0
Original Investigation |

Fasting Glucose Levels and Incident Diabetes Mellitus in Older Nondiabetic Adults Randomized to Receive 3 Different Classes of Antihypertensive Treatment:  A Report From the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) FREE

Joshua I. Barzilay, MD; Barry R. Davis, MD, PhD; Jeffrey A. Cutler, MD, MPH; Sara L. Pressel, MS; Paul K. Whelton, MD, MSC; Jan Basile, MD; Karen L. Margolis, MD, MPH; Stephen T. Ong, MD; Laurie S. Sadler, MD; John Summerson, MS; ALLHAT Collaborative Research Group
[+] Author Affiliations

Author Affiliations: Kaiser Permanente of Georgia and the Division of Endocrinology, Emory University School of Medicine, Atlanta (Dr Barzilay); Coordinating Center for Clinical Trials, University of Texas School of Public Health, Houston (Dr Davis and Ms Pressel); Division of Epidemiology and Clinical Applications, National Heart, Lung, and Blood Institute, Bethesda, Md (Dr Cutler); Tulane University Health Sciences Center, New Orleans, La (Dr Whelton); Ralph H. Johnson Veterans Affairs Medical Center, Medical University of South Carolina, Charleston (Dr Basile); Hennepin County Medical Center, Division of Clinical Epidemiology, Minneapolis, Minn (Dr Margolis); Ong Medical Center, Oxon Hill, Md (Dr Ong); Lipid Research Center, St Vincent Charity Hospital, Cleveland, Ohio (Dr Sadler); and Department of Public Health Sciences, Wake Forest University School of Medicine, Winston-Salem, NC (Mr Summerson).


Arch Intern Med. 2006;166(20):2191-2201. doi:10.1001/archinte.166.20.2191.
Text Size: A A A
Published online

Background  Elevated blood glucose levels are reported with thiazide-type diuretic treatment of hypertension. The significance of this finding is uncertain. Our objectives were to compare the effect of first-step antihypertensive drug therapy with thiazide-type diuretic, calcium-channel blocker, or angiotensin-converting enzyme inhibitor on fasting glucose (FG) levels and to determine cardiovascular and renal disease risks associated with elevated FG levels and incident diabetes mellitus (DM) in 3 treatment groups.

Methods  We performed post hoc subgroup analyses from the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) among nondiabetic participants who were randomized to receive treatment with chlorthalidone (n = 8419), amlodipine (n = 4958), or lisinopril (n = 5034) and observed for a mean of 4.9 years.

Results  Mean FG levels increased during follow-up in all treatment groups. At year 2, those randomized to the chlorthalidone group had the greatest increase (+8.5 mg/dL [0.47 mmol/L] vs +5.5 mg/dL [0.31 mmol/L] for amlodipine and +3.5 mg/dL [0.19 mmol/L] for lisinopril). The odds ratios for developing DM with lisinopril (0.55 [95% confidence interval, 0.43-0.70]) or amlodipine (0.73 [95% confidence interval, 0.58-0.91]) vs chlorthalidone at 2 years were significantly lower than 1.0 (P<.01). There was no significant association of FG level change at 2 years with subsequent coronary heart disease, stroke, cardiovascular disease, total mortality, or end-stage renal disease. There was no significant association of incident DM at 2 years with clinical outcomes, except for coronary heart disease (risk ratio, 1.64; P = .006), but the risk ratio was lower and nonsignificant in the chlorthalidone group (risk ratio, 1.46; P = .14).

Conclusions  Fasting glucose levels increase in older adults with hypertension regardless of treatment type. For those taking chlorthalidone vs other medications, the risk of developing FG levels higher than 125 mg/dL (6.9 mmol/L) is modestly greater, but there is no conclusive or consistent evidence that this diuretic-associated increase in DM risk increases the risk of clinical events.

Figures in this Article

Diabetes mellitus (DM) and hypertension (HTN) share several common antecedents, including obesity and insulin resistance.1 For this reason, treatment of one disorder may affect the other. For example, β-blocker or thiazide diuretic administration in nondiabetic adults may lead to raised blood glucose levels.2,3 Some have suggested that diuretic administration in people with DM may increase morbidity and mortality4; the opposite has been observed in large randomized trials.5 Others have suggested that administration of angiotensin-converting enzyme (ACE) inhibitors or angiotensin receptor blockers may result in lower rates of incident DM,1,69 perhaps contributing to protection against cardiovascular disease (CVD).10 Given the growing prevalence of DM, insulin resistance, and their associated CVD burden, elucidating these relationships is important.

The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT)11 was a multicenter randomized clinical trial designed to determine whether the occurrence of fatal coronary heart disease (CHD) or nonfatal myocardial infarction is lower for high-risk patients with hypertension treated initially with a calcium-channel blocker or an ACE inhibitor compared with a diuretic. No differences were found for CHD risk. However, chlorthalidone-based treatment resulted in significantly lower heart failure risk.11 In the present study, we conducted a post hoc analysis from the ALLHAT to investigate differences in elevated fasting glucose (FG) levels, change in FG levels, incident DM rates, and risk of CVD and renal disease by treatment assignment in participants without DM at baseline. To our knowledge, this is the largest study to examine these issues.

Details of the ALLHAT design have been published.12 In brief, the trial compared first-step treatment with a diuretic (chlorthalidone) with an ACE inhibitor (lisinopril) or a calcium-channel blocker (amlodipine). A fourth arm of the study, using an α-blocker (doxazosin) (n = 9061), was stopped early13 and is not considered in this report.

