Incidence of Diabetes and Gout in Hypertensive Patients During 8 Years of Follow-Up

Incidence of diabetes and gout in hypertensive patients during 8 years of follow-up

T Grodzicki1,2, A Palmer1 and CJ Bulpitt1 on behalf of the General Practice Hypertension Study Group 1Royal Postgraduate Medical School, London, UK; 2Jagiellonian University, Cracow, Poland

Hypertension has been shown to be closely associated mellitus was diagnosed in 5.0% of hypertensive and with metabolic disorders such as diabetes mellitus and 1.5% normotensive subjects. The diagnosis of gout was gout. The aim of this study was to establish the inci- made in 3.1% and 0.9%, respectively. Moreover, dia- dence of both these diseases in hypertensive patients betes mellitus occurred more frequently in both hyper- and their age-matched controls and to assess the risk tensive men and women, with diuretic treatment and associated with diuretic treatment. The results refer to without (respectively, men given diuretic: relative risk the data base of 2295 hypertensive and 2280 controls (RR) = 2.74, 95% conﬁdence interval (CI): 0.99, 7.5; and from the General Practice Hypertension Study Group men not given diuretic: RR = 2.25, 95% CI: 1.0, 5.3: who were screened for hypertension. The rescreened women RR = 4.03 95% CI: 1.5, 10.8, and RR = 3.47, 95% hypertensive subjects (n = 1190) and their controls CI: 1.4, 8.9) in comparison with their age-matched con- (n = 938) were followed for an average of 8 years. The trols. For gout the excess was conﬁrmed in hyperten- diagnosis of diabetes and gout was made by their gen- sive men with and without diuretic treatment (RR = 6.25 eral practitioners. After 8 years of follow-up, diabetes (95% CI: 2.4, 16.7) vs 3.93 (95% CI: 1.6, 9.7)).

Keywords: hypertension; gout; diabetes mellitus; diuretics

Introduction Design and methods Hypertension is closely associated with metabolic The results refer to the data base of 2295 hyperten- disorders such as diabetes mellitus and gout and sive and 2280 controls from the General Practice this relationship has been confirmed in multiple 1–3 Hypertension Study Group who were screened for studies in different populations. Despite certain 9 controversies the insulin resistance (X) syndrome hypertension from 1975 to 1979. Patients aged 18– has been widely accepted as a link factor.4 More- 65 years were identified from the age–sex registers over, the coexistence of hypertension, diabetes and and screened by practice nurses trained to use a ran- gout significantly increases the risk of cardiovascu- dom zero sphygmomanometer. Within each practice lar, renal and cerebrovascular complications.5 It has a normotensive control was selected for each patient also been shown that the use of diuretics in hyper- with hypertension matched for sex, age, date of tension reduces cardiovascular risk but induces det- screening and ethnic group. The hypertensive rimental changes in serum uric acid concentration patients already identified as hypertensives were and glucose metabolism that may be of clinical sig- categorised as ‘known hypertensives’. Patients were nificance.6 The report of the European Working classified as ‘new hypertensives’ if their diastolic Party on High Blood Pressure in the Elderly blood pressure (DBP) phase IV was 90 mm Hg or (EPWHE)6 demonstrated an excess in the incidence more at two of three visits. These two groups of of gout and diabetes in the diuretic group compared patients (n = 1190) and their controls (n = 938) were with placebo. This diabetogenic action of thiazides rescreened between 1983 and 1989, an average of 8 may be due to their ability to open calcium-activated years later. A patient was considered to be on hypo- potassium channels in pancreatic beta-cells7 or tensive agents at the time of rescreen if there was through stimulation of A cell secretion in the pan- any record of an antihypertensive drug, a beta-adre- creas.8 nergic blocker, a calcium channel blocker, or a The aim of this study was to establish the inci- diuretic, regardless of the reason for prescription. If dence of diabetes and gout in hypertensive patients a patient was not currently taking antihypertensive and their age-matched controls and to evaluate the drugs, the clinican recorded whether such treatment risk associated with diuretic therapy. had ever been prescribed. The diagnosis of diabetes mellitus and gout in both hypertensive (known and new) and control subjects was made by their gen- Correspondence: Professor CJ Bulpitt, Division of Geriatric Medi- eral practioners. cine, Royal Postgraduate Medical School, Hammersmith Hospi- tal, Du Cane Road, London W12 ONN, UK Diabetes and gout in hypertensive patients T Grodzicki et al 584 Table 1 Baseline characteristics of hypertensive subjects and their controls. Mean and s.e.m.

