#NephJC Chat
Tuesday Dec 7th 9 pm Eastern
Wednesday Dec 8th 9 pm IST
N Engl J Med 2021 Nov 5 doi: 10.1056/NEJMoa2110730. Online ahead of print.
Chlorthalidone for Hypertension in Advanced Chronic Kidney Disease
Rajiv Agarwal, Arjun D Sinha, Andrew E Cramer, Mary Balmes-Fenwick, Jazmyn H Dickinson, Fangqian Ouyang, Wanzhu Tu
PMID: 34739197
Introduction
Blood pressure control is often a challenge in patients with advanced chronic kidney disease (CKD). Once the GFR drops, the risk of hyperkalemia goes up (so RAAS blocker use decreases). RAAS blockers also often get blamed for those annoying creatinine bumps, and get stopped. (Needlessly. Can we all just embrace permissive hypercreatininemia?) Calcium channel blockers (CCBs) are fine - unless they cause edema. Thiazides don’t work in advanced CKD and are often switched over to loop diuretics which are not as great at BP lowering, though they are good at natriuresis (Holland et al, Arch Int Med 1979). That leaves second line drugs and their thin evidence base. This is why we find alpha blockers, clonidine and the dreaded hydralazine polluting the medication lists of our advanced CKD patients. But wait - why did we say thiazides don’t work in advanced CKD? Where does that come from and is it true?
Indeed, guidelines say that at GFRs south of 30, loop diuretics should preferentially be used (KDOQI, JNC 7 - PDF link). However, a systematic review from 2012 reported on a small number of RCTs (ranging from N = 7 to N = 19) that showed BP reductions in advanced CKD (Agarwal et al, J Am Soc HT 2012). But these were small trials from the pleistocene era. Additionally, thiazides cause a host of problems from the predictable hyponatremia (see NephJC summary), hypercalcemia, and hypokalemia to the more insidious increases in uric acid, glucose and lipids (Grimm 1981). Are these drugs even worth the hassle?
Let’s recall that among all drugs, despite these metabolic side effects, thiazides (or rather, thiazide-like diuretics), have demonstrated benefits not just in lowering BP, but target organ damage, in trials such as ALLHAT, MRC, SHEP, HYVET, and SPRINT. Notably, in all these trials, it was chlorthalidone (or indapamide), and not hydrochlorothiazide (HCTZ), or bendroflumethiazide, or metolazone. In a meta-analysis of the dose-response relationship, it was reported that the systolic BP showed a log-linear relationship with potency in this order: bendroflumethiazide>chlorthalidone>hydrochlorothiazide. The estimated dose of each drug predicted to reduce systolic BP by 10 mm Hg was 1.4, 8.6, and 26.4 mm Hg (Perterzan et al, Hypertension 2012). Additionally, chlorthalidone is longer acting, and at half the HCTZ dose, reduces 24-hour BP by an additional 5 mm Hg, and nighttime BP by an additional 7 mm Hg (Ernst et al Hypertension 2006). This makes chlorthalidone an ideal candidate to try in advanced CKD. It is also a tenth (or lower) the cost of the much lauded metolazone.
Based on this premise, Agarwal and colleagues conducted a pilot feasibility trial in 2014 in 14 patients and demonstrated that chlorthalidone safely lowered BP in advanced CKD (Agarwal et al. Am J Nephrol 2014). Click through and join us as we explore the CLICK trial.
The Study
Study design
The Chlorthalidone in Chronic Kidney Disease (CLICK) Trial was a double-click, I mean double blind, randomized, placebo-controlled trial of chlorthalidone in patients with advanced CKD and treated but poorly controlled hypertension. Patients were recruited from 3 hospitals in Indiana: the Indiana University Hospital, the Eskenazi Hospital, and the Richard L. Roudebush VA Medical Center.
Study population
Patients with stage 4 CKD (GFR 15-29 ml/minute/1.73 m2)
Uncontrolled hypertension based on 24-hour BP > 130/80 on at least one BP lowering drug (either an ACEi/ARB or a beta-blocker) and after standardization - more on that below.