ELIGIBILITY

Eligible participants were at least 55 years old and had a systolic blood pressure of 140 mm Hg or higher, and/or a diastolic blood pressure of 90 mm Hg or higher, and/or were taking HTN medication (<3 drugs) with a blood pressure of 160/100 mm Hg or lower at randomization, and had at least 1 additional CHD risk factor. The latter included previous myocardial infarction or stroke, other manifestations of atherosclerotic CVD, left ventricular hypertrophy, a history of type 2 DM, current smoking, or low high-density lipoprotein cholesterol level.

MEDICATIONS

Step-1 medications were formulated to look alike, so that identity of each agent was double masked. The dose of each step-1 blinded medication was titrated in an attempt to achieve a blood pressure lower than 140/90 mm Hg. If the blood pressure could not be controlled using the maximum dose of step-1 medication, open-label step-2 (reserpine, clonidine hydrochloride, or atenolol) and step-3 (hydralazine hydrochloride) medications were available. The choice and dose of step-2 and -3 medications were at the discretion of the investigator. Other drugs, including low doses of open-label step-1 drug classes, were permitted if clinically indicated. Potassium supplementation (potassium chloride) was mandated when a local recheck confirmed hypokalemia (potassium level <3.2 mEq/L).

LABORATORY TESTS

Baseline laboratory test results for glucose, lipids, creatinine, and potassium levels were obtained after an overnight fast (8 or more hours without food). Analyses were done in a certified laboratory. At years 2, 4, and 6, fasting total cholesterol and glucose levels were evaluated again. Serum potassium and creatinine levels were measured at 1 month and at years 1, 2, 4, and 6.

COHORTS FOR ANALYSIS

Baseline DM was defined by physician diagnosis (n = 12 063). Evidence confirming diabetic status was not systematically sought. Those entering the trial without a DM diagnosis but with an FG level higher than 125 mg/dL (6.9 mmol/L) were defined as newly diagnosed DM cases (n = 1038). Participants with newly diagnosed DM and known DM were excluded from analyses. Those with no baseline FG measurement or a baseline glucose level (nonfasting) of 110 mg/dL (6.1 mmol/L) or higher were also excluded (n = 1845) (Figure 1).

Place holder to copy figure label and caption
Figure 1.

Randomization and follow-up of participants in the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT)11 by baseline glycemic status for participants without diabetes mellitus at baseline. To convert fasting glucose to millimoles per liter, multiply by 0.0555. *Per ALLHAT database as of September 30, 2002.

Graphic Jump Location
END POINTS

The primary outcome was nonfatal myocardial infarction plus CHD death. Major prespecified secondary outcomes were (1) combined CHD (primary outcome, coronary revascularization, or hospitalized angina), (2) fatal and nonfatal stroke, (3) combined CVD (primary outcome, stroke, combined CHD, other nonhospitalized treated angina, heart failure [fatal, hospitalized, or treated but not hospitalized], or peripheral arterial disease), and (4) all-cause mortality. Development of end-stage renal disease—requiring dialysis and/or transplantation—was also a prespecified secondary end point.

STATISTICAL ANALYSIS

To compare baseline characteristics of participants assigned to amlodipine or lisinopril vs chlorthalidone, we used the χ2 and z-tests. To analyze effects of type of HTN treatment on development of incident DM, we used logistic regression models (limited to those with ≥1 follow-up FG measurements available) over 3 periods (0-2 years, 2-4 years, and 4-6 years) adjusting for age, sex, race, FG level at the beginning of the interval, body mass index, and cigarette smoking. We also adjusted for the following factors (which have been observed to influence FG levels): β-blocker treatment,2,14 hypokalemia,15,16 and statin treatment.17 Specific first-step antihypertensive medications were entered into analyses as randomization assignment to gauge intention-to-treat effect on DM development. Kaplan-Meier curves were plotted and Cox regression analyses done using “time to first fasting glucose level higher than 125 mg/dL (6.9 mmol/L).” These were similar to the “higher than 125 mg/dL (6.9 mmol/L) ever” results and are therefore not shown.

The effects of FG level at 2 years, change in FG level from baseline to 2 years, and incident DM at 2 years on subsequent clinical events were assessed using Cox regression models. Possible interactions of the FG level variables with treatment group were assessed. Given the many multivariate, subgroup, and interaction analyses performed, statistical significance at the .05 level should be interpreted with caution.

Of 18 411 participants without diabetes at baseline, 9802 (53.2%) had 1 or more FG evaluations during follow-up. Of those without available follow-up FG values, about half had glucose values that were not fasting. Baseline characteristics of those with and without follow-up FG values were similar except for race and sex. Participants without follow-up FG values were more likely to be black or female (data not shown). Those randomized to lisinopril treatment were less likely (P<.001) to have follow-up FG evaluations (51.0%) than those taking chlorthalidone or amlodipine (54.0% and 54.3%, respectively). The cohort for the present study includes only those participants for whom follow-up FG values were available.

Baseline characteristics of the nondiabetic participants with follow-up FG values according to medication assignment are listed in Table 1. There were a few small but statistically significant differences between groups with regard to the percentage of women, pulse rate, serum potassium and serum creatinine levels, and the distribution of baseline glucose levels.

Table Graphic Jump LocationTable 1. Baseline Characteristics by Treatment Group of ALLHAT11 Antihypertensive Component Participants Without Diabetes Mellitus at Baseline*
MEDICATION ADMINISTRATION

Blinded medication administration decreased over the course of follow-up (mostly in the first year), whereas administration of open-label medications increased steadily (Table 2). This was especially so for participants assigned to lisinopril (blinded medication use at year 4: lisinopril, 76.3%; amlodipine, 84.7%; chlorthalidone, 83.2%). Except for year 1, there was no difference between treatment groups with regard to atenolol administration. Those taking diuretics were more likely to be given potassium supplements, which increased from about 7% at year 1 to about 22% at year 5.

