Michael D. Mendoza, MD, MPH

ILLUSTRATIVE CASE

Two weeks ago, you informed your patient—a 53-year-old man with a body mass index of 28.4—that he has type 2 diabetes. Since then, he has seen a nutritionist and begun exercising regularly. His hemoglobin A1c (HbA1c) is 7.7%. You recommend that he begin taking metformin. The patient is worried about the potential for oral antidiabetic agents to cause hypoglycemia. He’s aware that many patients with diabetes monitor their blood sugar levels at home and wants to know if he should, too. You wonder whether it’s necessary, or even advisable, to initiate self-monitoring at this time.

For patients with type 2 diabetes, self-monitoring of blood glucose makes intuitive sense. Theoretically, it reinforces self-management behaviors, promotes adherence to the prescribed medication regimen, and leads to better glucose control. It seems obvious, too, that patients taking medications intended to lower blood sugar need to be aware of their glucose levels so they can take action to reduce the risk of complications.

But things that make sense intuitively do not always stand up to scrutiny. New high-quality evidence suggests that for those with newly diagnosed diabetes, self-monitoring of blood glucose may do more harm than good.

More questions than answers

While it is generally accepted that glucose self-monitoring is useful for those with insulin-treated type 2 diabetes,^2-4 evidence supporting the practice for patients with diabetes who do not require insulin is limited. Two recent meta-analyses of RCTs^5,6 found that self-monitoring of blood glucose achieves a statistically significant reduction of 0.4% in HbA1c; the quality of the studies, however, was limited. A well-designed RCT was needed, the researchers concluded, to settle questions about the value of self-monitoring.

The most recent Cochrane review⁷ of self-monitoring reached a similar conclusion: The reviewers called for additional research into the benefits of self-monitoring for patients with diabetes who do not need insulin. The reviewers also emphasized the need for information on patient-related outcomes such as quality of life, well-being, and satisfaction.

Are recommendations out of step?

Despite the lack of definitive evidence, the Department of Health and Human Services calls on us to increase the proportion of patients with diabetes who monitor their blood sugar at least once daily to 60% as part of its Healthy People 2010 initiative.⁸ The American Diabetes Association states that self-monitoring of blood glucose may help patients taking oral antidiabetic agents achieve glycemic goals.⁹ And the International Diabetes Federation recommends that self-monitoring of blood glucose be offered to all people with type 2 diabetes taking insulin or oral agents—and be part of the patient education that is given to all those who are newly diagnosed.¹⁰

But all of these groups may need to rethink their recommendations in light of the latest findings from the O’Kane RCT.

STUDY SUMMARY: Self-monitoring has little effect on glycemic control

O’Kane and colleagues conducted a prospective RCT comparing self-monitoring versus no monitoring among 184 people with newly diagnosed type 2 diabetes.¹ Patients were randomized to the self-monitoring or control group for 1 year, with clinic visits at 3-month intervals. Those who were already taking insulin or had engaged in self-monitoring of blood glucose were excluded.

At baseline, there was no significant difference in HbA1c, age, or sex between the 2 groups. Participants in both groups underwent identical diabetes education programs throughout the study period and received dietary and medical management based on the same treatment algorithm. Patients whose baseline HbA1c was >7.5% received metformin, followed by the sulfonylurea gliclazide if they did not reach target at the maximum dose of metformin. There was no significant difference in medication use at baseline or at 12 months.

Patients in the self-monitoring group were given glucose monitors and asked to record 4 fasting and 4 postprandial capillary blood glucose measurements per week. They were also taught to monitor and interpret blood glucose readings, and to respond appropriately to high or low readings.

At each follow-up visit, patients underwent blood tests for HbA1c, lipids, and electrolyte levels and completed questionnaires about treatment satisfaction, attitudes about diabetes, and levels of depression, anxiety, and well-being. Adherence to self-monitoring was verified by downloading meter readings. The dropout rate was low (2.2%), and adherence in the self-monitoring group was high. Study results were assessed using intent-to-treat analysis.

HbA1c fell in both the self-monitoring and control groups, with no significant differences at any point. The mean (standard deviation) value at 12 months was 6.9% (0.8%) in the self-monitoring group, compared to 6.9% (1.2%) in the control group, with a 95% confidence interval for the change in HbA1c of –0.25% to 0.38%. Throughout the study period, there was no difference in use of oral hypoglycemic medications or reported hypoglycemia.

Self-monitoring linked to depression

Measures of depression and anxiety were scored on a 100-point scale and compared to baseline measurements. At 12 months, participants in the self-monitoring group were more depressed, scoring 6% higher, on average, on the depression subscale of the well-being questionnaire (P=.01) than those in the control group. There was a trend toward increased anxiety in the self-monitoring group, but no significant differences in well-being, energy, or any of the other diabetes attitude subscales.

WHAT’S NEW: Less may be better

Because we emphasize self-management skills when we counsel patients about diabetes, it is surprising to learn that knowledge about glycemic control and blood sugar levels does not lead to better glycemic control. This RCT provides strong evidence that more information is not necessarily desirable, at least for patients with newly diagnosed type 2 diabetes who do not need insulin.

Depression is a known complication of diabetes. It affects an estimated 10% to 30% of patients with diabetes, who have double the odds of depression compared to people without diabetes.^11,12 Patients with depression and diabetes have poorer glycemic control,^13,14 an increased risk of complications,^15-17 a decreased quality of life,¹⁸ an increased disability burden,^19,20 and increased health care use and costs.^18,21,22 In addition, they face a significantly higher risk of death from all causes, beyond the risks associated with depression or diabetes alone.²³

CAVEATS: Patients on sulfonylureas may be an exception

This study used metformin as the initial oral medication, with sulfonylureas reserved for those who did not reach target glycemic control with maximum metformin therapy. The number of patients taking sulfonylureas was 11 in the self-monitoring group and 6 in the control group. Because hypoglycemia is a concern in patients taking sulfonylureas, there may be a role for self-monitoring of blood glucose in these patients.

Also of note: This study does not provide definitive evidence that self-monitoring of blood glucose causes harm. Although self-monitoring was associated with a 6% higher score on a depression subscale and a trend toward increased anxiety, overall satisfaction with treatment was similar in both groups. Additional studies are needed to better understand the relationship between self-monitoring and depression.

Self-monitoring may still be a good idea for certain patients, regardless of their diabetic medication regimen. When evaluating the potential benefits of self-monitoring of blood glucose, physicians should consider the individual’s predisposition to depression, among other concerns.

