for reducing AECOPD are not controversial, but the approach to AECOPD treatment is, particularly decisions about who gets an antibiotic and who doesn’t. Since antibiotic indications are tied to severity, using the Rome Proposal criteria may affect management in unpredictable ways. As such, it’s worth reviewing the data on antibiotics for AECOPD.
 

What do the data reveal?

To start, it’s important to note that GOLD doesn’t equate having an AECOPD with needing an antibiotic. I myself have conflated the diagnosis with the indication and thereby overprescribed. The bar for diagnosis is quite low. In previous GOLD summaries, any “change in respiratory symptoms” would warrant the AECOPD label. Although the Rome Proposal definition is more specific, it leaves room for liberal interpretation. It’s likely to have a greater effect on research than on clinical practice. My guess is that AECOPD prevalence doesn’t change.

Holley_Aaron_B_BETHESDA_web.jpg

The antibiotic hurdle is slightly higher than that for diagnosis but is equally open to interpretation. In part, that’s related to the inherent subjectivity of judging symptoms, sputum production, and changes in color, but it’s also because the data are so poor. The meta-analyses that have been used to establish the indications include fewer than 1000 patients spread across 10 to 11 trials. Thus, the individual trials are small, and the sample size remains nominal even after adding them together. The addition of antibiotics – and it doesn’t seem to matter which class, type, or duration – will decrease mortality and hospital length of stay. One study says these effects are limited to inpatients while the other does not. After reading GOLD 2013, GOLD 2023, and both the meta-analyses they used to support their recommendations, I’m still not sure who benefits. Do you have to be hospitalized? Is some sort of ventilatory support required? Does C-reactive protein help or not?

In accordance with the classic Anthonisen criteria, GOLD relies on sputum volume and color as evidence of a bacterial infection. Soon after GOLD 2023 was published, a meta-analysis found that sputum color isn’t particularly accurate for detecting bacterial infection. Because it doesn’t seem to matter which antibiotic class is used, I always thought we were using antibiotics for their magical, pleiotropic anti-inflammatory effects anyway. I didn’t think the presence of an actual bacterial infection was important. If I saw an infiltrate on chest x-ray, I’d change my diagnosis from AECOPD to community-acquired pneumonia (CAP) and switch to CAP coverage. I’ve been doing this so long that I swear it’s in a guideline somewhere, though admittedly I couldn’t find said guideline while reading for this piece.
 

Key takeaways

In summary, I believe that the guidance reflects the data, which is muddy. The Rome Proposal should be seen as just that – a framework for moving forward with AECOPD classification and antibiotic indications that will need to be refined over time as better data become available. In fact, they allow for a more objective, point-of-care assessment of severity that can be validated and tied to antibiotic benefits. The Rome criteria aren’t evidence-based; they’re a necessary first step toward creating the evidence.

In the meantime, if your AECOPD patients are hospitalized, they probably warrant an antibiotic. If they’re not, sputum changes may be a reasonable surrogate for a bacterial infection. Considerable uncertainty remains.

Aaron B. Holley, MD, is a professor of medicine at Uniformed Services University in Bethesda, Md., and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington. He reported conflicts of interest with Metapharm, CHEST College, and WebMD.

A version of this article first appeared on Medscape.com.

In late 2021, the Rome Proposal for diagnosing acute exacerbations of chronic obstructive pulmonary disease (AECOPD) and grading their severity was published. The 2023 Global Strategy for the Diagnosis, Management, and Prevention of Chronic Obstructive Lung Disease (GOLD) Report has adopted the Rome Proposal criteria. Given that an endorsement by GOLD is tantamount to acceptance by clinicians, researchers, and policymakers alike, I guess we’re all using them now.

Anyone who’s ever cared for patients with COPD knows that treatment and reduction of exacerbations is how we improve outcomes. AECOPD are associated with considerable morbidity, greater health care utilization and costs, and a long-term decline in lung function. While we hope our pharmacotherapies improve symptoms, we know they reduce AECOPD. If our pharmacotherapies have any impact on mortality, it’s probably via AECOPD prevention.

Methods for reducing AECOPD are not controversial, but the approach to AECOPD treatment is, particularly decisions about who gets an antibiotic and who doesn’t. Since antibiotic indications are tied to severity, using the Rome Proposal criteria may affect management in unpredictable ways. As such, it’s worth reviewing the data on antibiotics for AECOPD.
 

What do the data reveal?