Table Graphic Jump LocationTable 2. Administration of Blinded Study Drug, Open-Label Diuretics, ACE Inhibitors, Calcium-Channel Blockers, β-Blockers, Potassium Supplements, and Statins by Antihypertensive Treatment Group*
INCIDENT ELEVATED FG LEVELS

During follow-up, mean FG levels increased from baseline in all treatment groups (Table 3). During the first 2 years, those in the chlorthalidone group had a greater increase in mean FG level (8.5 mg/dL [0.47 mmol/L]) and in incident DM (9.3%) than those taking amlodipine (5.5 mg/dL [0.31 mmol/L] and 7.2%, respectively) (P<.001 for FG level; P = .008 for increase in incident DM). Those taking lisinopril had an even smaller increase in FG level (3.5 mg/dL [0.19 mmol/L]) (P<.001) and less incident DM (5.6%) (P<.001). From baseline to year 4, there were 1.5-mg/dL (0.08-mmol/L) and 4.0-mg/dL (0.22-mmol/L) smaller mean FG changes for those taking amlodipine (P = .07) and lisinopril (P<.001) than for those taking chlorthalidone. Overall, the risk of FG level ever being higher than 125 mg/dL (6.9 mmol/L) was 2.9% higher among participants taking chlorthalidone than for those taking amlodipine (P<.001) and 4.5% higher than for those taking lisinopril (P<.001).

Table Graphic Jump LocationTable 3. Follow-up Fasting Glucose Levels and Incident Diabetes Mellitus in Participants Without Diabetes Mellitus at Baseline by Treatment Group*

Almost all participants who developed incident DM at year 2 had FG level increases greater than 10 mg/dL (0.56 mmol/L), regardless of treatment group (Table 4). Among those who developed DM, there was little difference in median change in FG level between those treated with chlorthalidone and amlodipine (+40.0 vs +38.0 mg/dL [+2.2 vs +2.1 mmol/L]). Those treated with lisinopril had a lower median FG level change (+33.0 mg/dl [+1.8 mmol/L]).

Table Graphic Jump LocationTable 4. Baseline and 2-Year Change in Fasting Glucose Levels for Those With and Without Incident Diabetes Mellitus*
ODDS OF DEVELOPING DM

During the first 2 years, the odds for developing DM were lower for participants taking lisinopril (odds ratio [OR], 0.55 [95% confidence interval {CI}, 0.43-0.70]) or amlodipine (OR, 0.73 [95% CI, 0.58-0.91]) than for those taking chlorthalidone (Table 5). At years 4 and 6, assignment to amlodipine or lisinopril was still associated with lower odds of developing DM, but the ORs were no longer statistically significant. Higher initial FG level, systolic blood pressure, and body mass index were associated with higher odds of developing DM from baseline to year 2 and from year 2 to year 4. Sex and age had inconsistent relationships with incident DM. Treatment with atenolol, treatment with statin medications, and hypokalemia (potassium level <3.2 mEq/dL), whether or not requiring potassium supplementation, had no significant effect on the odds of developing DM; the OR for years 4 to 6 tended to be increased for those taking supplements.

Table Graphic Jump LocationTable 5. Logistic Regression Analysis of the Association of Baseline and Follow-up Factors With Incident Diabetes Mellitus in Those Without Diabetes Mellitus at Baseline*
ASSOCIATION OF CHANGE IN FG LEVEL WITH CVD AND RENAL END POINTS

There was no significant effect on any of the end points in the total group and in those treated with chlorthalidone or amlodipine when the effects of a change in FG level was considered a continuous variable (Figure 2). Among those treated with lisinopril, a 10-mg/dL (0.56-mmol/L) increase in FG level during the first years of follow-up was associated with a significantly increased risk of CHD (hazard ratio [HR], 1.09 [95% CI, 1.01-1.17]) and combined CVD (HR, 1.06 [95% CI, 1.00-1.13]). Interaction terms between change in FG level and treatment with amlodipine or lisinopril vs treatment with chlorthalidone were not significant. This suggests that the absence of an effect of change in FG level was present in all treatment groups.

Place holder to copy figure label and caption
Figure 2.

Cox regression models showing the hazard ratios (95% confidence intervals) associated with a 10-mg/dL (0.56-mmol/L) increase in fasting glucose level during the first 2 years of follow-up for subsequent cardiovascular disease and renal end points in participants without diabetes mellitus at baseline. CHD indicates coronary heart disease; CCVD, combined cardiovascular disease (ie, CHD death, nonfatal myocardial infarction, coronary revascularization procedures, stroke, hospitalization or treatment for angina or heart failure, or peripheral arterial disease involving hospitalization or outpatient revascularization); ESRD, end-stage renal disease. All hazard ratios controlled for treatment group (total cohort), 2-year blood pressure, age, race, sex, smoking status, baseline fasting glucose level, baseline body mass index, 2-year serum potassium level, and atenolol and statin administration at 2 years.

Graphic Jump Location
ASSOCIATION OF INCIDENT DM WITH OUTCOMES

The effects of incident DM (an FG level >125 mg/dL [6.9 mmol/L]) at year 2 of follow-up on study outcomes are shown in Figure 3. In the combined treatment groups, incident DM was associated with a statistically significant increased risk of CHD (HR, 1.64 [95% CI, 1.15-2.32]). Among those taking chlorthalidone, there was no significant increase in any outcome in association with incident DM, and except for stroke, the relative risk estimates were lower than for the other 2 treatment arms. Among those taking amlodipine, incident DM was associated with increased total mortality risk (HR, 1.92 [95% CI, 1.07-3.44]). Among those taking lisinopril, incident DM was associated with increased risk of CHD (HR, 2.23 [95% CI, 1.07-4.62]) and heart failure (HR, 3.66 [95% CI, 1.30-10.32]). Interaction terms between incident DM and treatment with amlodipine or lisinopril vs chlorthalidone were nonsignificant, suggesting that the effect of incident DM on end points was similar in all 3 treatment groups.