Men Women

Cases Controls Cases Controls

On a diuretic ϩϪ Ϫ ϩϪ Ϫ No. 155 437 468 221 377 470 SBP (mm Hg) 169.8*** 158.0 131.9# 173.5*** 163.4 131.4# (2.0) (0.91) (0.89) (1.81) (1.23) (0.84) DBP (IV) (mm Hg) 104.9*** 101.4 78.2# 103.9 102.3 78.0# (0.98) (0.45) (0.39) (0.89) (0.6) (0.4) Weight (kg) 80.8 81.6 70.5# 69.0 69.9 62.8# (0.58) (0.98) (0.6) (0.84) (0.74) (0.45) Height (m) 1.72 1.73 1.68# 1.60 1.60 1.61 (0.69) (0.33) (0.46) (0.48) (0.33) (0.32) BMI (kg/m2) 27.8* 26.9 25.0# 26.9 27.2 24.3# (0.31) (0.18) (0.18) (0.30) (0.27) (0.16)

***P Ͻ 0.001; *P Ͻ 0.05 vs non-diuretic group; #P Ͻ 0.01 vs both hypertensive groups. SBP = systolic blood pressure; DBP = diastolic blood pressure (phase IV); BMI = body mass index.

Statistical analysis was signiﬁcantly higher in both sexes in hypertensive patients when compared with their controls. The differences between groups were tested with This was true for both hypertensive groups although analysis of variance. The relationship between the there was a tendency towards a higher relative risk incidence of diabetes mellitus and gout and pres- of diabetes in patients treated with diuretic. The ence of hypertension and diuretic therapy was incidence of gout showed a similar pattern in men, determined using the Cox proportional Hazards although in women gout was very rare and no con- Model. The results are described as mean ± s.e.m. or clusions could be drawn. as otherwise stated.

Results Discussion Table 1 gives the baseline characteristics of hyper- We have demonstrated that diabetes mellitus occurs tensive cases and controls. The cases are divided more frequently in treated hypertensive patients in into those on a diuretic and those not on this treat- comparison with their age-matched controls (RR ment. As expected, hypertensive patients had higher 2.3–2.7 in men and 3.5–4.0 in women). Surpris- BP than their age-matched controls. Moreover, their ingly, the increase in incidence of diabetes was only body mass index (BMI) was signiﬁcantly lower in marginally increased by diuretic therapy. In the pre- normotensives in comparison with both hyperten- viously cited EPWHE trial6 a tendency toward more sive groups. Hypertensive men on diuretic therapy frequent elevated plasma glucose concentrations, were characterised by higher mean systolic and dias- diagnosis of diabetes mellitus and prescriptions for tolic pressures and BMI in comparison with hyper- hypoglycaemic drugs in the diuretic treated group tensive men not on a diuretic. In women, the hyper- was observed. Lind et al10 in 2–3 years observation tensive groups differed only by SBP. of 65 patients with essential hypertension showed Table 2 gives the incidence of diabetes and gout that a signiﬁcant decrease in insulin sensitivity was in these groups. The incidence of diabetes mellitus induced by treatment with beta-blockers and hydro-

Table 2 Incidence of diabetes (DM) and gout according to the diuretic treatment

Men Women

Cases Controls Cases Controls

On a diuretic ϩϪ Ϫ ϩϪ Ϫ No. 155 437 468 221 377 470 New DM (n)716811166 Incidence DM/1000 pt years 5.83 4.79 2.13 6.40 5.53 1.59 RR 2.74 2.25 1.00 4.03 3.47 1.00 95% CI 0.99–7.5 1.0–5.3 — 1.5–10.8 1.4–8.9 — New gout (n)12216222 Incidence gout/1000 pt years 10 6.29 1.6 1.16 0.69 0.53 RR 6.25 3.93 1.00 — — — 95% CI 2.4–16.7 1.6–9.7 — — — —