Exclusion criteria
Patients with a 24-hour ambulatory SBP ≥ 160 mm Hg or DBP ≥ 100 mm Hg,
History of stroke or myocardial infarction,
Hospitalization for heart failure within 12 weeks before randomization,
On high-dose loop diuretics (>200 mg of furosemide or >100 mg of torsemide daily),
Patients who received a thiazide or thiazide-like diuretic within 12 weeks of randomization.
Run-in period
Eligible patients measured home BP twice daily for a week, then entered a 2-week run-in period with placebo. During the run-in all antihypertensive regimens were standardized as follows:
Lisinopril (20-40 mg/day)
Losartan (50-100 mg/day)
Amlodipine (10 mg/day)
Atenolol (25-100 mg/day)
Torsemide (10-20 mg/day).
Thus all ACEi were replaced with lisinopril, all ARB were replaced with losartan, and so on. Participants were also instructed to follow a 100 mmol sodium diet, and underwent a 24-hour ABPM investigations at the end of run-in period to confirm ongoing eligibility (eGFR and ABPM).
Trial procedures
The entire trial lasted 12 weeks, with a final visit two weeks after treatment regimen discontinuation. At every visit, seated and standing BP were measured with a digital sphygmomanometer. The BP was measured in triplicate in close succession in both arms after an appropriately sized cuff was applied. Patients were instructed to measure BP at home in triplicate after 5 minutes of seated rest twice daily for 1 week before the scheduled clinic visit.
As indicated in figure S1, serum creatinine, electrolytes, glucose, renin, aldosterone, N-terminal pro–B-type natriuretic peptide (NT-pro BNP), urine albumin and creatinine levels were measured. Total body volume (the volume of air displaced by the body in a pressure-regulated closed chamber) was assessed by air-displacement plethysmography.
Intervention
Patients were randomly assigned in 1:1 ratio to receive chlorthalidone or matching placebo for a period of 12 weeks. The dose of chlorthalidone (or placebo) was doubled every 4 weeks if the home SBP ≥ 135 mm Hg or DBP ≥ 85 mm Hg, from 12.5 mg once daily to 25 mg once daily and finally to a maximum dose of 50 mg once daily. The dose was not increased if the patient had symptomatic orthostatic hypotension, hypercalcemia, hypokalemia (potassium level, <3 mmol/L), acute gout, or poorly controlled diabetes. In these instances, the dose was either maintained or decreased. The assigned drug was discontinued in case of rash, acute kidney injury or other adverse events. Patients were withdrawn from the trial, if the home BP ≥ 180/110 mm Hg after 1 week of dose increment or ≥ 160/100 mm Hg after 4 weeks of dose increment. The final trial visit was scheduled 2 weeks after the assigned regimen was discontinued and the patients were invited to follow-up annually for 3 years.
Outcomes
The primary outcome was change in 24-hour ambulatory SBP from baseline to 12 weeks. Secondary outcomes were changes from baseline to 12 weeks in the urinary albumin-to-creatinine ratio (ACR), NT-pro BNP level, plasma renin and aldosterone levels, and total body volume.
Analysis
The trial was powered for a 6 mm Hg difference between groups from baseline, thus needing 64 in each arm (128 total). Accounting for 20% discontinuation, the planned sample size was thus 160 patients. The analysis was based on an intention-to-treat (ITT) model with a mixed model for repeated measures, including 24 hour ABPM at baseline and 12 weeks as the response variables, and trial group, time, and use of loop diuretics as independent variables. Missing data was imputed (more on that later).
Funding
The trial was funded by the NHLBI and the Indiana Institute of Medical Research
Results
Of 2847 patients screened, 1519 were potentially eligible, 495 consented and 403 entered the run-in phase from which 160 were randomized. Out of a total of 160 patients, 81 were assigned to chlorthalidone and 79 to placebo group. A total of 140 patients completed the 12-week trial period.
At baseline, 60% were on loop diuretics and 99% were receiving an ACEi, an ARB or a beta-blocker. Beta-blocker use was most common, at about 75%, followed by CCBs (~ 65%) then RAS blockers (~ 60%).
About 75% were men (one of the hospitals was a Veterans Administration facility), and 40% were Black. The mean estimated GFR was 23 ml/min/1.73 m2. Diabetes was the commonest culprit. Baseline BP was ~ 140/70, median ACR was about 800, and 70% had an ACR > 300. The median 24-hour urinary sodium was just over 100 mmol/day.
