Quick viewing(Text Mode)

Diuretic Treatment of Hypertension

Diuretic Treatment of Hypertension

Diuretic Treatment of Hypertension

1 3 EHUD GROSSMAN, MD FABIO ANGELI, MD cerebrovascular , combination 2 4 PAOLO VERDECCHIA, MD, FACC, FAHA, FESC GIANPAOLO REBOLDI, MD, PHD, MSC 1 of a () and ARI SHAMISS, MD ACE inhibitor () reduced the risk of by 43% compared with pla- cebo. Perindopril alone, despite lowering systolic BP by 5 mmHg, decreased stroke lthough and thiazide-like multiple- combinations in patients risk only by a nonsignificant 5%. are indispensable in with resistant hypertension. The main Several studies attested to the supe- A fi the treatment of hypertension, their argument that will be discussed is the rior ef cacy of diuretic therapy over other role as first-line or even second-line drugs place of diuretics as first-line drugs or add- antihypertensive agents in reducing the is a provoking debate. on drugs in the context of the available risk for stroke (4–6,8,10,11). In the Sec- The European Society of / antihypertensive armamentarium. ond Australian National Pressure European Society of Hypertension (ESC/ The pro side of the controversy will Study (ANBP2) (10), fatal stroke occurred ESH) guidelines recommend that thiazide argue that diuretics should remain the two times more in patients treated with an diuretics should be considered as suitable preferred drugs for initial treatment in ACE inhibitor than in patients treated as b-blockers, antagonists, ACE many hypertensive patients, whereas the with a diuretic. In the Antihypertensive inhibitors, and receptor cons side will contend that emerging and Lipid-Lowering Treatment to Prevent blockers for the initiation and mainte- evidence from outcome-based studies is Attack Trial (ALLHAT) (4,5), chlor- nance of antihypertensive treatment (1). casting doubt on the role of these drugs as thalidone was superior to the a-blocker Another European position, en- first-line and even second-line antihyper- mesylate in the prevention of dorsed by the British Hypertension Soci- tensive treatment. stroke and was superior to the ACE inhib- ety, is that diuretics and itor in the prevention of stroke blockers should be first-line drugs in THE PRO SIDE—Lowering blood in black individuals. In the Medical Re- hypertensive patients aged $55 years pressure (BP) has been shown to reduce search Council (MRC) study in 1985, or black patients of any age, whereas the risk of cardiovascular (CV) morbidity bendrofluazide was documented to be al- ACE inhibitors (or angiotensin receptor and mortality. The main benefitoflow- most three times as efficacious as the blockers in the case of intolerance to ering BP is due to the reduction in the risk b-blocker hydrochloride in ACE inhibitors) should be first-line of stroke and (HF). In many preventing stroke (8). In the MRC trial drugs in hypertensive patients younger trials in which a reduction in CV events in elderly patients (6), hydrochlorothia- than 55 years of age (http://nice.org.uk/ was documented, antihypertensive ther- zide and reduced the risk of CG034guidance). apy was diuretic-based (3–8). stroke, whereas b-blockers failed to re- The Seventh Report of the Na- duce the risk of stroke despite a similar tional Committee (JNC VII) on Prevention, Effect of diuretic treatment on lowering of BP. In the International Nifed- Detection, Evaluation, and Treatment of stroke morbidity and mortality ipine GITS Study: Intervention as a Goal High recommends that In the era of placebo-controlled trials, in Hypertension Treatment (INSIGHT), thiazide diuretics should be preferred several studies attested to the efficacy of 25 mg plus amiloride drugs in “most” hypertensive patients, ei- diuretics in reducing stroke morbidity 2.5 were as effective as 30 mg for ther alone or combined with drugs from and mortality (6,7). In a recent published preventing stroke (12). other classes (2). study from China, indapamide given to In a large meta-analysis, including The present review does not intend to patients with a history of stroke or tran- 48,220 patients, Psaty et al. (13) found negate the important role of diuretics in sient ischemic attack reduced the risk of that high-dose diuretic therapy reduced certain groups of patients (blacks, - stroke by 31% (3). In the Perindopril the risk of stroke by 51%, whereas ther- sensitive patients, concomitant heart Protection Against Recurrent Stroke apy with b-blockers reduced the risk by failure) or to underestimate their role in Study (PROGRESS) (9) in patients with only 29% (P = 0.02). Klungel et al. (14) showed that among 1,237 single-drug ccccccccccccccccccccccccccccccccccccccccccccccccc users with no history of CV disease, the From the 1Internal D and Hypertension Unit, The Chaim Sheba Medical Center, Tel-Hashomer, adjusted risk of ischemic stroke was 2 to affiliated with the Sackler School of Medicine, Tel-Aviv University, Tel-Aviv, Israel; the 2Divisione di Me- 2 1/2 times higher among users of dicina, Ospedale di Assisi, Assisi, Italy; the 3Struttura Complessa di Cardiologia, Ospedale S. Maria della b-blockers, calcium antagonists, or ACE Misericordia, Perugia, Italy; and the 4Dipartimento di Medicina Interna, Università di Perugia, Perugia, Italy. inhibitors than among users of a diuretic Corresponding author: Paolo Verdecchia, [email protected]. alone. Interestingly, even in patients with This publication is based on the presentations at the 3rd World Congress on Controversies to Consensus in , and Hypertension (CODHy). The Congress and the publication of this supplement were CV disease, diuretics still conferred a lower made possible in part by unrestricted educational grants from AstraZeneca, Boehringer Ingelheim, Bristol- stroke risk than other drugs, although Myers Squibb, Daiichi Sankyo, Eli Lilly, Ethicon Endo-Surgery, Generex Biotechnology, F. Hoffmann-La the