CHALLENGES TO IMPLEMENTATION: Hard to forego a practice that everyone expects

Self-monitoring serves different purposes for different populations. Blood glucose levels, along with HbA1c, can guide clinicians in making treatment decisions. Knowing blood sugar levels may be educational or empowering to patients, and provides critical information if hypoglycemia is a concern. These considerations lead us to conclude that while self-monitoring is not indicated for all newly diagnosed diabetic patients, it should be considered in selected circumstances.

Because of the prevalence of self-monitoring of blood glucose, patients may see it as a key component of an optimal self-management regimen. It may be hard to convince patients with newly diagnosed diabetes otherwise—and to convince some clinicians that there is little benefit in recommending it. Again, clinical judgment is required. We suspect, however, that with the proper explanation, many patients will be relieved to learn that they will not have to prick their fingers regularly or record their blood glucose.

Acknowledgements

The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

This study was selected and evaluated using FPIN’s Priority Updates from the Research Literature (PURL) Surveillance System methodology. The criteria and findings leading to the selection of this study as a PURL can be accessed at www.jfponline.com/purls.

ILLUSTRATIVE CASE

Two weeks ago, you informed your patient—a 53-year-old man with a body mass index of 28.4—that he has type 2 diabetes. Since then, he has seen a nutritionist and begun exercising regularly. His hemoglobin A1c (HbA1c) is 7.7%. You recommend that he begin taking metformin. The patient is worried about the potential for oral antidiabetic agents to cause hypoglycemia. He’s aware that many patients with diabetes monitor their blood sugar levels at home and wants to know if he should, too. You wonder whether it’s necessary, or even advisable, to initiate self-monitoring at this time.

For patients with type 2 diabetes, self-monitoring of blood glucose makes intuitive sense. Theoretically, it reinforces self-management behaviors, promotes adherence to the prescribed medication regimen, and leads to better glucose control. It seems obvious, too, that patients taking medications intended to lower blood sugar need to be aware of their glucose levels so they can take action to reduce the risk of complications.

But things that make sense intuitively do not always stand up to scrutiny. New high-quality evidence suggests that for those with newly diagnosed diabetes, self-monitoring of blood glucose may do more harm than good.

More questions than answers

While it is generally accepted that glucose self-monitoring is useful for those with insulin-treated type 2 diabetes,^2-4 evidence supporting the practice for patients with diabetes who do not require insulin is limited. Two recent meta-analyses of RCTs^5,6 found that self-monitoring of blood glucose achieves a statistically significant reduction of 0.4% in HbA1c; the quality of the studies, however, was limited. A well-designed RCT was needed, the researchers concluded, to settle questions about the value of self-monitoring.

The most recent Cochrane review⁷ of self-monitoring reached a similar conclusion: The reviewers called for additional research into the benefits of self-monitoring for patients with diabetes who do not need insulin. The reviewers also emphasized the need for information on patient-related outcomes such as quality of life, well-being, and satisfaction.

Are recommendations out of step?

Despite the lack of definitive evidence, the Department of Health and Human Services calls on us to increase the proportion of patients with diabetes who monitor their blood sugar at least once daily to 60% as part of its Healthy People 2010 initiative.⁸ The American Diabetes Association states that self-monitoring of blood glucose may help patients taking oral antidiabetic agents achieve glycemic goals.⁹ And the International Diabetes Federation recommends that self-monitoring of blood glucose be offered to all people with type 2 diabetes taking insulin or oral agents—and be part of the patient education that is given to all those who are newly diagnosed.¹⁰

But all of these groups may need to rethink their recommendations in light of the latest findings from the O’Kane RCT.

STUDY SUMMARY: Self-monitoring has little effect on glycemic control

O’Kane and colleagues conducted a prospective RCT comparing self-monitoring versus no monitoring among 184 people with newly diagnosed type 2 diabetes.¹ Patients were randomized to the self-monitoring or control group for 1 year, with clinic visits at 3-month intervals. Those who were already taking insulin or had engaged in self-monitoring of blood glucose were excluded.

At baseline, there was no significant difference in HbA1c, age, or sex between the 2 groups. Participants in both groups underwent identical diabetes education programs throughout the study period and received dietary and medical management based on the same treatment algorithm. Patients whose baseline HbA1c was >7.5% received metformin, followed by the sulfonylurea gliclazide if they did not reach target at the maximum dose of metformin. There was no significant difference in medication use at baseline or at 12 months.

Patients in the self-monitoring group were given glucose monitors and asked to record 4 fasting and 4 postprandial capillary blood glucose measurements per week. They were also taught to monitor and interpret blood glucose readings, and to respond appropriately to high or low readings.

At each follow-up visit, patients underwent blood tests for HbA1c, lipids, and electrolyte levels and completed questionnaires about treatment satisfaction, attitudes about diabetes, and levels of depression, anxiety, and well-being. Adherence to self-monitoring was verified by downloading meter readings. The dropout rate was low (2.2%), and adherence in the self-monitoring group was high. Study results were assessed using intent-to-treat analysis.

HbA1c fell in both the self-monitoring and control groups, with no significant differences at any point. The mean (standard deviation) value at 12 months was 6.9% (0.8%) in the self-monitoring group, compared to 6.9% (1.2%) in the control group, with a 95% confidence interval for the change in HbA1c of –0.25% to 0.38%. Throughout the study period, there was no difference in use of oral hypoglycemic medications or reported hypoglycemia.

Self-monitoring linked to depression

Measures of depression and anxiety were scored on a 100-point scale and compared to baseline measurements. At 12 months, participants in the self-monitoring group were more depressed, scoring 6% higher, on average, on the depression subscale of the well-being questionnaire (P=.01) than those in the control group. There was a trend toward increased anxiety in the self-monitoring group, but no significant differences in well-being, energy, or any of the other diabetes attitude subscales.

WHAT’S NEW: Less may be better

Because we emphasize self-management skills when we counsel patients about diabetes, it is surprising to learn that knowledge about glycemic control and blood sugar levels does not lead to better glycemic control. This RCT provides strong evidence that more information is not necessarily desirable, at least for patients with newly diagnosed type 2 diabetes who do not need insulin.