Place holder to copy figure label and caption
Figure 3.

Cox regression models showing the hazard ratios (95% confidence intervals) associated with incident diabetes mellitus during the first 2 years of follow-up on subsequent cardiovascular disease and renal end points in those without diabetes mellitus at baseline. CHD indicates coronary heart disease; CCVD, combined cardiovascular disease (ie, CHD death, nonfatal myocardial infarction, coronary revascularization procedures, stroke, hospitalization or treatment for angina or heart failure, or peripheral arterial disease involving hospitalization or outpatient revascularization); ESRD, end-stage renal disease; NA, not applicable. All hazard ratios controlled for treatment group (total cohort), 2-year blood pressure, age, race, sex, smoking status, baseline fasting glucose level, baseline body mass index, 2-year serum potassium level, and atenolol and statin administration at 2 years.

Graphic Jump Location

In this analysis of nondiabetic older adults from the ALLHAT, FG levels increased steadily over a mean of 4.9 years, regardless of HTN treatment type. Participants randomized to treatment with chlorthalidone had a significantly higher FG level and DM incidence at years 2 and 4 of follow-up compared with those randomized to lisinopril or amlodipine treatment. These results suggest either that chlorthalidone has a detrimental effect on glucose metabolism, that lisinopril and amlodipine have a neutral or protective effect on glucose metabolism, or both.

Differences in mean FG level change between treatment groups were small (3.0 and 1.5 mg/dL [0.17 and 0.08 mmol/L] between chlorthalidone and amlodipine and 5.0 and 4.0 mg/dL [0.22 and 0.28 mmol/L] between chlorthalidone and lisinopril at years 2 and 4, respectively). These small differences translated into higher odds of DM at 2 years in those taking chlorthalidone compared with those taking amlodipine or lisinopril. Similar findings were noted in 2 other studies that reported FG level changes with different HTN medications.18,19 In those studies, the changes in FG levels were small (mean, 2.0-3.5 mg/dL [0.11-0.19 mmol/L]), though the relative risk of developing new-onset DM was significantly increased.

Unlike other studies that have examined the relationship of change in FG level with use of HTN medications, we examined the effect of such change on CVD and renal outcomes. We found that there was no effect of change in FG level on any study end point, neither for all treatment groups combined nor for the chlorthalidone group separately. Furthermore, for no outcome was the hazard ratio for a 10-mg/dL (0.56-mmol/L) change in FG level larger for chlorthalidone than for the other 2 arms. For incident DM, CHD was the only outcome with a statistically significant increased hazard ratio, and this did not differ significantly across the 3 treatment groups.

Our findings are consistent with other studies in that HRs for end points tended to be lower in those treated with chlorthalidone compared with the other groups. In the 14-year follow-up of the Systolic Hypertension in the Elderly Program,20 DM diagnosed during chlorthalidone therapy was not associated with a significant increase in CVD mortality (HR, 1.04 [95% CI, 0.75-1.46]) compared with DM that occurred in the absence of diuretic administration (HR, 1.56 [95% CI, 1.12-2.18]). In a second study, a 15-year follow-up of 686 middle-aged adults with hypertension treated with diuretics,21 incident DM did not have a significant effect on CVD mortality, whereas prevalent DM at baseline did. An explanation for these findings comes from studies that demonstrate that cessation of long-term use of thiazide diuretics is associated with prompt improvement in glucose levels.18 This suggests that diuretics lead to elevated glucose levels by mechanisms different from those associated with DM. Finally, our findings are consistent with a recent meta-analysis of ACE inhibitors and angiotensin receptor blockers.19 That analysis demonstrated that while ACE inhibitors and angiotensin receptor blockers decrease DM risk, they do not reduce odds of mortality or cardiovascular or cerebrovascular events when compared with thiazides and β-blockers.

The finding that chlorthalidone-induced hyperglycemia and DM do not result in increased CVD outcomes is similar in certain ways to the results of the Cardiac Arrhythmia Suppression Trial.22 In that study, a surrogate marker for the mechanism of action of the study medications—suppression of arrhythmias—did not translate into lower mortality. Indeed, the researchers found that the medications resulted in increased mortality.22 It follows that drugs may have a positive or negative effect on a surrogate outcome (such as arrhythmia suppression or FG levels >125 mg/dL [6.9 mmol/L]) that does not necessarily translate into a positive or negative effect on risk of clinical outcomes.

Low potassium levels did not significantly increase odds of DM development in our analyses, although in the stratum given potassium supplementation (suggesting that hypokalemia was persistent enough to trigger clinical action), odds for incident DM with hypokalemia were well above unity in late follow-up. Furthermore, as the trial progressed, use of potassium supplements increased sharply (doubling at 2-5 years). Concomitantly, the differential increase in FG levels and DM incidence decreased at years 4 and 6 compared with earlier in the trial. Prior studies suggest that hypokalemia is an important intermediary factor for DM development in persons treated with diuretics, either due to diminished insulin secretion23 or diminished insulin sensitivity.16 A systematic review of 59 treatment trials showed a moderately high negative correlation coefficient of mean change in serum potassium and mean change in blood glucose.24 The magnitude of the association was estimated to be about a 10-mg/dL (0.56-mmol/L) glucose level increase for a 1-mEq/L decrease in potassium. An accompanying editorial concluded that the proven value of thiazides as central drugs in antihypertensive regimens may be enhanced by careful attention to prevention or management of potassium losses.25

The present study has several strengths. Compared with other HTN trials, the ALLHAT is larger and provides much greater statistical power to recognize associations and differences between medications. Considerable attention was paid to quality assurance, and all laboratory tests were done in a certified central laboratory. Limitations should also be recognized. No FG measurements during follow-up were available for nearly half of the cohort, owing primarily to participants not fasting prior to venipuncture. Given the intrinsic variability of FG measurements, some misclassification of participants most probably occurred. This would have the effect of diluting our findings and making our estimates of incident glucose disorders conservative. We did not obtain measures of overall glucose control, such as glycosylated hemoglobin. Also, in older adults, as in this cohort, glucose abnormalities are often subtle. They are detected with the use of a 2-hour glucose challenge rather than by measurement of FG level.26 These abnormalities were not sought in the ALLHAT. It is therefore likely that many “nondiabetic” individuals had glucose abnormalities. In community practice–based trials like the ALLHAT, investigators seek to reduce clinic and participant burden so as to maximize participation. We also did not seek information on treatment of diabetes.

