Real‐World Evidence Supports Optimally Dosed Thiazide‐Type Diuretics As Preferred in Treatment Regimens of Older Adults with Hypertension

To the Editor: Makam and colleagues1 explored the clinical relevance of metabolic adverse events associated with thiazide diuretic antihypertensive therapy in high-risk older adults. We commend them for attempting to “validate results of clinical trials in real-world settings.” Unfortunately, the 2,120 veterans studied were considerably more selective than the hundreds of thousands of participants (including the Department of Veteran Affairs (VA) patients) randomized in relevant clinical trials conducted since the 1950s. They used national VA administrative data to study veterans aged 65 and older selected from a sample of 19,878 who met their inclusion and exclusion criteria within an available population of 35,865 with documented hypertension but untreated for 9 months or longer. Insufficient information was provided, including initial and achieved blood pressure (BP) levels, to explain why the 1,060 “cases” were started on thiazide diuretics and why the 18,186 “controls” remained untreated. The total number of veterans with hypertension (probably 1–2 million) was not provided, and the exceptional 77% BP control rates to less than 140/90 mmHg (http://www.va.gov/QUALITYOFCARE/initiatives/compare/high-blood-pressure-control.asp) were not mentioned. Those prescribed thiazide-type diuretics appeared similar to the matched 1,060 controls for the baseline variables reported (~12% had known malignancy), but propensity matching in an observational study design cannot control for unknown factors to the same degree as randomization in placebo-controlled clinical trials and does not provide the benefit of blinding to minimize ascertainment bias. In contrast to Makam and colleagues’1 focus on intermediate outcomes, selective hospitalizations, and emergency department visits, prospective randomized clinical trials have systematically collected clinical outcomes in a masked manner to assure complete and bias-free reporting of endpoints. In the Hypertension in the Very Elderly Trial,2 diuretic treatment (vs placebo) resulted in 64% lower risk of heart failure (HF) (P < .001), 21% lower risk of all-cause mortality (P = .02), 30% lower risk of stroke (P = .06), and fewer serious adverse events (P = .001), prompting early termination of the trial because of the apparent benefit of diuretic treatment. In the Systolic Hypertension in the Elderly Program,3 diuretic treatment (vs placebo) yielded 36% lower risk of stroke (P < .001), 27% lower risk of nonfatal myocardial infarction or coronary heart disease death (P = .01), and 49% lower risk of HF (P < .001).4 Based on the Second Australian National Blood Pressure Study (ANBP2)5 and a network meta-analysis,6 Makam and colleagues concluded that other antihypertensive classes are as effective as thiazides in reducing cardiovascular events in older adults, but the ANBP2 did not raise safety concerns regarding diuretics, and the authors of the network meta-analysis concluded that “Low-dose diuretics are the most effective first-line treatment for preventing the occurrence of cardiovascular disease morbidity and mortality. Clinical practice and treatment guidelines should reflect this evidence, and future trials should use low-dose diuretics as the standard for clinically useful comparisons.” In the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT),7 the thiazide-like diuretic chlorthalidone was compared with drugs from three classes of antihypertensive agents with a better metabolic profile (the angiotensin-converting enzyme inhibitor lisinopril, the calcium-channel blocker amlodipine, and the alpha-receptor blocker doxazosin) for prevention of all-cause mortality, cardiovascular disease (CVD) and renal failure. Participants were individuals with hypertension aged 55 and older (N = 42,418) at high risk of experiencing CVD events recruited from 623 mostly primary care practice clinics across the United States, Canada, Puerto Rico, and the U.S. Virgin Islands. Participants included 15,084 blacks, 8,072 Hispanics, 7,067 enrolled from 70 VA clinics, and 19,841 women. The only general medical exclusion was a known illness that might lead to death from a non-CVD during the trial. Discontinuation rates of assigned medications, including symptomatic adverse effects, were highest in the lisinopril arm (Table 1). Makam and colleagues cite the association between hypokalemia and mortality in ALLHAT8 without mentioning the statistically significant disparity (interaction P < .01) in hazard ratios across treatment arms, which strongly suggested that the greater mortality in the chlorthalidone arm was due to underlying conditions, such as malignancies, associated with loss of potassium and high mortality rather than a specific effect of the diuretic. Also, they failed to mention the association between hyperkalemia and CVD and that age-related declines in renal function make older adults susceptible to development of hyperkalemia, especially when treated with nonthiazide antihypertensives. It has been noted that ALLHAT was the only hypertension treatment trial with adequate power to detect moderate but important differences in several major clinical outcomes.9 Chlorthalidone (12.5–25 mg/d) was unsurpassed in preventing CVD and renal outcomes. It was superior to doxazosin (2–8 mg/d), lisinopril (10–40 mg/d), and amlodipine (2.5–10 mg/d) in preventing new-onset HF and superior to doxazosin and lisinopril in preventing stroke (for lisinopril, in blacks only).4, 10 Laboratory abnormalities associated with diuretic treatment of hypertension are uncommon, usually easy to manage, and typically of no serious consequence. The cost in lives lost, disability, and medical care provision resulting from stroke and HF make use of properly dosed diuretics (chlorthalidone 12.5–25 mg/d or equivalent doses of other thiazide-type diuretics) preferred in treatment regimens of most individuals with hypertension. This work was supported in part by Contracts NO1-HC-35130 and HHSN268201100036C from the National Heart, Lung, and Blood Institute of the National Institutes of Health. Conflicts of Interest: W.C. Cushman has received honoraria from Takeda, Novartis, and Daiichi-Sankyo. Other authors report no relationships relevant to the contents of this paper to disclose. Author Contributions: All authors contributed to this paper. Sponsor's Role: This work reflects the views of the authors and does not necessarily reflect the positions of the National Heart, Lung, and Blood Institute or the National Institute on Aging, the National Institutes of Health, or the U.S. Department of Health and Human Services.