Depression is a known complication of diabetes. It affects an estimated 10% to 30% of patients with diabetes, who have double the odds of depression compared to people without diabetes.^11,12 Patients with depression and diabetes have poorer glycemic control,^13,14 an increased risk of complications,^15-17 a decreased quality of life,¹⁸ an increased disability burden,^19,20 and increased health care use and costs.^18,21,22 In addition, they face a significantly higher risk of death from all causes, beyond the risks associated with depression or diabetes alone.²³

CAVEATS: Patients on sulfonylureas may be an exception

This study used metformin as the initial oral medication, with sulfonylureas reserved for those who did not reach target glycemic control with maximum metformin therapy. The number of patients taking sulfonylureas was 11 in the self-monitoring group and 6 in the control group. Because hypoglycemia is a concern in patients taking sulfonylureas, there may be a role for self-monitoring of blood glucose in these patients.

Also of note: This study does not provide definitive evidence that self-monitoring of blood glucose causes harm. Although self-monitoring was associated with a 6% higher score on a depression subscale and a trend toward increased anxiety, overall satisfaction with treatment was similar in both groups. Additional studies are needed to better understand the relationship between self-monitoring and depression.

Self-monitoring may still be a good idea for certain patients, regardless of their diabetic medication regimen. When evaluating the potential benefits of self-monitoring of blood glucose, physicians should consider the individual’s predisposition to depression, among other concerns.

CHALLENGES TO IMPLEMENTATION: Hard to forego a practice that everyone expects

Self-monitoring serves different purposes for different populations. Blood glucose levels, along with HbA1c, can guide clinicians in making treatment decisions. Knowing blood sugar levels may be educational or empowering to patients, and provides critical information if hypoglycemia is a concern. These considerations lead us to conclude that while self-monitoring is not indicated for all newly diagnosed diabetic patients, it should be considered in selected circumstances.

Because of the prevalence of self-monitoring of blood glucose, patients may see it as a key component of an optimal self-management regimen. It may be hard to convince patients with newly diagnosed diabetes otherwise—and to convince some clinicians that there is little benefit in recommending it. Again, clinical judgment is required. We suspect, however, that with the proper explanation, many patients will be relieved to learn that they will not have to prick their fingers regularly or record their blood glucose.

Acknowledgements

The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

This study was selected and evaluated using FPIN’s Priority Updates from the Research Literature (PURL) Surveillance System methodology. The criteria and findings leading to the selection of this study as a PURL can be accessed at www.jfponline.com/purls.

ILLUSTRATIVE CASE

Two weeks ago, you informed your patient—a 53-year-old man with a body mass index of 28.4—that he has type 2 diabetes. Since then, he has seen a nutritionist and begun exercising regularly. His hemoglobin A1c (HbA1c) is 7.7%. You recommend that he begin taking metformin. The patient is worried about the potential for oral antidiabetic agents to cause hypoglycemia. He’s aware that many patients with diabetes monitor their blood sugar levels at home and wants to know if he should, too. You wonder whether it’s necessary, or even advisable, to initiate self-monitoring at this time.

For patients with type 2 diabetes, self-monitoring of blood glucose makes intuitive sense. Theoretically, it reinforces self-management behaviors, promotes adherence to the prescribed medication regimen, and leads to better glucose control. It seems obvious, too, that patients taking medications intended to lower blood sugar need to be aware of their glucose levels so they can take action to reduce the risk of complications.

But things that make sense intuitively do not always stand up to scrutiny. New high-quality evidence suggests that for those with newly diagnosed diabetes, self-monitoring of blood glucose may do more harm than good.

More questions than answers

While it is generally accepted that glucose self-monitoring is useful for those with insulin-treated type 2 diabetes,^2-4 evidence supporting the practice for patients with diabetes who do not require insulin is limited. Two recent meta-analyses of RCTs^5,6 found that self-monitoring of blood glucose achieves a statistically significant reduction of 0.4% in HbA1c; the quality of the studies, however, was limited. A well-designed RCT was needed, the researchers concluded, to settle questions about the value of self-monitoring.

The most recent Cochrane review⁷ of self-monitoring reached a similar conclusion: The reviewers called for additional research into the benefits of self-monitoring for patients with diabetes who do not need insulin. The reviewers also emphasized the need for information on patient-related outcomes such as quality of life, well-being, and satisfaction.

Are recommendations out of step?

Despite the lack of definitive evidence, the Department of Health and Human Services calls on us to increase the proportion of patients with diabetes who monitor their blood sugar at least once daily to 60% as part of its Healthy People 2010 initiative.⁸ The American Diabetes Association states that self-monitoring of blood glucose may help patients taking oral antidiabetic agents achieve glycemic goals.⁹ And the International Diabetes Federation recommends that self-monitoring of blood glucose be offered to all people with type 2 diabetes taking insulin or oral agents—and be part of the patient education that is given to all those who are newly diagnosed.¹⁰

But all of these groups may need to rethink their recommendations in light of the latest findings from the O’Kane RCT.

STUDY SUMMARY: Self-monitoring has little effect on glycemic control

O’Kane and colleagues conducted a prospective RCT comparing self-monitoring versus no monitoring among 184 people with newly diagnosed type 2 diabetes.¹ Patients were randomized to the self-monitoring or control group for 1 year, with clinic visits at 3-month intervals. Those who were already taking insulin or had engaged in self-monitoring of blood glucose were excluded.

At baseline, there was no significant difference in HbA1c, age, or sex between the 2 groups. Participants in both groups underwent identical diabetes education programs throughout the study period and received dietary and medical management based on the same treatment algorithm. Patients whose baseline HbA1c was >7.5% received metformin, followed by the sulfonylurea gliclazide if they did not reach target at the maximum dose of metformin. There was no significant difference in medication use at baseline or at 12 months.

Patients in the self-monitoring group were given glucose monitors and asked to record 4 fasting and 4 postprandial capillary blood glucose measurements per week. They were also taught to monitor and interpret blood glucose readings, and to respond appropriately to high or low readings.

At each follow-up visit, patients underwent blood tests for HbA1c, lipids, and electrolyte levels and completed questionnaires about treatment satisfaction, attitudes about diabetes, and levels of depression, anxiety, and well-being. Adherence to self-monitoring was verified by downloading meter readings. The dropout rate was low (2.2%), and adherence in the self-monitoring group was high. Study results were assessed using intent-to-treat analysis.

HbA1c fell in both the self-monitoring and control groups, with no significant differences at any point. The mean (standard deviation) value at 12 months was 6.9% (0.8%) in the self-monitoring group, compared to 6.9% (1.2%) in the control group, with a 95% confidence interval for the change in HbA1c of –0.25% to 0.38%. Throughout the study period, there was no difference in use of oral hypoglycemic medications or reported hypoglycemia.

