#NephJC Chat
Tuesday Dec 17th, 2024 at 9pm Eastern (AEDT = Oct 23rd, 12pm) on Bluesky
Ann Intern Med. 2024 Jul;177(7):953-963.
doi: 10.7326/M23-3236. Epub 2024 Jul 2.
Angiotensin-Converting Enzyme Inhibitors or Angiotensin-Receptor Blockers for Advanced Chronic Kidney Disease
PMID: 38950402
Introduction
Angiotensin-converting enzyme inhibitors (ACEis) and angiotensin-receptor blockers (ARBs) have been the standard of care for decades in proteinuric CKD to slow the progression to ESKD. However, evidence supporting the initiation of these renin-angiotensin inhibitors (RASi) in the higher-risk subgroup of CKD stage 4 is surprisingly limited, compared to the earlier CKD stages. In fact we lack large dedicated trials looking at those with eGFR 15-30, so would benefit from a meta-analysis of RCTs which recruited some patients like this, to see whether the benefit remains even when starting these medications in CKD this advanced.
In fact, a previous observational study supported the discontinuation of RASi in patients with advanced CKD (Ahmed AK et. al, NDT 2010). The authors of this small study (52 patients with CKD 4 and 5) concluded that stopping ACEi/ARB delayed the onset of kidney replacement therapy (KRT), but observational studies cannot prove a causal relationship to outcomes. By contrast, the much higher quality STOP-ACEi trial was a multicenter, open-label trial that randomized 411 patients with advanced CKD to continue or discontinue therapy with RASi, with a follow-up of 3 years (Bhandari S, et al NEJM, 2022). The STOP-ACEi trial did not find any benefit of stopping ACEi (or ARBs) in advanced CKD, with no meaningful improvement of eGFR. Indeed, the discontinuation arm had a 6% higher risk of needing kidney failure replacement therapy (KFRT), as well as a higher risk of CV events, which though not statistically significant was enough to confirm that RASi should not be routinely discontinued. Ultimately, the decision to continue or discontinue RASi should be made in the context of the individual patient’s level of proteinuria, blood pressure control, tolerability, and cardiovascular risk profile. An interesting discussion about the STOP ACEi study can be found in this NephJC blog post and the Freely Filtered podcast.
So, why are we revisiting RASi in advanced CKD? Well if the STOP-ACEi trial showed that routine discontinuation of RASi came with a signal for harm, how about initiation of RASi looking for benefit? This is a particularly pertinent question at present, as now armed with multiple effective CKD therapies many healthcare systems are seeking out their highest-risk patients to target for CKD medication optimisation. Unfortunately only small-scale trials in the 1990s and early 2000s (Hou FF, et al NEJM, 2006; Ihle BU et al AJKD, 1996; Kamper AL et al AJH, 1992) tried to address the issue of RASi initiation in patients with advanced CKD, and though these studies were generally supportive, decades on the momentum is only just turning away from studying routine discontinuation. As we initiate our new medications of SGLT2i, GLP1-RA, and nsMRAs, we are left with the fact that the majority of these recent CKD trials also excluded patients with an eGFR <20 or <25 ml/ml/1.72 m2. The following meta-analysis attempts to at least find an answer to the vitally important question regarding whether in CKD stage 4 starting RASi, the backbone of our ESKD therapeutic strategy, is a harmful, neutral or beneficial move.
Author’s note, a small issue with nomenclature:
What’s in a name? Kidney replacement therapy (KRT), renal replacement therapy (RRT), and KFRT (kidney failure replacement therapy) all are seemingly interchangeable. This article has chosen the latter, KFRT, as its primary designation for patients with kidney failure on dialysis or transplantation. Nephrology has been burdened with the use of "nephro,” "renal,” and “kidney” sprinkled throughout various journals and books, much to the chagrin of KDIGO and many non-US nephrologists. For the sake of this discussion, we will use the ETLA (extended three-letter acronym) of KFRT for consistency with this paper.