RR = relative risk; CI = confidence interval. Diabetes and gout in hypertensive patients T Grodzicki et al 585 chlorothiazide after 4 –6 months and persisted after 3 Duncan B et al. Association of fasting insulin with 2–3 years of treatment. In another long-term obser- clustering of metabolic disturbances in African Amer- vation of hypertensive patients a significant increase icans and Whites (abstract). Diabetes 1994; 43: (Suppl in glucose level was restricted to patients taking a 1): 150A. thiazide diuretic.11 The dose-dependent influence of 4 Haffner SM. Epidemiology of hypertension and insulin resistance syndrome. J Hypertens 1997; 15 (Suppl 1): thiazides on hepatic insulin metabolism was dem- 12 S25–S30. onstrated by Harper et al in a double-blind ran- 5 Levy D, Kannel WB. Cardiovascular risks: new insights domised crossover study of two 12-week treatment from Framingham. Am Heart J 1988; 116: 266–272. periods with either low dose (1.25 mg) or conven- 6 Fletcher A et al. Risks and benefits in the trial of the tional dose (5.0 mg) bendrofluazide. The hyperglyca- European Working Party on High Blood Pressure in the emic effect of diuretics appears to be partly related Elderly. J Hypertens 1991; 9: 225–230. to potassium loss since impairment of glucose toler- 7 Pickkers P, Schachter M, Sever PS, Hughes AD. Do ance is most marked in those in whom serum potass- KCa channels and carbonic anhydrase play a role in ium decreased13 and supplementation of potassium thiazide-induced hyperglycaemia? J Hum Hypertens prevents glucose intolerance in thiazide-treated 1995; 9: 691–693. 14,15 8 Hermansen K, Schmitz O, Mogensen CE. Effects of a patients. thiazide diuretic (hydroflumethiazide) and a loop In our study the incidence of gout was very low diuretic (bumetanide) on the endocrine pancreas: stud- in women, but hypertension in men and no diuretic ies in vitro. Metabolism 1985; 34: 784–789. therapy was associated with a four-fold increase in 9 Mayhew SR. Hypertension screening in general prac- gout. Impaired tubular secretion of urate has been tice. Report on behalf of the General Practitioner suggested as a potential mechanism of hyperuricae- Hypertension Study Group. JR Coll Gen Pract 1983; 33: mia in essential hypertension.16 The concentration 434–437. of serum uric acid in hypertension has been shown 10 Lind L, Pollare T, Berne C, Lithell H. Long-term meta- to be related to plasma urea17 and any impairment bolic effects of antihypertensive drugs. Am Heart J of renal function would be expected to increase 1994; 128: 1177–1183. 11 Curzio JL et al. Risk factor modification in hyperten- serum uric acid. Nevertheless, gout is rare in the 18 sion-success and failure. J Hum Hypertens 1987; 1: presence of renal failure and a four-fold increase 131–136. in gout could not be predicted and needs to be con- 12 Harper R et al. A comparison of the effects of low- and firmed. conventional-dose thiazide diuretic on insulin action Hypertensive men on diuretic treatment had a six- in hypertensive patients with NIDDM. Diabetologia fold increase in incidence of gout. The negative 1995; 38: 853–859. influence of thiazide diuretics on the elimination of 13 Amery A et al. Glucose intolerance during diuretic uric acid and related increase in serum concen- therapy in elderly hypertensive patients. A second tration has been previously described.6,19,20 report from the European Working Party on high blood Dueymes,19 in a 3-month study in 946 patients with pressure in the elderly (EWPHE). Postgrad Med J 1986; 62: 919–924. hypertension showed that serum uric acid levels 14 Helderman JH et al. Prevention of the glucose intoler- increased significantly in 5.5% of patients. Similarly ance of thiazide diuretics by maintenance of body pot- 20 Plante et al after 48 weeks of therapy with hyd- assium. Diabetes 1983; 32: 106–111. rochlorothiazide demonstrated that serum uric acid 15 Langford HG et al. The effect of thiazide therapy on levels rose from 6.5 to 8.7 mg/dl. The detrimental glucose, insulin and cholesterol metabolism and of effects of thiazides on uric acid elimination further glucose on potassium: results of a cross-sectional raises the four-fold increase in incidence of gout in study in patients from the Hypertension Detection and non-diuretic treated hypertensive men. Follow-up Program. J Hum Hypertens 1990; 4: 491– 500. 16 Tykarski A. Evaluation of renal handling of uric acid Acknowledgements in essential hypertension: hyperuricemia related to decreased urate secretion. Nephron 1991; 59(3): 364– Members of the General Practice Hypertension 368. Study Group include: IC Bradley, JS Broxton, AJ 17 Bulpitt CJ. Serum uric acid in hypertensive patients. Davis, AE Fletcher, PL Ganvir, NK Gostick, SR Br Heart J 1975; 37: 1210–1215. Mayhew, D Mukerji, F Shepherd and FO Wells. 18 Bywaters EGL, Glynn LE. Connective tissue disorders. In: Thompson RHS, Wootton IDP (eds) Biochemical Disorders in Human Disease, 3rd edn. Churchill: Lon- References don, 1970, p 735. 1 Tan MH, Maclean DR. Epidemiology of diabetes mel- 19 Dueymes JM. Clinical update: spironolactone and alti- litus in Canada. Clin Invest Med 1995; 18: 240–246. zide as monotherapy in systemic hypertension. Am J 2 el Mugamer IT, Ali Zayat AS, Hossain MM, Pugh RN. Cardiol 1990; 65: 20K–23K. Diabetes, obesity and hypertension in urban and rural 20 Plante GE, Dessurault DL. Hypertension in elderly people of bedouin origin in the United Arab Emirates. patients. A comparative study between indapamide J Trop Med Hyg 1995; 98: 407–415. and hydrochlorothiazide. Am J Med 1988; 84: 98–103.