Self-monitoring linked to depression

Measures of depression and anxiety were scored on a 100-point scale and compared to baseline measurements. At 12 months, participants in the self-monitoring group were more depressed, scoring 6% higher, on average, on the depression subscale of the well-being questionnaire (P=.01) than those in the control group. There was a trend toward increased anxiety in the self-monitoring group, but no significant differences in well-being, energy, or any of the other diabetes attitude subscales.

WHAT’S NEW: Less may be better

Because we emphasize self-management skills when we counsel patients about diabetes, it is surprising to learn that knowledge about glycemic control and blood sugar levels does not lead to better glycemic control. This RCT provides strong evidence that more information is not necessarily desirable, at least for patients with newly diagnosed type 2 diabetes who do not need insulin.

Depression is a known complication of diabetes. It affects an estimated 10% to 30% of patients with diabetes, who have double the odds of depression compared to people without diabetes.^11,12 Patients with depression and diabetes have poorer glycemic control,^13,14 an increased risk of complications,^15-17 a decreased quality of life,¹⁸ an increased disability burden,^19,20 and increased health care use and costs.^18,21,22 In addition, they face a significantly higher risk of death from all causes, beyond the risks associated with depression or diabetes alone.²³

CAVEATS: Patients on sulfonylureas may be an exception

This study used metformin as the initial oral medication, with sulfonylureas reserved for those who did not reach target glycemic control with maximum metformin therapy. The number of patients taking sulfonylureas was 11 in the self-monitoring group and 6 in the control group. Because hypoglycemia is a concern in patients taking sulfonylureas, there may be a role for self-monitoring of blood glucose in these patients.

Also of note: This study does not provide definitive evidence that self-monitoring of blood glucose causes harm. Although self-monitoring was associated with a 6% higher score on a depression subscale and a trend toward increased anxiety, overall satisfaction with treatment was similar in both groups. Additional studies are needed to better understand the relationship between self-monitoring and depression.

Self-monitoring may still be a good idea for certain patients, regardless of their diabetic medication regimen. When evaluating the potential benefits of self-monitoring of blood glucose, physicians should consider the individual’s predisposition to depression, among other concerns.

CHALLENGES TO IMPLEMENTATION: Hard to forego a practice that everyone expects

Self-monitoring serves different purposes for different populations. Blood glucose levels, along with HbA1c, can guide clinicians in making treatment decisions. Knowing blood sugar levels may be educational or empowering to patients, and provides critical information if hypoglycemia is a concern. These considerations lead us to conclude that while self-monitoring is not indicated for all newly diagnosed diabetic patients, it should be considered in selected circumstances.

Because of the prevalence of self-monitoring of blood glucose, patients may see it as a key component of an optimal self-management regimen. It may be hard to convince patients with newly diagnosed diabetes otherwise—and to convince some clinicians that there is little benefit in recommending it. Again, clinical judgment is required. We suspect, however, that with the proper explanation, many patients will be relieved to learn that they will not have to prick their fingers regularly or record their blood glucose.

Acknowledgements

The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

This study was selected and evaluated using FPIN’s Priority Updates from the Research Literature (PURL) Surveillance System methodology. The criteria and findings leading to the selection of this study as a PURL can be accessed at www.jfponline.com/purls.

ILLUSTRATIVE CASE

Your new patient is a 57-year-old African American man. His blood pressure is 150/95 mm Hg, fasting glucose 115 mg/dL, body mass index 32, and triglycerides 155 mg/dL; he is on no prior medications. During the course of his care you diagnose hypertension with metabolic syndrome and decide to recommend an antihypertensive. Thiazide-type diuretics are your standard initial therapy, but this patient has metabolic syndrome, and you know that certain antihypertensive agents have a more favorable metabolic profile than thiazide diuretics. Furthermore, metabolic differences among races have been touted as reason to use other agents in black patients. Should you recommend a thiazide diuretic, or another agent?

Until now, we’ve had no simple approach to treating hypertension in patients with metabolic syndrome—and half or more of our hypertensive patients over the age of 55 have this disorder.

Now, however, we can base decisions on clinical outcomes data from a subgroup analysis of the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT).¹ This new subgroup analysis supports use of thiazide-type diuretics in these patients—particularly in black patients—despite the more favorable metabolic profile of calcium channel blockers, alpha-blockers, and angiotensin-converting enzyme (ACE) inhibitors.

Cost is no longer as big a factor as it once was, now that ACE inhibitors and alpha-blockers, as well as thiazide diuretics, are available generically.

Does a better metabolic profile improve outcomes?

We have had reason to be concerned about the metabolic adverse effects of thiazide-type diuretics in the past. Studies published before this ALLHAT subgroup analysis showed that hydrochlorothiazide for essential hypertension had adverse effects on potassium, glucose, and lipid metabolism. Some speculated that these changes aggravate the metabolic changes in early diabetes^2-4 and may contribute to increased coronary heart disease risk.^5,6

FIGURE
Outcomes favor thiazides

We know many physicians who have adopted thiazide-type diuretics as the first-line treatment for hypertension in metabolic syndrome, but until now, data have been inadequate to support this decision.

A subgroup analysis from the ALLHAT¹ concludes: “The ALLHAT findings fail to support the preference for calcium channel blockers, alpha-blockers, or angiotensin-converting enzyme inhibitors compared with thiazide-type diuretics in patients with the metabolic syndrome, despite their more favorable metabolic profiles. This was particularly true for black participants.”

ACE inhibitors and ARBs

The metabolic benefits of ACE inhibitors and angiotensin-receptor blockers (ARBs) are widely known. In comparative studies prior to this ALLHAT subgroup analysis, ACE inhibitors were thought to be preferable to thiazide diuretics and beta-blockers for patients with obesity⁷ or the metabolic syndrome.⁸ These agents also protect against diabetic nephropathy.⁹ Other studies attribute additional vascular benefits to ACE inhibitors,¹⁰ beyond that of lowering blood pressure.

A 2005 meta-analysis by Abuissa et al¹¹ showed that ACE inhibitors and ARBs were associated with significant reductions in the incidence of newly diagnosed diabetes, which, in turn, might lead to reduced heart disease. That report concluded that use of ACE inhibitors or ARBs should be considered in patients with prediabetic conditions such as metabolic syndrome, hypertension, impaired fasting glucose, family history of diabetes, obesity, congestive heart failure, or coronary heart disease.

