Lisinopril vs. Atenolol for hypertension in the dialysis unit
Background and Rationale
A study reporting the relative effectiveness of lisinopril over atenolol for hypertension in the hemodialysis unit was presented in the 2013 Kidney Week meeting. The HDPAL (NCT00582114) was a parallel group, open label randomized control study conducted between 2005-2013 in a single center that was sponsored by IU and supported by NIDDK. The aim of HDPAL (per the clinicialtrials.gov entry) was to “directly test the hypothesis that an initial strategy of lisinopril-based therapy will be more effective than atenolol-based therapy in causing regression of LVH over one year in patients with hemodialysis hypertension despite similar degree of BP reduction”. The justification for the study was provided by the high prevalence (>80%) of hypertension in the dialysis unit and the associated morbidity of stroke and left ventricular hypertrophy, while the choice of interventions tested in HDPAL was motivated by prior pilot studies performed by the PI in the late 1990s. These studies showed that atelolol and lisinopril are both able to reduce monitored ambulatory blood pressure (ABPM) to a similar extent i.e. by ~20/11 when dosed three days a week after dialysis in a supervised fashion.
Study Procedures and Outcomes
HDPAL randomized HD patients with echocardiographically documented LVH to either atenolol (initial dose: 25mg TIW, titrated to a maximum of 100 mg by doubling the dose q2wks) or lisinopril (initial dose: 10 mg TIW titrated to 40 mg by doubling the dose q2wk). To be eligible to participate, patients with LVH had to either have an ABPM > 135/75 (after a UF trial), or not be receiving any antihypertensives if they did not want to participate in a UF trial. Felodipine (10 mg) could be added once the maximum dose of the study medications had been reached and other anti-hypertensive agents could be added after felodipine if ABPM did not decline below 140/90.
The primary outcome for the study was regression of LVH by echocardiographic criteria at 12 months, while secondary outcomes to be assessed were regression of LVH by echo at six months and the adjusted (for age, gender, ABPM) index of LV mass/BSA from baseline to month 12.
Patients would have echocardiographic assessments at baseline and again at 6 and 12 months, as well as ABPM at baseline and 3, 6, and 12 months and were also asked to self monitor their BP.
What did the study show?
The study enrolled 200 (mostly African American) patients, but it was stopped prematurely for safety reasons so that only 104 patients completed follow-up (58 in the atenolol and 46 in the lisinopril arm). At the time the study was stopped, patients on atenolol had numerically higher reductions in LVH but this finding did not reach statistical significance. The safety signals were ubiquitous for patients on lisinopril:
- Excess number of serious cardiovascular events(IRR: 2.36, 95% CI 1.36 to 4.23, P=0.001)
- Excess major cardiovascular events (IRR 2.29 95%CI: 1.07-5.21 p=0.02)
- More frequent hospitalization for all causes (IRR: 1.61 p<0 .01="" li="">
- More frequent hospitalization for congestive heart failure (IRR 3.13 p=0.02)
- More likely to develop hyperkalemia (IRR 3.38 p=0.05)
- More likely to suffer a hypertensive crisis (IRR 3.81 p=0.03)
On the other hand, atenolol was more efficacious in reducing BP, irrespective of the method of assessment (ambulatory v.s. home self-monitoring) by 3.5 and 6.3 mmHg respectively). These patterns of suboptimal BP control by lisinopril were noted despite more aggressive fluid removal and a larger number of additional blood pressure medications.
Discussion
As the study was terminated prematurely, the primary and secondary end points could not be fully assessed and the superiority of beta blockers versus ACE inhibitors remains an open question. On the other hand, the reported patterns of blood pressure reduction and the cardiovascular safety signal were surprising given the established track record of ACEis in reversing LVH and the poor relative efficacy of atenolol against other agents in the non-ESRD population (Cochrane Database Systemic Review, though not everyone agrees with this interpretation e.g. Blood Pressure, 2007 and BMJ, 2009). However tempting it may be to attribute these findings to a play of chance, or undifferentiated secular trends (e.g. the study was registered with Clinical Trials.gov in 2005, yet only 104 patients had a year of follow-up by 2013) the possibility that atenolol is indeed a better drug than lisinopril for dialysis patients with LVH should be entertained. Working under this hypothesis, there are at least two possible explanations for the apparent benefit of atenolol on blood pressure control:
- A ghost of studies past: A number of studies have suggested that African American patients with cardiac disease (either LVH or systolic dysfunction) may respond better to beta blockade compared to RAAS inhibition (a pattern seen in the Losartan Intervention for Endpoint Reduction – LIFE study), or even receive no benefit from ACEs compared to whites (e.g. SOLVD). Considering that the mostly African American participants in HDPAL were on the ultimate “diuretic” (dialysis to dry weight), an intervention that diminishes the relative (in)effectiveness of beta blockers in the non-ESRD African American population, one could hypothesize that the results of this study may reflect a racial benefit of beta-blockers for African American with LVH. It would be interesting to see whether the investigators of the HPDAL broke down the results according to race, as a hypothesis generating analysis.
- Pharmacokinetics: Atenolol and lisinopril pharmacokinetics on dialysis differ and these differences may explain the inferior blood pressure control in home and ABPM recordings. While both agents are efficiently cleared by hemodialysis, with apparent intradialytic half life between 3.5-5 hrs (atenolol: BJCP, 1980, Arch Toxicol Suppl, 1980 and Eur J Clin Pharmacol, 1981, Hemodialysis International 2013 and lisinopril BJCP, 1988) there are important differences in the time to the peak (~4hrs with atenolol, 8-44 hrs with lisinopril) and rebound kinetics (larger rebound with atenolol). Based on these considerations one could hypothesize that patients receiving atenolol in the dialysis unit would spend a much shorter period of time under-medicated as a result of the faster absorption of the drug and possibly its higher rebound. This could explain the larger time difference in BP noted in ABPM recordings and home BP measurements. To the extent the investigators obtained blood levels, it might be possible to explore the pharmacokinetic hypothesis by correlating free drug concentrations to ABPM or home BP recordings. To the extent that pharmacokinetics play a role in explaining the HDPAL results, one could consider using alternative RAAS inhibitors (e.g ramipril) that are exhibit more comparable kinetic behavior to atenolol.
Implications for clinical practice
This is an interesting pilot report about therapeutic intervention to control hypertension in dialysis, a common problem for our patients. Current approaches to this problem are unsatisfactory, judging from the frequency with which nephrologists switch agents (BMC Nephrol. 2013) in the unit. Far from definitively proving the superiority of beta blockers over ACE inhibitors due to the limitations of a prematurely stopped and thus underpowered study, HDPAL adds some important information that could help clinicians choose blood pressure medications for their dialysis patients. In particular, HDPAL suggests that the perceived inferiority of beta blockers in the non-ESRD population may not apply to dialysis patients. Though it is customary to say that further studies will be needed (and in fact we do need them!), clinicians managing hypertensive dialysis patients should lead by example and consider applying the HDPAL protocol in the context of “n=1” trials. These studies “consider an individual patient as the sole unit of observation” to investigate the efficacy or side-effect profiles of different interventions (Per Med. 2011). In particular, rather than switching antihypertensives around in no systematic pattern, the nephrologist working with the patient under a shared decision making paradigm carry a structured evaluation of lisinopril (the most commonly prescribed ACEi, used in 20.9% of dialysis patients) vs. atenolol (the least prescribed beta blocker, but still used in 7.2% of patients) correlating home blood pressure (or ABPM if available) in a 4 period crossover fashion. By pooling multiple such studies it may be possible to fill in the knowledge gap that HDPAL tried to fill.
