Diuretics are one of the most important classes of drugs used in hypertension. In simplistic terms, the basic premise for their use is that they reduce extracellular fluid volume and increase sodium excretion, which leads to reductions in blood pressure. Historically, the degree of efficacy in lowering blood pressure was considered directly proportional to the dose. Diuretics have been used in hypertension for 60 years and they are now appreciated as a much more complex class of drugs, and their role in therapy continues to evolve. Previous guidelines endorsed them as first-line therapy, whereas the most recent 2014 Hypertension Guidelines suggested they were now one of five drug classes acceptable as initial agents. Regardless of where and when they are employed, it is universally agreed that diuretics remain a mainstay in the pharmacologic management of hypertension.
This chapter will initially focus on thiazides (hydrochlorothiazide [HCTZ], indapamide, metolazone), and thiazide-like (chlorthalidone) agents because they account for the vast majority of diuretic use in primary hypertension and are the only diuretic agents to demonstrate efficacy in reducing hypertensive-related morbidity and mortality.
Loop diuretics (bumetanide, furosemide, or torsemide) have a more specialized role in hypertension, particularly for patients with kidney disease, low glomerular filtration rate, or with accompanying edematous disorders. Potassium-sparing diuretics (especially spironolactone) have been shown to be effective for patients with resistant or difficult to control blood pressure. These agents will also be briefly discussed in this chapter.
Pharmacology
Thiazides and Thiazide-Like Diuretics
These diuretics act primarily within the early distal convoluted tubule to reduce the reabsorption of sodium and chloride. Thiazide and thiazide-like diuretics must be secreted into the renal tubule, which is reduced with significant renal impairment. Although some antihypertensive activity remains even in the presence of significant renal impairment, most clinicians switch to loop diuretics when glomerular filtration rate (GFR) is less than 20 to 30 mL/min. The exact GFR threshold at which efficacy is lost is not well studied, and the switch to loop diuretics is based largely on the theory that there is a ceiling effect of thiazides in chronic kidney disease (CKD) that is controlled by several factors, including the reduced delivery of filtered solute and drug to the distal tubule site of action, and the fact that only a small amount of sodium reabsorption occurs in the distal tubule even under normal circumstances. Chlorthalidone, perhaps because of its long-acting nature, has been shown to remain effective at usual doses in patients with uncontrolled hypertension and chronic kidney disease and may be a preferred choice in this setting.
Thiazide, thiazide-like and loop diuretics (below) all result in initial volume depletion which then stimulates the renin, aldosterone, angiotensin system (RAAS). Sodium depletion can also lead to increased serum aldosterone. In extreme cases, these effects can lead to a decreased antihypertensive effect and in some cases resistant hypertension. Diuretics are more effective when given first in a regimen rather than second, probably because they prime the RAAS. These effects make the addition of a RAAS blocker an attractive combination to improve blood pressure (BP) control.
Hydrochlorothiazide (HCTZ) is the most commonly prescribed thiazide diuretic. However, chlorthalidone use has increased in recent years in large part because of its longer duration of action and indirect evidence suggesting it may be superior to HCTZ in reducing morbidity from hypertension. Better outcomes with chlorthalidone are plausible as a result of differences in the pharmacokinetic properties of these two agents. Chlorthalidone has a much longer duration of action, and is nearly twice as potent as HCTZ ( Table 22.1 ). Chlorthalidone is different from other diuretics because it heavily compartmentalizes into red blood cells by binding to carbonic anhydrase and then slowly “backleaks” into serum. This backleaking leads to an equilibrium between the amount of drug bound to carbonic anhydrase in the red blood cell compartment and the amount of free drug available in the plasma compartment. This theorized depot effect presumably results in a prolonged, low-level diuresis which sustains its antihypertensive action and mitigates the rebound antinaturetic period occurring when the plasma level of the diuretic falls below the threshold for diuresis.
