American Thoracic Society (ATS): International Conference 2016

SAN FRANCISCO – The opioid overdose crisis in the United States is now plainly evident in intensive care units (ICUs), finds a study of hospitals in 44 states conducted between 2009 and 2015.

During the study period, ICU admissions for opioid overdoses increased by almost half, investigators reported in a session and related press briefing an international conference of the American Thoracic Society. Furthermore, ICU deaths from this cause roughly doubled.

Susan London/Frontline Medical News

“This means the opioid use epidemic has probably reached a new level of crisis,” said lead investigator Jennifer P. Stevens, MD, an instructor in medicine at Harvard Medical School, and an adult intensive care physician at Beth Israel Deaconess Medical Center, both in Boston. “And this means that in spite of everything that we can do in the ICU – keeping them alive on ventilators, doing life support, doing acute dialysis, doing round-the-clock care, round-the-clock board-certified intensivist care – we are still not able to make a difference in that mortality.”

Dr. Stevens added that any ICU admission for overdose from opioids is a preventable admission. “So if we have an increase in mortality of this population, we have a number of patients who have preventable deaths in our ICU,” she said.

Efforts to track this epidemic on a national level are important, she said, and the U.S. Centers for Disease Control and Prevention has been investigating opioid overdoses in some cities, including Boston, as they would any epidemic.

The factors driving the observed trends could not be determined from the study data, Dr. Stevens said. But state-specific patterns that show, for example, higher baseline rates and greater increases over time in ICU admissions for opioid overdose in Massachusetts and Indiana may be a starting point for investigation.

Certain practices in the ICU may also be inadvertently contributing. “I imagine that a patient who comes in with an opioid overdose can cause harm to themselves in a number of ways, and the things that we try to do to help them might cause harm in other ways as well,” she said. “So in an effort to try to maintain them in a safe, ventilated state, we might give them a ton of sedation that then prolongs their time on the ventilator. That’s sort of a simple example of how the two could intersect to have a multiplicative effect of harm.”

The idea for the study arose because ICU staff anecdotally noticed an uptick in admissions for opioid use disorder. “Not only were we seeing more people coming in, but we were seeing sicker people coming in, and with the associated tragedy that comes with a lot of young people coming in with opioid use disorder,” Dr. Stevens said. “We wanted to see if this was happening nationally... We asked, is this epidemic now reaching the most technologically advanced parts of our health care system?”

The investigators studied hospitals providing data to Vizient (formerly the University HealthSystem Consortium) between 2009 and 2015. The included hospitals – about 200 for each study year – were predominantly urban and university affiliated, but representation of community hospitals increased during the study period.

Ultimately, analyses were based on a total of 28.2 million hospital discharges of patients aged 18 years or older, which included 4.9 million ICU admissions.

Results reported at the meeting showed that 27,325 patients were admitted to the study hospitals’ ICUs with opioid overdose during the study period, as ascertained from billing codes.

Opioid overdose was seen in 45 patients per 10,000 ICU admissions in 2009 but rose to 65 patients per 10,000 ICU admissions in 2015, a 46% increase.

Furthermore, ICU deaths due to opioid overdose rose by 87% during the same time period, and mortality among patients admitted to the unit with overdose rose at a pace of 0.5% per month.

“This is somewhat unusual because a lot of times, when we are admitting more people to our ICUs or examining [a trend] further, mortality actually goes down. This is partly because maybe we are doing more for them and we are taking care of them in an aggressive way. But it’s also because we are admitting less sick people because we are more aware of the issue,” Dr. Stevens said. “And we saw the opposite of this – we saw that the mortality was going up.”

The use of billing data was a specific means but not a sensitive means of identifying opioid overdoses, she noted. Therefore, the observed values are likely underestimates of these outcomes.

Addressing the opioid overdose epidemic will require a multifaceted approach, according to Dr. Stevens, who disclosed that she had no relevant conflicts of interest.

“Folks are doing very impressive work in the community trying to make sure EMTs and other first responders have access to the tools that they need in those settings,” she said. “But one thing we haven’t approached before is the care that we provide in the ICU, and maybe that’s a space that we need to think more prospectively about.”

SAN FRANCISCO – The opioid overdose crisis in the United States is now plainly evident in intensive care units (ICUs), finds a study of hospitals in 44 states conducted between 2009 and 2015.

During the study period, ICU admissions for opioid overdoses increased by almost half, investigators reported in a session and related press briefing an international conference of the American Thoracic Society. Furthermore, ICU deaths from this cause roughly doubled.

Susan London/Frontline Medical News

“This means the opioid use epidemic has probably reached a new level of crisis,” said lead investigator Jennifer P. Stevens, MD, an instructor in medicine at Harvard Medical School, and an adult intensive care physician at Beth Israel Deaconess Medical Center, both in Boston. “And this means that in spite of everything that we can do in the ICU – keeping them alive on ventilators, doing life support, doing acute dialysis, doing round-the-clock care, round-the-clock board-certified intensivist care – we are still not able to make a difference in that mortality.”

Dr. Stevens added that any ICU admission for overdose from opioids is a preventable admission. “So if we have an increase in mortality of this population, we have a number of patients who have preventable deaths in our ICU,” she said.

Efforts to track this epidemic on a national level are important, she said, and the U.S. Centers for Disease Control and Prevention has been investigating opioid overdoses in some cities, including Boston, as they would any epidemic.

The factors driving the observed trends could not be determined from the study data, Dr. Stevens said. But state-specific patterns that show, for example, higher baseline rates and greater increases over time in ICU admissions for opioid overdose in Massachusetts and Indiana may be a starting point for investigation.

Certain practices in the ICU may also be inadvertently contributing. “I imagine that a patient who comes in with an opioid overdose can cause harm to themselves in a number of ways, and the things that we try to do to help them might cause harm in other ways as well,” she said. “So in an effort to try to maintain them in a safe, ventilated state, we might give them a ton of sedation that then prolongs their time on the ventilator. That’s sort of a simple example of how the two could intersect to have a multiplicative effect of harm.”

The idea for the study arose because ICU staff anecdotally noticed an uptick in admissions for opioid use disorder. “Not only were we seeing more people coming in, but we were seeing sicker people coming in, and with the associated tragedy that comes with a lot of young people coming in with opioid use disorder,” Dr. Stevens said. “We wanted to see if this was happening nationally... We asked, is this epidemic now reaching the most technologically advanced parts of our health care system?”

The investigators studied hospitals providing data to Vizient (formerly the University HealthSystem Consortium) between 2009 and 2015. The included hospitals – about 200 for each study year – were predominantly urban and university affiliated, but representation of community hospitals increased during the study period.

Ultimately, analyses were based on a total of 28.2 million hospital discharges of patients aged 18 years or older, which included 4.9 million ICU admissions.

Results reported at the meeting showed that 27,325 patients were admitted to the study hospitals’ ICUs with opioid overdose during the study period, as ascertained from billing codes.

Opioid overdose was seen in 45 patients per 10,000 ICU admissions in 2009 but rose to 65 patients per 10,000 ICU admissions in 2015, a 46% increase.