Calcium-channel blockers and alpha-blockers

Calcium-channel blockers and alpha-blockers also do not appear to have the adverse metabolic effects of thiazides, and have also been advocated over beta-blockers and thiazides for hypertensive patients with metabolic syndrome.^12-15

Racial differences

In a consensus statement developed before the findings from the ALLHAT subgroup analysis were available, it was noted that racial differences in metabolic syndrome may make the selection of antihypertensive agents particularly important in African American patients.¹⁶

ALLHAT and JNC7 recommendations

The 2002 ALLHAT demonstrated that chlorthalidone (a thiazide-type diuretic) is superior to lisinopril, amlodipine, and doxazosin in preventing 1 or more major forms of cardiovascular disease. No difference was observed, however, for fatal coronary heart disease, nonfatal myocardial infarction, or all-cause mortality.¹⁷ These findings persisted in sub-group analyses stratified by race, diabetic status, and level of renal function, but ALLHAT did not identify patients with metabolic syndrome a priori.

The ALLHAT influenced the 2003 Joint National Commission VII (JNC7) Report, which recommends thiazide diuretics for first-line treatment of hypertension in the absence of compelling indications to begin an alternative antihypertensive agent.¹⁸

Special consideration, but no recommendation. The JNC7 Report mentions the metabolic syndrome as a special consideration, but does not explicitly recommend a first-line therapy other than thiazides.

Anecdotally, we know many physicians who have adopted thiazide-type diuretics as the first-line treatment for hypertension in metabolic syndrome, but until now, data have been inadequate to support this decision.

STUDY SUMMARY: Chlorthalidone outcomes were equivalent or better

Wright and colleagues analyzed a sub-group¹ of the ALLHAT cohort, which consisted of 42,418 participants, aged ≥55, with hypertension and at least 1 other cardiovascular risk factor ( FIGURE ).

Patients were randomly assigned to therapy with chlorthalidone, amlodipine, lisinopril, or doxazosin. After randomization, if patients failed to reach the target blood pressure (<140/90 mm Hg) with their assigned therapy, they were started on atenolol, clonidine, or reserpine. If they required a third agent, they received hydralazine. The doxazosin arm was stopped early due to increased stroke and heart failure risk.

The ALLHAT was well done and designed for adequate power to evaluate clinical outcomes in racial subgroups, as well as the general population.

Outcomes were compared by race in hypertensive patients with and without metabolic syndrome.

A total of 23,077 (54%) patients met all criteria; 12,818 were black, 7327 (57%) of whom had metabolic syndrome.

Not surprisingly in a study of this size, the expected metabolic effects of all 4 antihypertensive agents were detected. Patients taking chlorthalidone had higher glucose levels (1–4 mg/dL) and higher levels of cholesterol, although these higher glucose and cholesterol levels were not statistically significant for all comparisons over time and between different drugs.

Outcomes in the chlorthalidone group were equivalent or superior to the 3 other therapies, generally. This pattern held true regardless of race ( TABLE ):

Heart failure rates were significantly higher in patients with metabolic syndrome across all treatments compared with chlorthalidone.

Combined cardiovascular disease rates were higher with lisinopril and doxazosin compared with chlorthalidone.

Stroke rates were higher among black participants only in the lisinopril group.

TABLE
Number needed to treat to prevent blood pressure-related adverse outcomes in patients with hypertension and metabolic syndrome

WHAT’S NEW: Most effective, least expensive

First-line use of thiazide diuretics for hypertension gained major support from the findings of the first ALLHAT report, published in 2002. A year later, JNC7 supported the practice. Yet questions have persisted about whether the choice of initial antihypertensive agent in patients with metabolic syndrome warrants special consideration.

The difference for one patient is small, but when you consider the high prevalence of hypertension, the cumulative benefit at a population level is significant. This subgroup analysis confirms that there is no harm, and potentially a small benefit, in using chlorthalidone as a first-line agent for treating hypertension in patients with metabolic syndrome, regardless of race—despite the measurable and presumably adverse effects of diuretic agents on metabolic measurements.

How large is the benefit of first-line thiazides, overall?

Although, statistically, the relative risks (RR) are not large, the sheer number of patients means that there is significant benefit to the selection of thiazides as first-line treatment in most patients.

CAVEATS: Is stroke a concern? Was follow-up sufficient?

In this study, the only finding of harm in the diuretic group was an increased risk for stroke compared with amlodipine among non-black patients with metabolic syndrome. While this finding does raise some uncertainty, we still think that, on balance, thiazides are the most beneficial, even in this subgroup, as there was a larger benefit in preventing heart failure.

Another theoretical possibility is that follow-up was too short to demonstrate harm from the metabolic effects of thiazides. However, the metabolic effects of thiazides are very small and we believe that the evidence of benefit shown during this study period easily outweighs any such theoretical harms.

We also assume that hydrochlorothiazide, a commonly prescribed thiazide, has the same benefits as chlorthalidone, the medication studied.

Most ALLHAT participants with metabolic syndrome already had diabetes: 67.6% of black participants and 51.8% of non-black participants. Another sub-group analysis of the ALLHAT studied patients with metabolic syndrome without diabetes, and found similar results.¹⁹

Of note, lisinopril reduced the onset of diabetes over 5 years (number needed to treat [NNT]=22.2), at the cost of increased heart failure (RR=1.31; 95% confidence interval [CI], 1.04-1.64) and combined cardiovascular disease (RR=1.19; 95% CI, 1.07-1.32). This potentially confounds the claim that thiazides are effective in preventing diabetes, since so many people had it to begin with.

The criteria for metabolic syndrome did not include waist circumference, which is the National Cholesterol Education Program definition. The World Health Organization definition, however, does allow substitution of BMI. Purists would have you believe waist circumference is necessary. In practice, we have come to use BMI as an adequate surrogate. Some say it has, in fact, replaced waist circumference.

CHALLENGES TO IMPLEMENTATION: Inertia

Few interventions are as simple as this. Thiazide diuretics are well tolerated, need to be taken only once daily, and are inexpensive. Because generics are available, little to no pharmaceutical marketing is done to promote their use. The major barriers to implementing this practice may be overcoming clinical inertia, and the message of pharmaceutical marketing on behalf of the more expensive alternatives.

Acknowledgement

We acknowledge Sofia Medvedev, PhD, University HealthSystem Consortium, Oak Brook, IL for analysis of the National Ambulatory Medical Care Survey data.

PURLs methodology

This study was selected and evaluated using FPIN’s Priority Updates from the Research Literature (PURL) Surveillance System methodology. The criteria and findings leading to the selection of this study as a PURL can be accessed at here.