| Drug | Percent Absorbed | Onset (Hours) | Peak (Hours) | Half-Life (Hours) | Duration (Hours) | Evidence-Based Dose a (Mg/Day) | Number of Doses per Day b |
|---|---|---|---|---|---|---|---|
| Thiazide-Like | |||||||
| Bendroflumethiazide | 90-100 | 2 | 4 | 3-4 | 6-12 | 10 | 1 |
| Chlorthalidone | 65 | 2-3 | 2-6 | 45-60 | 48-72 | 12.5-25 | 1 |
| Hydrochlorothiazide | 65-75 | 2 | 4-6 | 8-15 | 12-16 | 25-50 | 1-2 |
| Indapamide | 90 | 1-2 | 2 | 15-20 | 24-36 | 1.25 | 1 |
| Potassium-Sparing | |||||||
| Amiloride | 20 | 2 | 6-10 | 6-9 | 24 | 5-10 | 1 |
| Eplerenone | 70 | 1-2 | 2 | 4-6 | 24 | 25-50 | 1 |
| Spironolactone | 90 | 24-48 | 48-72 | 48-72 | 24-36 c | 12.5-50 | 1 |
| Triamterene | >80 | 2-4 | 6-8 | 3 | 12-16 | 100-200 | 2 |
| Loop | |||||||
| Bumetanide | 72-96 | 0.5-1 | 1-2 | 1-2 | 4-6 | 0.5-2 | 1-2 |
| Furosemide | 10-100 | 0.5-1 | 6-8 | 1.5-2 | 6-8 | 40-80 | 2 |
| Torsemide | 80 | 0.5-1 | 1-2 | 3.5 | 6-8 | 5-10 | 1-2 |
a Only the thiazide-like agents have been demonstrated to reduce morbidity and mortality.
b Daily doses for a sustained antihypertensive effect
Table 22.1 compares the pharmacokinetics and pharmacodynamics of the most commonly used diuretics. The comparative antihypertensive effects of equivalent doses of chlorthalidone and HCTZ has only recently been examined. Ernst et al. compared 50 mg of HCTZ with 25 mg of chlorthalidone in how each influenced both office and 24-hour ambulatory BP monitoring values. In spite of similar reductions in clinic BP, 24-hour monitoring revealed a significantly lower nighttime BP with chlorthalidone even though the dose of HCTZ was twice that of chlorthalidone. Peterzan et al. performed a meta-analysis and found that the estimated dose of bendroflumethiazide, chlorthalidone, or HCTZ to reduce systolic BP by 10 mm Hg was 1.4, 8.6, and 26.4 mg, respectively, and there was no evidence of a difference in maximum reduction of systolic BP (SBP) by high doses of different thiazides. Potency series for diastolic BP, serum potassium, and urate were similar to those seen for SBP. Their data suggests that chlorthalidone is three times as potent as HCTZ. These and other data also indicate that HCTZ should ideally be given twice daily compared with once daily with chlorthalidone.
Potassium-Sparing Agents
Potassium-sparing agents can be divided into those that antagonize aldosterone (spironolactone and eplerenone) and those independent of aldosterone (amiloride and triamterene). The latter are in the class of epithelial sodium channel blockers. Potassium-sparing agents are not primary monotherapies for hypertension; however, they have been used to counter the potassium wasting effects, and/or the increases in aldosterone following the use of other diuretics. More recently, amiloride and spironolactone have been shown to be highly effective at achieving better blood pressure when combined with other agents for patients with resistant or difficult to control blood pressure. This latter effect makes these agents particularly important for the small percentage of patients who cannot achieve BP control despite the use of multiple medications.
All of the agents in this class inhibit sodium absorption in the distal tubule and the collecting duct. With the reduction in sodium/potassium ATPase, potassium secretion is reduced. This effect can lead to hyperkalemia and further limit use in patients with reduced renal function and in some with heart failure. These agents also reduce the excretion of calcium and magnesium.