difference was considerably smaller. Roche, Janssen-Cilag, Johnson & Johnson, Novo Nordisk, Medtronic, and Pfizer. The recent Avoiding Cardiovascular DOI: 10.2337/dc11-s246 © 2011 by the American Diabetes Association. Readers may use this article as long as the work is properly Events Through Combination Therapy cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.org/ in Patients Living With Systolic Hyper- licenses/by-nc-nd/3.0/ for details. tension (ACCOMPLISH) trial showed care.diabetesjournals.org DIABETES CARE, VOLUME 34, SUPPLEMENT 2, MAY 2011 S313 Diuretics in hypertension that a combination of the ACE inhibitor studies showed the efficacy of diuretics and that this protective effect disappears with hydrochlorothiazide was in reducing CV morbidity and mortality within 4 months after use is discontinued. less effective in lowering the risk of the in the elderly (6,7,16,19). In the Systolic Thus, in addition to their use to lower BP, predefinedprimaryendpointsthanthe Hypertension in the Elderly Program thiazide plays a major role in the preven- combination of benazepril with (SHEP) (7), chlorthalidone reduced in el- tion of and fractures. (15). However, analysis of the benefitfor derly patients with isolated systolic hy- Diuretic therapy can transform non- the individual components of the pri- pertensiontherateoftotalstrokeby dippers to dippers and thereby offer an mary end points showed that, for stroke 36%, the rate of major CV events by additional therapeutic advantage of re- prevention, hydrochlorothiazide and 32%, and the rate of all-cause mortality ducing the risk of CV complications (25). amlodipine were the same. Thus, for stroke by 13%. We have shown in a meta-analysis prevention, a diuretic is superior to some that in the elderly, diuretics are more ef- Diuretic-induced elevations antihypertensive agents. fective than b-blockers in lowering BP Several studies showed that use of thia- (20). Moreover, only diuretics reduced zide diuretic increases glucose levels Effect of diuretic treatment on HF the risk of coronary heart disease and all- (4,12,26), but in these studies, the second Thiazide diuretic is very effective in pre- cause mortality (20). The ALLHAT study, drug was a b-blocker that impaired glu- venting the development of HF in hyper- which showed superiority of diuretics cose . The tensive patients. In a large meta-analysis over other antihypertensive agents in Risk in Communities (ARIC) study as- that included 18 long-term placebo- some secondary end points (see above), sessed the incidence of new-onset diabe- controlled randomized trials, high-dose was not defined as a study of the elderly, tes (NOD) after 3 and 6 years in 12,550 diuretic therapy reduced the risk of HF by but 57.5% of the participants were age adults who did not have diabetes. Patients 83% and low-dose diuretic reduced the $65 years; therefore, this study is con- who received thiazide diuretics were not risk of HF by 42% (13). In the Hyperten- sidered a study in the elderly (4,5). The at greater risk for the subsequent devel- sion in the Very Elderly Trial (HYVET), only exception was the ANBP2 study, in opment of diabetes than the subjects with indapamide reduced the rate of HF by which treatment with an ACE inhibitor in hypertension who were not receiving any 64% in very elderly patients with hyper- older subjects, particularly men, led to antihypertensive therapy (27). In this tension (16). In INSIGHT, diuretic was better outcomes than treatment with di- study, only subjects with hypertension more effective than nifedipine in prevent- uretic agents, despite similar reductions who were taking b-blockers had a 28% ing nonfatal HF (12). In the ALLHAT of BP (10). It is noteworthy that the design higher risk of subsequent diabetes. In study, chlorthalidone was superior to of the ANBP2 study was less rigorous than the ACCOMPLISH study, the effects of doxazosin, lisinopril, and amlodipine in other studies, since it was a prospective, the two treatment arms on glucose levels preventing HF (4,5). The data were vali- randomized, open-label, blinded-endpoint were not reported (15). It is likely that the dated after a rigorous evaluation of all (PROBE) study that is open to bias. In the use of diuretics with an ACE inhibitor did hospitalized HF events (17). In a subanal- ANBP2 study, only 83% of the partici- not adversely affect glucose metabolism, ysis of ALLHAT, chlorthalidone was su- pants received their assigned treatment, as we have previously shown (28). If a perior to the other agents in preventing only 58% of participants were randomly high-dose diuretic has a negative effect HF in participants with the metabolic syn- assigned to an ACE inhibitor, and 62% of on glucose metabolism, it may be related drome and in patients with diabetes those assigned to a diuretic were still re- to (29–31). Analysis of the (18). One of the arguments against the ceiving assigned treatment at the end of SHEP data showed that each 0.5 mEq/L findings of the ALLHAT study was that the study (10). In the recent HYVET (16), decrease in serum during the the achieved BP in the chlorthalidone indapamide reduced the rate of stroke, 1st year of treatment was associated with a arm was lower than the achieved BP in coronary heart disease, HF, and all-cause 45% higher adjusted diabetes risk (32). the other treatment arms. However, anal- mortality. It is noteworthy that in the pilot, Potassium supplementation or combina- yses using achieved BP levels as time- HYVET participants received either di- tion of thiazide with ACE inhibitor or dependent covariates in a Cox proportional uretic or ACE inhibitor or placebo, and potassium-sparing agents might prevent hazard regression model showed that af- only diuretics reduced the risk of stroke, thiazide-induced diabetes (33). The com- ter adjustment for BP, the differences in whereas ACE