The Study
Methods
This is a systematic review and individual participant-level meta-analysis. The systematic review was registered in PROSPERO (an international database of prospectively registered systematic reviews in health and social care) and a one-stage, fixed-effects meta-analysis approach was used, pooling individual participant data from multiple studies into a single model to obtain results.
A systematic review was done of the literature from the CKD-EPI CT Consortium from January 1946 to December 2020. The original search included trials that enrolled adults with a baseline eGFR >15 ml/min/1.73m2, were published in English, had at least 12 months of follow-up, and measured serum creatinine within the first 12 months after trial enrollment. An updated bridge analysis with the same inclusion criteria was performed for trials between January 2021 to December 2023, excluding trials that tested the effect of ACEis or ARB on COVID-19 outcomes.
Seventy-one trials were available for potential inclusion. Trials that did not test either ACEi or ARB interventions, compared ACEis or ARBs against other renin-angiotensin-aldosterone inhibitor interventions, that did not enroll participants with eGFR below 30 mL/min/1.73m2, or that did not ascertain the onset of KFRT as an endpoint were excluded. They also excluded trials that focused on heart failure populations, where other indications for ACEis or ARBs may exist, and those that tested dual versus single renin angiotensin system blockade.
Participants of each individual trial were classified as receiving the intervention (ACEi or ARB) versus a comparator (placebo or other antihypertensive).
Study Outcomes
Primary Outcome
The primary kidney outcome in this study was the onset of KFRT (defined as receiving dialysis or a kidney transplant).
Secondary Outcomes
Death before the onset of KFRT and a composite outcome of KFRT or death.
Covariates were selected based on the association with the progression of CKD and the availability across all trials. Those included were: age, sex, race, body mass index, eGFR, albuminuria, diabetes status, baseline systolic BP, and history of coronary artery disease.
Statistical Analysis
They used Cox proportional hazards regression models to examine the association between the randomized assignment to ACEi or ARB therapy and risk for onset of KFRT. Primary analyses were based on the intention-to-treat principle and were adjusted to account for potential imbalances in baseline characteristics. Subgroup analyses explored the effect of treatment initiation with an ACEi or ARB versus placebo and versus other antihypertensive agents.
Funding
National Institute of Health. The funders had no role in the design or conduct of the study.
Results
Included Trials
Of the 71 trials identified by the original CKD-EPI CT systematic search, 18 met the inclusion criteria. An updated bridge search was done (trials between January 2021 and December 2023) but none met the inclusion criteria.
Supplemental Figure 1. Flow diagram of search strategy, from E Ku et al, Ann Intern Med. 2024
The design and baseline characteristics of the 18 trials are described in Table 1. Notably the last trial is from 2007. In most trials the comparator was placebo, and in 6 of them it was another antihypertensive agent. Patients from all KDIGO albuminuria stages were included in the meta-analyses, although the median urinary ACR of included patients from most studies was in the severely increased albuminuria range (A3, >300 mg/g). Twelve studies did not include patients with diabetes.
Table 1. Design of the Included Trials and Characteristics of the Participants included for analysis who had a baseline eGFR 15 ml/min/1.732, from E Ku et al, Ann Intern Med. 2024
Study Population
Baseline characteristics of the 1739 trial participants included for individual analysis are shown overall in table 1 and separately in trials whose comparator was placebo versus other antihypertensive agents in table 2. The mean age of participants was 55 years; 52 % were women and 15 % were Black. The median eGFR of the included participants was 23 ml/min/1.73m2 , and 32% had eGFR below 20 ml/min/1.73^2. The median ACR was 1212 mg/g (137 mg/mmol), and 85 % had severe albuminuria (A3, ≥ 300 mg/g). Mean systolic BP was 155.4 mmHg (these trials all predate SPRINT and the intensive BP management practiced now). Studies that contributed with more participants had more balanced characteristics; however, some imbalances were noted by sex, race, and baseline SBP. As you may note in table 2 below, RASi patients and comparator groups were not even because not all trials were randomized 1:1. One trial was 1:1:1 (irbesartan, amlodipine, and placebo), and AASK was randomized at 2:2:1 (ramipril, metoprolol, and amlodipine).