Spironolactone has two active metabolites, 7α-thiomethylspirolactone and canrenone. These metabolites result in a slow onset of action and a very long half-life. Spironolactone has been shown to cause impressive reductions in SBP (20 to 30 mm Hg) when combined with other agents for patients with resistant or difficult to control blood pressure. This latter effect makes these agents particularly important for the small percentage of patients, including African Americans, who often cannot achieve BP control despite the use of multiple medications. One randomized, crossover study evaluated spironolactone, doxazosin, and bisoprolol in 335 subjects with uncontrolled hypertension despite maximal doses of 3 medications. The average reduction in home SBP by spironolactone was significantly greater than placebo (−8.70 mm Hg [95% confidence interval {CI} −9.72 to −7.69]; p < 0.0001), or compared with doxazosin (−4.03 [−5.04 to −3.02]; p < 0.0001) or bisoprolol (−4.48 [−5.50 to −3.46]; p < 0.0001). Although spironolactone has not been shown to lower morbidity or mortality in hypertension, it did reduce mortality 30% when added to standard therapy for patients with heart failure (HF).
Eplerenone is an mineralocorticoid receptor antagonist with greater selectivity for the aldosterone receptor and less for androgen and progesterone receptors leading to less gynecomastia than with spironolactone. Eplerenone is a weak diuretic but it has antihypertensive effects similar to angiotensin converting-enzyme (ACE)-inhibitors and calcium channel blockers (CCBs). Eplerenone has also been shown to be effective for patients with resistant hypertension. Eplerenone reduced morbidity and mortality in patients with a recent myocardial infarction (MI) and left ventricular dysfunction or HF, as well as in systolic HF and mild symptoms.
Finerenone is another mineralocorticoid receptor antagonist currently being evaluated in clinical trials. It appears to cause less hyperkalemia than spironolactone or eplerenone, especially in patients with chronic kidney disease. One study evaluated finerenone in doses of 7.5, 10, 15, and 20 mg daily in patients with diabetes and high or very high albuminuria. There was a dose-dependent reduction in urine albumin-to-creatinine ratio with finerenone at 90 days compared with baseline with 7.5, 10, 15, and 20 mg per day groups (7.5 mg/day, 0.79 [90% CI, 0.68 to 0.91; p = 0.004]; 0.76 [90% CI, 0.65 to 0.88; p = 0.001]; 0.67 [90% CI, 0.58 to 0.77; p < 0.001]; 0.62 [90% CI, 0.54 to 0.72; p < 0.001], respectively). The secondary outcome discontinuation as a result of hyperkalemia was not observed in the placebo and finerenone 10 mg per day groups, whereas the incidences in the finerenone 7.5, 15, and 20 mg per day groups were 2.1%, 3.2%, and 1.7%, respectively. These investigators concluded that the use of finerenone in diabetic patients with nephropathy, many of whom received RAAS-blocking therapy, reduced the urinary albumin-creatinine ratio with minimal risk of hyperkalemia.
Amiloride is actively secreted in the proximal tubule and blocks sodium excretion. Amiloride is cleared extensively by the kidney and accumulates in patients with CKD and can cause hyperkalemia. If amiloride therapy is needed in patients with CKD, the dose should be reduced or the dosing frequency decreased. Amiloride was compared with spironolactone in blacks who still had uncontrolled hypertension despite a diuretic and calcium channel blocker. The addition of these drugs resulted in reductions in systolic and diastolic blood pressures (mm Hg) that were, respectively, 9.8 and 3.4 for amiloride ( p < 0.001) and 4.6 ( p = 0.006) and 1.8 for spironolactone ( p = 0.07). These authors concluded that treatment with either amiloride or spironolactone provided an additional reduction in blood pressure in blacks already receiving conventional antihypertensive therapy. Amiloride alone or in combination with HCTZ was evaluated to determine their effects on serum potassium and glucose. These investigators found that equipotent doses on BP of the combination of amiloride with hydrochlorothiazide prevented glucose intolerance and improved BP control compared with monotherapy with either drug alone.