In the chlorthalidone group, the mean dose received per day was 11.5 mg at 4 weeks after the initiation of the regimen, 18.3 mg at 8 weeks and 23.1 mg at 12 weeks.
Office and Home BP
There was a 10 - 12 mm Hg decrease in office SBP, compared to a 2 - 5 mm Hg increase in the placebo group. Home BP also was similarly lowered (figure S4) at about 11 mm Hg lower.
Primary outcome: Change in 24-hour ambulatory BP from baseline to 12 weeks
This is shown in table 2 (below). About 20 patients (of 160) had missing values and these were imputed (see table S2 for complete case and imputed value comparisons). The change below is not the crude change, it represents the adjusted change from the mixed models method (see analysis section above).
The percentage of patients who had a nocturnal dip in SBP of >10% from the daytime value did not change substantially from baseline to 12 weeks in both the groups.
The subgroup data can be seen below, and demonstrates a consistent effect in all subgroups.
Other Outcomes
Both body weight and body volume decreased significantly. Changes in plasma renin and aldosterone rose with the chlorthalidone, while NT-proBNP fell. Both of these changes reversed after discontinuation.
Other Renal Outcomes
There was a significant change in albuminuria with chlorthalidone
There was also an accompanying decrease in eGFR with chlorthalidone, as expected. Did that represent some more real AKI? Read on.
Adverse events
Adverse events occurred more with chlorthalidone therapy than in the placebo group as shown below.
Everything metabolic that we know with chlorthalidone was worse: more hyponatremia, hypokalemia, hyperglycemia. Some more orthostatic hypotension and dizziness. Also some more AKI. In 5 patients the adverse events were serious enough to lead to discontinuation (4 in chlorthalidone and 1 in placebo). Serious adverse events leading to hospitalization were similar - 8 with chlorthalidone, 11 with placebo.
Observational follow-up study
49 patients (29 in the placebo and 20 in the chlorthalidone group) had a decrease in the estimated GFR to < 10 ml/min/1.73 m2, underwent long-term dialysis, or died. The hazard ratio (chlorthalidone vs. placebo) for a decrease in the estimated GFR to below 10 ml/min/1.73 m2, long-term dialysis, or death, with adjustment for use of loop diuretics at baseline, was 0.63 (95% CI, 0.36 to 1.12).
Discussion
In summary, despite previous misconceptions, chlorthalidone does reduce blood pressure in advanced CKD, compared to placebo. This is accompanied by a reduction in weight, and albuminuria. It is also accompanied by an increase in almost all possible metabolic side effects. Thus chlorthalidone is a potent weapon, and has to be used carefully - even more so in advanced CKD.
Limitations
The trial was relatively small and had an underrepresentation of women (22%). Very few patients were Asian or Hispanic.
The change in BP is huge, but a surrogate outcome.
Strengths
This was powered for a BP reduction, to demonstrate that BP lowering does occur and that thiazides do not lose their potency with decreasing GFR. That was clearly demonstrated. Indeed it was not powered for clinical outcomes, nevertheless the reduction in BP and albuminuria is promising. The mechanistic effects (renin, aldo, body weight and volume, NT-proBNP) also support that chlorthalidone works at these GFRs. The use of home and ambulatory BP in this setting was neat.
Conclusion
This trial provides pretty conclusive and robust data to support chlorthalidone use in this population - there is no need to switch to a loop, and if need be, chlorthalidone can be initiated in this population.
Can this be extrapolated to other thiazides? If we are looking at HCTZ - one will need higher doses, perhaps 50 to 75 mg HCTZ, and the nighttime BP lowering might not be the same. Indapamide is the other option which is somewhere in between HCTZ and chlorthalidone. Lastly, this trial does not tell us about the comparative effects between HCTZ and chlorthalidone. Epidemiological studies in this area are susceptible to selection bias (sicker people will more likely get chlorthalidone) and other flaws, and we should wait for the ongoing Diuretics Comparison project for an answer to that question.
Summary by M. Aarthi
Nephrologist, Chennai, India
NSMC Intern, Class of 2021