of Cardiovascular Disease Medications and Mortality in People with Older Onset Diabetes Use

Bartbara E. K Kleu MD, MPH, Scot E. Moss, MA, and Ronald Klein, MD, MPH

Intrducion Diabetes mellitus increases the risk of death in several populations in the United States.' -6Much of the increase is attnbuted to cardiovascular disease. People with diabetes are more likely to have cardiovascular disease, including high blood pressure.7'8 Thus, in addition to therapy for their diabetes, these people may also be taking medication to lower blood pressure and for other cardiovascular problems. Warram has reported excess mortality associated with diuretic usage in patients followed at the Joslin Clinic.9 We explore the relationship between use of cardiovascular medications and mortality in people with diabetes diagnosed after 30 years of age in the population participating in the Wisconsin Epidemiologic Study of Diabetic Retinopathy.

Meto&ds The population has been described in

previous reports.10-'2 Briefly, 10 135 diabetic persons were identified from the records of physicians providing primary care in an 11-county area in southern Wisconsin from July 1, 1979, through June 30, 1980. A probability sample of persons diagnosed as having diabetes after age 30 (n = 1780) was invited to the examination phase of the study from 1980 to 1982 (1370 participated). Pertinent parts of the examination included measurements of weight, height, blood pressure,'3 glycosylated hemoglobin,'4 and gross urine protein (Labstix). An interview included questions about medications for control of hyperglycemia and blood pressure and diuretic agents. We asked for the number of aspirin used in the 30 days before the baseline examination, and about smoking, angina, myocardial infarction, and stroke. Questions about medication usage were verified by a physician's report. Participants, relatives, and contact persons have been called annually to determine vital status. Newspaper obituaries were reviewed daily. Annually, a request was made to the Wisconsin Center

for Health Statistics, Section of Vital Statistics, for death certificate information for these persons. Names of persons who had been lost to follow-up were submitted for matching against death records. Wisconsin death records through 1988 have been searched. Information on persons who have moved out of Wisconsin and were suspected of being deceased and on those lost to follow-up was submitted to the National Death Index for matching against national death data. The time interval of interest for this study ranges from the examination date through the date of death or December 31, 1988. Only deaths confirmed by a death certificate are included in the analyses. Cause-specific mortality analyses are based on the underlying cause of death or any mention on the death certificate. People were classified as having a cardiovascular disease at examination if they reported that they had angina or had had a heart attack or stroke that was verifiedby a physician. Peoplewere classified as having never smoked if they had smoked fewer than 100 cigarettes in their lifetime. Proteinuria was classified as a urine protein concentration of 0.30 g/l or more as measured by Labstix. The Cox proportional hazards model was used to evaluate the effect on mortality of a medication while controlling for other risk factors.15 The Statistical Analysis System was used for estimating the model parameters.16

Remd There were 605 (44.2%) confirmed deaths (Table 1). Heart disease (codes The authors are with the Department of Ophthalmology, University of Wisconsin, Madison. Requests for reprints should be sent to Barbara E. K. Klein, MD, MPH, Department of Ophthalmology, University of Wisconsin, Madison, 600 Highland Avenue, Madison, WI 53792. This paper was submitted to the Journal August 19, 1991, and accepted with revisions January 6, 1992.

August 1992, Vol. 82, No. 8

Public Health Briefs

insulin (no or yes), systolic and diastolic blood pressures, glycosylated hemoglobin (as a percentage of hemoglobin), sex, body mass index (weight/height2), smoking status (current, past, or never smoked), and presence of proteinuria. The latter was included in an attempt to control for the presence of renal dsease. The odds of death as derived from proportional hazards analyses appear in Table 3. The use of loop diuretics was associated with all causes of mortality and with mortality attributed to diabetes, heart disease, and ischemic heart disease. Use of thiazide diuretics was associated with death due to heart diseae, other than ischemic, and with stroke. Potassium sparing diuretics were associated with all other causes of mortality. These analyses were repeated after excluding those who had proteinuria. There was no change in the pattern of siifant associations between use of loop diuretics and death. Also, major drug combinationswere examined. These included use of a beta blocker or potassium sparing diuretic with a thiazide diuretic and use of a beta blocker or potassium sparing diuretic with a loop diuretic. In all cases, except for the association of loop diuretics with death due to other heart disease, the effect of loop and thiazide diuretics remained signiicant after controlling for use of beta blockers or potassium sparing diuretics and interactions with these agents.

Discussion Patients with diabetes mellitus are at

391-398, 402, and 404-429 of the International Classification of Diseases, 9th version) was listed as the underlying cause of 49.9%o of deaths. Table 2 lists the frequencies of antihypertensive and diuretic medications, aspirin, and tolbutamide use. August 1992, Vol. 82, No. 8

Many subjects took more than one of these medications. The odds of death for use of these medications were computed. Other relevant characteristics included in the analyses were age, duration of diabetes, use of

increased risk of death due to cardiovascular diseases.17 One might anticipate that the presence of these diseases would necessitate the use of any of the medications of interest in this paper. Therefore, in order to attempt to control for this "confounding by indication," we included a history of cardiovascular disease in our analyses. For each cause of death cate-