Triamterene is a weak antihypertensive so it is typically combined with HCTZ to minimize hypokalemia and hypomagnesemia. Triamterene is metabolized to an active metabolite and both accumulate in patients with CKD. Triamterene would rarely be necessary in CKD but if used the dosage should be reduced to prevent hyperkalemia. Triamterene is a potential nephrotoxin, and is associated with formation of crystals, nephrolithiasis, and interstitial nephritis. It can cause acute kidney injury (AKI) when given with other potentially nephrotoxic drugs, such as nonsteroidal antiinflammatory agents, perhaps as a result of increased renal vascular resistance and reduced renal blood flow.
Loop Diuretics
Loop diuretics primarily act on the ascending limb of the loop of Henle to inhibit sodium and chloride reabsorption. Furosemide is historically the most commonly prescribed loop diuretic. However, furosemide has erratic absorption and unpredictable bioavailability. Bumetanide and torsemide have more predictable absorption and longer durations of action, and may be preferred over furosemide.
Loop diuretics are not a primary therapy in uncomplicated hypertensive patients and thus do not have the same evidence as thiazides in lowering hypertensive-related morbidity and mortality. Loop diuretics are relatively weak antihypertensive agents. The pharmacologic basis for this lies primarily in their short duration of action; when the blood level falls below the diuretic threshold, it prompts a compensatory period of postdose sodium retention that is a natural response designed to mitigate extracellular fluid loss. Other adaptive responses contribute, but the net result limits their antihypertensive effect. Loop diuretics have an important role in hypertensive patients who have chronic kidney disease and/or accompanying edematous disorders such as heart failure, and nephrotic syndrome. In these situations, thiazides may have more limited antihypertensive efficacy and may result in inadequate diuresis of the underlying edema. In some isolated situations, loops can be considered for use in combination with a thiazide in a therapeutic strategy known as sequential nephron blockade, but the risks of electrolyte imbalances and hypoperfusion necessitate use of these combinations only in rare circumstances.
Clinical Trials
Previous U.S. guidelines had suggested that diuretics were the preferred drug class for the initial treatment of hypertension. However, the 2014 National Guideline Committee did not feel there was sufficient evidence to suggest that diuretics be preferred and, instead, listed them as one of five preferred drug classes.
Several thiazide-type diuretics and/or other agents in combination with diuretics have been studied as initial therapy ( Table 22.2 ). It is interesting to observe that of the trials with neutral or negative results, HCTZ was used, whereas those with more favorable outcomes used chlorthalidone. The HCTZ studies with favorable outcomes typically used doses of 50 mg daily or more and frequently administered HCTZ twice daily. Several of the modern diuretic trials establishing the role of diuretic therapy in hypertension are further discussed later.
| Clinical Trials With Hydrochlorothiazide-Based Regimens | |||
|---|---|---|---|
| Trial (Year Published) | Regimen a | Population | Outcome |
| Oslo Hypertension Study (1982) | Hydrochlorothiazide (HCTZ) 50 mg/day (36%), HCTZ 50 mg/day + propranolol 320 mg/day (26%), HCTZ 50 mg/day + methyldopa 1000 mg/day (20%), or other drugs (18%) compared with no treatment | Men aged 40-49 years (406 treated and 379 untreated); Blood pressure (BP) (baseline): 156/97 mm Hg; 17/10 mm Hg reduction (systolic blood pressure/diastolic blood pressure [SBP/DBP]) vs. untreated | More noncoronary events in untreated ( p < 0.001); more coronary events in treated (14) compared with untreated (3), ( p < 0.01) |