Finally, our findings cannot be extrapolated beyond 5 years. Change in FG level and incident DM were evaluated at 2 years; after that, the average duration of follow-up for the analyses was about 3 years. The effects of hyperglycemia on atherosclerosis progression may occur over many decades. For example, data from the Diabetes Control and Complications Trial27 of type 1 DM suggest that effects of elevated glucose levels on atherosclerosis, as measured by carotid intima-media thickness, occur after 10 years. Increase in glucose levels associated with such change was greater in that study, and participants had established DM (compared with our participants, who developed new-onset DM). Another report indicates that increased FG levels secondary to HTN treatment (mostly with diuretics and β-blockers) over more than 15 years were associated with increased CHD risk,28 but the study failed to adjust for effects of HTN on rising glucose levels.29 Another HTN study30 reported a nearly 3-fold increase in CVD events with new-onset DM compared with not developing DM, but this study was based on a small number of events (n = 63) in patients who were taking combinations of medications. Whereas it found that new-onset DM was associated with increased CHD risk (similar to what we found), it did not find an independent association of diuretic administration with subsequent CHD events (again similar to our findings). A long-term follow-up study of the ALLHAT participants who developed DM while taking diuretics is planned.

In conclusion, FG levels increase in older adults with hypertension regardless of treatment with chlorthalidone, amlodipine, or lisinopril. For those taking chlorthalidone vs other treatments, the risk of developing FG levels higher than 125 mg/dL (6.9 mmol/L) is greatest, but there is no conclusive or consistent evidence that this diuretic-associated increase in DM increases the risk of clinical events.

Correspondence: Barry R. Davis, MD, PhD, University of Texas School of Public Health, Coordinating Center for Clinical Trials, 1200 Herman Pressler St, Suite E801, Houston, TX 77030 (barry.r.davis@uth.tmc.edu).

Accepted for Publication: July 25, 2006.

Author Contributions: Dr Davis had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Barzilay and Cutler. Acquisition of data: Davis, Pressel, Margolis, and Summerson. Analysis and interpretation of data: Barzilay, Davis, Cutler, Pressel, Whelton, Basile, Margolis, Ong, and Sadler. Drafting of the manuscript: Barzilay, Cutler, Pressel, and Sadler. Critical revision of the manuscript for important intellectual content: Barzilay, Davis, Whelton, Basile, Margolis, Ong, and Summerson. Statistical analysis: Davis. Obtained funding: Davis, Cutler, and Pressel. Administrative, technical, and material support: Davis, Cutler, Pressel, Whelton, and Ong. Study supervision: Barzilay, Cutler, and Basile.

Financial Disclosure: Dr Barzilay has received honoraria from Takeda. Dr Davis has served as a paid consultant and on advisory boards for GlaxoSmithKline, Merck, Takeda, BioMarin, Proctor & Gamble, and Bristol-Myers Squibb. Dr Whelton has received honoraria from Pfizer. Dr Basile has received honoraria from Abbott Laboratories, Astra Zeneca, Boehringer Ingelheim, Forest Pharmaceuticals, Merck, Novartis, Pfizer, Reddy, Sankyo, and SmithKline Beechman/Glaxo Wellcome and has received grants from Boehringer Ingelheim. Dr Ong has received grants from Abbott Laboratories, AstraZeneca, Aventis, Bertek, Boehringer Ingelheim, DepoMed, and Novartis. Dr Sadler has received honoraria from Abbott Laboratories and has stock ownership in Medco, Merck, and Wyeth.

Funding/Support: This study was supported by contract NO1-HC-35130 with the National Heart, Lung, and Blood Institute (NHLBI). The ALLHAT investigators received contributions of study medications from Pfizer (amlodipine and doxazosin), AstraZeneca (atenolol and lisinopril), and Bristol-Myers Squibb (pravastatin), and financial support from Pfizer.

Role of the Sponsor: The NHLBI sponsored the study and was involved in all aspects other than direct operations of the study centers. This included collection, analysis, and interpretation of the data in addition to the decision to submit the manuscript for publication.

Group Information: A list of the ALLHAT Collaborative Research Group members was published in JAMA (2002;288:2981-2997).