Furthermore, ICU deaths due to opioid overdose rose by 87% during the same time period, and mortality among patients admitted to the unit with overdose rose at a pace of 0.5% per month.

“This is somewhat unusual because a lot of times, when we are admitting more people to our ICUs or examining [a trend] further, mortality actually goes down. This is partly because maybe we are doing more for them and we are taking care of them in an aggressive way. But it’s also because we are admitting less sick people because we are more aware of the issue,” Dr. Stevens said. “And we saw the opposite of this – we saw that the mortality was going up.”

The use of billing data was a specific means but not a sensitive means of identifying opioid overdoses, she noted. Therefore, the observed values are likely underestimates of these outcomes.

Addressing the opioid overdose epidemic will require a multifaceted approach, according to Dr. Stevens, who disclosed that she had no relevant conflicts of interest.

“Folks are doing very impressive work in the community trying to make sure EMTs and other first responders have access to the tools that they need in those settings,” she said. “But one thing we haven’t approached before is the care that we provide in the ICU, and maybe that’s a space that we need to think more prospectively about.”

SAN FRANCISCO – The opioid overdose crisis in the United States is now plainly evident in intensive care units (ICUs), finds a study of hospitals in 44 states conducted between 2009 and 2015.

During the study period, ICU admissions for opioid overdoses increased by almost half, investigators reported in a session and related press briefing an international conference of the American Thoracic Society. Furthermore, ICU deaths from this cause roughly doubled.

Susan London/Frontline Medical News

“This means the opioid use epidemic has probably reached a new level of crisis,” said lead investigator Jennifer P. Stevens, MD, an instructor in medicine at Harvard Medical School, and an adult intensive care physician at Beth Israel Deaconess Medical Center, both in Boston. “And this means that in spite of everything that we can do in the ICU – keeping them alive on ventilators, doing life support, doing acute dialysis, doing round-the-clock care, round-the-clock board-certified intensivist care – we are still not able to make a difference in that mortality.”

Dr. Stevens added that any ICU admission for overdose from opioids is a preventable admission. “So if we have an increase in mortality of this population, we have a number of patients who have preventable deaths in our ICU,” she said.

Efforts to track this epidemic on a national level are important, she said, and the U.S. Centers for Disease Control and Prevention has been investigating opioid overdoses in some cities, including Boston, as they would any epidemic.

The factors driving the observed trends could not be determined from the study data, Dr. Stevens said. But state-specific patterns that show, for example, higher baseline rates and greater increases over time in ICU admissions for opioid overdose in Massachusetts and Indiana may be a starting point for investigation.

Certain practices in the ICU may also be inadvertently contributing. “I imagine that a patient who comes in with an opioid overdose can cause harm to themselves in a number of ways, and the things that we try to do to help them might cause harm in other ways as well,” she said. “So in an effort to try to maintain them in a safe, ventilated state, we might give them a ton of sedation that then prolongs their time on the ventilator. That’s sort of a simple example of how the two could intersect to have a multiplicative effect of harm.”

The idea for the study arose because ICU staff anecdotally noticed an uptick in admissions for opioid use disorder. “Not only were we seeing more people coming in, but we were seeing sicker people coming in, and with the associated tragedy that comes with a lot of young people coming in with opioid use disorder,” Dr. Stevens said. “We wanted to see if this was happening nationally... We asked, is this epidemic now reaching the most technologically advanced parts of our health care system?”

The investigators studied hospitals providing data to Vizient (formerly the University HealthSystem Consortium) between 2009 and 2015. The included hospitals – about 200 for each study year – were predominantly urban and university affiliated, but representation of community hospitals increased during the study period.

Ultimately, analyses were based on a total of 28.2 million hospital discharges of patients aged 18 years or older, which included 4.9 million ICU admissions.

Results reported at the meeting showed that 27,325 patients were admitted to the study hospitals’ ICUs with opioid overdose during the study period, as ascertained from billing codes.

Opioid overdose was seen in 45 patients per 10,000 ICU admissions in 2009 but rose to 65 patients per 10,000 ICU admissions in 2015, a 46% increase.

Furthermore, ICU deaths due to opioid overdose rose by 87% during the same time period, and mortality among patients admitted to the unit with overdose rose at a pace of 0.5% per month.

“This is somewhat unusual because a lot of times, when we are admitting more people to our ICUs or examining [a trend] further, mortality actually goes down. This is partly because maybe we are doing more for them and we are taking care of them in an aggressive way. But it’s also because we are admitting less sick people because we are more aware of the issue,” Dr. Stevens said. “And we saw the opposite of this – we saw that the mortality was going up.”

The use of billing data was a specific means but not a sensitive means of identifying opioid overdoses, she noted. Therefore, the observed values are likely underestimates of these outcomes.

Addressing the opioid overdose epidemic will require a multifaceted approach, according to Dr. Stevens, who disclosed that she had no relevant conflicts of interest.

“Folks are doing very impressive work in the community trying to make sure EMTs and other first responders have access to the tools that they need in those settings,” she said. “But one thing we haven’t approached before is the care that we provide in the ICU, and maybe that’s a space that we need to think more prospectively about.”

SAN FRANCISCO – The work environment for nurses and the physician staffing model in the intensive care unit influence patients’ likelihood of acquiring ventilator-associated pneumonia (VAP), based on a cohort study of 25 ICUs.

Overall, each 1-point increase in the score for the nurse work environment – indicating that nurses had a greater sense of playing an important role in patient care – was unexpectedly associated with a roughly sixfold higher rate of VAP among the ICU’s patients, according to data reported in a session and press briefing at an international conference of the American Thoracic Society. However, additional analyses showed that the rate of VAP was higher in closed units where a board-certified critical care physician (intensivist) managed and led care rather than an open unit where care is shared.

“We think that the organization of the ICU is actually influencing nursing practice, which is a really novel finding,” commented first author Deena Kelly Costa, PhD, RN, of the University of Michigan School of Nursing in Ann Arbor. “In closed ICUs, when you have a board-certified physician and an ICU team managing and leading care, even if the work environment is better, nurses may not feel as empowered to standardize their care or practice.”

“ICU nurses are the ones who are primarily responsible for VAP preventive practices: they keep the head of the bed higher than 45 degrees, they conduct oral care, they conduct (patient) surveillance. ICU physicians are involved with writing the orders and ventilator setting management. So how these providers work together could theoretically influence the risk for patients developing VAP,” Dr. Costa said.

“We need to be thinking a little bit more critically about not only the care that’s happening at the bedside... but also at an organizational level. How are these providers organized, and can we work together to improve patient outcomes?”

“I’m not suggesting that we get rid of all closed ICUs because I don’t think that’s the solution,” Dr. Costa maintained. “I think from an administrative perspective, we need to be considering what’s the organization of these clinicians and this unit, and [in a context-specific manner], how can we improve it for better patient outcomes? That may be both working on improving the work environment and making the nurses feel more empowered, or it could be potentially considering other staffing models.”

Some data have already linked a more favorable nurse work environment and the presence of a board-certified critical care physician independently with better patient outcomes in the ICU. But studies of their joint impact are lacking.

The investigators performed a secondary, unit-level analysis of nurse survey data collected during 2005 and 2006 in ICUs in southern Michigan.