| European Working Party on High Blood Pressure in the Elderly (EWPHBPE) (1985) | HCTZ 25-50 mg + triamterene 50-100 mg; methyldopa 500-2000 mg could be added, compared with placebo | 840 patients over 60 years blood pressure (BP; baseline): 183/101 mm Hg; BP (active treatment): 148/85 mm Hg; BP (placebo): 167/90 mm Hg | Cardiovascular (CV) mortality reduced 27% ( p = 0.037), cardiac mortality 38% ( p = 0.036) in active treatment group; no change in total mortality ( p = 0.41) |
| Medical Research Council (MRC) (1992) | HCTZ 25-50 mg + amiloride 2.5-5 mg vs. atenolol 50 mg vs. placebo | 4396 patients 65-74 years BP (baseline): 185/91 mm Hg; BP (diuretic): 150/78 mm Hg; BP (atenolol): 152/78 mm Hg | 31% fewer strokes ( p = 0.04) in HCTZ group, 44% fewer coronary events ( p = 0.0009), and 35% fewer CV events ( p = 0.0005) compared with placebo. No significant reductions in outcomes for atenolol group. |
| Multicenter Isradipine Diuretic Atherosclerosis Study (MIDAS) (1996) | HCTZ 12.5-25 mg BID vs. Isradipine 2.5-5 mg two times daily; open label enalapril could be added | 883 patients; mean age = 58 years. BP with HCTZ decreased from 149/96 mm Hg to 130/82 mm Hg; BP with isradipine decreased from 151/97 mm Hg to 135/84 mm Hg | Fewer major vascular events (3.2% vs. 5.7%, p = 0.07) and fewer nonmajor vascular events (5.2% vs. 9.1%, p = 0.02) with HCTZ than isradipine |
| International Nifedipine GITS Study (INSIGHT) (2000) | HCTZ 25-50 mg + amiloride 50-100 mg vs. nifedipine GITS 30-60 mg and atenolol or enalapril could be added | 6321 patients aged 55-80 years; similar BP decline in both groups from 173/99 mm Hg to 138/82 mm Hg | No difference overall in combined (primary and secondary) endpoints. However, there were more fatal myocardial infarctions (16 vs. 5, odds ratio [OR] 3.2, p = 0.017) and nonfatal heart failure (24 vs. 11, OR 2.2, p = 0.028) with nifedipine compared with the diuretic. |
| Second Australian National Blood Pressure Study (ANBP-2) (2003) | Randomized to either HCTZ or enalapril (doses were adjusted by family practitioner and not reported) | 6083 patients aged 65 to 84 years: BP (baseline) 168/91 mm Hg; BP (HCTZ): 144/81 mm Hg; BP (enalapril): 145/81 mm Hg | Significantly fewer CV events or deaths in enalapril group compared with HCTZ (hazard ratio [HR] = 0.89, p = 0.05). |
| Avoiding Cardiovascular Events through Combination Therapy in Patients Living with Systolic Hypertension (ACCOMPLISH) (2008) | Randomized to amlodipine 5-10 mg + benazepril 20-40 mg or benazepril 20-40 mg + HCTZ 12.5-25 mg | 11,462 patients ≥ 55 years: SBP ≥ 160 mm Hg | Relative risk (RR) 0.8 (0.72-0.90) for composite of CV mortality/morbidity for amlodipine-benazepril group vs. benazepril-HCTZ group ( p = 0.0002) |
| Trials With Chlorthalidone-Based Regimens | |||
| Hypertension Detection and Follow-up Program (HDFP) (1979) | Stepped care (SC) using chlorthalidone as Step 1 compared with regular care (RC) consisting of multiple regimens and doses | 10,940 patients age 30-69 years; BP (baseline): 159/101 mm Hg; reported DBP only: RC 89 mm Hg; SC 84 mm Hg | 17% lower mortality in SC than RC groups ( p < 0.01). Total stroke was lower in the SC than RC groups ( p < 0.01). |
| Multiple Risk Factor Intervention Trial (MRFIT) (1990) | Randomized to special intervention (SI) or usual care (UC). Step 1 included either HCTZ or chlorthalidone (50-100 mg daily). Mid-study protocol changed HCTZ to chlorthalidone. | 8012 hypertensive men; BP (baseline): 141/91 mm Hg; BP (UC group): 130/86 mm Hg; BP (SI group): 122/81 mm Hg | Mortality reduced 36% ( p = 0.07) and coronary heart disease (CHD) reduced 50% ( p = 0.0001) in SI group vs. UC. Early mortality was 44% higher in SI clinics using HCTZ vs. UC but was 28% lower following a change from HCTZ to chlorthalidone ( p = 0.04 comparing the two time periods) |