To start, it’s important to note that GOLD doesn’t equate having an AECOPD with needing an antibiotic. I myself have conflated the diagnosis with the indication and thereby overprescribed. The bar for diagnosis is quite low. In previous GOLD summaries, any “change in respiratory symptoms” would warrant the AECOPD label. Although the Rome Proposal definition is more specific, it leaves room for liberal interpretation. It’s likely to have a greater effect on research than on clinical practice. My guess is that AECOPD prevalence doesn’t change.

Holley_Aaron_B_BETHESDA_web.jpg

The antibiotic hurdle is slightly higher than that for diagnosis but is equally open to interpretation. In part, that’s related to the inherent subjectivity of judging symptoms, sputum production, and changes in color, but it’s also because the data are so poor. The meta-analyses that have been used to establish the indications include fewer than 1000 patients spread across 10 to 11 trials. Thus, the individual trials are small, and the sample size remains nominal even after adding them together. The addition of antibiotics – and it doesn’t seem to matter which class, type, or duration – will decrease mortality and hospital length of stay. One study says these effects are limited to inpatients while the other does not. After reading GOLD 2013, GOLD 2023, and both the meta-analyses they used to support their recommendations, I’m still not sure who benefits. Do you have to be hospitalized? Is some sort of ventilatory support required? Does C-reactive protein help or not?

In accordance with the classic Anthonisen criteria, GOLD relies on sputum volume and color as evidence of a bacterial infection. Soon after GOLD 2023 was published, a meta-analysis found that sputum color isn’t particularly accurate for detecting bacterial infection. Because it doesn’t seem to matter which antibiotic class is used, I always thought we were using antibiotics for their magical, pleiotropic anti-inflammatory effects anyway. I didn’t think the presence of an actual bacterial infection was important. If I saw an infiltrate on chest x-ray, I’d change my diagnosis from AECOPD to community-acquired pneumonia (CAP) and switch to CAP coverage. I’ve been doing this so long that I swear it’s in a guideline somewhere, though admittedly I couldn’t find said guideline while reading for this piece.
 

Key takeaways

In summary, I believe that the guidance reflects the data, which is muddy. The Rome Proposal should be seen as just that – a framework for moving forward with AECOPD classification and antibiotic indications that will need to be refined over time as better data become available. In fact, they allow for a more objective, point-of-care assessment of severity that can be validated and tied to antibiotic benefits. The Rome criteria aren’t evidence-based; they’re a necessary first step toward creating the evidence.

In the meantime, if your AECOPD patients are hospitalized, they probably warrant an antibiotic. If they’re not, sputum changes may be a reasonable surrogate for a bacterial infection. Considerable uncertainty remains.

Aaron B. Holley, MD, is a professor of medicine at Uniformed Services University in Bethesda, Md., and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington. He reported conflicts of interest with Metapharm, CHEST College, and WebMD.

A version of this article first appeared on Medscape.com.

If you’re looking for it, you’ll find fatigue almost everywhere. It’s so common that it hides in plain sight, never dealt with because it’s present for good reason: the inevitable consequence of age, whatever disease you’re treating, poor lifestyle choices, and the daily grind of twenty-first–century life. Its impact is so ubiquitous and pernicious that it’s considered acceptable.

Is it though? After all, fatigue can be debilitating. Not every symptom is worthy of a chronic syndrome bearing its name. Furthermore, what if its relationship to the disease you’re treating is bidirectional? What if we actually paid attention, asked about it, and expended energy trying to relieve it? Could we improve quality of life and other outcomes too?

Outside of sleep medicine, I see little focus on fatigue among pulmonologists. This despite the existing data on fatigue related to sarcoidosis, chronic obstructive pulmonary disease (COPD), and interstitial lung disease. Even when we do pay it lip service, “addressing” fatigue or sleep is essentially a euphemism for ordering a sleep study.

As with fatigue, if you look for obstructive sleep apnea, it’ll be there, although with OSA, it’s related to the incredibly low, nonevidence-based threshold the American Academy of Sleep Medicine has established for making the diagnosis. With continuous positive airway pressure (CPAP) in hand, the patient has a new disease to worry about and a difficult behavioral change (wearing, cleaning, and resupplying their CPAP equipment) to make. Too often, the CPAP isn’t used – or is – and the fatigue persists. But it’s okay, because we followed somebody’s guideline.