In all, 462 nurses working in 25 ICUs completed the Practice Environment Scale of the Nursing Work Index, on which averaged summary scores range between 1 (unfavorable) and 4 (favorable). The scale captures environmental factors such as the adequacy of resources for nurses, support from their managers, and their level of involvement in hospital policy decisions.

The rate of VAP during the same period was assessed using data from more than 1,000 patients from each ICU.

The summary nurse work environment score averaged 2.69 points in the 21 ICUs that had a closed physician staffing model and 2.62 points in the 4 ICUs that had an open physician staffing model. The respective rates of VAP were 7.5% and 2.5%.

In adjusted analysis among all 25 ICUs, each 1-point increase in an ICU’s Practice Environment Scale score was associated with a sharply higher rate of VAP on the unit (adjusted incidence rate ratio, 5.76; P = .02).

However, there was a strong interaction between the score and physician staffing model (P less than .001). In open ICUs, as the score rose, the rate of VAP fell (from about 16% to 5%), whereas in closed ICUs, as the score rose, so did the rate of VAP (from about 3% to 14%).

Dr. Costa disclosed that she had no relevant conflicts of interest. The parent survey was funded by the Blue Cross Blue Shield Foundation of Michigan.

SAN FRANCISCO – The work environment for nurses and the physician staffing model in the intensive care unit influence patients’ likelihood of acquiring ventilator-associated pneumonia (VAP), based on a cohort study of 25 ICUs.

Overall, each 1-point increase in the score for the nurse work environment – indicating that nurses had a greater sense of playing an important role in patient care – was unexpectedly associated with a roughly sixfold higher rate of VAP among the ICU’s patients, according to data reported in a session and press briefing at an international conference of the American Thoracic Society. However, additional analyses showed that the rate of VAP was higher in closed units where a board-certified critical care physician (intensivist) managed and led care rather than an open unit where care is shared.

“We think that the organization of the ICU is actually influencing nursing practice, which is a really novel finding,” commented first author Deena Kelly Costa, PhD, RN, of the University of Michigan School of Nursing in Ann Arbor. “In closed ICUs, when you have a board-certified physician and an ICU team managing and leading care, even if the work environment is better, nurses may not feel as empowered to standardize their care or practice.”

“ICU nurses are the ones who are primarily responsible for VAP preventive practices: they keep the head of the bed higher than 45 degrees, they conduct oral care, they conduct (patient) surveillance. ICU physicians are involved with writing the orders and ventilator setting management. So how these providers work together could theoretically influence the risk for patients developing VAP,” Dr. Costa said.

“We need to be thinking a little bit more critically about not only the care that’s happening at the bedside... but also at an organizational level. How are these providers organized, and can we work together to improve patient outcomes?”

“I’m not suggesting that we get rid of all closed ICUs because I don’t think that’s the solution,” Dr. Costa maintained. “I think from an administrative perspective, we need to be considering what’s the organization of these clinicians and this unit, and [in a context-specific manner], how can we improve it for better patient outcomes? That may be both working on improving the work environment and making the nurses feel more empowered, or it could be potentially considering other staffing models.”

Some data have already linked a more favorable nurse work environment and the presence of a board-certified critical care physician independently with better patient outcomes in the ICU. But studies of their joint impact are lacking.

The investigators performed a secondary, unit-level analysis of nurse survey data collected during 2005 and 2006 in ICUs in southern Michigan.

In all, 462 nurses working in 25 ICUs completed the Practice Environment Scale of the Nursing Work Index, on which averaged summary scores range between 1 (unfavorable) and 4 (favorable). The scale captures environmental factors such as the adequacy of resources for nurses, support from their managers, and their level of involvement in hospital policy decisions.

The rate of VAP during the same period was assessed using data from more than 1,000 patients from each ICU.

The summary nurse work environment score averaged 2.69 points in the 21 ICUs that had a closed physician staffing model and 2.62 points in the 4 ICUs that had an open physician staffing model. The respective rates of VAP were 7.5% and 2.5%.

In adjusted analysis among all 25 ICUs, each 1-point increase in an ICU’s Practice Environment Scale score was associated with a sharply higher rate of VAP on the unit (adjusted incidence rate ratio, 5.76; P = .02).

However, there was a strong interaction between the score and physician staffing model (P less than .001). In open ICUs, as the score rose, the rate of VAP fell (from about 16% to 5%), whereas in closed ICUs, as the score rose, so did the rate of VAP (from about 3% to 14%).

Dr. Costa disclosed that she had no relevant conflicts of interest. The parent survey was funded by the Blue Cross Blue Shield Foundation of Michigan.

SAN FRANCISCO – The work environment for nurses and the physician staffing model in the intensive care unit influence patients’ likelihood of acquiring ventilator-associated pneumonia (VAP), based on a cohort study of 25 ICUs.

Overall, each 1-point increase in the score for the nurse work environment – indicating that nurses had a greater sense of playing an important role in patient care – was unexpectedly associated with a roughly sixfold higher rate of VAP among the ICU’s patients, according to data reported in a session and press briefing at an international conference of the American Thoracic Society. However, additional analyses showed that the rate of VAP was higher in closed units where a board-certified critical care physician (intensivist) managed and led care rather than an open unit where care is shared.

“We think that the organization of the ICU is actually influencing nursing practice, which is a really novel finding,” commented first author Deena Kelly Costa, PhD, RN, of the University of Michigan School of Nursing in Ann Arbor. “In closed ICUs, when you have a board-certified physician and an ICU team managing and leading care, even if the work environment is better, nurses may not feel as empowered to standardize their care or practice.”

“ICU nurses are the ones who are primarily responsible for VAP preventive practices: they keep the head of the bed higher than 45 degrees, they conduct oral care, they conduct (patient) surveillance. ICU physicians are involved with writing the orders and ventilator setting management. So how these providers work together could theoretically influence the risk for patients developing VAP,” Dr. Costa said.

“We need to be thinking a little bit more critically about not only the care that’s happening at the bedside... but also at an organizational level. How are these providers organized, and can we work together to improve patient outcomes?”

“I’m not suggesting that we get rid of all closed ICUs because I don’t think that’s the solution,” Dr. Costa maintained. “I think from an administrative perspective, we need to be considering what’s the organization of these clinicians and this unit, and [in a context-specific manner], how can we improve it for better patient outcomes? That may be both working on improving the work environment and making the nurses feel more empowered, or it could be potentially considering other staffing models.”

Some data have already linked a more favorable nurse work environment and the presence of a board-certified critical care physician independently with better patient outcomes in the ICU. But studies of their joint impact are lacking.

The investigators performed a secondary, unit-level analysis of nurse survey data collected during 2005 and 2006 in ICUs in southern Michigan.

In all, 462 nurses working in 25 ICUs completed the Practice Environment Scale of the Nursing Work Index, on which averaged summary scores range between 1 (unfavorable) and 4 (favorable). The scale captures environmental factors such as the adequacy of resources for nurses, support from their managers, and their level of involvement in hospital policy decisions.

The rate of VAP during the same period was assessed using data from more than 1,000 patients from each ICU.