inhibitors did not reduce the bination of thiazide with an- risk of stroke and HF between treatment risk of stroke, despite a similar reduction tagonist may not only prevent NOD but arms remained statistically significant in BP (21). Thus, it seems that for elderly also improve BP control (34). It seems (18). In the ACCOMPLISH trial, the patients, a diuretic should remain the drug that not all diuretics are equal in regard combination of benazepril with hydro- of choice. to the effect on insulin resistance. Leonetti was as effective as the com- et al. (35) showed that indapamide does bination of benazepril with amlodipine in Additional advantages of diuretics not have a deleterious effect on glucose preventing HF (15). Thus, it is clear that Several studies showed that diuretics pre- tolerance. The effects of diuretic-induced diuretic is very effective and may be supe- vent the development of osteoporosis and glucose elevation on long-term CV risk rior to other agents in preventing new- reduce the risk of hip fractures (22–24). were reported in several studies. Verdecchia onset HF in hypertensive patients. In a randomized double-blind 2-year et al. (36) reported a nearly threefold trial, Reid et al. (24) showed that hydro- higher CV disease risk after 16 years Diuretics in the elderly chlorothiazide slowed cortical loss of follow-up in treated patients with hy- Hypertension is much more common in in normal postmenopausal women. pertension (54% treated with diuretics) the elderly, and in this age–group, iso- Schoofs et al. (23) showed in a prospec- who developed NOD; no relationship lated systolic hypertension is particularly tive population-based cohort study that was seen between diuretic usage and CV common. Several placebo-controlled thiazide protects against hip fractures events. In post hoc subgroup analyses of

S314 DIABETES CARE, VOLUME 34, SUPPLEMENT 2, MAY 2011 care.diabetesjournals.org Grossman and Associates the ALLHAT data, there was no significant pill combination could be titrated to 25 and flexibility of choosing among avail- association of fasting glucose level change and 10 mg, respectively. Although the able antihypertensive drugs on the basis at 2 years with subsequent coronary heart dose of amlodipine in the trial was similar of several considerations, including effi- disease, stroke, CV disease, total mortality, to that demonstrating favorable outcomes cacy, tolerability, compelling indications, or end-stage renal disease. There was no in other outcome trials, the dose range for contraindications, race, and cost. significant association of incident diabetes hydrochlorothiazide (12.5–25 mg) was at 2 years with clinical outcomes, except lower than the dose range (equivalent to Diuretics as first-line drugs: for coronary heart disease (risk ratio 1.64; 25–50 mg) used in trials demonstrating outcome-based studies P = 0.006), but the risk ratio was lower and benefits of thiazide on CV outcome ALLHAT was perceived as the trial that nonsignificant in the chlorthalidone (4,6,19). Information on supplemented conclusively demonstrated the superior- group (risk ratio 1.46; P =0.14)(37). antihypertensive agents was not reported, ity of diuretics over other classes of Analysis of the 14.3 years of follow-up but the recommended supplementary antihypertensive drugs. ALLHAT was de- from the SHEP revealed that incident di- drugs were a- and b-blockers, for which signed to test the hypothesis that the abetes during the trial among participants effects on CV outcomes are inferior. Of combined incidence of fatal CHD and randomized to placebo was associated note, a small but significant BP difference nonfatal will be with a .50% increase in CV mortality (0.9 mmHg systolic and 1.1 mmHg dia- lower by 16% in hypertensive patients but not in individuals randomized to the stolic; P , 0.001 for both) was recorded receiving a calcium antagonist (amlodi- diuretic (38). Thus, diuretic-induced glu- between the two arms of treatment favor- pine), an ACE inhibitor (lisinopril), or an cose changes may underline lesser prog- ing the ACE inhibitor–calcium antagonist a adrenergic blocker (doxazosin) as first- nostic significance. combination. The right conclusion of the line therapy than in subjects treated with ACCOMPLISH study is that hydrochloro- chlorthalidone as first-line therapy. The Other disadvantages of diuretics thiazide in a dose of #25 mg/day may be study enrolled 42,418 high-risk patients Diuretics may induce some metabolic less effective in preventing CV disease aged $55 years, and 35% were black (4). alterations that are harmful. The most than a full dose of amlodipine. The doxazosin arm was prematurely common metabolic derangement is hy- The results of this study raised the stopped because of a significantly higher ponatremia, which appears to be partic- question whether all thiazide-type diu- incidence of HF. ularlycommoninelderlywomen(39). retics are equal. Several successful di- It is frequently forgotten that the This side effect can be prevented by use uretic studies used chlorthalidone in a ALLHAT study failed to demonstrate its of a low to medium dose of diuretics and dose of up to 25 mg/day (4,5,7,41,42). A primary goal because the incidence of the by instructing patients to limit fluid in- meta-analysis of trials done until 2004 re- primary end point did not show any take. The deleterious effects of thiazide ported similar clinical CV outcomes statistical differences between the chlor- on lipid profile are mainly observed in across the class (43). However, these thalidone group and any other treatment the short term and almost disappear in studies used doses of these agents that group (6-year event rate: long-term studies (40). were higher than the 12.5–25 mg/day 11.5%, amlodipine 11.3%, lisinopril dose of hydrochlorothiazide used in the 11.4%). Compared with chlorthalidone, Role