The American Thoracic Society just published a research statement on cancer-related fatigue. It is comprehensive and highlights the high prevalence and poor recognition of cancer-related fatigue. The authors note that among cancers, those of the lung are associated with a higher comorbid disease burden, older age, and cigarette smoking. All these factors make patients with lung cancer particularly prone to fatigue. Interactions between these factors, lung cancer histology, and specific chemotherapy regimens are poorly understood. True to its title, the “research statement” serves more as a call to action than an evidence-based blueprint for diagnosis and management.

The cancer-related fatigue data that does exist suggests treatment starts with recognition followed by a focus on sleep, exercise, and nutrition. This should surprise no one. The data on fatigue in general (not specific to cancer-related fatigue) shows that although fatigue is not synonymous with poor quality or insufficient sleep, sleep is usually a major factor. The cancer-related conditions affecting sleep include anxiety, depression, insufficient sleep, insomnia, medication side effects, and OSA. The intersecting web is complex, but across underlying conditions (cancer or otherwise), the quickest most efficient method for mitigating fatigue is optimizing sleep.

Exercise and nutrition are also important. Again, across disease processes (interstitial lung disease, COPD, lung cancer, and so on), no drug comes close to aerobic exercise for reducing symptoms, including fatigue. If an exercise prescription could be delivered in pill-form, it’d be a blockbuster. But it can’t be, and the ATS lung cancer–related fatigue research statement nicely outlines the evidence for increased activity levels and the barriers to obtaining support and compliance. As is the case with exercise, support for improving nutrition is limited by cost, access, and patient education.

Perhaps most importantly, sleep, exercise, and nutrition require time for counseling and a behavior change for the physician and patient. Both are in short supply, and commitment is always ephemeral. Incentivization could perhaps be re-structured, but the ATS document notes this will be challenging. With respect to pulmonary rehabilitation (about 50% of patients with lung cancer have comorbid COPD), for example, reimbursement is poor, which serves as a disincentive. Their suggestions? Early integration and repeated introduction to rehabilitation and exercise concepts. Sounds great.

In summary, in my opinion, fatigue doesn’t receive the attention level commensurate with its impact. It’s easy to understand why, but I’m glad the ATS is highlighting the problem. Unbeknownst to me, multiple cancer guidelines already recommend screening for fatigue. The recent sarcoidosis treatment guideline published by the European Respiratory Society dedicated a PICO (Patients, Intervention, Comparison, Outcomes) to the topic and recommended exercise (pulmonary rehabilitation). That said, consensus statements on COPD mention it only in passing in relation to severe disease and end-of-life care, and idiopathic pulmonary fibrosis guidelines ignore it entirely. So, recognition is improving, but we’ve got ways to go.

139578_Holley_Aaron_web.jpg

Dr. Holley is professor of medicine at Uniformed Services University, Bethesda, Md., and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington. He disclosed ties with Metapharm, CHEST College, and WebMD.

A version of this article originally appeared on Medscape.com.

If you’re looking for it, you’ll find fatigue almost everywhere. It’s so common that it hides in plain sight, never dealt with because it’s present for good reason: the inevitable consequence of age, whatever disease you’re treating, poor lifestyle choices, and the daily grind of twenty-first–century life. Its impact is so ubiquitous and pernicious that it’s considered acceptable.

Is it though? After all, fatigue can be debilitating. Not every symptom is worthy of a chronic syndrome bearing its name. Furthermore, what if its relationship to the disease you’re treating is bidirectional? What if we actually paid attention, asked about it, and expended energy trying to relieve it? Could we improve quality of life and other outcomes too?

Outside of sleep medicine, I see little focus on fatigue among pulmonologists. This despite the existing data on fatigue related to sarcoidosis, chronic obstructive pulmonary disease (COPD), and interstitial lung disease. Even when we do pay it lip service, “addressing” fatigue or sleep is essentially a euphemism for ordering a sleep study.

As with fatigue, if you look for obstructive sleep apnea, it’ll be there, although with OSA, it’s related to the incredibly low, nonevidence-based threshold the American Academy of Sleep Medicine has established for making the diagnosis. With continuous positive airway pressure (CPAP) in hand, the patient has a new disease to worry about and a difficult behavioral change (wearing, cleaning, and resupplying their CPAP equipment) to make. Too often, the CPAP isn’t used – or is – and the fatigue persists. But it’s okay, because we followed somebody’s guideline.