The summary nurse work environment score averaged 2.69 points in the 21 ICUs that had a closed physician staffing model and 2.62 points in the 4 ICUs that had an open physician staffing model. The respective rates of VAP were 7.5% and 2.5%.

In adjusted analysis among all 25 ICUs, each 1-point increase in an ICU’s Practice Environment Scale score was associated with a sharply higher rate of VAP on the unit (adjusted incidence rate ratio, 5.76; P = .02).

However, there was a strong interaction between the score and physician staffing model (P less than .001). In open ICUs, as the score rose, the rate of VAP fell (from about 16% to 5%), whereas in closed ICUs, as the score rose, so did the rate of VAP (from about 3% to 14%).

Dr. Costa disclosed that she had no relevant conflicts of interest. The parent survey was funded by the Blue Cross Blue Shield Foundation of Michigan.

SAN FRANCISCO – Hospitalized patients who have a change in the medical residents responsible for their care are more likely to die, finds a retrospective cohort study of roughly a quarter million discharges from Veterans Affairs medical centers.

A monthly change in resident care was associated with 9%-20% higher adjusted odds of death during the hospital stay and after discharge, investigators reported in a poster discussion session and press briefing at an international conference of the American Thoracic Society. Analyses suggested that such transitions accounted for 718 additional deaths in the hospital alone during the 6-year study period.

“These are very strong findings,” said Dr. Joshua L. Denson, a fellow in the divisions of pulmonary sciences and critical care medicine at the University of Colorado, Aurora.

The study results represent an important initial step in bringing the problem to light, he said. “Handoffs shift to shift have been looked at, but not this end-of-month, more permanent switching, which I think is a much more substantial transition in care.”

The factors driving the increased mortality are unclear, according to Dr. Denson; however, “when you go on to a new service [as a resident] ... you are now responsible for 20 new people all of a sudden that night.” Therefore, these transitions can be a hectic time characterized by reduced communication and inefficient discharges. In addition, the incoming residents lack familiarity with their new patients’ particulars.

“The handoffs are definitely not preventable, so this is something that has to be dealt with,” he maintained. The study’s findings hint at several possible areas for improvement.

None of the 10 residency programs surveyed provided formal education for monthly resident handoffs, focusing instead on handoffs at shift changes, and most programs lacked a standard procedure, with just one requiring that the handoff be done in person. The programs also varied greatly in their staggering of handoffs – separating transitions of interns (first-year residents) and higher-level residents by at least a few days – to minimize impact.

Despite the absence of outcomes data in this area, some hospitals are forging ahead with their own interventions intended to smooth care transitions, Dr. Denson reported. “In at least two hospitals that I’ve worked in, they are implementing what is called a warm handoff,” he explained. “Basically, a resident from the prior rotation comes the next day and rounds with the new team so he can tell them, ‘Oh, this guy looks a little worse today, you may want to watch him,’ or ‘He looks a little better.’ ”

In the study, conducted while Dr. Denson was chief resident at the NYU School of Medicine, he and his colleagues analyzed data from 10 university-affiliated Veterans Affairs hospitals and internal medicine residency programs that provided their residents’ schedules. Analyses were based on a total of 230,701 discharges of adult medical patients between July 2008 and June 2014.

Hospitalized patients were categorized as having a transition in resident care if they were admitted before the date of an end-of-month house staff transition in care and were discharged in the week after it.

In unadjusted analyses, patients who had a transition of care – whether of intern only, resident only, or both – had significantly higher odds of inpatient mortality and of 30-day mortality and 90-day postdischarge mortality, compared with counterparts who did not have the corresponding transition of care.

In adjusted analyses, patients who had an intern transition still had higher odds of in-hospital mortality (odds ratio, 1.14). In addition, patients had persistently elevated odds of 30-day mortality and 90-day postdischarge mortality if they had an intern transition (odds ratios, 1.20 and 1.17, respectively), a resident transition (1.15 and 1.14), or both (1.10 and 1.09).

The findings “suggest possibly a level-of-training effect to these transitions, as it’s the most inexperienced people that have the higher rate of mortality,” noted Dr. Denson, who disclosed that he had no relevant conflicts of interest. “Interns, being the first-years, tend to carry the bulk of the work in most hospitals, which is an interesting paradigm in our organization. And that may be a good explanation for why we are seeing this.”

SAN FRANCISCO – Hospitalized patients who have a change in the medical residents responsible for their care are more likely to die, finds a retrospective cohort study of roughly a quarter million discharges from Veterans Affairs medical centers.

A monthly change in resident care was associated with 9%-20% higher adjusted odds of death during the hospital stay and after discharge, investigators reported in a poster discussion session and press briefing at an international conference of the American Thoracic Society. Analyses suggested that such transitions accounted for 718 additional deaths in the hospital alone during the 6-year study period.

“These are very strong findings,” said Dr. Joshua L. Denson, a fellow in the divisions of pulmonary sciences and critical care medicine at the University of Colorado, Aurora.

The study results represent an important initial step in bringing the problem to light, he said. “Handoffs shift to shift have been looked at, but not this end-of-month, more permanent switching, which I think is a much more substantial transition in care.”

The factors driving the increased mortality are unclear, according to Dr. Denson; however, “when you go on to a new service [as a resident] ... you are now responsible for 20 new people all of a sudden that night.” Therefore, these transitions can be a hectic time characterized by reduced communication and inefficient discharges. In addition, the incoming residents lack familiarity with their new patients’ particulars.

“The handoffs are definitely not preventable, so this is something that has to be dealt with,” he maintained. The study’s findings hint at several possible areas for improvement.

None of the 10 residency programs surveyed provided formal education for monthly resident handoffs, focusing instead on handoffs at shift changes, and most programs lacked a standard procedure, with just one requiring that the handoff be done in person. The programs also varied greatly in their staggering of handoffs – separating transitions of interns (first-year residents) and higher-level residents by at least a few days – to minimize impact.

Despite the absence of outcomes data in this area, some hospitals are forging ahead with their own interventions intended to smooth care transitions, Dr. Denson reported. “In at least two hospitals that I’ve worked in, they are implementing what is called a warm handoff,” he explained. “Basically, a resident from the prior rotation comes the next day and rounds with the new team so he can tell them, ‘Oh, this guy looks a little worse today, you may want to watch him,’ or ‘He looks a little better.’ ”

In the study, conducted while Dr. Denson was chief resident at the NYU School of Medicine, he and his colleagues analyzed data from 10 university-affiliated Veterans Affairs hospitals and internal medicine residency programs that provided their residents’ schedules. Analyses were based on a total of 230,701 discharges of adult medical patients between July 2008 and June 2014.

Hospitalized patients were categorized as having a transition in resident care if they were admitted before the date of an end-of-month house staff transition in care and were discharged in the week after it.

In unadjusted analyses, patients who had a transition of care – whether of intern only, resident only, or both – had significantly higher odds of inpatient mortality and of 30-day mortality and 90-day postdischarge mortality, compared with counterparts who did not have the corresponding transition of care.

In adjusted analyses, patients who had an intern transition still had higher odds of in-hospital mortality (odds ratio, 1.14). In addition, patients had persistently elevated odds of 30-day mortality and 90-day postdischarge mortality if they had an intern transition (odds ratios, 1.20 and 1.17, respectively), a resident transition (1.15 and 1.14), or both (1.10 and 1.09).