of diuretics as an add-on ACCOMPLISH study. Recent data sug- the relative risks were 0.98 (95% CI 0.90– therapy gest that chlorthalidone is 1.5- to 2-fold 1.07) for amlodipine and 0.99 (95% CI Recently, two large prospective studies more potent in lowering BP than hydro- 0.91–1.08) for lisinopril. Furthermore, cast doubt on the role of as chlorothiazide (44). Thus, to achieve the all-cause mortality did not differ between an add-on therapy (15,26). The Anglo- beneficial effect with diuretics, one the groups (4). Scandinavian Cardiac Outcomes Trial should use hydrochlorothiazide in a The only significant differences in (ASCOT) compared the b-blocker dose of up to at least 37.5 mg/day. An- ALLHATemergedintheanalysisof with the calcium antagonist amlodipine. other thiazide-like diuretic that is less dis- some secondary end points. The risk of A thiazide was added to atenolol and an cussed is indapamide. This agent has less HF, which in itself was not a prespecified ACE inhibitor was added to amlodipine on metabolic parameters secondary end point, but just a compo- when BP did not reach the goal. The pri- than other diuretics (45,46), is more ef- nent of a secondary end point (named mary end points were not significantly fective than in reducing left ven- “combined ” and different between the two regimens, but tricular mass (47), is equivalent to consisting of CHD + stroke + revascular- fewer individuals on the amlodipine- enalapril in reducing microalbuminuria ization procedures + + HF [hospi- based regimen had fatal and nonfatal (48), and is effective in reducing CV mor- talized or treated] + peripheral arterial stroke, total CV events and procedures, bidity and mortality in clinical trials disease), was 38% higher with amlodi- and all-cause mortality. From this study, (3,9,16,49). Thus, the use of indapamide pine and 15% with lisinopril than with we can only learn that atenolol is less ef- as a leading diuretic agent may be worth- chlortalidone (both P , 0.01). Further- fective than amlodipine, but we cannot while. more, the risk of stroke, a prespecified blame the diuretic in the worse outcome. secondary end point, was 7% lower with The ACCOMPLISH trial was stopped THE CON SIDE—There is no evidence amlodipine than with chlorthalidone (P = early when the data were clear that the from systematic overviews and meta- NS) and 15% higher with lisinopril than single pill combination of ACE inhibitor analyses that thiazide diuretics are superior with chlortalidone (P = 0.02). with calcium antagonist was superior to to other classes of antihypertensive drugs These results were attributed to the the combination of ACE inhibitor with a in reducing CV risk (50). These results en- lower systolic BP in the patients allocated diuretic (15). The amlodipine and hydro- dorse the position of the European guide- to chlorthalidone compared with lisinopril chlorothiazide components of the single lines, which leave to the doctor the choice (2 mmHg, P , 0.001) and amlodipine care.diabetesjournals.org DIABETES CARE, VOLUME 34, SUPPLEMENT 2, MAY 2011 S315 Diuretics in hypertension

(0.8 mmHg, P = 0.03) (4). However, Diuretics as second-line drugs: These data may relegate thiazide diuretics .90% of ALLHAT patients were receiv- outcome-based studies to third-line therapy. However, since the ing an , a diuretic ASCOT-BPLA (Blood Pressure–Lowering study population was composed of com- in most cases, at the time of randomiza- Arm) was a multicenter randomized con- plicated patients with hypertension and tion, when they abruptly abandoned the trolled trial conducted in 19,257 hyper- prior history of CHD, diabetes, or organ previous agents and were fully switched tensive patients aged 40–79 years who damage, it is unclear to what extent these to trials drugs. Thus, patients allocated had at least three other CV risk factors. findings can be extrapolated to less un- to drugs different from chlorthalidone Patients were randomized to a first-line complicated hypertensive subjects. were more likely to regain fluids, with po- treatment with either atenolol or amlodi- tential rapid disclosure of signs and pine. In the case of lack of BP control, Diuretics and new-onset diabetes symptoms of heart failure. Consistent bendroflumethiazide was added to ateno- Diuretics increase the risk of NOD. In a with this view is the early divergence of lol and perindopril to amlodipine. Hence, network meta-analysis of 22 clinical trials, the Kaplan-Meier curved after randomi- the trial compared an “old-drug” strategy ACE inhibitors, angiotensin receptor zation. However, subsequent post hoc (b-blocker alone or with a diuretic) with a blockers, calcium channel blockers, and analyses with validation of HF events “new-drug” strategy (calcium antagonist placebo were associated with a signifi- and adjustment for pre-entry diuretic alone or with an ACE inhibitor). The trial cantly lesser risk of NOD compared with use seemed to confirm the original re- was stopped prematurely after 5.5 years diuretics (51). The risk of NOD did not sults (17). Thus, ALLHAT did not meet because of statistically significant lower differ between diuretics and b-blockers its primary goal, and the evidence of incidence of all-cause mortality, CV mor- (51). superiority of chlortalidone over compa- tality, and other important secondary end Importantly, NOD was not a prespe- rators was based on the analysis of sec- points in the new-drug strategy group. The cified primary end point in any of these ondary end points. Consequently, the primary end point, a composite of nonfatal trials. Diuretic-induced hypokalemia is enthusiastic statement that the “verdict myocardial infarction and fatal CHD, did believed to be one possible cause of the from ALLHAT is that thiazide diuretics not differ between the groups (HR 0.90, rise in glucose (52), perhaps through an are the