Table 2. Characteristics of Participants With Advanced CKD, by Randomized Assignment in the Overall Cohort and Separately by Type of Comparator, from E Ku et al, Ann Intern Med. 2024
Primary Outcome
During a median follow-up of 34 months, 36 % developed KFRT and just under 8 % of patients died. In adjusted analyses, ACEi or ARB treatment initiation resulted in a 34% lower risk for progression to KFRT than the comparator group. Findings were similar in the subgroup analysis.
Figure 1. Cumulative incidence of KFRT in unadjusted and adjusted analysis, by treatment group, for the overall cohort, place-controlled trials, and other antihypertensive agents, from E Ku et al, Ann Intern Med. 2024
The following plot shows the adjusted hazard ratios of ACEi or ARB treatment initiation versus comparator on the risk for KFRT in each trial and in the pooled analysis.
Figure 2. Forest plot of the effect of ACEi or ARB treatment initiation on the risk for KFRT in the overall cohort, placebo-controlled trials, and other antihypertensive agent comparator trials, from E Ku et al, Ann Intern Med. 2024
Secondary Outcomes
The authors did not find a benefit of ACEi or ARB treatment initiation in terms of risk for death in adjusted analysis. Findings were consistent in subgroup analysis.
Table 3. Event Rates and the Risk for Onset of KFRT and Death Before KFRT, from E Ku et al, Ann Intern Med. 2024
Subgroup Analyses
The following plot shows the effect of ACEi/ARB versus comparator on the risk of KFRT by risk factors of interest in the subgroup analysis, including age, eGFR, ACR, diabetes status, and ACEi/ARB dosing protocol in the parent trial. The results were quite consistent.
Supplement Figure 4B. Forest plot of the effect of ACEi or ARB initiation on the risk of KFRT in adjusted group analysis, from E Ku et al, Ann Intern Med. 2024
Discussion
As renal specialists, we've all had that moment—receiving a patient with advanced CKD through a referral so late that it feels more like a “best wishes” card than a meaningful handoff. These patients often arrive with severe albuminuria and uncontrolled hypertension, leaving us with little more than hope and the medical equivalent of duct tape to slow progression while preparing for the inevitable and inexorable onset of dialysis. RASi are supposed to be our trusty sidekick in CKD management, but their use in advanced stages is backed by evidence so scarce it feels like trying to build a skyscraper with a single brick. And here's the kicker: despite decades of studies and the glaring importance of this question, no nephrologist has thought, "Hey, maybe a meta-analysis would be a good idea." It’s almost ironical—key studies have been around for over 20 years, yet we've collectively ignored the elephant in the nephrology room.
In the 18 included trials that tested treatment initiation with ACEis or ARBs versus placebo or other antihypertensive agents, the risk of progression to KFRT was reduced by 34%, which is a huge win for the positive impact of RASi even in CKD stage 4. Nephrologists should be encouraged that the treatments we have are effective, and the challenge remains to get our patients to consistently take them. While we are excited about our new armamentarium it is also fantastic that our cheapest medication has such efficacy in CKD stage 4, which will be particularly relevant internationally in places where the newer treatments remain unaffordable. As you'd expect there are a lot of familiar names in the included trials - the big three: AASK, RENAAL, and IDNT for example - contributed approximately 2/3s of events for the meta-analysis. Finally, the effects of ACEis or ARBs did not vary by age, eGFR, or history of diabetes for the outcome of KFRT or death.
These findings are particularly striking when you consider just how high-risk these patients were. With a mean eGFR of 22 ml/min/1.73m², heavy proteinuria, and an average baseline blood pressure of 155 mmHg, they were clearly teetering on the edge of kidney failure. Using the Kidney Failure Risk Equation (KFRE), and assuming North American male patients with these values, the estimated risk of kidney failure is staggering—37% at 2 years and 76% at 5 years. Against this backdrop, the hazard ratio for preventing ESKD in these studies is comparable to what we’ve seen with flozins in CKD4, such as in EMPA-KIDNEY’s subgroup analysis or the advanced CKD data from DAPA-CKD (Chertow GM, JASN 2021) (acknowledging that the flozins had ‘more work to do’ as most patients were already on RASi, event rates were generally lower and the follow-up time was generally shorter). On the other hand, a similar magnitude of benefit may not be seen in non-proteinuric and lower risk patients.