Padwal  RLaupacis  A Antihypertensive therapy and incidence of type 2 diabetes: a systemic review. Diabetes Care 2004;27247- 255
PubMed
Gress  TWNieto  FJShahar  EWofford  MRBrancati  FL Hypertension and antihypertensive therapy as risk factors for type 2 diabetes mellitus: Atherosclerosis Risk in Communities Study. N Engl J Med 2000;342905- 912
PubMed
Mykkanen  LKuusisto  JPyorala  KLaakso  MHaffner  SM Increased risk of non–insulin-dependent diabetes mellitus in elderly hypertensive subjects. J Hypertens 1994;121425- 1432
PubMed
Warram  JHLaffel  LMValsania  PChristlieb  ARKrolewski  AS Excess mortality associated with diuretic therapy in diabetes mellitus. Arch Intern Med 1991;1511350- 1356
PubMed
Curb  JDPressel  SLCutler  JA  et al. Systolic Hypertension in the Elderly Program Cooperative Research Group, Effect of diuretic-based antihypertensive treatment on cardiovascular disease risk in older diabetic patients with isolated systolic hypertension. JAMA 1996;2761886- 1892
PubMed
Yusuf  SGerstein  HHoogwerf  B  et al.  Ramipril and the development of diabetes. JAMA 2001;2861882- 1885
PubMed
Lindholm  LHIbsen  HBorch-Johnsen  K  et al.  Risk of new-onset diabetes in the Losartan Intervention For Endpoint reduction in hypertension study. J Hypertens 2002;201879- 1886
PubMed
Vermes  EDucharme  ABourassa  MGLessard  MWhite  MTardif  J-C Enalapril reduces the incidence of diabetes in patients with chronic heart failure: insight from the studies of left ventricular dysfunction (SOLVD). Circulation 2003;1071291- 1296
PubMed
Dahlöf  BSever  PSPoulter  NR  et al.  Prevention of cardiovascular events with an antihypertensive regimen of amlodipine adding perindopril as required versus atenolol adding dendroflumethiazide as required, in the Anglo-Scandinavian Cardiac Outcomes Trial-Blood Pressure Lowering Arm (ASCOT-BPLA): a multicentre randomized controlled trial. Lancet 2005;366895- 906
PubMed
Gerstein  HC Reduction of cardiovascular events and microvascular complications in diabetes with ACE inhibitor treatment: HOPE and MICRO-HOPE. Diabetes Metab Res Rev 2002;18 ((suppl 3)) S82- S85
PubMed
ALLHAT Officers and Coordinators for the ALLHAT Collaborative Research Group, Major outcomes in high-risk hypertensive patients randomized to angiotensin-converting enzyme inhibitor or calcium channel blocker vs diuretic: the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). JAMA 2002;2882981- 2997
PubMed
Davis  BRCutler  JAGordon  DJ  et al.  Rationale and design for the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). Am J Hypertens 1996;9342- 360
PubMed
The ALLHAT Officers and Coordinators for the ALLHAT Collaborative Research Group, Major cardiovascular events in hypertensive patients randomized to doxazosin vs chlorthalidone: the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). JAMA 2000;2831967- 1975
PubMed
Gurwitz  JHBohn  RLGlynn  RJMonane  MMogun  HAvorn  J Antihypertensive drug therapy and the initiation of treatment for diabetes mellitus. Ann Intern Med 1993;118273- 278
PubMed
Helderman  JHElahi  DAnderson  DK  et al.  Prevention of the glucose intolerance of thiazide diuretics by maintenance of body potassium. Diabetes 1983;32106- 111
PubMed
Andersson  OKGudbrandsson  TJamerson  K Metabolic adverse effects of thiazide diuretics: the importance of normokalemia. J Intern Med Suppl 1991;73589- 96
PubMed
Freeman  DJNorrie  JSattar  N  et al.  Pravastatin and the development of diabetes mellitus: evidence for a protective treatment effect in the West of Scotland Coronary Prevention Study. Circulation 2001;103357- 362
PubMed
Kostis  JBWilson  ACFreudenberger  RSCosgrove  NMPressel  SLDavis  BR Long-term effect of diuretic-based therapy on fatal outcomes in subjects with isolated systolic hypertension with and without diabetes. Am J Cardiol 2005;9529- 35
PubMed
Samuelsson  OPennert  KAndersson  O  et al.  Diabetes mellitus and raised serum triglyceride concentration in treated hypertension: are they of prognostic importance? observational study. BMJ 1996;313660- 663
PubMed
Murphy  MBKohner  ELewis  PJSchumer  BDollery  CT Glucose intolerance in hypertensive patients treated with diuretics; a fourteen-year follow-up. Lancet 1982;21293- 1295
PubMed
Gillespie  ELWhite  CMKardas  MLindberg  MColeman  CI The impact of ACE inhibitors or angiotensin II type 1 receptor blockers on the development of new-onset type 2 diabetes. Diabetes Care 2005;282261- 2266
PubMed
The Cardiac Arrhythmia Suppression Trial (CAST) Investigators, Preliminary report: effect of encainide and flecainide on mortality in a randomized trial of arrhythmia suppression after myocardial infarction. N Engl J Med 1989;321406- 412
PubMed
Rowe  JWTobin  JDRosa  RMAndres  R Effect of experimental potassium deficiency on glucose and insulin metabolism. Metabolism 1980;29498- 502
PubMed
Zillich  AJGarg  JSanjib  BBakris  GLCarter  BL Thiazide diuretics, potassium, and the development of diabetes: a quantitative review. Hypertension 2006;48219- 224
PubMed
Cutler  JA Thiazide-associated glucose abnormalities: prognosis, etiology, and prevention: is potassium balance the key? Hypertension 2006;48198- 200
PubMed
Barzilay  JISpiekerman  CFWahl  PW  et al.  Cardiovascular disease in older adults with glucose disorders: comparison of American Diabetes Association criteria for diabetes mellitus with WHO criteria. Lancet 1999;354622- 625
PubMed
Nathan  DMLachin  JCleary  P  et al. Diabetes Control and Complications Trial; Epidemiology of Diabetes Interventions and Complications Research Group, Intensive diabetes therapy and carotid intima-media thickness in type 1 diabetes mellitus. N Engl J Med 2003;3482294- 2303
PubMed
Dunder  KLind  LZethelius  BBerglund  LLithell  H Increase in blood glucose concentration during antihypertensive treatment as a predictor of myocardial infarction: population based cohort study. BMJ 2003;326681- 686
PubMed
Petrie  JRMcDougall  CAl-Mamari  A  et al.  Controversies from observational data: dangers for evidence-based prescribing. http://bmj.bmjjournals.com/cgi/eletters/326/7391/681. Accessed July 2003
Verdecchia  PReboldi  GAngeli  F  et al.  Adverse prognostic significance of new diabetes in treated hypertensive subjects. Hypertension 2004;43963- 969
PubMed

Figures

Place holder to copy figure label and caption
Figure 1.