The findings “suggest possibly a level-of-training effect to these transitions, as it’s the most inexperienced people that have the higher rate of mortality,” noted Dr. Denson, who disclosed that he had no relevant conflicts of interest. “Interns, being the first-years, tend to carry the bulk of the work in most hospitals, which is an interesting paradigm in our organization. And that may be a good explanation for why we are seeing this.”

SAN FRANCISCO – Hospitalized patients who have a change in the medical residents responsible for their care are more likely to die, finds a retrospective cohort study of roughly a quarter million discharges from Veterans Affairs medical centers.

A monthly change in resident care was associated with 9%-20% higher adjusted odds of death during the hospital stay and after discharge, investigators reported in a poster discussion session and press briefing at an international conference of the American Thoracic Society. Analyses suggested that such transitions accounted for 718 additional deaths in the hospital alone during the 6-year study period.

“These are very strong findings,” said Dr. Joshua L. Denson, a fellow in the divisions of pulmonary sciences and critical care medicine at the University of Colorado, Aurora.

The study results represent an important initial step in bringing the problem to light, he said. “Handoffs shift to shift have been looked at, but not this end-of-month, more permanent switching, which I think is a much more substantial transition in care.”

The factors driving the increased mortality are unclear, according to Dr. Denson; however, “when you go on to a new service [as a resident] ... you are now responsible for 20 new people all of a sudden that night.” Therefore, these transitions can be a hectic time characterized by reduced communication and inefficient discharges. In addition, the incoming residents lack familiarity with their new patients’ particulars.

“The handoffs are definitely not preventable, so this is something that has to be dealt with,” he maintained. The study’s findings hint at several possible areas for improvement.

None of the 10 residency programs surveyed provided formal education for monthly resident handoffs, focusing instead on handoffs at shift changes, and most programs lacked a standard procedure, with just one requiring that the handoff be done in person. The programs also varied greatly in their staggering of handoffs – separating transitions of interns (first-year residents) and higher-level residents by at least a few days – to minimize impact.

Despite the absence of outcomes data in this area, some hospitals are forging ahead with their own interventions intended to smooth care transitions, Dr. Denson reported. “In at least two hospitals that I’ve worked in, they are implementing what is called a warm handoff,” he explained. “Basically, a resident from the prior rotation comes the next day and rounds with the new team so he can tell them, ‘Oh, this guy looks a little worse today, you may want to watch him,’ or ‘He looks a little better.’ ”

In the study, conducted while Dr. Denson was chief resident at the NYU School of Medicine, he and his colleagues analyzed data from 10 university-affiliated Veterans Affairs hospitals and internal medicine residency programs that provided their residents’ schedules. Analyses were based on a total of 230,701 discharges of adult medical patients between July 2008 and June 2014.

Hospitalized patients were categorized as having a transition in resident care if they were admitted before the date of an end-of-month house staff transition in care and were discharged in the week after it.

In unadjusted analyses, patients who had a transition of care – whether of intern only, resident only, or both – had significantly higher odds of inpatient mortality and of 30-day mortality and 90-day postdischarge mortality, compared with counterparts who did not have the corresponding transition of care.

In adjusted analyses, patients who had an intern transition still had higher odds of in-hospital mortality (odds ratio, 1.14). In addition, patients had persistently elevated odds of 30-day mortality and 90-day postdischarge mortality if they had an intern transition (odds ratios, 1.20 and 1.17, respectively), a resident transition (1.15 and 1.14), or both (1.10 and 1.09).

The findings “suggest possibly a level-of-training effect to these transitions, as it’s the most inexperienced people that have the higher rate of mortality,” noted Dr. Denson, who disclosed that he had no relevant conflicts of interest. “Interns, being the first-years, tend to carry the bulk of the work in most hospitals, which is an interesting paradigm in our organization. And that may be a good explanation for why we are seeing this.”

SAN FRANCISCO – Delaying initiation of renal replacement therapy in critically ill patients with severe acute kidney injury appears to be not only safe but beneficial, according to a randomized controlled trial conducted in France.

The trial, known as Artificial Kidney Initiation in Kidney Injury (AKIKI), was conducted among 620 adult patients from 31 intensive care units. The investigators were led by Dr. Stéphane Gaudry of Assistance Publique–Hôpitaux de Paris.

The death rate did not differ significantly between groups assigned to an early versus a late initiation strategy, according to results presented at an international conference of the American Thoracic Society and simultaneously published (N Engl J Med. 2016 May 15. doi: 10.1056/NEJMoa1603017).

Moreover, nearly half of the patients in the delayed initiation group were able to avoid renal replacement therapy. And they were less likely to develop bloodstream infections and had more rapid onset of diuresis (heralding recovery of renal function) than did peers in whom the therapy was initiated early.

“Our study should not be interpreted as suggesting that a ‘wait and see’ approach is safe for all patients. Indeed, careful surveillance is mandatory when deciding to delay renal-replacement therapy in patients with severe acute kidney injury so that any complications will be detected and renal-replacement therapy initiated without delay,” the investigators concluded. “In our trial, delaying the initiation of therapy allowed many patients to recover from acute kidney injury without embarking on such a treatment course.”

Further, the “findings may not be generalizable, because more than 50% of the patients in our trial received intermittent hemodialysis as the first method of therapy and only 30% of the patients received continuous renal-replacement therapy as the sole method (with no intermittent dialysis at any time).”

The author of an accompanying editorial, Dr. Ravindra L. Mehta of the University of California, San Diego, lists some caveats in interpreting the trial’s findings as support for the delayed initiation strategy.

For example, he notes, the longer time to initiation with the delayed strategy contributed to worsening of metabolic and clinical status in the patients who ultimately did need therapy; the study did not assess the development of chronic kidney disease; and the types of renal replacement therapy selected for patients seem “surprising” as the majority put on this therapy needed vasopressors.

“The findings highlight a need for dynamic risk-stratification tools to identify patients who will not need renal-replacement therapy for management of their acute kidney injury,” Dr. Mehta concluded, noting that ongoing studies should help inform management in this area. “Meanwhile, we should focus on the timely application of renal-replacement therapy while considering individual patient characteristics, process-of-care elements, and logistics to achieve therapeutic goals …”

Patients were eligible for the trial if they had severe acute kidney injury, defined as Kidney Disease: Improving Global Outcomes (KDIGO) stage 3; required mechanical ventilation, catecholamine infusion, or both; and did not have a potentially life-threatening complication directly related to renal failure.

In those assigned to the early strategy, renal replacement therapy was started immediately after randomization. In those assigned to the delayed strategy, it was started if any of several criteria was met: severe hyperkalemia, metabolic acidosis, pulmonary edema, blood urea nitrogen level higher than 112 mg/dL, or oliguria for more than 72 hours after randomization. The specific type of renal replacement therapy was left up to each study site.

The median time between randomization and initiation of renal replacement therapy was 2 hours in the early strategy group and 57 hours in the delayed strategy group.

The estimated 60-day mortality rate – the trial’s primary outcome – was 48.5% with early initiation of therapy and 49.7% with delayed initiation, a nonsignificant difference.