preferred initial treatment of hy- 95% CI 0.79–1.02, P = 0.105) (26). impaired insulin secretion by pancreatic pertension” was excessive. Furthermore, Although the benefits of amlodipine b-cells. Also diuretic-induced hyperuri- results obtained with chlorthalidone and perindopril over atenolol and bend- cemia was associated with impaired glu- cannot be extrapolated to hydrochloro- roflumethiazide appeared to be largely cose tolerance. thiazide or other thiazide diuretics. The driven by the 2.7 mmHg greater reduc- The controversy surrounding the is- duration of the antihypertensive effect of tion in systolic BP, this study clearly sue of NOD in treated hypertensive sub- chlorthalidone is significantly longer than demonstrated that a new-drug strategy is jects is not focused on the diabetogenic that of hydrochlorothiazide, as evidenced superior to an old-drug strategy in pa- effect of diuretics and b-blockers, which by 24-h ambulatory BP monitoring (44). tients with complicated hypertension or is taken for granted (1), but on the con- The results of ALLHAT are consistent with associated risk factors. troversial interpretation of the few data on the INSIGHT study, which failed to detect ACCOMPLISH was a double-blind the prognostic impact of NOD induced by outcome differences between a diuretic randomized study in which 11,506 pa- these drugs. (hydrochlorothiazide plus amiloride) tients with hypertension complicated by In a cohort study from our group, and a calcium antagonist (nifedipine in a organ damage or associated with diabetes NOD portended a risk for subsequent CV long-acting gastrointestinal transport sys- or overt CV disease were randomized to disease that was not dissimilar from that tem) in 6,321 hypertensive patients aged either benazepril plus amlodipine or be- of previously known diabetes. Notably, 55–80 years. Again, nonfatal HF (a sec- nazepril plus hydrochlorothiazide as first- plasma glucose at entry and diuretic ondary end point) was less frequent in step treatment. The trial was prematurely treatment at the follow-up visit were in- the diuretic group than in the calcium an- stoppedafterameanfollow-upof36 dependent predictors of NOD (36). In a tagonist group (P = 0.028) (12). months because the boundary of the pre- post hoc analysis of the Antihy- Another major study that failed to specified stopping rule was exceeded. The pertensive Long-Term Use Evaluation demonstrate the superiority of diuretics risk of primary composite end point ( (VALUE) study, the hypertensive subjects over comparators was the ANBP2 trial. from CV causes or nonfatal CV disease) was who developed NOD showed a 43% This was a randomized open-label study 20% lower with benazepril-amlodipine higher risk of cardiac morbidity when between diuretics and ACE inhibitors than with benazepril-hydrochlorothiazide compared with individuals who did not conducted in 6,083 elderly subjects with (HR 0.80, 95% CI 0.72–0.90, P , develop diabetes (53). NOD was associ- hypertension. The ACE inhibitor enala- 0.001). Also, the composite secondary ated with a marginally higher risk of myo- pril and the diuretic hydrochlorothiazide end point (death from CV causes, non- cardial infarction (P = 0.057) and a were recommended as initial therapy, but fatal myocardial infarction, and nonfatal significantly higher risk of congestive the final choice of the specificagentwas stroke) was less frequent in the benazepril- HF (P = 0.017) (53). These findings are left to investigators, who were family amlodipine group than in the benazepril- consistent with a report from the Ongoing practitioners. The primary end point of hydrochlorothiazide group (HR 0.79, 95% Alone and In Combination the study, a composite of CV morbidity CI 0.67–0.92, P = 0.002) (15). With Global End Point Trial and all-cause mortality, was marginally The ACCOMPLISH study is unique (ONTARGET), in which NOD was asso- less frequent in the ACE inhibitor group in its design by substantiating the supe- ciated with a 74% excess risk of conges- than in the diuretic group (hazard ratio riority of a fixed combination of ACE tive HF requiring hospitalization (54). [HR] 0.89, 95% CI 0.79–1.00; P = 0.05) inhibitor plus amlodipine over a fixed com- Individuals who are skeptical about (10). bination of ACE inhibitor plus diuretic. the adverse prognostic value of NOD

S316 DIABETES CARE, VOLUME 34, SUPPLEMENT 2, MAY 2011 care.diabetesjournals.org Grossman and Associates argue that NOD failed to translate into a because of the failure by most random- Although often neglected by doctors, prognostic disadvantage in most trials. ized trials to disclose a significant associ- these symptoms may be troubling in el- In the ALLHAT study, the higher in- ation between NOD and outcome. NOD, derly subjects. cidence of NOD in the chlorthalidone whether or not induced by drugs, remains Diuretics may cause several other group did not translate into a prognostic an important adverse prognostic marker adverse reactions, potentially leading to burden in this group, and a similar that should be prevented. We suggested discontinuation. Hypokalemia was sug- situationoccurredinotherstudies that in subjects at increased risk of NOD gested as a potential trigger of arrhyth- (12). However, a frequently forgotten (impaired fasting glucose, obesity, meta- mias and sudden cardiac death (60), consideration is that different risks of bolic syndrome), diuretics and b-block- although its impact is now less than in NOD are unlikely to translate into differ- ers should 1) be used cautiously, with the the past because of the widespread use ent risks of hard end points in the setting lowest effective dose and plasma glucose of low-dose thiazides, potassium-sparing of available mega-trials. We estimated periodically checked, and 2) be avoided in diuretics, and combinations with ACE in- that one CV event specifically