Randomization and follow-up of participants in the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT)11 by baseline glycemic status for participants without diabetes mellitus at baseline. To convert fasting glucose to millimoles per liter, multiply by 0.0555. *Per ALLHAT database as of September 30, 2002.

Graphic Jump Location
Place holder to copy figure label and caption
Figure 2.

Cox regression models showing the hazard ratios (95% confidence intervals) associated with a 10-mg/dL (0.56-mmol/L) increase in fasting glucose level during the first 2 years of follow-up for subsequent cardiovascular disease and renal end points in participants without diabetes mellitus at baseline. CHD indicates coronary heart disease; CCVD, combined cardiovascular disease (ie, CHD death, nonfatal myocardial infarction, coronary revascularization procedures, stroke, hospitalization or treatment for angina or heart failure, or peripheral arterial disease involving hospitalization or outpatient revascularization); ESRD, end-stage renal disease. All hazard ratios controlled for treatment group (total cohort), 2-year blood pressure, age, race, sex, smoking status, baseline fasting glucose level, baseline body mass index, 2-year serum potassium level, and atenolol and statin administration at 2 years.

Graphic Jump Location
Place holder to copy figure label and caption
Figure 3.

Cox regression models showing the hazard ratios (95% confidence intervals) associated with incident diabetes mellitus during the first 2 years of follow-up on subsequent cardiovascular disease and renal end points in those without diabetes mellitus at baseline. CHD indicates coronary heart disease; CCVD, combined cardiovascular disease (ie, CHD death, nonfatal myocardial infarction, coronary revascularization procedures, stroke, hospitalization or treatment for angina or heart failure, or peripheral arterial disease involving hospitalization or outpatient revascularization); ESRD, end-stage renal disease; NA, not applicable. All hazard ratios controlled for treatment group (total cohort), 2-year blood pressure, age, race, sex, smoking status, baseline fasting glucose level, baseline body mass index, 2-year serum potassium level, and atenolol and statin administration at 2 years.

Graphic Jump Location

Tables

Table Graphic Jump LocationTable 1. Baseline Characteristics by Treatment Group of ALLHAT11 Antihypertensive Component Participants Without Diabetes Mellitus at Baseline*
Table Graphic Jump LocationTable 2. Administration of Blinded Study Drug, Open-Label Diuretics, ACE Inhibitors, Calcium-Channel Blockers, β-Blockers, Potassium Supplements, and Statins by Antihypertensive Treatment Group*
Table Graphic Jump LocationTable 3. Follow-up Fasting Glucose Levels and Incident Diabetes Mellitus in Participants Without Diabetes Mellitus at Baseline by Treatment Group*
Table Graphic Jump LocationTable 4. Baseline and 2-Year Change in Fasting Glucose Levels for Those With and Without Incident Diabetes Mellitus*
Table Graphic Jump LocationTable 5. Logistic Regression Analysis of the Association of Baseline and Follow-up Factors With Incident Diabetes Mellitus in Those Without Diabetes Mellitus at Baseline*