ILLUSTRATIVE CASE

Your new patient is a 57-year-old African American man. His blood pressure is 150/95 mm Hg, fasting glucose 115 mg/dL, body mass index 32, and triglycerides 155 mg/dL; he is on no prior medications. During the course of his care you diagnose hypertension with metabolic syndrome and decide to recommend an antihypertensive. Thiazide-type diuretics are your standard initial therapy, but this patient has metabolic syndrome, and you know that certain antihypertensive agents have a more favorable metabolic profile than thiazide diuretics. Furthermore, metabolic differences among races have been touted as reason to use other agents in black patients. Should you recommend a thiazide diuretic, or another agent?

Until now, we’ve had no simple approach to treating hypertension in patients with metabolic syndrome—and half or more of our hypertensive patients over the age of 55 have this disorder.

Now, however, we can base decisions on clinical outcomes data from a subgroup analysis of the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT).¹ This new subgroup analysis supports use of thiazide-type diuretics in these patients—particularly in black patients—despite the more favorable metabolic profile of calcium channel blockers, alpha-blockers, and angiotensin-converting enzyme (ACE) inhibitors.

Cost is no longer as big a factor as it once was, now that ACE inhibitors and alpha-blockers, as well as thiazide diuretics, are available generically.

Does a better metabolic profile improve outcomes?

We have had reason to be concerned about the metabolic adverse effects of thiazide-type diuretics in the past. Studies published before this ALLHAT subgroup analysis showed that hydrochlorothiazide for essential hypertension had adverse effects on potassium, glucose, and lipid metabolism. Some speculated that these changes aggravate the metabolic changes in early diabetes^2-4 and may contribute to increased coronary heart disease risk.^5,6

FIGURE
Outcomes favor thiazides

We know many physicians who have adopted thiazide-type diuretics as the first-line treatment for hypertension in metabolic syndrome, but until now, data have been inadequate to support this decision.

A subgroup analysis from the ALLHAT¹ concludes: “The ALLHAT findings fail to support the preference for calcium channel blockers, alpha-blockers, or angiotensin-converting enzyme inhibitors compared with thiazide-type diuretics in patients with the metabolic syndrome, despite their more favorable metabolic profiles. This was particularly true for black participants.”

ACE inhibitors and ARBs

The metabolic benefits of ACE inhibitors and angiotensin-receptor blockers (ARBs) are widely known. In comparative studies prior to this ALLHAT subgroup analysis, ACE inhibitors were thought to be preferable to thiazide diuretics and beta-blockers for patients with obesity⁷ or the metabolic syndrome.⁸ These agents also protect against diabetic nephropathy.⁹ Other studies attribute additional vascular benefits to ACE inhibitors,¹⁰ beyond that of lowering blood pressure.

A 2005 meta-analysis by Abuissa et al¹¹ showed that ACE inhibitors and ARBs were associated with significant reductions in the incidence of newly diagnosed diabetes, which, in turn, might lead to reduced heart disease. That report concluded that use of ACE inhibitors or ARBs should be considered in patients with prediabetic conditions such as metabolic syndrome, hypertension, impaired fasting glucose, family history of diabetes, obesity, congestive heart failure, or coronary heart disease.

Calcium-channel blockers and alpha-blockers

Calcium-channel blockers and alpha-blockers also do not appear to have the adverse metabolic effects of thiazides, and have also been advocated over beta-blockers and thiazides for hypertensive patients with metabolic syndrome.^12-15

Racial differences

In a consensus statement developed before the findings from the ALLHAT subgroup analysis were available, it was noted that racial differences in metabolic syndrome may make the selection of antihypertensive agents particularly important in African American patients.¹⁶

ALLHAT and JNC7 recommendations

The 2002 ALLHAT demonstrated that chlorthalidone (a thiazide-type diuretic) is superior to lisinopril, amlodipine, and doxazosin in preventing 1 or more major forms of cardiovascular disease. No difference was observed, however, for fatal coronary heart disease, nonfatal myocardial infarction, or all-cause mortality.¹⁷ These findings persisted in sub-group analyses stratified by race, diabetic status, and level of renal function, but ALLHAT did not identify patients with metabolic syndrome a priori.

The ALLHAT influenced the 2003 Joint National Commission VII (JNC7) Report, which recommends thiazide diuretics for first-line treatment of hypertension in the absence of compelling indications to begin an alternative antihypertensive agent.¹⁸

Special consideration, but no recommendation. The JNC7 Report mentions the metabolic syndrome as a special consideration, but does not explicitly recommend a first-line therapy other than thiazides.

Anecdotally, we know many physicians who have adopted thiazide-type diuretics as the first-line treatment for hypertension in metabolic syndrome, but until now, data have been inadequate to support this decision.

STUDY SUMMARY: Chlorthalidone outcomes were equivalent or better

Wright and colleagues analyzed a sub-group¹ of the ALLHAT cohort, which consisted of 42,418 participants, aged ≥55, with hypertension and at least 1 other cardiovascular risk factor ( FIGURE ).

Patients were randomly assigned to therapy with chlorthalidone, amlodipine, lisinopril, or doxazosin. After randomization, if patients failed to reach the target blood pressure (<140/90 mm Hg) with their assigned therapy, they were started on atenolol, clonidine, or reserpine. If they required a third agent, they received hydralazine. The doxazosin arm was stopped early due to increased stroke and heart failure risk.

The ALLHAT was well done and designed for adequate power to evaluate clinical outcomes in racial subgroups, as well as the general population.

Outcomes were compared by race in hypertensive patients with and without metabolic syndrome.

A total of 23,077 (54%) patients met all criteria; 12,818 were black, 7327 (57%) of whom had metabolic syndrome.

Not surprisingly in a study of this size, the expected metabolic effects of all 4 antihypertensive agents were detected. Patients taking chlorthalidone had higher glucose levels (1–4 mg/dL) and higher levels of cholesterol, although these higher glucose and cholesterol levels were not statistically significant for all comparisons over time and between different drugs.

Outcomes in the chlorthalidone group were equivalent or superior to the 3 other therapies, generally. This pattern held true regardless of race ( TABLE ):

Heart failure rates were significantly higher in patients with metabolic syndrome across all treatments compared with chlorthalidone.

Combined cardiovascular disease rates were higher with lisinopril and doxazosin compared with chlorthalidone.

Stroke rates were higher among black participants only in the lisinopril group.