The American Thoracic Society just published a research statement on cancer-related fatigue. It is comprehensive and highlights the high prevalence and poor recognition of cancer-related fatigue. The authors note that among cancers, those of the lung are associated with a higher comorbid disease burden, older age, and cigarette smoking. All these factors make patients with lung cancer particularly prone to fatigue. Interactions between these factors, lung cancer histology, and specific chemotherapy regimens are poorly understood. True to its title, the “research statement” serves more as a call to action than an evidence-based blueprint for diagnosis and management.

The cancer-related fatigue data that does exist suggests treatment starts with recognition followed by a focus on sleep, exercise, and nutrition. This should surprise no one. The data on fatigue in general (not specific to cancer-related fatigue) shows that although fatigue is not synonymous with poor quality or insufficient sleep, sleep is usually a major factor. The cancer-related conditions affecting sleep include anxiety, depression, insufficient sleep, insomnia, medication side effects, and OSA. The intersecting web is complex, but across underlying conditions (cancer or otherwise), the quickest most efficient method for mitigating fatigue is optimizing sleep.

Exercise and nutrition are also important. Again, across disease processes (interstitial lung disease, COPD, lung cancer, and so on), no drug comes close to aerobic exercise for reducing symptoms, including fatigue. If an exercise prescription could be delivered in pill-form, it’d be a blockbuster. But it can’t be, and the ATS lung cancer–related fatigue research statement nicely outlines the evidence for increased activity levels and the barriers to obtaining support and compliance. As is the case with exercise, support for improving nutrition is limited by cost, access, and patient education.

Perhaps most importantly, sleep, exercise, and nutrition require time for counseling and a behavior change for the physician and patient. Both are in short supply, and commitment is always ephemeral. Incentivization could perhaps be re-structured, but the ATS document notes this will be challenging. With respect to pulmonary rehabilitation (about 50% of patients with lung cancer have comorbid COPD), for example, reimbursement is poor, which serves as a disincentive. Their suggestions? Early integration and repeated introduction to rehabilitation and exercise concepts. Sounds great.

In summary, in my opinion, fatigue doesn’t receive the attention level commensurate with its impact. It’s easy to understand why, but I’m glad the ATS is highlighting the problem. Unbeknownst to me, multiple cancer guidelines already recommend screening for fatigue. The recent sarcoidosis treatment guideline published by the European Respiratory Society dedicated a PICO (Patients, Intervention, Comparison, Outcomes) to the topic and recommended exercise (pulmonary rehabilitation). That said, consensus statements on COPD mention it only in passing in relation to severe disease and end-of-life care, and idiopathic pulmonary fibrosis guidelines ignore it entirely. So, recognition is improving, but we’ve got ways to go.

139578_Holley_Aaron_web.jpg

Dr. Holley is professor of medicine at Uniformed Services University, Bethesda, Md., and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington. He disclosed ties with Metapharm, CHEST College, and WebMD.

A version of this article originally appeared on Medscape.com.

If you’re looking for it, you’ll find fatigue almost everywhere. It’s so common that it hides in plain sight, never dealt with because it’s present for good reason: the inevitable consequence of age, whatever disease you’re treating, poor lifestyle choices, and the daily grind of twenty-first–century life. Its impact is so ubiquitous and pernicious that it’s considered acceptable.

Is it though? After all, fatigue can be debilitating. Not every symptom is worthy of a chronic syndrome bearing its name. Furthermore, what if its relationship to the disease you’re treating is bidirectional? What if we actually paid attention, asked about it, and expended energy trying to relieve it? Could we improve quality of life and other outcomes too?

Outside of sleep medicine, I see little focus on fatigue among pulmonologists. This despite the existing data on fatigue related to sarcoidosis, chronic obstructive pulmonary disease (COPD), and interstitial lung disease. Even when we do pay it lip service, “addressing” fatigue or sleep is essentially a euphemism for ordering a sleep study.

As with fatigue, if you look for obstructive sleep apnea, it’ll be there, although with OSA, it’s related to the incredibly low, nonevidence-based threshold the American Academy of Sleep Medicine has established for making the diagnosis. With continuous positive airway pressure (CPAP) in hand, the patient has a new disease to worry about and a difficult behavioral change (wearing, cleaning, and resupplying their CPAP equipment) to make. Too often, the CPAP isn’t used – or is – and the fatigue persists. But it’s okay, because we followed somebody’s guideline.