associated subjects with BP normalized by different hibitors or angiotensin receptor blockers. with NOD may be prevented for every classes of antihypertensive drugs. Muscle may cause suspicion of 385–449 subjects treated with new, hypokalemia. is another rather than old (diuretics, b-blockers), Diuretics and lipids insidious side effect of diuretics and is antihypertensive drugs for ;4 years Thiazide diuretics increase total choles- particularly frequent in elderly women af- (55). Consequently, even large trials terol and HDL by 5–7%. In a ter prolonged use of the drug. Hyperuri- such as ALLHAT may be under-powered meta-analysis of 474 studies (57), diuret- cemia is a dose-dependent effect that may to detect the adverse prognostic impact ics increased cholesterol and triglyceride lead to acute gouty arthritis. of NOD (55). levels, and the rise in total cholesterol was Consistent with this view, in the paralleled by a rise in LDL cholesterol. CONCLUSIONS—Thiazide-type di- ALLHAT study, the incidence of coronary The rise in cholesterol was dose depen- uretics are at least as effective as b-blockers, artery disease was 64% higher (95% CI dent and greater in blacks. A reduction 15–233) in the subjects who developed in the HDL cholesterol levels was noted calcium antagonists, and ACE inhibitors in reducing CV outcomes. Thiazide di- NOD in the first 2 years of follow-up only in patients with diabetes. The poten- uretics are particularly effective in pre- than in those who did not (56). However, tially adverse prognostic impact of in- when the chlorthalidone, amlodipine, creased total and LDL cholesterol in the venting stroke and HF in hypertensive patients. These drugs are very effective in and lisinopril groups were analyzed sep- very long term may be underestimated by the elderly and very elderly patients. The arately, the excess risk of coronary artery the relatively short duration (generally 3– combined use of thiazide-like diuretics disease was significant only in the lisino- 5 years) of available intervention trials. with aldosterone antagonists may be pril group (HR 2.23, 95% CI 1.07–4.62) When dealing with a young hypertensive worthwhile. Thus, diuretics should re- and not in the other groups, although the patient, it is unlikely that an expected per- main the leading agents in the manage- P value for the interaction term was not sistent elevation of total and LDL choles- ment of hypertension. However, the significant (P = 0.21). The power of the terol over decades may be beneficial. statement that these drugs are superior chlorthalidone and amlodipine groups to other drugs in almost all patients might have been inadequate to detect Diuretics and the the adverse impact of NOD on coronary The long-term use of diuretics was asso- with hypertension is not supported by superiority studies. Particularly in the artery disease demonstrated in the total ciated with an increased risk of renal cell younger hypertensive subjects, the ben- ALLHAT cohort. We argued that the carcinoma. In a meta-analysis, Grossman efit of diuretics as first-line antihyper- lesser BP reduction in the lisinopril group et al. (58) found a 55% higher odds (95% tensive drugs should be weight against compared with the other groups might CI 42–71%, P , 0.00001) of renal cell the risk of unwanted effects in the have allowed NOD to unveil its adverse carcinoma in patients treated with diuret- long term. This holds particularly true prognostic impact (55). ics compared with diuretic nonusers. The in subjects at high risk of developing In line with this interpretation, in a renal tubular cells, which are the main diabetes. post hoc analysis of the SHEP study, NOD target of diuretics, are also the site of or- was associated with a higher risk of all- igin of malignancy. The association be- – cause mortality (HR 1.35, 95% CI 1.05 tween diuretic treatment and renal cell Acknowledgments—This study was sup- 1.72) and CV mortality (HR 1.56, 95% CI carcinoma is a potentially important issue ported in part by the Fondazione Umbra 1.12–2.18) in the placebo group, not in that requires solid confirmation in larger Cuore e Ipertensione–ONLUS, Perugia, Italy. the active treatment group (38). Because studies. No potential conflicts of interest relevant to the SHEP population was composed by this article were reported. elderly hypertensive patients at high risk Discontinuation of diuretics of events in the short term, the favorable Discontinuation of diuretics is ;83% prognostic impact of BP reduction in more likely than discontinuation of ACE References the active treatment group may have out- inhibitors (59). By causing increased pro- 1. Mancia G, De Backer G, Dominiczak A, weighed the adverse prognostic impact of duction of , diuretics may increase et al. 2007 guidelines for the manage- ment of arterial hypertension: the Task NOD. the urinary frequency. Overactive bladder fi Force for the Management of Arterial In conclusion, we should refrain from de ned as a syndrome consisting of ur- Hypertension of the European Society of underestimating the adverse prognostic gency, with or without incontinence, usu- Hypertension (ESH) and of the European impact of NOD induced by diuretics and ally associated with nocturia, is common Society of Cardiology (ESC). J Hypertens b-blockers, alone or combined, solely in older subjects treated with diuretics. 2007;25:1105–1187 care.diabetesjournals.org DIABETES CARE, VOLUME 34, SUPPLEMENT 2, MAY 2011 S317 Diuretics in hypertension