References

Padwal  RLaupacis  A Antihypertensive therapy and incidence of type 2 diabetes: a systemic review. Diabetes Care 2004;27247- 255
PubMed
Gress  TWNieto  FJShahar  EWofford  MRBrancati  FL Hypertension and antihypertensive therapy as risk factors for type 2 diabetes mellitus: Atherosclerosis Risk in Communities Study. N Engl J Med 2000;342905- 912
PubMed
Mykkanen  LKuusisto  JPyorala  KLaakso  MHaffner  SM Increased risk of non–insulin-dependent diabetes mellitus in elderly hypertensive subjects. J Hypertens 1994;121425- 1432
PubMed
Warram  JHLaffel  LMValsania  PChristlieb  ARKrolewski  AS Excess mortality associated with diuretic therapy in diabetes mellitus. Arch Intern Med 1991;1511350- 1356
PubMed
Curb  JDPressel  SLCutler  JA  et al. Systolic Hypertension in the Elderly Program Cooperative Research Group, Effect of diuretic-based antihypertensive treatment on cardiovascular disease risk in older diabetic patients with isolated systolic hypertension. JAMA 1996;2761886- 1892
PubMed
Yusuf  SGerstein  HHoogwerf  B  et al.  Ramipril and the development of diabetes. JAMA 2001;2861882- 1885
PubMed
Lindholm  LHIbsen  HBorch-Johnsen  K  et al.  Risk of new-onset diabetes in the Losartan Intervention For Endpoint reduction in hypertension study. J Hypertens 2002;201879- 1886
PubMed
Vermes  EDucharme  ABourassa  MGLessard  MWhite  MTardif  J-C Enalapril reduces the incidence of diabetes in patients with chronic heart failure: insight from the studies of left ventricular dysfunction (SOLVD). Circulation 2003;1071291- 1296
PubMed
Dahlöf  BSever  PSPoulter  NR  et al.  Prevention of cardiovascular events with an antihypertensive regimen of amlodipine adding perindopril as required versus atenolol adding dendroflumethiazide as required, in the Anglo-Scandinavian Cardiac Outcomes Trial-Blood Pressure Lowering Arm (ASCOT-BPLA): a multicentre randomized controlled trial. Lancet 2005;366895- 906
PubMed
Gerstein  HC Reduction of cardiovascular events and microvascular complications in diabetes with ACE inhibitor treatment: HOPE and MICRO-HOPE. Diabetes Metab Res Rev 2002;18 ((suppl 3)) S82- S85
PubMed
ALLHAT Officers and Coordinators for the ALLHAT Collaborative Research Group, Major outcomes in high-risk hypertensive patients randomized to angiotensin-converting enzyme inhibitor or calcium channel blocker vs diuretic: the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). JAMA 2002;2882981- 2997
PubMed
Davis  BRCutler  JAGordon  DJ  et al.  Rationale and design for the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). Am J Hypertens 1996;9342- 360
PubMed
The ALLHAT Officers and Coordinators for the ALLHAT Collaborative Research Group, Major cardiovascular events in hypertensive patients randomized to doxazosin vs chlorthalidone: the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). JAMA 2000;2831967- 1975
PubMed
Gurwitz  JHBohn  RLGlynn  RJMonane  MMogun  HAvorn  J Antihypertensive drug therapy and the initiation of treatment for diabetes mellitus. Ann Intern Med 1993;118273- 278
PubMed
Helderman  JHElahi  DAnderson  DK  et al.  Prevention of the glucose intolerance of thiazide diuretics by maintenance of body potassium. Diabetes 1983;32106- 111
PubMed
Andersson  OKGudbrandsson  TJamerson  K Metabolic adverse effects of thiazide diuretics: the importance of normokalemia. J Intern Med Suppl 1991;73589- 96
PubMed
Freeman  DJNorrie  JSattar  N  et al.  Pravastatin and the development of diabetes mellitus: evidence for a protective treatment effect in the West of Scotland Coronary Prevention Study. Circulation 2001;103357- 362
PubMed
Kostis  JBWilson  ACFreudenberger  RSCosgrove  NMPressel  SLDavis  BR Long-term effect of diuretic-based therapy on fatal outcomes in subjects with isolated systolic hypertension with and without diabetes. Am J Cardiol 2005;9529- 35
PubMed
Samuelsson  OPennert  KAndersson  O  et al.  Diabetes mellitus and raised serum triglyceride concentration in treated hypertension: are they of prognostic importance? observational study. BMJ 1996;313660- 663
PubMed
Murphy  MBKohner  ELewis  PJSchumer  BDollery  CT Glucose intolerance in hypertensive patients treated with diuretics; a fourteen-year follow-up. Lancet 1982;21293- 1295
PubMed
Gillespie  ELWhite  CMKardas  MLindberg  MColeman  CI The impact of ACE inhibitors or angiotensin II type 1 receptor blockers on the development of new-onset type 2 diabetes. Diabetes Care 2005;282261- 2266
PubMed
The Cardiac Arrhythmia Suppression Trial (CAST) Investigators, Preliminary report: effect of encainide and flecainide on mortality in a randomized trial of arrhythmia suppression after myocardial infarction. N Engl J Med 1989;321406- 412
PubMed
Rowe  JWTobin  JDRosa  RMAndres  R Effect of experimental potassium deficiency on glucose and insulin metabolism. Metabolism 1980;29498- 502
PubMed
Zillich  AJGarg  JSanjib  BBakris  GLCarter  BL Thiazide diuretics, potassium, and the development of diabetes: a quantitative review. Hypertension 2006;48219- 224
PubMed
Cutler  JA Thiazide-associated glucose abnormalities: prognosis, etiology, and prevention: is potassium balance the key? Hypertension 2006;48198- 200
PubMed
Barzilay  JISpiekerman  CFWahl  PW  et al.  Cardiovascular disease in older adults with glucose disorders: comparison of American Diabetes Association criteria for diabetes mellitus with WHO criteria. Lancet 1999;354622- 625
PubMed
Nathan  DMLachin  JCleary  P  et al. Diabetes Control and Complications Trial; Epidemiology of Diabetes Interventions and Complications Research Group, Intensive diabetes therapy and carotid intima-media thickness in type 1 diabetes mellitus. N Engl J Med 2003;3482294- 2303
PubMed
Dunder  KLind  LZethelius  BBerglund  LLithell  H Increase in blood glucose concentration during antihypertensive treatment as a predictor of myocardial infarction: population based cohort study. BMJ 2003;326681- 686
PubMed
Petrie  JRMcDougall  CAl-Mamari  A  et al.  Controversies from observational data: dangers for evidence-based prescribing. http://bmj.bmjjournals.com/cgi/eletters/326/7391/681. Accessed July 2003
Verdecchia  PReboldi  GAngeli  F  et al.  Adverse prognostic significance of new diabetes in treated hypertensive subjects. Hypertension 2004;43963- 969
PubMed

Correspondence

CME
Meets CME requirements for:
Browse CME for all U.S. States
Accreditation Information
The American Medical Association is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians. The AMA designates this journal-based CME activity for a maximum of 1 AMA PRA Category 1 CreditTM per course. Physicians should claim only the credit commensurate with the extent of their participation in the activity. Physicians who complete the CME course and score at least 80% correct on the quiz are eligible for AMA PRA Category 1 CreditTM.
Note: You must get at least of the answers correct to pass this quiz.
You have not filled in all the answers to complete this quiz
The following questions were not answered:
Sorry, you have unsuccessfully completed this CME quiz with a score of
The following questions were not answered correctly:
Commitment to Change (optional):
Indicate what change(s) you will implement in your practice, if any, based on this CME course.
Your quiz results:
The filled radio buttons indicate your responses. The preferred responses are highlighted
For CME Course: A Proposed Model for Initial Assessment and Management of Acute Heart Failure Syndromes
Indicate what changes(s) you will implement in your practice, if any, based on this CME course.
NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).
Submit a Comment

Multimedia

Some tools below are only available to our subscribers or users with an online account.

Web of Science® Times Cited: 123

Related Content

Customize your page view by dragging & repositioning the boxes below.

See Also...
Articles Related By Topic
Related Topics
PubMed Articles