Fully 49% of patients in the delayed strategy group never received renal replacement therapy. In addition, patients in this group were half as likely as were peers in the early initiation group to develop a bloodstream infection (5% vs. 10%), and they had more rapid onset of diuresis (P less than .001).

The groups were essentially the same with respect to the rate of gastrointestinal bleeding and the lengths of stay in the intensive care unit and in the hospital.

Dr. Gaudry disclosed that he received grant support from the French Ministry of Health during the study, and from XENIOS France outside the research. The trial was supported by a grant from Programme Hospitalier de Recherche Clinique National, 2012 (AOM12456), funded by the French Ministry of Health..

SAN FRANCISCO – Delaying initiation of renal replacement therapy in critically ill patients with severe acute kidney injury appears to be not only safe but beneficial, according to a randomized controlled trial conducted in France.

The trial, known as Artificial Kidney Initiation in Kidney Injury (AKIKI), was conducted among 620 adult patients from 31 intensive care units. The investigators were led by Dr. Stéphane Gaudry of Assistance Publique–Hôpitaux de Paris.

The death rate did not differ significantly between groups assigned to an early versus a late initiation strategy, according to results presented at an international conference of the American Thoracic Society and simultaneously published (N Engl J Med. 2016 May 15. doi: 10.1056/NEJMoa1603017).

Moreover, nearly half of the patients in the delayed initiation group were able to avoid renal replacement therapy. And they were less likely to develop bloodstream infections and had more rapid onset of diuresis (heralding recovery of renal function) than did peers in whom the therapy was initiated early.

“Our study should not be interpreted as suggesting that a ‘wait and see’ approach is safe for all patients. Indeed, careful surveillance is mandatory when deciding to delay renal-replacement therapy in patients with severe acute kidney injury so that any complications will be detected and renal-replacement therapy initiated without delay,” the investigators concluded. “In our trial, delaying the initiation of therapy allowed many patients to recover from acute kidney injury without embarking on such a treatment course.”

Further, the “findings may not be generalizable, because more than 50% of the patients in our trial received intermittent hemodialysis as the first method of therapy and only 30% of the patients received continuous renal-replacement therapy as the sole method (with no intermittent dialysis at any time).”

The author of an accompanying editorial, Dr. Ravindra L. Mehta of the University of California, San Diego, lists some caveats in interpreting the trial’s findings as support for the delayed initiation strategy.

For example, he notes, the longer time to initiation with the delayed strategy contributed to worsening of metabolic and clinical status in the patients who ultimately did need therapy; the study did not assess the development of chronic kidney disease; and the types of renal replacement therapy selected for patients seem “surprising” as the majority put on this therapy needed vasopressors.

“The findings highlight a need for dynamic risk-stratification tools to identify patients who will not need renal-replacement therapy for management of their acute kidney injury,” Dr. Mehta concluded, noting that ongoing studies should help inform management in this area. “Meanwhile, we should focus on the timely application of renal-replacement therapy while considering individual patient characteristics, process-of-care elements, and logistics to achieve therapeutic goals …”

Patients were eligible for the trial if they had severe acute kidney injury, defined as Kidney Disease: Improving Global Outcomes (KDIGO) stage 3; required mechanical ventilation, catecholamine infusion, or both; and did not have a potentially life-threatening complication directly related to renal failure.

In those assigned to the early strategy, renal replacement therapy was started immediately after randomization. In those assigned to the delayed strategy, it was started if any of several criteria was met: severe hyperkalemia, metabolic acidosis, pulmonary edema, blood urea nitrogen level higher than 112 mg/dL, or oliguria for more than 72 hours after randomization. The specific type of renal replacement therapy was left up to each study site.

The median time between randomization and initiation of renal replacement therapy was 2 hours in the early strategy group and 57 hours in the delayed strategy group.

The estimated 60-day mortality rate – the trial’s primary outcome – was 48.5% with early initiation of therapy and 49.7% with delayed initiation, a nonsignificant difference.

Fully 49% of patients in the delayed strategy group never received renal replacement therapy. In addition, patients in this group were half as likely as were peers in the early initiation group to develop a bloodstream infection (5% vs. 10%), and they had more rapid onset of diuresis (P less than .001).

The groups were essentially the same with respect to the rate of gastrointestinal bleeding and the lengths of stay in the intensive care unit and in the hospital.

Dr. Gaudry disclosed that he received grant support from the French Ministry of Health during the study, and from XENIOS France outside the research. The trial was supported by a grant from Programme Hospitalier de Recherche Clinique National, 2012 (AOM12456), funded by the French Ministry of Health..

SAN FRANCISCO – Delaying initiation of renal replacement therapy in critically ill patients with severe acute kidney injury appears to be not only safe but beneficial, according to a randomized controlled trial conducted in France.

The trial, known as Artificial Kidney Initiation in Kidney Injury (AKIKI), was conducted among 620 adult patients from 31 intensive care units. The investigators were led by Dr. Stéphane Gaudry of Assistance Publique–Hôpitaux de Paris.

The death rate did not differ significantly between groups assigned to an early versus a late initiation strategy, according to results presented at an international conference of the American Thoracic Society and simultaneously published (N Engl J Med. 2016 May 15. doi: 10.1056/NEJMoa1603017).

Moreover, nearly half of the patients in the delayed initiation group were able to avoid renal replacement therapy. And they were less likely to develop bloodstream infections and had more rapid onset of diuresis (heralding recovery of renal function) than did peers in whom the therapy was initiated early.

“Our study should not be interpreted as suggesting that a ‘wait and see’ approach is safe for all patients. Indeed, careful surveillance is mandatory when deciding to delay renal-replacement therapy in patients with severe acute kidney injury so that any complications will be detected and renal-replacement therapy initiated without delay,” the investigators concluded. “In our trial, delaying the initiation of therapy allowed many patients to recover from acute kidney injury without embarking on such a treatment course.”

Further, the “findings may not be generalizable, because more than 50% of the patients in our trial received intermittent hemodialysis as the first method of therapy and only 30% of the patients received continuous renal-replacement therapy as the sole method (with no intermittent dialysis at any time).”

The author of an accompanying editorial, Dr. Ravindra L. Mehta of the University of California, San Diego, lists some caveats in interpreting the trial’s findings as support for the delayed initiation strategy.

For example, he notes, the longer time to initiation with the delayed strategy contributed to worsening of metabolic and clinical status in the patients who ultimately did need therapy; the study did not assess the development of chronic kidney disease; and the types of renal replacement therapy selected for patients seem “surprising” as the majority put on this therapy needed vasopressors.

“The findings highlight a need for dynamic risk-stratification tools to identify patients who will not need renal-replacement therapy for management of their acute kidney injury,” Dr. Mehta concluded, noting that ongoing studies should help inform management in this area. “Meanwhile, we should focus on the timely application of renal-replacement therapy while considering individual patient characteristics, process-of-care elements, and logistics to achieve therapeutic goals …”

Patients were eligible for the trial if they had severe acute kidney injury, defined as Kidney Disease: Improving Global Outcomes (KDIGO) stage 3; required mechanical ventilation, catecholamine infusion, or both; and did not have a potentially life-threatening complication directly related to renal failure.