| Systolic Hypertension in the Elderly (SHEP) (1991) | Randomized to chlorthalidone 12.5-25 mg and could add atenolol 25 mg or reserpine 0.05 mg vs. placebo | 4736 patients, mean age 72 years; BP (baseline): 170/76 mm Hg; BP (chlorthalidone): 144/68 mm Hg; BP (placebo): 155/71 mm Hg | Significant reduction in stroke for chlorthalidone compared with placebo (RR = 0.64, p = 0.0003); 32% fewer combined nonfatal and fatal CV events in chlorthalidone group. Heart failure reduced 54% in chlorthalidone group compared with placebo. |
| Verapamil in Hypertension and Atherosclerosis Study (VHAS) | Chlorthalidone 25 mg (707 patients) vs. verapamil slow release 240 mg daily (707 patients) Captopril 25 mg daily could be added to either regimen | 1414 patients mean age 53.9 years; BP (baseline):169/102 mm Hg; BP reduced by −29/17 mm Hg (chlorthalidone) and −28/17 mm Hg (verapamil) | After 2 years of follow-up; no difference in fatal plus nonfatal events 43 vs. 42. |
| Treatment of Mild Hypertension Study (TOMHS) (1998) | Randomized to chlorthalidone 15-30 mg/day (136 patients), acebutolol 400 mg/day (132 patients), doxazosin 1-2 mg/day (134 patients), amlodipine 5 mg/day (131 patients), or enalapril 5 mg/day (135 patients) vs. placebo; chlorthalidone could be added to placebo if nutritional-hygienic intervention did not control BP | Men and women 45 to 69 years of age; BP (baseline): 140/91 mm Hg; largest reduction in systolic BP in chlorthalidone group (−17.7 mm Hg) | Rate of all clinical events was 11.1% in combined active treatments vs. 16.2% for placebo ( p = 0.03); no consistent differences among active treatments for LV mass, lipid levels, or other outcomes |
| Antihypertensive and Lipid-lowering Treatment to Prevent Heart Attack Trial (ALLHAT) (2002) | Randomized to chlorthalidone 12.5-25 mg, lisinopril 10-40 mg, amlodipine 2.5-10 mg, or doxazosin 2-8 mg | 42,424 patients over 55 years; BP (baseline): 146/84 mm Hg; BP (chlorthalidone): 134/75 mm Hg; BP (amlodipine) 135/75 mm Hg; BP (lisinopril) 136/75 mm Hg | Doxazosin arm discontinued because of significantly higher stroke ( p = 0.04), heart failure ( p = <0.001), or combined CVD ( p < 0.001) compared with chlorthalidone. No differences in primary outcome for lisinopril or amlodipine compared with chlorthalidone. Higher risk of heart failure with amlodipine (RR = 1.38) or lisinopril (RR = 1.19) compared with chlorthalidone. Combined CVD (RR = 1.10) and stroke (RR = 1.15) were higher with lisinopril than chlorthalidone. |
| Treatment of Isolated Systolic Hypertension (SHELL) (2003) | Chlorthalidone 12.5-25 mg vs. lacidipine 4-6 mg daily. Fosinopril 10 mg daily or other ACE inhibitor could be added. | 1882 subjects, mean age 72 years and baseline BP 178/87 mm Hg; BP after 32 months was reduced −37/8 mm Hg with chlorthalidone vs. −38/8 mm Hg with lacidipine | Composite primary endpoint hazard ratio (HR) = 1.01 (0.75-1.36, p = 0.94). All-cause mortality 122 chlorthalidone vs. 145 lacidipine HR = 1.23 (0.97-1.57, p = 0.09). |
| Clinical Trial With Indapamide-Based Regimen | |||
| Hypertension in the Very Elderly Trial (HYVET) (2008) | Indapamide sustained release 1.5 mg could add perindopril 2-4 mg vs. placebo | 3845 subjects age ≥80 years (mean 83.5 years) with SBP >160 mm Hg and DBP 90-110 mm Hg | Fatal or nonfatal CV events 33.7/1000 person years for indapamide versus 50.6 for indapamide ( p < 0.01, HR 0.66 [0.53-0.82]) |
| Clinical Trials With Bendroflumethiazide-Based Regimens | |||
| Metoprolol Atherosclerosis Prevention in Hypertensives (MAPHY) (1988) | Metoprolol vs. HCTZ 50 mg daily or bendroflumethiazide 5 mg daily | 1609 men aged 40-64 years BP (baseline): 167/107; BP (treatment): 142/89 | Metoprolol treated subjects had significantly lower CV mortality ( p = 0.012), CHD mortality ( p = 0.048), stroke mortality ( p = 0.043), or total mortality ( p = 0.28) compared with diuretics |
| Medical Research Council (MRC) trial (1985) | Bendroflumethiazide 10 mg vs. propranolol 240 mg daily or vs. placebo | 17,354 subjects, mean age 35-64 years with SBP < 200 mm Hg and DBP < 90 mm Hg | No difference between drugs in combined CV events, coronary events or mortality. |
a Starting and end-titration dose range given for diuretic used.