2. Chobanian AV, Bakris GL, Black HR, et al. or diuretic in the International Nifedipine an antihypertensive regimen of amlodi- The Seventh Report of the Joint National GITS study: Intervention as a Goal in Hy- pine adding perindopril as required ver- Committee on Prevention, Detection, pertension Treatment (INSIGHT). Lancet sus atenolol adding bendroflumethiazide Evaluation, and Treatment of High Blood 2000;356:366–372 as required, in the Anglo-Scandinavian Pressure: the JNC 7 report. JAMA 2003; 13. Psaty BM, Smith NL, Siscovick DS, et al. Cardiac Outcomes Trial-Blood Pressure 289:2560–2572 Health outcomes associated with anti- Lowering Arm (ASCOT-BPLA): a multi- 3. Liu L, Wang Z, Gong L, et al. Blood hypertensive used as first-line centre randomised controlled trial. Lancet pressure reduction for the secondary agents: a and meta- 2005;366:895–906 prevention of stroke: a Chinese trial and a analysis. JAMA 1997;277:739–745 27. Gress TW, Nieto FJ, Shahar E, Wofford systematic review of the literature. Hy- 14. Klungel OH, Heckbert SR, Longstreth WT MR, Brancati FL. Hypertension and anti- pertens Res 2009;32:1032–1040 Jr, et al. Antihypertensive drug therapies hypertensive therapy as risk factors for 4. ALLHAT Officers and Coordinators for and the risk of ischemic stroke. Arch In- mellitus: Atherosclerosis the ALLHAT Collaborative Research tern Med 2001;161:37–43 Risk in Communities Study. N Engl J Med Group. The Antihypertensive and Lipid- 15. Jamerson K, Weber MA, Bakris GL, et al. 2000;342:905–912 Lowering Treatment to Prevent Heart Benazepril plus amlodipine or hydro- 28. Shamiss A, Carroll J, Peleg E, Grossman E, Attack Trial. Major outcomes in high- chlorothiazide for hypertension in high- Rosenthal T. The effect of enalapril with risk hypertensive patients randomized to risk patients. N Engl J Med 2008;359: and without hydrochlorothiazide on in- angiotensin-converting 2417–2428 sulin sensitivity and other metabolic ab- or vs. diuretic: 16. Beckett NS, Peters R, Fletcher AE, et al. normalities of hypertensive patients with the Antihypertensive and Lipid-Lowering Treatment of hypertension in patients 80 NIDDM. Am J Hypertens 1995;8:276–281 Treatment to Prevent Heart Attack years of age or older. N Engl J Med 2008; 29. Plavinik FL, Rodrigues CI, Zanella MT, Trial (ALLHAT). JAMA 2002;288:2981– 358:1887–1898 Ribeiro AB. Hypokalemia, glucose in- 2997 17. Einhorn PT, Davis BR, Massie BM, et al. tolerance, and hyperinsulinemia during 5. ALLHAT Collaborative Research Group. The Antihypertensive and Lipid Lowering diuretic therapy. Hypertension 1992;19 Major cardiovascular events in hyperten- Treatment to Prevent Heart Attack Trial (Suppl.):II26–II29 sive patients randomized to doxazosin vs. (ALLHAT) Heart Failure Validation Study: 30. Amery A, Berthaux P, Bulpitt C, et al. chlorthalidone: the Antihypertensive and diagnosis and prognosis. Am Heart J 2007; Glucose intolerance during diuretic ther- Lipid-Lowering Treatment to Prevent Heart 153:42–53 apy: results of trial by the European Attack Trial (ALLHAT). JAMA 2000;283: 18. Wright JT Jr, Probstfield JL, Cushman Working Party on Hypertension in the 1967–1975 WC, et al. ALLHAT findings revisited in Elderly. Lancet 1978;1:681–683 6. MRC Working Party. Medical Research the context of subsequent analyses, other 31. Langford HG, Cutter G, Oberman A, Council trial of treatment of hypertension trials, and meta-analyses. Arch Intern Med Kansal P, Russell G. The effect of thiazide in older adults: principal results. BMJ 2009;169:832–842 therapy on glucose, insulin and choles- 1992;304:405–412 19. Amery A, Birkenhäger W, Brixko P, et al. terol metabolism and of glucose on po- 7. SHEP Cooperative Research Group. Pre- Mortality and morbidity results from the tassium: results of a cross-sectional study vention of stroke by antihypertensive European Working Party on High Blood in patients from the Hypertension De- drug treatment in older persons with Pressure in the Elderly trial. Lancet 1985; tection and Follow-up Program. J Hum isolated systolic hypertension: final results 1:1349–1354 Hypertens 1990;4:491–500 of the Systolic Hypertension in the Elderly 20. Messerli FH, Grossman E, Goldbourt U. 32. Shafi T, Appel LJ, Miller ER 3rd, Klag MJ, Program (SHEP). JAMA 1991;265:3255– Are beta-blockers efficacious as first-line Parekh RS. Changes in serum potassium 3264 therapy for hypertension in the elderly? A mediate thiazide-induced diabetes. Hy- 8. Medical Research Council Working Party. systematic review. JAMA 1998;279: pertension 2008;52:1022–1029 MRC trial of treatment of mild hyperten- 1903–1907 33. Jeunemaitre X, Charru A, Chatellier G, sion: principal results. Br Med J (Clin Res 21. Bulpitt CJ, Beckett NS, Cooke J, et al. et al. Long-term metabolic effects of spi- Ed) 1985;291:97–104 Results of the pilot study for the Hyper- ronolactone and thiazides combined with 9. PROGRESS Collaborative Group. Rand- tension in the Very Elderly Trial. J Hy- potassium-sparing agents for treatment omised trial of a perindopril-based blood- pertens 2003;21:2409–2417 of . Am J Cardiol pressure-lowering regimen among 6,105 22. Bolland MJ, Ames RW, Horne AM, Orr- 1988;62:1072–1077 individuals with previous stroke or tran- Walker BJ, Gamble GD, Reid IR. The effect 34. Sharabi Y, Adler E, Shamis A, Nussinovitch N, sient ischaemic attack. Lancet 2001;358: of treatment with a thiazide diuretic for Markovitz A, Grossman E. Efficacy of add- 1033–1041 4 years on bone density in normal post- on aldosterone receptor blocker in un- 10. Wing LM, Reid CM, Ryan P, et al. A menopausal women. Osteoporos Int controlled hypertension. Am J Hypertens comparison of outcomes with angiotensin- 2007;18:479–486 2006;19:750–755 converting–enzyme inhibitors and di- 23. Schoofs MW, van der Klift M, Hofman A, 35. Leonetti G, Rappelli A, Salvetti A, Scapellato L. uretics for hypertension in the elderly. et al. Thiazide diuretics and the risk for Long-term effects of indapamide: final re- N Engl J Med 2003;348:583–592 hip fracture. Ann Intern Med 2003;139: sults of a two-year Italian multicenter study 11. Hansson L, Lindholm LH, Niskanen L, 476–482 in systemic hypertension. Am J Cardiol et al. Effect of angiotensin-converting- 24. Reid IR, Ames RW, Orr-Walker BJ, et al. 