The American Thoracic Society just published a research statement on cancer-related fatigue. It is comprehensive and highlights the high prevalence and poor recognition of cancer-related fatigue. The authors note that among cancers, those of the lung are associated with a higher comorbid disease burden, older age, and cigarette smoking. All these factors make patients with lung cancer particularly prone to fatigue. Interactions between these factors, lung cancer histology, and specific chemotherapy regimens are poorly understood. True to its title, the “research statement” serves more as a call to action than an evidence-based blueprint for diagnosis and management.

The cancer-related fatigue data that does exist suggests treatment starts with recognition followed by a focus on sleep, exercise, and nutrition. This should surprise no one. The data on fatigue in general (not specific to cancer-related fatigue) shows that although fatigue is not synonymous with poor quality or insufficient sleep, sleep is usually a major factor. The cancer-related conditions affecting sleep include anxiety, depression, insufficient sleep, insomnia, medication side effects, and OSA. The intersecting web is complex, but across underlying conditions (cancer or otherwise), the quickest most efficient method for mitigating fatigue is optimizing sleep.

Exercise and nutrition are also important. Again, across disease processes (interstitial lung disease, COPD, lung cancer, and so on), no drug comes close to aerobic exercise for reducing symptoms, including fatigue. If an exercise prescription could be delivered in pill-form, it’d be a blockbuster. But it can’t be, and the ATS lung cancer–related fatigue research statement nicely outlines the evidence for increased activity levels and the barriers to obtaining support and compliance. As is the case with exercise, support for improving nutrition is limited by cost, access, and patient education.

Perhaps most importantly, sleep, exercise, and nutrition require time for counseling and a behavior change for the physician and patient. Both are in short supply, and commitment is always ephemeral. Incentivization could perhaps be re-structured, but the ATS document notes this will be challenging. With respect to pulmonary rehabilitation (about 50% of patients with lung cancer have comorbid COPD), for example, reimbursement is poor, which serves as a disincentive. Their suggestions? Early integration and repeated introduction to rehabilitation and exercise concepts. Sounds great.

In summary, in my opinion, fatigue doesn’t receive the attention level commensurate with its impact. It’s easy to understand why, but I’m glad the ATS is highlighting the problem. Unbeknownst to me, multiple cancer guidelines already recommend screening for fatigue. The recent sarcoidosis treatment guideline published by the European Respiratory Society dedicated a PICO (Patients, Intervention, Comparison, Outcomes) to the topic and recommended exercise (pulmonary rehabilitation). That said, consensus statements on COPD mention it only in passing in relation to severe disease and end-of-life care, and idiopathic pulmonary fibrosis guidelines ignore it entirely. So, recognition is improving, but we’ve got ways to go.

139578_Holley_Aaron_web.jpg

Dr. Holley is professor of medicine at Uniformed Services University, Bethesda, Md., and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington. He disclosed ties with Metapharm, CHEST College, and WebMD.

A version of this article originally appeared on Medscape.com.

Clinicians working within the U.S. health care system order CTs; it’s just what we do, and we do it a lot. This isn’t necessarily bad, but an inevitable byproduct is the pandemic of incidental findings. One underrecognized but frequent “incidentaloma” on CT is an interstitial lung abnormality (ILA). The Fleischner Society defines an ILA as honeycombing, traction bronchiectasis, parenchymal distortions, and reticular abnormalities that take up more than 5% of a particular lung zone in a patient without a clinical diagnosis of interstitial lung disease (ILD). In essence, ILAs are both a radiographic and a clinical diagnosis.

ILAs are common. Depending on population characteristics, ILAs occur at a prevalence of up to 10%. With the advent of lung cancer screening and advances in CT technology, we’re now inundated with detailed images of lung parenchyma in older smokers who are at high risk for respiratory disease. The resulting opportunity for early identification of disease is as exciting as the risk for overdiagnosis, excessive testing, and unnecessary treatment is frightening. Early diagnosis remains critical for preventing irreversible respiratory disease. But as with any disease process, when we attempt to detect pathology before it has become apparent, the line between benign change and true abnormality is blurred.

Such is the challenge with ILAs. Past studies have shown an association between ILAs and morbidity and mortality, but considerable uncertainty persists over what the ILAs represent and how they should be managed. A recent study published in the American Journal of Respiratory and Critical Care Medicine  provides some clarity. The authors used data from the COPDGene cohort to correlate ILAs with lung testing, and functional and respiratory outcomes. As with other studies, they found that approximately 10% of the COPDGene patients that they examined had ILAs on CT and half of those met their criteria for “suspected ILD.” Suspected ILD was defined radiographically (definite fibrosis) and on lung function testing (abnormal forced vital capacity [FVC] or diffusing capacity of the lungs for carbon monoxide [DLCO]). The patients with suspected ILD had worse clinical outcomes; being a Black individual, pack-years of smoking, and GOLD stage on spirometry were independently associated with suspected ILD.