That said, it’s worth noting that most of these trials titrated RASi doses to maximize the effect, so these HRs aren’t what you should expect from starting someone on, say, a quasi-homeopathic dose of 1.25 mg of ramipril. While our current standard of care in CKD4 has come a long way, making it less likely to see this same magnitude of benefit today, these results are still impressive and underscore the potential impact of RASi in high-risk populations. The data remind us of the importance of tailoring aggressive treatment to those most at risk, where the payoff can be significant.
All of the trials included in this meta-analysis are decades old (old enough that you won’t find the term ‘CKD’ used in most of them), and the question being asked is of such clinical relevance, that it is hard to believe this analysis has never been undertaken before - the authors must be congratulated on this. Breaking of strongly held beliefs and paradigms is often difficult, in this case amongst nephrologists who are still on the ‘routine discontinuation of RASi in CKD4’ mindset, but there is no denying based upon these findings, that the case for actually initiating RASi in CKD stage 4 is compelling. One wonders if doing the STOP-ACEi trial at all could actually have been justified if this data was available first.
This year’s KDIGO CKD guidelines misfile this study under "continuation of RASi in CKD4" (practice point 3.6.7: “Continue ACEi or ARB in people with CKD even when the eGFR falls below 30 ml/min/1.73m²”, missing the far more important point: this is the strongest evidence we have that initiation of RASi in CKD4 is beneficial. KDIGO also states, “Consider reducing or discontinuing RASi to reduce uraemic symptoms when eGFR <15 ml/min/1.73m²,” which feels increasingly out of step. STOP-ACEi showed discontinuing RASi didn’t help eGFR, so it’s hard to imagine symptoms improving meaningfully with this approach. The KDIGO workgroup may have been influenced by the (excellent) target trial emulation study which suggested possible CV benefit but renal harm of continuing RASi in advanced CKD (Fu et al, JASN 2021). More to the point, 32% of patients in the present analysis had eGFR between 15-20 ml/min/1.73m², and the reduction in dialysis events was substantial. Instead of lingering on when to stop RASi, KDIGO should focus on encouraging appropriate initiation in CKD4. Admittedly, no study pushed initiation into CKD5, so caution is warranted there—but these findings at least open the door to exploring benefits even in the lower teens with the right patient selection.
Figure 21. Algorithm for monitoring potassium and eGFR after RASi. KDIGO CKD 2024.
Strengths
They included a large number of patients with CKD 4, and over 50% of the included patients were female (more impressive than our modern DKD trials)
Individual-level pooled meta-analysis
Many of the trials were against an active antihypertensive comparator, and the results were consistent within these trials, so we can say it is the specific magic of RASi rather than just having lower blood pressure per se, that is driving lower KFRT rates.
Limitations
Individual participant data for hyperkalemia or acute kidney injury were not available in the CKD-EPI CT and small trials did not report the rate of these events. However, newer treatments for hyperkalemia, in combination with significant benefits from RASi, suggest treating the potassium rather than stopping the ACEi/ARB.
Non-fatal cardiovascular endpoints and cardiovascular death were not available in many trials. However, numerically fewer deaths were seen in the RASi group, so even if there was no significant reduction in overall deaths, there was no signal that we should be concerned about RASi use.
Most patients had severely increased albuminuria (A3 category).
A few trials had an open-label design, which may have introduced some bias. However, most of the large trials that contributed to a majority of the events were not open-label design.
With a mean age of 55, one might suppose RASi in this analysis to be better tolerated than in older cohorts, so caution should be applied when extrapolating these results to your more elderly patients.
Conclusion
In this era of evolving treatments for the prevention of KFRT, it is evident that patients benefit not only from the continuation but even the initiation of RASi in CKD stage 4. We are never going to see more RASi RCTs, so this is the best evidence we’ll ever have, and it is impressive, to show the imperative to start these time-tested therapeutics for these high-risk patients.