1990;65:67H–71H enzyme inhibition compared with con- Hydrochlorothiazide reduces loss of cor- 36. Verdecchia P, Reboldi G, Angeli F, et al. ventional therapy on cardiovascular tical bone in normal postmenopausal Adverse prognostic significance of new morbidity and mortality in hypertension: women: a randomized controlled trial. diabetes in treated hypertensive subjects. the Prevention Project (CAPPP) Am J Med 2000;109:362–370 Hypertension 2004;43:963–969 randomised trial. Lancet 1999;353:611– 25. Uzu T, Kimura G. Diuretics shift circadian 37. Barzilay JI, Davis BR, Cutler JA, et al. 616 rhythm of blood pressure from nondipper Fasting glucose levels and incident di- 12. Brown MJ, Palmer CR, Castaigne A, to dipper in essential hypertension. Cir- abetes mellitus in older nondiabetic adults et al. Morbidity and mortality in patients culation 1999;100:1635–1638 randomized to receive 3 different classes randomised to double-blind treatment 26. Dahlöf B, Sever PS, Poulter NR, et al. of antihypertensive treatment: a report from with a long-acting calcium-channel blocker Prevention of cardiovascular events with the Antihypertensive and Lipid-Lowering

S318 DIABETES CARE, VOLUME 34, SUPPLEMENT 2, MAY 2011 care.diabetesjournals.org Grossman and Associates

Treatment to Prevent Heart Attack Trial on ambulatory and office blood pressure. 52. Zillich AJ, Garg J, Basu S, Bakris GL, (ALLHAT). Arch Intern Med 2006;166: Hypertension 2006;47:352–358 Carter BL. Thiazide diuretics, potassium, 2191–2201 45. Ames RP. A comparison of blood lipid and and the development of diabetes: a quan- 38. Kostis JB, Wilson AC, Freudenberger RS, blood pressure responses during the titative review. Hypertension 2006;48: Cosgrove NM, Pressel SL, Davis BR; SHEP treatment of systemic hypertension with 219–224 Collaborative Research Group. Long-term indapamide and with thiazides. Am J 53. Aksnes TA, Kjeldsen SE, Rostrup M, effect of diuretic-based therapy on fatal Cardiol 1996;77:12b–16b Omvik P, Hua TA, Julius S. Impact of new- outcomes in subjects with isolated systolic 46. Gerber A, Weidmann P, Bianchetti MG, onset diabetes mellitus on cardiac outcomes hypertension with and without diabetes. et al. Serum lipoproteins during treat- in the Valsartan Antihypertensive Long- Am J Cardiol 2005;95:29–35 ment with the antihypertensive agent term Use Evaluation (VALUE) trial pop- 39. Sharabi Y, Illan R, Kamari Y, et al. Diuretic indapamide. Hypertension 1985;7:II164– ulation. Hypertension 2007;50:467–473 induced hyponatraemia in elderly hyper- II169 54. Held C, Gerstein HC, Yusuf S, et al. Glu- tensive women. J Hum Hypertens 2002; 47. Gosse P, Sheridan DJ, Zannad F, et al. cose levels predict hospitalization for 16:631–635 Regression of left congestive heart failure in patients at high 40. Lakshman MR, Reda DJ, Materson BJ, in hypertensive patients treated with in- cardiovascular risk. Circulation 2007; Cushman WC, Freis ED; Department of dapamide SR 1.5 mg versus enalapril 20 115:1371–1375 Veterans Affairs Cooperative Study Group mg: the LIVE study. J Hypertens 2000;18: 55. Verdecchia P, Angeli F, Reboldi G. New- on Antihypertensive Agents. Diuretics 1465–1475 onset diabetes, antihypertensive treat- and beta-blockers do not have adverse 48. Marre M, Puig JG, Kokot F, et al. Equiv- ment, and outcome. Hypertension 2007; effects at 1 year on plasma lipid and li- alence of indapamide SR and enalapril on 50:459–460 poprotein profiles in men with hyperten- microalbuminuria reduction in hyper- 56. Barzilay JI, Cutler JA, Davis BR. Antihy- sion. Arch Intern Med 1999;159:551–558 tensive patients with type 2 diabetes: the pertensive and risk of di- 41. Multiple Risk Factor Intervention Trial NESTOR Study. J Hypertens 2004;22: abetes mellitus. Curr Opin Nephrol Research Group. Mortality after 10 1/2 1613–1622 Hypertens 2007;16:256–260 years for hypertensive participants in the 49. Patel A, MacMahon S, Chalmers J, et al. 57. Kasiske BL, Ma JZ, Kalil RS, Louis TA. Multiple Risk Factor Intervention Trial. Effects of a fixed combination of peri- Effects of antihypertensive therapy on se- Circulation 1990;82:1616–1628 ndopril and indapamide on macrovascular rum lipids. Ann Intern Med 1995;122: 42. Hypertension Detection and Follow-up and microvascular outcomes in patients 133–141 Program Cooperative Group. Five-year with type 2 diabetes mellitus (the ADVANCE 58. Grossman E, Messerli FH, Goldbourt U. findings of the hypertension detection trial): a randomised controlled trial. Lancet Does diuretic therapy increase the risk of and follow-up program. I. Reduction in 2007;370:829–840 renal cell carcinoma? Am J Cardiol 1999; mortality of persons with high blood 50. Turnbull F; Blood Pressure Lowering 83:1090–1093 pressure, including mild hypertension. Treatment Trialists’ Collaboration. Effects 59.CorraoG,ZambonA,ParodiA,etal. JAMA 1979;242:2562–2571 of different blood-pressure-lowering reg- Discontinuation of and changes in drug 43. Psaty BM, Lumley T, Furberg CD. imens on major cardiovascular events: therapy for hypertension among newly- Meta-analysis of health outcomes of results of prospectively-designed over- treated patients: a population-based study chlorthalidone-based vs. nonchlorthalidone- views of randomised trials. Lancet 2003; in Italy. J Hypertens 2008;26:819–824 based low-dose diuretic therapies. JAMA 362:1527–1535 60. Siegel D, Hulley SB, Black DM, et al. Di- 2004;292:43–44 51. Elliott WJ, Meyer PM. Incident diabetes in uretics, serum and intracellular electrolyte 44. Ernst ME, Carter BL, Goerdt CJ, et al. clinical trials of antihypertensive drugs: levels, and ventricular in hy- Comparative antihypertensive effects of a network meta-analysis. Lancet 2007; pertensive men. JAMA 1992;267:1083– hydrochlorothiazide and chlorthalidone 369:201–207 1089

care.diabetesjournals.org DIABETES CARE, VOLUME 34, SUPPLEMENT 2, MAY 2011 S319

Top View