In those assigned to the early strategy, renal replacement therapy was started immediately after randomization. In those assigned to the delayed strategy, it was started if any of several criteria was met: severe hyperkalemia, metabolic acidosis, pulmonary edema, blood urea nitrogen level higher than 112 mg/dL, or oliguria for more than 72 hours after randomization. The specific type of renal replacement therapy was left up to each study site.

The median time between randomization and initiation of renal replacement therapy was 2 hours in the early strategy group and 57 hours in the delayed strategy group.

The estimated 60-day mortality rate – the trial’s primary outcome – was 48.5% with early initiation of therapy and 49.7% with delayed initiation, a nonsignificant difference.

Fully 49% of patients in the delayed strategy group never received renal replacement therapy. In addition, patients in this group were half as likely as were peers in the early initiation group to develop a bloodstream infection (5% vs. 10%), and they had more rapid onset of diuresis (P less than .001).

The groups were essentially the same with respect to the rate of gastrointestinal bleeding and the lengths of stay in the intensive care unit and in the hospital.

Dr. Gaudry disclosed that he received grant support from the French Ministry of Health during the study, and from XENIOS France outside the research. The trial was supported by a grant from Programme Hospitalier de Recherche Clinique National, 2012 (AOM12456), funded by the French Ministry of Health..

SAN FRANCISCO – Mechanically ventilated patients in the intensive care unit (ICU) have poorer outcomes if extubated during the night instead of during the day, finds a retrospective cohort study reported at an international conference of the American Thoracic Society.

Overall, 20.1% of the nearly 98,000 adult patients studied were extubated during nighttime hours, between 7:00 p.m. and 7:00 a.m., according to data presented in a session and a related press conference.

Compared with patients extubated during daytime hours, patients extubated during nighttime hours had higher rates of ICU and hospital death, with the absolute difference ranging from 1.0% to 5.1%. Additionally, among those mechanically ventilated for at least 12 hours, nighttime extubation was associated with an absolute 2% increase in the risk of reintubation.

“I think this is the first large-scale study that looks at a practice that, although not as common as we thought it was, is still done about a fifth of the time and even with decreasing rates, is not a rare practice on our units,” commented lead author Dr. Hayley B. Gershengorn of the department of medicine (critical care) and the Saul R. Korey department of neurology at the Albert Einstein College of Medicine, New York.

“As we have increasing staffing [overnight] and maybe an increasing push to move people through our ICUs, we need to probably take some care because although we can’t demonstrate a causal link, it is quite concerning, this consistent finding of increased mortality and reintubation in these folks,” she said.

There are several possible reasons for the observed heightened risks of death and reintubation with nighttime extubation that could not be fully explored in the study, Dr. Gershengorn said.

“We were not able to identify the indication for extubation or discontinuation of mechanical ventilation. So one of the concerns that we have is that it’s probably more common that folks unintentionally extubate themselves or someone unintentionally extubates them overnight, when staffing is less,” she explained. “The other part, which we tried to adjust for but we don’t have perfect data on, is what is the staffing overnight,” including factors such as the ratio of nurses to patients and how many units an intensivist is covering, not just whether he or she is present.

“In terms of the reintubation risk being higher in the [group with longer duration of mechanical ventilation], the question I have is whether or not there is less comfort with somebody looking less well when there is less staff around, and whether or not there may be a quicker trigger to reintubate them if they don’t look so great,” she said.

The majority of intubated patients are unlikely to improve enough physiologically to prompt nighttime extubation rather than waiting until daytime, according to Dr. Gershengorn. But there are at least two groups whom clinicians might want to extubate at night.

One group is those who underwent elective surgery during the day. “They are waiting to come out of anesthesia, and the plan is to discontinue mechanical ventilation at the time that that occurs,” she explained. Another group is those who are agitated on the ventilator, require more sedation than usual, and suddenly awake at night. “These patients are really hard to keep comfortable. I can [sedate them] again and try this problem all over again tomorrow morning, or I can just bite the bullet and pull the tube out,” she said.

The investigators analyzed data from the Project IMPACT critical care medicine database, in which data are prospectively collected for benchmarking purposes. In all, they studied 97,844 mechanically ventilated adults from 165 medical and surgical ICUs across the United States between 2000 and 2009.

Results showed that nighttime extubation was more common among elective surgical patients, those coming from the operating room or a postanesthesia care unit, and those mechanically ventilated for less than 12 hours.

In a finding that Dr. Gershengorn described as surprising, there was a temporal trend by which the adjusted proportion of extubations performed at night actually decreased in more recent years during the study period.

The investigators next looked at outcomes among 10,279 propensity-matched pairs of patients, one member of the pair having been extubated during the night and the other having been extubated during the day.

Among those mechanically ventilated for less than 12 hours, nighttime extubation was associated with higher ICU mortality (5.6% vs. 4.6%; P = .025) and hospital mortality (8.3% vs. 7.0%; P = .014). Findings were inconsistent for length of stay, with nighttime extubation associated with a shorter ICU stay but a longer hospital stay.

Among patients mechanically ventilated for 12 hours or longer, those extubated during the night had a higher rate of reintubation (14.6% vs. 12.4%; P less than .001), as well as higher ICU mortality (11.2% vs. 6.1%; P less than .001) and hospital mortality (16.0% vs. 11.1%; P less than .001). Lengths of stay did not differ by extubation time of day in this group.

In sensitivity analyses, findings were similar when the definition of nighttime extubation was altered to the hours of midnight to 5 a.m. and when analyses were restricted to nonpalliative patients, according to Dr. Gershengorn, who disclosed that she had no relevant conflicts of interest.

SAN FRANCISCO – Mechanically ventilated patients in the intensive care unit (ICU) have poorer outcomes if extubated during the night instead of during the day, finds a retrospective cohort study reported at an international conference of the American Thoracic Society.

Overall, 20.1% of the nearly 98,000 adult patients studied were extubated during nighttime hours, between 7:00 p.m. and 7:00 a.m., according to data presented in a session and a related press conference.

Compared with patients extubated during daytime hours, patients extubated during nighttime hours had higher rates of ICU and hospital death, with the absolute difference ranging from 1.0% to 5.1%. Additionally, among those mechanically ventilated for at least 12 hours, nighttime extubation was associated with an absolute 2% increase in the risk of reintubation.

“I think this is the first large-scale study that looks at a practice that, although not as common as we thought it was, is still done about a fifth of the time and even with decreasing rates, is not a rare practice on our units,” commented lead author Dr. Hayley B. Gershengorn of the department of medicine (critical care) and the Saul R. Korey department of neurology at the Albert Einstein College of Medicine, New York.

“As we have increasing staffing [overnight] and maybe an increasing push to move people through our ICUs, we need to probably take some care because although we can’t demonstrate a causal link, it is quite concerning, this consistent finding of increased mortality and reintubation in these folks,” she said.

There are several possible reasons for the observed heightened risks of death and reintubation with nighttime extubation that could not be fully explored in the study, Dr. Gershengorn said.

