Hydrochlorothiazide––Teaching an Old Drug New Tricks

In 1905, when Korotkoff first described the auscultatory technique for measurement of brachial blood pressure (BP), little would he have known how great a clinical significance this would achieve over the next 100 years.1 It is now abundantly clear that elevated brachial BP is directly associated with myocardial infarction, heart failure, vascular disease, renal failure, atrial fibrillation, and stroke.2 Attempts at lowering this elevated peripheral blood pressure (pBP) have been unequivocally linked to a decrease in mortality and adverse outcomes.3 This led to an explosion in antihypertensive drug development, and randomized trials have addressed the comparative efficacies of these various drugs often with mixed results 4, 5, 6 As a result, current guidelines do not strongly advocate one antihypertensive agent over any other, apart from recommending against initial monotherapy with β‐blockers for hypertension.7 Based on the 2003 Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7) guidelines, many physicians often initiate treatment with thiazide diuretics (TDs), given their long proven safety and efficacy, minimal side effects, and favorable cost‐benefit ratio.8

Although the Korotkoff method has been widely adopted, and essentially unchanged in method from its original description, this technique measures only the brachial or peripheral BP, which is often discordant from measured central aortic blood pressure (cBP).9 The pressure waveform in the proximal aorta is the summation of the incident forward ejection wave in systole and the backward reflected wave from zones of impedance mismatch at the muscular peripheral arteries and arterioles. With stiffening of the central aorta, the incident and reflected waves have a higher pulse wave velocity, resulting in pathologic early reflection to the central aorta in systole (as opposed to diastole), thereby imposing additional load on the left ventricle. The amplitude of the reflected wave is also dependent on the degree of vasodilation of the peripheral arteries and degree of impedance mismatch (which can be affected by vasodilator drugs). With further age‐related stiffening of the central arteries, there is less impedance mismatching between the elastic central aorta and the normally higher resistance peripheral arteries. This results in less reflection and more distal transmission of the pulsatile load down the brachial and peripheral arteries, which can increase distal pulsatile loading and brachial pulse pressure.10

As opposed to invasive catheterization, applanation tonometry over the radial, femoral, and carotid arteries has emerged as a novel noninvasive technique for the determination of cBP and waveform analysis. The cBP is more reflective of the hemodynamic load imposed on the heart, cerebrovascular system, and kidneys, with many studies demonstrating the relationship between surrogate cerebrovascular and cardiovascular endpoints and increased central pulse pressure, central systolic pressure, augmentation index (AIx) (a measure of the load imposed by reflected waves from the periphery), and pulse wave velocity (a measure of arterial stiffness).11, 12 In an individual patient, all these complex mechanisms lead to varying relationships between central and peripheral pulse pressures that are modified in poorly understood ways by different antihypertensive therapeutic strategies. This varying response of cBP to therapy is a potential explanation for the benefit of amlodipine±benazepril compared with both TD±benazepril (Avoiding Cardiovascular Events Through Combination Therapy in Patients Living With Systolic Hypertension [ACCOMPLISH] trial) and atenolol±TD (Anglo‐Scandinavian Cardiac Outcomes Trial [ASCOT]), despite largely similar reductions in pBP.4, 6, 13 In fact, a nonrandomized substudy of ASCOT––the Conduit Artery Function Evaluation (CAFE) trial––showed lower cBP with amlodipine±benazepril compared with atenolol+TD, despite similar pBP.13 The difference in cBP in CAFE, however, was likely largely driven by the slower heart rate and prolonged ejection time in the atenolol group (with an increased likelihood for augmentation of the prolonged incident systolic wave by reflected waves), as opposed to the effect of the TD itself.14

On this background, Moura and colleagues15 sought to demonstrate noninferiority of the commonly prescribed TD to other antihypertensive agents, in regards to lowering of cBP. They performed a cross‐sectional subgroup analysis of 1194 patients from the CARTaGENE cohort and limited their analysis to nondiabetic hypertensive patients on monotherapy. The CARTaGENE is a prospective genomic population‐based study sampling a random 1% of Quebec province, and thus is well representative of the general population. The groups were well matched overall apart from a higher proportion of women in the TD arm, and the presence of more comorbidities with a longer duration of hypertension in the β‐blocker arm (which, as a result, was a sicker population at baseline).

Applanation tonometry of the radial artery was used to noninvasively determine cBP, heart rate, and AIx, but pulse wave velocity (an index of aortic stiffness) was not determined. Overall, central systolic and pulse pressures were not significantly different between TDs and the other drug groups, and this remained within their defined parameters of noninferiority after adjustment for multiple confounders, even in the subgroup of patients taking angiotensin‐converting enzyme (ACE) inhibitors/angiotensin receptor blockers (ARBs). At baseline, mean BP was lower in the β‐blocker group because of a lower diastolic pressure. The ACE inhibitor (but not ARB) group also demonstrated less reflective wave amplification of central BP as determined by lower AIx; however, on regression analysis, combining the ACE inhibitor and ARB groups, this remained within the defined realm of noninferiority. Analysis of TDs vs calcium channel blockers (CCBs) alone would have been interesting to see whether noninferiority of TDs remained true, given the theoretical possibility that lower cBP in the amlodipine arm of ACCOMPLISH may have contributed to the reported benefit in that trial compared with hydrochlorothiazide (HCTZ).4 Longitudinal follow‐up of this CARTaGENE cohort with documented central and peripheral BPs will provide important natural history information on the prognostic value of different baseline levels of cBP stratified by type of antihypertensive therapy.

Currently, chlorthalidone should be considered the standard of care for TD, compared with HCTZ, based on its stronger evidence base, longer duration of action with improved nocturnal blood pressure control, and improvement in cardiovascular outcomes.16 However, HCTZ continues to be the major TD prescribed, as also seen in this study, where 76% of patients taking thiazide monotherapy were prescribed HCTZ, with the remaining 24% receiving indapamide. Whether this study's observed similarity of cBP reduction with HCTZ compared with other drug classes also applies to chlorthalidone is unclear and requires further study. Given the cross‐sectional nature of this study, one cannot rule out that greater baseline differences and variable responses to antihypertensive therapy, prior to study entry, may have led to confounding in central hemodynamics at the time of cross‐sectional sampling. Drug assignment was also not randomized and the β‐blocker group in particular was sicker at baseline, with longer duration of treatment and more comorbidities, which potentially resulted in lower peripheral and central BPs. Adjusted regression analyses were performed separately only for TDs with ACE inhibitors and ARBs, making individual comparisons to β‐blockers and CCBs less robust.

Despite its limitations, this study provides reassurance that the widespread prescription of HCTZ is largely justified from the stand point of lowering cBP, with demonstrated noninferiority to other antihypertensive agents. The results need to be interpreted with caution given its nonrandomized observational design, but, on the basis of this data, nondiabetic mildly hypertensive patients taking HCTZ may be expected to demonstrate equivalent cBP response to ACE inhibitors, ARBs, and potentially also β‐blockers and CCBs. Whether this ultimately translates into equivalence in outcomes remains to be determined. There is abundant evidence from nonrandomized trials and intermediate endpoints to suggest that targeting cBP may potentially lead to improvement in hard clinical outcomes. Future double‐blind randomized controlled trials are urgently needed to help answer this important question regarding the most prevalent disease of our time.

Disclosure

The authors report no relevant conflicts of interest.