This type of study is urgently needed. Given their high prevalence, we’re in dire need of a valid model for risk stratifying ILAs. The authors of this study have moved us closer, but we’ve still got a long way to go. The study has significant limitations. First, although patients with previous documentation of ILD were excluded from COPDGene, no formal, multidisciplinary assessment was performed; therefore, some of the patients labeled as having ILA probably had diagnosable ILD. Their possible inclusion would falsely increase the prevalence of clinically important ILAs and exaggerate the relationship between ILAs and clinical outcomes.

The rhetorical gymnastics performed throughout the paper are necessary yet problematic. “Suspected ILD” is not a recognized diagnosis and the definition is therefore arbitrary. To the extent that “suspected ILD” requires an abnormality on spirometry or DLCO, one could argue it’s the lung function changes and not the radiographic findings that are driving the differences. In fact, “suspected ILD” was defined by lung function more often than radiographic criteria (16% had definite fibrosis on CT, 57% had an abnormal FVC, and 67% had an abnormal DLCO). Patients with ILAs without suspected ILD had outcomes that weren’t statistically different from those with no ILAs at all, implying that the lung testing and not the ILA is the better discriminator. Regardless, this leads us back to where we started before this paper was published: ILAs require lung function testing and referral to a pulmonologist for proper risk stratification. An accompanying editorial highlights these and other limitations.

One particular problem that isn’t addressed by the authors or the editorial is their findings on race. The authors concluded that Black persons with ILAs are more likely to have “suspected ILD.” However, their definition suffers from an insidious form of incorporation bias generated by the way they handled their DLCO reference values. The Global Lung Function Initiative equations they used were derived exclusively from White persons. In accordance with the recent American Thoracic Society/European Respiratory Society (ATS/ERS) statement on lung testing, the authors did not apply a fixed correction factor to adjust for race. Without such an adjustment, Black persons would be biased toward having lower percent predicted values for DLCO. In short, self-identified Black individuals would be more likely to have a predicted DLCO of less than 70% and to therefore meet criteria for “suspected ILD.” The resulting effects on biologic plausibility, causal inference, and the strength of the relationship between “suspected ILD” and clinical outcomes will vary by whether the association between race and lung function is considered a product of inherent biologic variability or a result of external (socioeconomic and environmental) effects.

In summary, ILAs remain a challenge for radiologists, primary care providers, pulmonologists, and anyone else who orders a CT of the lungs. Despite its limitations, I believe the recently published paper pushes us forward conceptually. Perhaps its most important contribution is showing that 50% of ILAs are clinically insignificant by definition. This offers further reassurance that a subset of ILAs can be dismissed. Now, all we need is an easy, cost-effective, and efficient way to identify this subset.
 

Dr. Holley is professor of medicine at Uniformed Services University in Bethesda, Md., and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington. He covers a wide range of topics in pulmonary, critical care, and sleep medicine. He disclosed ties to Metapharm Inc., CHEST College, and WebMD. A version of this article originally appeared on Medscape.com.

Clinicians working within the U.S. health care system order CTs; it’s just what we do, and we do it a lot. This isn’t necessarily bad, but an inevitable byproduct is the pandemic of incidental findings. One underrecognized but frequent “incidentaloma” on CT is an interstitial lung abnormality (ILA). The Fleischner Society defines an ILA as honeycombing, traction bronchiectasis, parenchymal distortions, and reticular abnormalities that take up more than 5% of a particular lung zone in a patient without a clinical diagnosis of interstitial lung disease (ILD). In essence, ILAs are both a radiographic and a clinical diagnosis.

ILAs are common. Depending on population characteristics, ILAs occur at a prevalence of up to 10%. With the advent of lung cancer screening and advances in CT technology, we’re now inundated with detailed images of lung parenchyma in older smokers who are at high risk for respiratory disease. The resulting opportunity for early identification of disease is as exciting as the risk for overdiagnosis, excessive testing, and unnecessary treatment is frightening. Early diagnosis remains critical for preventing irreversible respiratory disease. But as with any disease process, when we attempt to detect pathology before it has become apparent, the line between benign change and true abnormality is blurred.

