Current issues of ACP Journal Club are published in Annals of Internal Medicine


Diuretic therapy for hypertension was associated with increased mortality in patients with diabetes

ACP J Club. 1991 Nov-Dec;115:93. doi:10.7326/ACPJC-1991-115-3-093

Source Citation

Warram JH, Laffel LM, Valsania P, Christlieb AR, Krolewski AS. Excess mortality associated with diuretic therapy in diabetes mellitus. Arch Intern Med. 1991 Jul;151:1350-6.


Abstract title formerly " Diabetic therapy for hypertension was associated with increased mortality in patients with diabetes"


To examine the effect of antihypertensive treatment on mortality in patients with diabetes who had a range of associated risk factors.


Re-analysis of data from 1972 to 1979 for a cohort followed at 4-month intervals in a multicenter clinical trial (Diabetic Retinopathy Study).


15 clinical centers in the United States.


The 1758 patients who were enrolled in the original trial that assessed laser therapy for the prevention of blindness had a confirmed diagnosis of severe diabetic retinopathy. This analysis from the main study included 759 white patients, aged 35 to 69 years, with unimpaired renal function (serum creatinine level < 115 µmol/L for women and < 133 µmol/L for men). Patients were subdivided into those with proteinuria (reading of ≥ 1+ at baseline or at 1 y of follow-up; n = 263) and those without proteinuria (n = 496). Patients were followed for a mean of 3.2 years (median, 4.5 y); 42 patients (5.5%) were lost to follow-up.

Assessment of risk factors

Smoking history, duration of and treatment for diabetes, diuretic other hypertensive therapy, blood pressure (BP), serum creatinine and cholesterol levels, plasma glucose, and urine protein were recorded at baseline and at yearly intervals. Hypertensive patients were defined as those who received antihypertensive therapy or had diastolic BP ≥ 90 mm Hg at the 1-year visit.

Main outcome measure

All-cause mortality.

Main results

139 (18%) of 759 patients died (77 [29%] in the proteinuria group and 62 [13%] in the nonproteinuria group, P < 0.001) . By logistic regression, risk for death was increased with proteinuria (relative odds [RO] {odds ratio} 2.5, P< 0.001), with age ≥ 50 years (RO 1.7, P = 0.02), and by cholesterol ≥ 6.0 mmol/L (RO 1.8, P = 0.006). RO of death for untreated hypertensive patients compared with normotensive patients was 1. 6 (P = 0.13); risk for death increased for hypertensive patients treated with diuretics only (RO 5.1, P < 0.001), for treatment with diuretics and other agents (RO 3.4, P< 0.001), and for treatment with other agents only (RO 2.1, P = 0.02). The excess mortality was higher for hypertensive patients treated with diuretics only than for untreated hypertensive patients, even though the former group had the lowest BP with treatment (relative risk 3.2, 95% CI 1.7 to 6.2, P < 0.001).


Treatment of hypertension with diuretic agents, particularly when used alone, was associated with an increase in mortality among patients with diabetes mellitus and normal renal function.

Sources of funding: National Institutes of Health and the Joslin Diabetes Center.

Address for article reprint: Dr. J.H. Warram, Section on Epidemiology and Genetics, Research Division, Joslin Diabetes Center, 1 Joslin Place, Boston, MA 02215, USA.


The authors concluded that excess mortality among hypertensive patients with diabetes was attributable mainly to diuretic administration. Because this study has a retrospective analytic design, it is not possible to clarify the mechanism for this association. The authors feel, however, that their data support 2 major alternative hypotheses, a metabolic cause or the "J-curve" effect (i.e., BP that is higher or lower than optimal may increase the risk of mortality).

Although previous studies support the authors' concepts (1, 2), these findings may apply only to some patients with diabetes (i.e., white patients from 35 to 69 years of age without major renal impairment). The authors downplay other risk factors and state that the logistic regression analysis adjusts for these factors, but their results reached greater significance in the subset of patients with proteinuria—an independent cardiovascular risk factor in persons with and without diabetes. This subset also had an increased prevalence of other risk factors (e.g., age, serum cholesterol) compared with their nonproteinuric counterparts.

This study was not able to provide data on metabolic disturbances from medication, the prevalence of underlying cardiac disease, and the dose or form of diuretic used. This is especially relevant because of lower diuretic doses prescribed currently.

Whatever the possible reason for this disturbing association, whether physiologic or secondary to a selection bias, the clinician is torn between recommendations as well as benefits of aggressive blood pressure reduction and potential hazards of antihypertensive therapy. Clarifying the risk/benefit ratio of differing doses and classes of pharmacologic agents and delineating subpopulations of patients with diabetes who are at risk deserves further study.

Robert Bloomfield, MD
Carolyn Pedley, MDBowman Gray School of MedicineWinston-Salem, North Carolina, USA


1. Klein R, Moss SE, Klein BE, DeMets DL. Relation of ocular and systemic factors to survival in diabetes. Arch Intern Med. 1989;149:266-72.

2. Walker WG, Hermann J, Yin D, Murphy RP, Patz A. Diuretics accelerate diabetic nephropathy in hypertensive insulin dependent and non-insulin dependent subjects. Trans Assoc Am Physicians. 1987;100:305-15.


This association between diuretic therapy and excess mortality probably is also valid for patients without diabetes (3).

Hertzel G. Gerstein, MD
McMaster UniversityHamilton, Ontario, Canada

3. Hoes AW, Grobbee DE, Lubsen J, et al. Diuretics, beta-blockers, and the risk for sudden cardiac death in hypertensive patients. Ann Intern Med. 1995;123:481-7.