TABLE
Number needed to treat to prevent blood pressure-related adverse outcomes in patients with hypertension and metabolic syndrome

WHAT’S NEW: Most effective, least expensive

First-line use of thiazide diuretics for hypertension gained major support from the findings of the first ALLHAT report, published in 2002. A year later, JNC7 supported the practice. Yet questions have persisted about whether the choice of initial antihypertensive agent in patients with metabolic syndrome warrants special consideration.

The difference for one patient is small, but when you consider the high prevalence of hypertension, the cumulative benefit at a population level is significant. This subgroup analysis confirms that there is no harm, and potentially a small benefit, in using chlorthalidone as a first-line agent for treating hypertension in patients with metabolic syndrome, regardless of race—despite the measurable and presumably adverse effects of diuretic agents on metabolic measurements.

How large is the benefit of first-line thiazides, overall?

Although, statistically, the relative risks (RR) are not large, the sheer number of patients means that there is significant benefit to the selection of thiazides as first-line treatment in most patients.

CAVEATS: Is stroke a concern? Was follow-up sufficient?

In this study, the only finding of harm in the diuretic group was an increased risk for stroke compared with amlodipine among non-black patients with metabolic syndrome. While this finding does raise some uncertainty, we still think that, on balance, thiazides are the most beneficial, even in this subgroup, as there was a larger benefit in preventing heart failure.

Another theoretical possibility is that follow-up was too short to demonstrate harm from the metabolic effects of thiazides. However, the metabolic effects of thiazides are very small and we believe that the evidence of benefit shown during this study period easily outweighs any such theoretical harms.

We also assume that hydrochlorothiazide, a commonly prescribed thiazide, has the same benefits as chlorthalidone, the medication studied.

Most ALLHAT participants with metabolic syndrome already had diabetes: 67.6% of black participants and 51.8% of non-black participants. Another sub-group analysis of the ALLHAT studied patients with metabolic syndrome without diabetes, and found similar results.¹⁹

Of note, lisinopril reduced the onset of diabetes over 5 years (number needed to treat [NNT]=22.2), at the cost of increased heart failure (RR=1.31; 95% confidence interval [CI], 1.04-1.64) and combined cardiovascular disease (RR=1.19; 95% CI, 1.07-1.32). This potentially confounds the claim that thiazides are effective in preventing diabetes, since so many people had it to begin with.

The criteria for metabolic syndrome did not include waist circumference, which is the National Cholesterol Education Program definition. The World Health Organization definition, however, does allow substitution of BMI. Purists would have you believe waist circumference is necessary. In practice, we have come to use BMI as an adequate surrogate. Some say it has, in fact, replaced waist circumference.

CHALLENGES TO IMPLEMENTATION: Inertia

Few interventions are as simple as this. Thiazide diuretics are well tolerated, need to be taken only once daily, and are inexpensive. Because generics are available, little to no pharmaceutical marketing is done to promote their use. The major barriers to implementing this practice may be overcoming clinical inertia, and the message of pharmaceutical marketing on behalf of the more expensive alternatives.

Acknowledgement

We acknowledge Sofia Medvedev, PhD, University HealthSystem Consortium, Oak Brook, IL for analysis of the National Ambulatory Medical Care Survey data.

PURLs methodology

This study was selected and evaluated using FPIN’s Priority Updates from the Research Literature (PURL) Surveillance System methodology. The criteria and findings leading to the selection of this study as a PURL can be accessed at here.

ILLUSTRATIVE CASE

Your new patient is a 57-year-old African American man. His blood pressure is 150/95 mm Hg, fasting glucose 115 mg/dL, body mass index 32, and triglycerides 155 mg/dL; he is on no prior medications. During the course of his care you diagnose hypertension with metabolic syndrome and decide to recommend an antihypertensive. Thiazide-type diuretics are your standard initial therapy, but this patient has metabolic syndrome, and you know that certain antihypertensive agents have a more favorable metabolic profile than thiazide diuretics. Furthermore, metabolic differences among races have been touted as reason to use other agents in black patients. Should you recommend a thiazide diuretic, or another agent?

Until now, we’ve had no simple approach to treating hypertension in patients with metabolic syndrome—and half or more of our hypertensive patients over the age of 55 have this disorder.

Now, however, we can base decisions on clinical outcomes data from a subgroup analysis of the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT).¹ This new subgroup analysis supports use of thiazide-type diuretics in these patients—particularly in black patients—despite the more favorable metabolic profile of calcium channel blockers, alpha-blockers, and angiotensin-converting enzyme (ACE) inhibitors.

Cost is no longer as big a factor as it once was, now that ACE inhibitors and alpha-blockers, as well as thiazide diuretics, are available generically.

Does a better metabolic profile improve outcomes?

We have had reason to be concerned about the metabolic adverse effects of thiazide-type diuretics in the past. Studies published before this ALLHAT subgroup analysis showed that hydrochlorothiazide for essential hypertension had adverse effects on potassium, glucose, and lipid metabolism. Some speculated that these changes aggravate the metabolic changes in early diabetes^2-4 and may contribute to increased coronary heart disease risk.^5,6

FIGURE
Outcomes favor thiazides

We know many physicians who have adopted thiazide-type diuretics as the first-line treatment for hypertension in metabolic syndrome, but until now, data have been inadequate to support this decision.

A subgroup analysis from the ALLHAT¹ concludes: “The ALLHAT findings fail to support the preference for calcium channel blockers, alpha-blockers, or angiotensin-converting enzyme inhibitors compared with thiazide-type diuretics in patients with the metabolic syndrome, despite their more favorable metabolic profiles. This was particularly true for black participants.”

ACE inhibitors and ARBs

The metabolic benefits of ACE inhibitors and angiotensin-receptor blockers (ARBs) are widely known. In comparative studies prior to this ALLHAT subgroup analysis, ACE inhibitors were thought to be preferable to thiazide diuretics and beta-blockers for patients with obesity⁷ or the metabolic syndrome.⁸ These agents also protect against diabetic nephropathy.⁹ Other studies attribute additional vascular benefits to ACE inhibitors,¹⁰ beyond that of lowering blood pressure.

A 2005 meta-analysis by Abuissa et al¹¹ showed that ACE inhibitors and ARBs were associated with significant reductions in the incidence of newly diagnosed diabetes, which, in turn, might lead to reduced heart disease. That report concluded that use of ACE inhibitors or ARBs should be considered in patients with prediabetic conditions such as metabolic syndrome, hypertension, impaired fasting glucose, family history of diabetes, obesity, congestive heart failure, or coronary heart disease.