An interesting finding was reported by investigators conducting the Multiple Risk Factor Intervention Trial (MRFIT). Patients were randomized to either special intervention (SI) or usual care (UC). The initial therapy for hypertensive patients in the SI group was either HCTZ or chlorthalidone in a dose range of 50 to 100 mg without specification on frequency of dosing. Of note, the choice of diuretic was made locally by the clinic staff. The study followed 8012 men for 6.9 years and found a trend in favor of the SI group compared with the UC group but the difference was not statistically significant. However, 6 years into the trial the investigators observed that in the nine clinics predominantly using HCTZ, the mortality rate was 44% higher in the SI group compared with the UC group. The opposite was true in the six clinics that predominantly used chlorthalidone, where mortality in the SI group was more favorable compared with the UC group. As a result, the MRFIT Data Safety Monitoring Board changed the protocol to exclusively use chlorthalidone (daily maximum dose of 50 mg) in the SI group. In the clinics initially using HCTZ, that had a 44% higher mortality in the SI group, the trend reversed after the protocol change and they then had a 28% lower risk compared with UC at the end of the study ( p = 0.04 comparing coronary heart disease [CHD] mortality at the two time periods).
The MRFIT investigators proposed several possible explanations for the finding that mortality was more favorable in the SI group at 10.5 years but not at 6.9 years of follow-up including: a possible time delay in risk reduction that required longer follow-up to observe the effect or, alternatively, that the change in the protocol to switch to chlorthalidone produced the more favorable effect observed toward the end of the trial. Our comparative trial (discussed below) between HCTZ and chlorthalidone showing better 24-hour BP control with chlorthalidone may be one explanation for the MRFIT study findings that has been endorsed by other investigators.
Several studies have compared thiazide-type diuretics with other drug classes, including ALLHAT, the Second Australian National Blood Pressure Study (ANBP2), and the Avoiding Cardiovascular Events in Combination Therapy in Patients Living with Systolic Hypertension (ACCOMPLISH) trial. The ALLHAT was a randomized, double-blind, active-controlled antihypertensive treatment trial in 42,418 patients assigned to a chlorthalidone, an ACE-inhibitor (lisinopril), a CCB (amlodipine), or an alpha blocker (doxazosin). The doxazosin arm was terminated early after 3.3 years of follow-up when higher rates of heart failure were observed when compared with chlorthalidone. After an average of 4.9 years of follow-up, chlorthalidone was at least as beneficial as the comparator drugs in lowering BP and preventing cardiovascular (CV) and renal outcomes and was superior for preventing HF (versus each comparator arm), combined CV events (versus α-blocker and ACE-inhibitor arms), and stroke (versus ACE inhibitor [African Americans only] and α-blocker).