Such is the challenge with ILAs. Past studies have shown an association between ILAs and morbidity and mortality, but considerable uncertainty persists over what the ILAs represent and how they should be managed. A recent study published in the American Journal of Respiratory and Critical Care Medicine  provides some clarity. The authors used data from the COPDGene cohort to correlate ILAs with lung testing, and functional and respiratory outcomes. As with other studies, they found that approximately 10% of the COPDGene patients that they examined had ILAs on CT and half of those met their criteria for “suspected ILD.” Suspected ILD was defined radiographically (definite fibrosis) and on lung function testing (abnormal forced vital capacity [FVC] or diffusing capacity of the lungs for carbon monoxide [DLCO]). The patients with suspected ILD had worse clinical outcomes; being a Black individual, pack-years of smoking, and GOLD stage on spirometry were independently associated with suspected ILD.

This type of study is urgently needed. Given their high prevalence, we’re in dire need of a valid model for risk stratifying ILAs. The authors of this study have moved us closer, but we’ve still got a long way to go. The study has significant limitations. First, although patients with previous documentation of ILD were excluded from COPDGene, no formal, multidisciplinary assessment was performed; therefore, some of the patients labeled as having ILA probably had diagnosable ILD. Their possible inclusion would falsely increase the prevalence of clinically important ILAs and exaggerate the relationship between ILAs and clinical outcomes.

The rhetorical gymnastics performed throughout the paper are necessary yet problematic. “Suspected ILD” is not a recognized diagnosis and the definition is therefore arbitrary. To the extent that “suspected ILD” requires an abnormality on spirometry or DLCO, one could argue it’s the lung function changes and not the radiographic findings that are driving the differences. In fact, “suspected ILD” was defined by lung function more often than radiographic criteria (16% had definite fibrosis on CT, 57% had an abnormal FVC, and 67% had an abnormal DLCO). Patients with ILAs without suspected ILD had outcomes that weren’t statistically different from those with no ILAs at all, implying that the lung testing and not the ILA is the better discriminator. Regardless, this leads us back to where we started before this paper was published: ILAs require lung function testing and referral to a pulmonologist for proper risk stratification. An accompanying editorial highlights these and other limitations.

One particular problem that isn’t addressed by the authors or the editorial is their findings on race. The authors concluded that Black persons with ILAs are more likely to have “suspected ILD.” However, their definition suffers from an insidious form of incorporation bias generated by the way they handled their DLCO reference values. The Global Lung Function Initiative equations they used were derived exclusively from White persons. In accordance with the recent American Thoracic Society/European Respiratory Society (ATS/ERS) statement on lung testing, the authors did not apply a fixed correction factor to adjust for race. Without such an adjustment, Black persons would be biased toward having lower percent predicted values for DLCO. In short, self-identified Black individuals would be more likely to have a predicted DLCO of less than 70% and to therefore meet criteria for “suspected ILD.” The resulting effects on biologic plausibility, causal inference, and the strength of the relationship between “suspected ILD” and clinical outcomes will vary by whether the association between race and lung function is considered a product of inherent biologic variability or a result of external (socioeconomic and environmental) effects.

In summary, ILAs remain a challenge for radiologists, primary care providers, pulmonologists, and anyone else who orders a CT of the lungs. Despite its limitations, I believe the recently published paper pushes us forward conceptually. Perhaps its most important contribution is showing that 50% of ILAs are clinically insignificant by definition. This offers further reassurance that a subset of ILAs can be dismissed. Now, all we need is an easy, cost-effective, and efficient way to identify this subset.
 

Dr. Holley is professor of medicine at Uniformed Services University in Bethesda, Md., and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington. He covers a wide range of topics in pulmonary, critical care, and sleep medicine. He disclosed ties to Metapharm Inc., CHEST College, and WebMD. A version of this article originally appeared on Medscape.com.

Clinicians working within the U.S. health care system order CTs; it’s just what we do, and we do it a lot. This isn’t necessarily bad, but an inevitable byproduct is the pandemic of incidental findings. One underrecognized but frequent “incidentaloma” on CT is an interstitial lung abnormality (ILA). The Fleischner Society defines an ILA as honeycombing, traction bronchiectasis, parenchymal distortions, and reticular abnormalities that take up more than 5% of a particular lung zone in a patient without a clinical diagnosis of interstitial lung disease (ILD). In essence, ILAs are both a radiographic and a clinical diagnosis.

