Reconsidering Alpha-Blockade for the Management of Hypertension in Patients with Chronic Kidney Disease

Between 60% and 90% of individuals with chronic kidney disease (CKD) have hypertension, with higher prevalence of hypertension as kidney function worsens.¹ Compared with the general population, a disproportionate number of patients with CKD have treatment-resistant hypertension (elevated blood pressure despite treatment with at least three antihypertensives or requiring a fourth medication to achieve blood pressure control)² and refractory hypertension (uncontrolled hypertension on five or more antihypertensives).³ Thus, patients with CKD often require multiple classes of medications to control their hypertension, though frequently have restricted options due to elevated risk of adverse effects. Most commonly, patients with advanced CKD are susceptible to hyperkalemia, limiting the use of renin-angiotensin-aldosterone system inhibiting medications as kidney disease progresses, and potentially requiring the use of third- or fourth-line agents.¹ The distinctive challenges of treating hypertension in patients with CKD merit a better understanding of the risks and benefits associated with less-commonly used antihypertensive agents, such as alpha-blockers, in this patient population.

In the current issue of AJKD, Hundemer and colleagues⁴ investigate the association of alpha-blockade with clinical outcomes (progression of CKD, cardiovascular events, and mortality) and safety events typically attributed to this class of antihypertensives (hospitalization for hypotension, syncope, falls, and fractures) across CKD stages. The authors leveraged robust, province-wide clinical and prescription dispensing data on all Ontario residents ≥66 years of age. Over a maximum of three years of follow-up, they observed that initiation of alpha-blockade was associated with an 8% lower risk of cardiovascular events (hazard ratio [HR] 0.92, 95% confidence interval [CI] 0.89, 0.95), 11% lower risk of death (HR 0.89, 95% CI 0.84, 0.94), 14% higher risk of ≥30% decline in eGFR (HR 1.14, 95% CI 1.08, 1.21), and 28% higher risk of end stage kidney disease (HR 1.28, 95% CI 1.13, 1.44) compared with initiation of non-alpha-blocking antihypertensives. In subgroup analyses, the lower risk of death was only present among patients with CKD, and those patients with an eGFR <30 mL/min/1.73m² had the lowest risk of death when treated with alpha-blockade compared with non-alpha-blocking antihypertensives (HR 0.71, 95% CI 0.64–0.80). The results were similar across several sensitivity analyses, including upon restricting the cohort to individuals not on monotherapy, on 3–4 antihypertensives, or without prior cardiac disease. With regard to safety events, use of alpha-blockade was associated with an elevated risk of hospitalization for syncope (among all participants) and hypotension (among those with higher eGFRs).

The current study is strengthened by the authors’ approach to addressing confounding using high dimensional propensity score matching. Propensity score matching is an increasingly popular approach to addressing confounding in cohort studies. In standard propensity score matching, a score is calculated based on the likelihood of study participants to be in either exposure group.⁵ The score is then applied to generate a matched cohort that balances known clinical characteristics across exposure groups. This approach can be a helpful way to address measured (or known) confounders by indication for initiating different pharmacologic agents, such as when discrete clinical indications exist to select one agent over another. Despite clear benefits for addressing confounding, all matching approaches are prone to important limitations and misspecifications. Standard propensity score matching is particularly susceptible to inadequate matching, which can result in imbalance of some covariates across exposure groups. It is also susceptible to bias due to difficult-to-anticipate effects of investigator-selected covariates on the identification of matched pairs, which can result in inadvertently excluding participants with certain risk profiles, thus reducing the generalizability of the resulting cohort.⁵ In the current study, Hundemer et al. used high dimensional propensity score matching, a data-driven approach to deriving a large number of potential covariates (in this instance, 207 covariates) for inclusion in propensity score models.⁶ Due to its multitiered, empirical approach, high dimensional propensity score matching can help to overcome some of the biases and inadequate matching that can occur using more standard covariate selection approaches.⁶ Nonetheless, high dimensional propensity score matching is still susceptible to many of the methodologic misspecifications that can occur with any matching approach. While some authors propose that high dimensional propensity score matching can overcome unmeasured confounding in retrospective cohort studies, this has not been adequately substantiated.

This study provides unique insights into the management of hypertension in CKD, given that largescale trials evaluating the safety and cardiovascular effectiveness of alpha-blockers (i.e. the Antihypertensive and Lipid-Lowering treatment to prevent Heart Attack Trial [ALLHAT] and Anglo-Scandinavian Cardiac Outcomes Trial [ASCOT]) excluded patients with clinically significant CKD.^{7, 8} ALLHAT⁷ and ASCOT⁸ demonstrated poor cardiac outcomes with alpha-blockers; however several observational studies have shown the contrary. For example, in a propensity score-matched cohort study of United States veterans with heart failure, in which 13–14% of patients had moderate to severe CKD, alpha-blocker use was associated with fewer heart failure admissions.⁹ Additionally, higher non-selective alpha-blockade doses and lower eGFR were both associated with lower mortality risk. Kidney outcomes were not assessed.

The physiologic basis for lower cardiovascular risk with alpha-blockade may be explained by inhibition of part of the neurohormonal activation pathway that promotes development and progression of heart failure.¹⁰ In the COMET trial,¹¹ carvedilol was superior to metoprolol in reducing all-cause mortality in patients with chronic heart failure. The partial alpha-blocking effect of carvedilol may prevent cardiac remodeling by reducing systemic vascular resistance.¹² Furthermore, alpha-blockers have been associated with improvement in lipid profiles and insulin resistance,¹³ which may mediate cardiovascular risk reduction. Regarding the higher risk of CKD progression among those who received alpha-blockers, the current study as well as a prior study published by the same authors demonstrated higher incidence of hospitalizations for hypotension and syncope in older patients using alpha-blockers compared with other antihypertensive agents.¹⁴ This risk of hypotension may be the basis of progression of CKD. Low blood pressures can reduce afterload and prevent cardiac remodeling over time¹⁰ but can also result in renal hypoperfusion, particularly in patients with stiffer vasculature (e.g., those with underlying CKD). However, the authors’ prior and current studies were limited by a lack of blood pressure data;^{4, 14} hypotension was defined using diagnostic codes from inpatient hospitalizations in both studies. Thus, clinically significant hypotension and orthostasis not resulting in hospitalization would have been missed. That said, alpha-blockers are often well-tolerated in the general population. The “first-dose effect,” which can cause sudden, severe, symptomatic orthostatic hypotension, may be bypassed by using the extended-release dosage formulation and bedtime dosing.¹⁵

In summary, Hundermer et al.’s study⁴ sheds light on the potential cardiovascular benefit of alpha-blockers for the treatment of hypertension in older adults with CKD, balanced with elevated risk of CKD progression. The methodologic approach strengthens the authors’ conclusions about these relationships but cannot overcome unmeasured confounding. Lack of access to information on blood pressure control highlights the importance of future mechanistic investigations into these relationships, as outpatient hypotension could contribute to the observed outcomes. As nephrologists, we are hyperaware that cardiovascular diseases are the number one cause of mortality in patients with CKD. If these agents might mitigate that risk, perhaps we should give them stronger consideration when grasping for add-on therapies for the treatment of hypertension in our patients.