The ANBP2 was an open-label trial in 6083 subjects treated with either diuretic-based therapy (primarily HCTZ) or ACE inhibitor (enalapril recommended). Cardiovascular events were lower in the ACE group (relative risk [RR], 0.89; 95% CI, 0.79 to 1.00) but this difference was right at the threshold of statistical significance ( p = 0.05). Differences between this study and ALLHAT were that the latter had far more African Americans and 8 times as many CV events compared with ANBP2. Because ANBP2 was an open-label study with agents selected by the individual practitioners, it is not possible to determine if evidence-based doses of HCTZ were used.
The ACCOMPLISH trial included 11,462 high-risk patients and was stopped early after an average of 42 months of follow-up. Amlodipine/benazepril had an RR of 0.8 (95% CI, 0.72 to 0.90; p = 0.0002) for major fatal and nonfatal CV events when compared with HCTZ/benazepril despite nearly identical office BPs. This study has been criticized because the composite endpoint did not include HF, which is a critical hypertension endpoint found to be reduced by 505 to 68% in diuretic-based regimens. There are other design features that may help explain differences between ALLHAT and ACCOMPLISH trials. First, benazepril does not have consistent 24-hour BP coverage. Because both ACCOMPLISH arms included benazepril, the comparison is essentially between HCTZ and amlodipine. The study used HCTZ in suboptimal doses of only 12.5 to 25 mg once daily, which at best has only an 8- to 15-hour duration of action. As displayed in Table 22.2 and recommended by national guidelines, the HCTZ dose should be twice as high and perhaps given twice daily. Amlodipine is one of the longest acting antihypertensives with a half-life of 38 to 50 hours and provides definitive 24-hour coverage. One explanation is that HCTZ did not provide sufficient 24-hour BP coverage. However, the ACCOMPLISH investigators subsequently reported that in a subset of subjects who received 24-hour BP monitoring, the BP values were similar in the two arms.
An indapamide-based regimen has been shown to lower hypertension-related mortality. The Hypertension in the Very Elderly Trial (HYVET) was a randomized, double-blind, placebo-controlled study that compared sustained release indapamide (1.5 mg) with or without perindopril (2 to 4 mg) to placebo in 3845 patients over the age of 80 years. The target BP was less than 150/80 mm Hg and at 21 years there was a 21% reduction in all-cause death (95% CI, 4 to 35; p = 0.02), a 39% reduction in fatal stroke (95% CI, 1 to 62; p = 0.05), a 64% reduction in fatal and nonfatal HF (95% CI, 42 to 78; p < 0.001), and a 34% reduction in any CV event (95% CI, 18 to 47; p < 0.001). The use of long-acting compounds (sustained-release indapamide and perindopril) would suggest that nighttime BP may have been effectively reduced and led to such positive outcomes.
These and other studies suggest that diuretics are particularly effective in the elderly and African-American patients. For these reasons, it is often necessary to add a thiazide-type diuretic, or loop agent in the presence of moderate CKD, to achieve good BP control in these populations.
The studies described above led the members of the 2014 national BP guideline committee to conclude that the scientific evidence suggests that thiazide-type diuretics, ACE inhibitors, and CCBs are equally effective as initial therapy. There is also insufficient evidence to suggest chlorthalidone is different from HCTZ on critical CV endpoints. However, when all of the diuretic-based outcome trials are examined, those that used chlorthalidone have all been significantly different from placebo or other therapy. Studies that used HCTZ have had mixed results, with approximately half finding benefit and half finding either no benefit or inferior results to other drug therapy. For these reasons, we and other experts believe that chlorthalidone is the preferred thiazide diuretic. If HCTZ is used, it should be given twice daily. The only time when HCTZ once daily can be justified is within a combination regimen that clearly has 24-hour BP coverage, ideally demonstrated by ambulatory BP monitoring in each specific patient.
Loop diuretics and potassium-sparing agents have not been studied to evaluate their effects on morbidity or mortality. Therefore, loop diuretics should not be selected unless renal function is so poor, or edema so significant, that thiazide-type diuretics are ineffective. Potassium-sparing agents should not be used alone and would generally be reserved for cases where they can be combined with a thiazide-type diuretic.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