ILAs are common. Depending on population characteristics, ILAs occur at a prevalence of up to 10%. With the advent of lung cancer screening and advances in CT technology, we’re now inundated with detailed images of lung parenchyma in older smokers who are at high risk for respiratory disease. The resulting opportunity for early identification of disease is as exciting as the risk for overdiagnosis, excessive testing, and unnecessary treatment is frightening. Early diagnosis remains critical for preventing irreversible respiratory disease. But as with any disease process, when we attempt to detect pathology before it has become apparent, the line between benign change and true abnormality is blurred.

Such is the challenge with ILAs. Past studies have shown an association between ILAs and morbidity and mortality, but considerable uncertainty persists over what the ILAs represent and how they should be managed. A recent study published in the American Journal of Respiratory and Critical Care Medicine  provides some clarity. The authors used data from the COPDGene cohort to correlate ILAs with lung testing, and functional and respiratory outcomes. As with other studies, they found that approximately 10% of the COPDGene patients that they examined had ILAs on CT and half of those met their criteria for “suspected ILD.” Suspected ILD was defined radiographically (definite fibrosis) and on lung function testing (abnormal forced vital capacity [FVC] or diffusing capacity of the lungs for carbon monoxide [DLCO]). The patients with suspected ILD had worse clinical outcomes; being a Black individual, pack-years of smoking, and GOLD stage on spirometry were independently associated with suspected ILD.

This type of study is urgently needed. Given their high prevalence, we’re in dire need of a valid model for risk stratifying ILAs. The authors of this study have moved us closer, but we’ve still got a long way to go. The study has significant limitations. First, although patients with previous documentation of ILD were excluded from COPDGene, no formal, multidisciplinary assessment was performed; therefore, some of the patients labeled as having ILA probably had diagnosable ILD. Their possible inclusion would falsely increase the prevalence of clinically important ILAs and exaggerate the relationship between ILAs and clinical outcomes.

The rhetorical gymnastics performed throughout the paper are necessary yet problematic. “Suspected ILD” is not a recognized diagnosis and the definition is therefore arbitrary. To the extent that “suspected ILD” requires an abnormality on spirometry or DLCO, one could argue it’s the lung function changes and not the radiographic findings that are driving the differences. In fact, “suspected ILD” was defined by lung function more often than radiographic criteria (16% had definite fibrosis on CT, 57% had an abnormal FVC, and 67% had an abnormal DLCO). Patients with ILAs without suspected ILD had outcomes that weren’t statistically different from those with no ILAs at all, implying that the lung testing and not the ILA is the better discriminator. Regardless, this leads us back to where we started before this paper was published: ILAs require lung function testing and referral to a pulmonologist for proper risk stratification. An accompanying editorial highlights these and other limitations.

One particular problem that isn’t addressed by the authors or the editorial is their findings on race. The authors concluded that Black persons with ILAs are more likely to have “suspected ILD.” However, their definition suffers from an insidious form of incorporation bias generated by the way they handled their DLCO reference values. The Global Lung Function Initiative equations they used were derived exclusively from White persons. In accordance with the recent American Thoracic Society/European Respiratory Society (ATS/ERS) statement on lung testing, the authors did not apply a fixed correction factor to adjust for race. Without such an adjustment, Black persons would be biased toward having lower percent predicted values for DLCO. In short, self-identified Black individuals would be more likely to have a predicted DLCO of less than 70% and to therefore meet criteria for “suspected ILD.” The resulting effects on biologic plausibility, causal inference, and the strength of the relationship between “suspected ILD” and clinical outcomes will vary by whether the association between race and lung function is considered a product of inherent biologic variability or a result of external (socioeconomic and environmental) effects.

In summary, ILAs remain a challenge for radiologists, primary care providers, pulmonologists, and anyone else who orders a CT of the lungs. Despite its limitations, I believe the recently published paper pushes us forward conceptually. Perhaps its most important contribution is showing that 50% of ILAs are clinically insignificant by definition. This offers further reassurance that a subset of ILAs can be dismissed. Now, all we need is an easy, cost-effective, and efficient way to identify this subset.
 

Dr. Holley is professor of medicine at Uniformed Services University in Bethesda, Md., and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington. He covers a wide range of topics in pulmonary, critical care, and sleep medicine. He disclosed ties to Metapharm Inc., CHEST College, and WebMD. A version of this article originally appeared on Medscape.com.