American Journal of Public Health

1143

Public Health Briiefs gory, history of cardiovascular disease was a signiicant variable. However, use of at least one diuretic agent added significant information for most of the models. Loop agents were the agents most frequently included, but thiazide diuretics were also sinificantly associated with an increased risk of death. The differences between the analyses indicating one class of diuretics as important compared with another may be a reflection of the frequency of their use rather than the particular drug effect. In clinical practice, nephrologists may avoid using thiazides because of their effects on glycemia and uricemia. There are some studies that have indicated reduced mortality and morbidity in hypertensive subjects who receive medical treatment,18 but an association between diuretic use and arrhythmias has been reported.19-21 Whether the use of these agents with diabetic subjects truly predisposes to death cannot be determined from these analyses. However, it is probably wise for clinicians caring for such patients to be especially attentive to blood chemistries. Also, it has been suggested that the use of diuretics in patients with diabetic nephropathy may accelerate that condition.22 In these anablses we evaluated a possible relationship of mortality to use of tolbutamide. This drug had been implicated as a cause of increased mortality in the University Group Diabetes Prgram3 Its use was relatively infrequent in our study population (Table 2); it did not contrbute significantly to describing mortality. O

Acknofwledgments This research was supported by National Institutes of Health-National Eye Institute Grant EY 03083 to Ronald Klein. We are grateful to Fred Krantz, Wiscon-

sin Center for Health Statistics, for providing death certificate information on Wisconsin residents; to Betty Sullivan and Peggy Peterson, Wisconsin Center for Health Statistics, for coding certificates of out-of-state residents; to Stacy Meuer and Ane Weber for data processing; and to Julie K. Olson and Kristine A. Tway for assistance with the manuscript.

References 1. Moss SE, Klein R, Klein BEK. Causespecific mortality in a population-based study of diabetes. Am J Public Heakh 1991;81:1158-1162. 2. Kessleril. Mortalityexperience of diabetic patients. A twenty-six year follow-up study. Am JMed 1976;51:715-724. 3. Palumbo PJ, Elveback LR, Chu C-P, Connolly DC, Kurland LT. Diabetes mellitus: incidence, prevalence, survivorship, and causes of death in Rochester, Minnesota, 1945-1970. Diabetes. 1976;25:566-573. 4. Herman JB, Medalie JH, Goldbourt U. Differences in cardiovascular morbidity and mortality between previously known and newly diagnosed adult diabetes. Dia-

betokgia. 1977;13:229-234. 5. Krolewski AS, Cyzy KA, Janeczko D, Kopczynsid J. Mortality from cardiovascular disease among diabetics. Diabetologia. 1977;13:345-350. 6. KannelWB, McGee DL. Diabetes andglucose tolerance as risk factors for cardiovascular disease: The Framingham Study. Diabetes Care. 1979;2:120-126. 7. Klein R, Klein BEK, Moss SE, DeMets DL. Blood pressure and hypertension in diabetes. Am J EpidemioL 1985;122:7589. 8. Dupree EA, Meyer MB. Role of risk factors in complications of diabetes mellitus. Am JEpidemioL 1980;112:100-112. 9. Warram JH, Laffel LMD, Valsania P, Christleib AR, Krolewski AS. Excess mortality associated with diuretic therapy in diabetes mellitus. Arch Intern Med 1991; 151:1350-1356. 10. Klein R, Klein BEK, Moss SE, DeMets DL, Kaufman I, Voss PS. Prevalence of diabetes mellitus in southern Wisconsin. Am JEpidenioLW 1984;119:54-61. 11. Klein R, Klein BEK, Moss SE, Davis MD, DeMets DL. The Wisconsin Epide-

miologic Study of Diabetic Retinopathy II. Prevalence and risk of diabetic retinopathy when age at diagnosis is less than 30 years. Arch OphthalmoL 1984;102:520526. 12. Klein R, Klein BEK, Moss SE, Davis MD, DeMets DL. The Wisconsin Epidemiologic Study of Diabetic Retinopathy III. Prevalence and risk of diabetic retinopathywhen age at diagnosis is 30 or more years. Arch OphthamnoL 1984;102:527532. 13. Hypertension Detection and Follow-up Program Cooperative Group. The hypertension detection and follow-up program. Prev Med 1976;5:207-215. 14. Isolab. Quick Step, Fast Hemogoin Test System Akron, Ohio: Isolab; 1981:1-8. 15. Cox DR. Regression models and life tables (with discussion).JRoyal Stat SocB. 1972; 34:187-220. 16. SAS Institute Inc. SUGI Supplmental Library User's Guide Version 5 Edition. Cary, NC: SAS Institute Inc; 1986:437466. 17. Kleinman JC, Donahue RP, Harris MI, Finucane FF, Madans JH, Brock DB. Mortality among diabetics in a national sample. Am JEpidemioL 1988;128:389-401. 18. Hartford M, Wendelhad I, Berglund G, Wallentin I, Ljungman S, Wikstrand J. Cardiovascular and renal effects of longterm antihypertensive treatment. JAMA. 1988;259:2553-2557. 19. Hollan OB, Nixon JV, Kuhnert L. Diuretic-induced ventricular ectopic activity.Am JMed 1981;70:762-768. 20. Kuller LH, Hulley SB, Cohen JD, Neaton J. Unexpected effects of treating hypertension in men with electocdigraphic abnormalities: a critical analysis. Cirulation

1986;73:114-123. 21. Caralis PV, Perez-Stable E. Electrolyte abnormalities and ventricular arrhythmia.

Dnrgs. 1986;31(suppl 4):85-100.

22. 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; 100:305-315. 23. University Group Diabetes Program. Mortality results. Diabetes. 1970;19(suppl

2):789-830.

Use of cardiovascular disease medications and mortality in people with older onset diabetes.

Mortality follow-up for a cohort defined in 1980 with diabetes diagnosed at 30 years of age or older has been completed through 1988. History of medic...
2MB Sizes 0 Downloads 0 Views