Calcium-channel blockers and alpha-blockers

Calcium-channel blockers and alpha-blockers also do not appear to have the adverse metabolic effects of thiazides, and have also been advocated over beta-blockers and thiazides for hypertensive patients with metabolic syndrome.^12-15

Racial differences

In a consensus statement developed before the findings from the ALLHAT subgroup analysis were available, it was noted that racial differences in metabolic syndrome may make the selection of antihypertensive agents particularly important in African American patients.¹⁶

ALLHAT and JNC7 recommendations

The 2002 ALLHAT demonstrated that chlorthalidone (a thiazide-type diuretic) is superior to lisinopril, amlodipine, and doxazosin in preventing 1 or more major forms of cardiovascular disease. No difference was observed, however, for fatal coronary heart disease, nonfatal myocardial infarction, or all-cause mortality.¹⁷ These findings persisted in sub-group analyses stratified by race, diabetic status, and level of renal function, but ALLHAT did not identify patients with metabolic syndrome a priori.

The ALLHAT influenced the 2003 Joint National Commission VII (JNC7) Report, which recommends thiazide diuretics for first-line treatment of hypertension in the absence of compelling indications to begin an alternative antihypertensive agent.¹⁸

Special consideration, but no recommendation. The JNC7 Report mentions the metabolic syndrome as a special consideration, but does not explicitly recommend a first-line therapy other than thiazides.

Anecdotally, we know many physicians who have adopted thiazide-type diuretics as the first-line treatment for hypertension in metabolic syndrome, but until now, data have been inadequate to support this decision.

STUDY SUMMARY: Chlorthalidone outcomes were equivalent or better

Wright and colleagues analyzed a sub-group¹ of the ALLHAT cohort, which consisted of 42,418 participants, aged ≥55, with hypertension and at least 1 other cardiovascular risk factor ( FIGURE ).

Patients were randomly assigned to therapy with chlorthalidone, amlodipine, lisinopril, or doxazosin. After randomization, if patients failed to reach the target blood pressure (<140/90 mm Hg) with their assigned therapy, they were started on atenolol, clonidine, or reserpine. If they required a third agent, they received hydralazine. The doxazosin arm was stopped early due to increased stroke and heart failure risk.

The ALLHAT was well done and designed for adequate power to evaluate clinical outcomes in racial subgroups, as well as the general population.

Outcomes were compared by race in hypertensive patients with and without metabolic syndrome.

A total of 23,077 (54%) patients met all criteria; 12,818 were black, 7327 (57%) of whom had metabolic syndrome.

Not surprisingly in a study of this size, the expected metabolic effects of all 4 antihypertensive agents were detected. Patients taking chlorthalidone had higher glucose levels (1–4 mg/dL) and higher levels of cholesterol, although these higher glucose and cholesterol levels were not statistically significant for all comparisons over time and between different drugs.

Outcomes in the chlorthalidone group were equivalent or superior to the 3 other therapies, generally. This pattern held true regardless of race ( TABLE ):

Heart failure rates were significantly higher in patients with metabolic syndrome across all treatments compared with chlorthalidone.

Combined cardiovascular disease rates were higher with lisinopril and doxazosin compared with chlorthalidone.

Stroke rates were higher among black participants only in the lisinopril group.

TABLE
Number needed to treat to prevent blood pressure-related adverse outcomes in patients with hypertension and metabolic syndrome

WHAT’S NEW: Most effective, least expensive

First-line use of thiazide diuretics for hypertension gained major support from the findings of the first ALLHAT report, published in 2002. A year later, JNC7 supported the practice. Yet questions have persisted about whether the choice of initial antihypertensive agent in patients with metabolic syndrome warrants special consideration.

The difference for one patient is small, but when you consider the high prevalence of hypertension, the cumulative benefit at a population level is significant. This subgroup analysis confirms that there is no harm, and potentially a small benefit, in using chlorthalidone as a first-line agent for treating hypertension in patients with metabolic syndrome, regardless of race—despite the measurable and presumably adverse effects of diuretic agents on metabolic measurements.

How large is the benefit of first-line thiazides, overall?

Although, statistically, the relative risks (RR) are not large, the sheer number of patients means that there is significant benefit to the selection of thiazides as first-line treatment in most patients.

CAVEATS: Is stroke a concern? Was follow-up sufficient?

In this study, the only finding of harm in the diuretic group was an increased risk for stroke compared with amlodipine among non-black patients with metabolic syndrome. While this finding does raise some uncertainty, we still think that, on balance, thiazides are the most beneficial, even in this subgroup, as there was a larger benefit in preventing heart failure.

Another theoretical possibility is that follow-up was too short to demonstrate harm from the metabolic effects of thiazides. However, the metabolic effects of thiazides are very small and we believe that the evidence of benefit shown during this study period easily outweighs any such theoretical harms.

We also assume that hydrochlorothiazide, a commonly prescribed thiazide, has the same benefits as chlorthalidone, the medication studied.

Most ALLHAT participants with metabolic syndrome already had diabetes: 67.6% of black participants and 51.8% of non-black participants. Another sub-group analysis of the ALLHAT studied patients with metabolic syndrome without diabetes, and found similar results.¹⁹

Of note, lisinopril reduced the onset of diabetes over 5 years (number needed to treat [NNT]=22.2), at the cost of increased heart failure (RR=1.31; 95% confidence interval [CI], 1.04-1.64) and combined cardiovascular disease (RR=1.19; 95% CI, 1.07-1.32). This potentially confounds the claim that thiazides are effective in preventing diabetes, since so many people had it to begin with.

The criteria for metabolic syndrome did not include waist circumference, which is the National Cholesterol Education Program definition. The World Health Organization definition, however, does allow substitution of BMI. Purists would have you believe waist circumference is necessary. In practice, we have come to use BMI as an adequate surrogate. Some say it has, in fact, replaced waist circumference.

CHALLENGES TO IMPLEMENTATION: Inertia

Few interventions are as simple as this. Thiazide diuretics are well tolerated, need to be taken only once daily, and are inexpensive. Because generics are available, little to no pharmaceutical marketing is done to promote their use. The major barriers to implementing this practice may be overcoming clinical inertia, and the message of pharmaceutical marketing on behalf of the more expensive alternatives.

Acknowledgement

We acknowledge Sofia Medvedev, PhD, University HealthSystem Consortium, Oak Brook, IL for analysis of the National Ambulatory Medical Care Survey data.

PURLs methodology

This study was selected and evaluated using FPIN’s Priority Updates from the Research Literature (PURL) Surveillance System methodology. The criteria and findings leading to the selection of this study as a PURL can be accessed at here.