“We were not able to identify the indication for extubation or discontinuation of mechanical ventilation. So one of the concerns that we have is that it’s probably more common that folks unintentionally extubate themselves or someone unintentionally extubates them overnight, when staffing is less,” she explained. “The other part, which we tried to adjust for but we don’t have perfect data on, is what is the staffing overnight,” including factors such as the ratio of nurses to patients and how many units an intensivist is covering, not just whether he or she is present.

“In terms of the reintubation risk being higher in the [group with longer duration of mechanical ventilation], the question I have is whether or not there is less comfort with somebody looking less well when there is less staff around, and whether or not there may be a quicker trigger to reintubate them if they don’t look so great,” she said.

The majority of intubated patients are unlikely to improve enough physiologically to prompt nighttime extubation rather than waiting until daytime, according to Dr. Gershengorn. But there are at least two groups whom clinicians might want to extubate at night.

One group is those who underwent elective surgery during the day. “They are waiting to come out of anesthesia, and the plan is to discontinue mechanical ventilation at the time that that occurs,” she explained. Another group is those who are agitated on the ventilator, require more sedation than usual, and suddenly awake at night. “These patients are really hard to keep comfortable. I can [sedate them] again and try this problem all over again tomorrow morning, or I can just bite the bullet and pull the tube out,” she said.

The investigators analyzed data from the Project IMPACT critical care medicine database, in which data are prospectively collected for benchmarking purposes. In all, they studied 97,844 mechanically ventilated adults from 165 medical and surgical ICUs across the United States between 2000 and 2009.

Results showed that nighttime extubation was more common among elective surgical patients, those coming from the operating room or a postanesthesia care unit, and those mechanically ventilated for less than 12 hours.

In a finding that Dr. Gershengorn described as surprising, there was a temporal trend by which the adjusted proportion of extubations performed at night actually decreased in more recent years during the study period.

The investigators next looked at outcomes among 10,279 propensity-matched pairs of patients, one member of the pair having been extubated during the night and the other having been extubated during the day.

Among those mechanically ventilated for less than 12 hours, nighttime extubation was associated with higher ICU mortality (5.6% vs. 4.6%; P = .025) and hospital mortality (8.3% vs. 7.0%; P = .014). Findings were inconsistent for length of stay, with nighttime extubation associated with a shorter ICU stay but a longer hospital stay.

Among patients mechanically ventilated for 12 hours or longer, those extubated during the night had a higher rate of reintubation (14.6% vs. 12.4%; P less than .001), as well as higher ICU mortality (11.2% vs. 6.1%; P less than .001) and hospital mortality (16.0% vs. 11.1%; P less than .001). Lengths of stay did not differ by extubation time of day in this group.

In sensitivity analyses, findings were similar when the definition of nighttime extubation was altered to the hours of midnight to 5 a.m. and when analyses were restricted to nonpalliative patients, according to Dr. Gershengorn, who disclosed that she had no relevant conflicts of interest.

SAN FRANCISCO – Mechanically ventilated patients in the intensive care unit (ICU) have poorer outcomes if extubated during the night instead of during the day, finds a retrospective cohort study reported at an international conference of the American Thoracic Society.

Overall, 20.1% of the nearly 98,000 adult patients studied were extubated during nighttime hours, between 7:00 p.m. and 7:00 a.m., according to data presented in a session and a related press conference.

Compared with patients extubated during daytime hours, patients extubated during nighttime hours had higher rates of ICU and hospital death, with the absolute difference ranging from 1.0% to 5.1%. Additionally, among those mechanically ventilated for at least 12 hours, nighttime extubation was associated with an absolute 2% increase in the risk of reintubation.

“I think this is the first large-scale study that looks at a practice that, although not as common as we thought it was, is still done about a fifth of the time and even with decreasing rates, is not a rare practice on our units,” commented lead author Dr. Hayley B. Gershengorn of the department of medicine (critical care) and the Saul R. Korey department of neurology at the Albert Einstein College of Medicine, New York.

“As we have increasing staffing [overnight] and maybe an increasing push to move people through our ICUs, we need to probably take some care because although we can’t demonstrate a causal link, it is quite concerning, this consistent finding of increased mortality and reintubation in these folks,” she said.

There are several possible reasons for the observed heightened risks of death and reintubation with nighttime extubation that could not be fully explored in the study, Dr. Gershengorn said.

“We were not able to identify the indication for extubation or discontinuation of mechanical ventilation. So one of the concerns that we have is that it’s probably more common that folks unintentionally extubate themselves or someone unintentionally extubates them overnight, when staffing is less,” she explained. “The other part, which we tried to adjust for but we don’t have perfect data on, is what is the staffing overnight,” including factors such as the ratio of nurses to patients and how many units an intensivist is covering, not just whether he or she is present.

“In terms of the reintubation risk being higher in the [group with longer duration of mechanical ventilation], the question I have is whether or not there is less comfort with somebody looking less well when there is less staff around, and whether or not there may be a quicker trigger to reintubate them if they don’t look so great,” she said.

The majority of intubated patients are unlikely to improve enough physiologically to prompt nighttime extubation rather than waiting until daytime, according to Dr. Gershengorn. But there are at least two groups whom clinicians might want to extubate at night.

One group is those who underwent elective surgery during the day. “They are waiting to come out of anesthesia, and the plan is to discontinue mechanical ventilation at the time that that occurs,” she explained. Another group is those who are agitated on the ventilator, require more sedation than usual, and suddenly awake at night. “These patients are really hard to keep comfortable. I can [sedate them] again and try this problem all over again tomorrow morning, or I can just bite the bullet and pull the tube out,” she said.

The investigators analyzed data from the Project IMPACT critical care medicine database, in which data are prospectively collected for benchmarking purposes. In all, they studied 97,844 mechanically ventilated adults from 165 medical and surgical ICUs across the United States between 2000 and 2009.

Results showed that nighttime extubation was more common among elective surgical patients, those coming from the operating room or a postanesthesia care unit, and those mechanically ventilated for less than 12 hours.

In a finding that Dr. Gershengorn described as surprising, there was a temporal trend by which the adjusted proportion of extubations performed at night actually decreased in more recent years during the study period.

The investigators next looked at outcomes among 10,279 propensity-matched pairs of patients, one member of the pair having been extubated during the night and the other having been extubated during the day.

Among those mechanically ventilated for less than 12 hours, nighttime extubation was associated with higher ICU mortality (5.6% vs. 4.6%; P = .025) and hospital mortality (8.3% vs. 7.0%; P = .014). Findings were inconsistent for length of stay, with nighttime extubation associated with a shorter ICU stay but a longer hospital stay.

Among patients mechanically ventilated for 12 hours or longer, those extubated during the night had a higher rate of reintubation (14.6% vs. 12.4%; P less than .001), as well as higher ICU mortality (11.2% vs. 6.1%; P less than .001) and hospital mortality (16.0% vs. 11.1%; P less than .001). Lengths of stay did not differ by extubation time of day in this group.

In sensitivity analyses, findings were similar when the definition of nighttime extubation was altered to the hours of midnight to 5 a.m. and when analyses were restricted to nonpalliative patients, according to Dr. Gershengorn, who disclosed that she had no relevant conflicts of interest.