Angiotensin-converting enzyme inhibitors and angiotensin receptor blockers in the perioperative period: withdraw or continue? A multicentre randomised controlled trial

Editor

Many surgical patients use renin–angiotensin system inhibitors (RASi), including angiotensin-converting enzyme inhibitors and angiotensin receptor blockers. Although there is limited evidence, current guidelines recommend discontinuation before surgery in patients at risk of hypotension.1, 2, 3 This lack of evidence remains a crucial knowledge gap in perioperative medicine and leads to varying practices among hospitals and individual physicians.^4,5 We therefore aimed to compare the effect of temporary discontinuation vs continuation of RASi on acute kidney injury (AKI). This trial was ended prematurely because of slow recruitment and therefore did not generate substantial evidence. Nevertheless, it provided crucial lessons that we wish to share as they might be relevant to future trials in perioperative medicine.

We conducted a parallel, open-label, multicentre, randomised controlled trial (ClinicalTrials.gov: NCT04506372). All participants provided written informed consent and the local ethics committees approved the study protocol before patient recruitment (Medical Research Ethics Committee Utrecht; 20-027/D and University Health Network Toronto Research Ethics Board; 21-5815). Adult patients using RASi who were scheduled for elective major surgery with an anticipated postoperative length of hospital stay of at least 2 nights were eligible for inclusion. Patients were randomised to discontinue RASi therapy 1 day before surgery, and to resume on the second day after surgery. Participants assigned to perioperative continuation of RASi were instructed to continue this drug throughout the perioperative period. Perioperative care beyond randomisation was at the discretion of the treating physicians. The primary outcome was postoperative AKI. Based on previous studies in comparable populations and local retrospective data, we estimated the incidence of the primary outcome at 10%. To detect a reduction in AKI from 10% to 7% in the intervention group (discontinuing RASi), with a type 1 error rate of 0.05, a power of 80%, and accounting for 10% loss to follow-up, 3200 patients were required (n = 1600 per arm) (Supplemental File S1).

Between September 2020 and January 2023, we randomised 318 of the 2,613 patients who met the inclusion criteria and ultimately included 305 in the analysis (Figure 1). Their median age was 68 (IQR 61–75) yr, with 63% of the patients being male. The incidence of AKI was 4%, and no difference was observed between the two groups. However, because this study was terminated early, it was underpowered to detect the anticipated difference in the primary outcome. No differences were observed for any of the other outcomes (Supplemental File S2).

Fig 1.

Study flowchart. RASi, renin–angiotensin system inhibitors.

The trial was terminated because of a continuing low inclusion rate after multiple attempts to increase inclusions, and in consultation with the data safety monitoring board and the grant provider. There were several reasons for the low inclusion rate. During and after the COVID-19 pandemic and subsequent personnel shortages, a substantial reduction in the number of surgical procedures occurred, and patients were discharged sooner to overcome shortage of available (staffed) hospital beds. In addition, surgeries were relocated to other hospitals. Further, in the design phase of the trial, a larger number of hospitals intended to participate. However, by the eventual start of the trial, many hospitals were not able to participate because of not resolving logistical constraints related to the pandemic. In response to the low inclusion rate, we broadened the inclusion criteria and tried to recruit more participating centres. Despite these changes, centres that initially seemed interested considered the study not feasible because the study procedures (randomising individuals to an intervention outside the operating room area) or required resources (i.e. an established research infrastructure).

Importantly, several lessons were learned from this trial. First, we studied a simple intervention (continue or discontinue RASi therapy around surgery), and our trial followed a pragmatic design. As a substantial number of the target group was on RASi therapy, we expected that inclusion of patients would be relatively easy.^4,6 However, of 2613 patients who were eligible in the six participating hospitals in 2.5 yr, only 318 (12%) were randomised (Fig. 1) (i.e. an average of 20 patients per hospital per year). For comparison, the STOP-or-NOT trial, in which patients were randomly assigned to either continue or discontinue RASi therapy before surgery, recruited 2222 patients in 40 hospitals over 5 yr (i.e. an average of 11 inclusions per hospital per year).⁷ Similarly, the SPACE trial comparing the impact of continuing or not RASi therapy included 262 patients in six hospitals in 4 yr, again an average of 11 per year per hospital.⁸ Although recruitment in these trials was interrupted by the COVID-19 pandemic, these low inclusion rates are concerning examples for the previously described and persisting decrease in recruitment rates in clinical trials.⁹ Determination of relevant differences in outcomes that occur at incidences below 10% require large-sized trials. If even a simple intervention results in only 10–20 inclusions per hospital per year, trials will become difficult to manage and extremely expensive. Although suggestions to overcome the challenges in conducting large-sized trials in perioperative medicine were pointed out previously, successful completion of such trials is still challenging. In hindsight, a different design (e.g. cluster randomisation of hospitals instead of individual patients or applying practice-preference randomised consent) might have enhanced study feasibility.¹⁰

Second, the observed incidence of AKI in this study was 4%, notably lower than the anticipated 10%. This might be explained by selection bias caused by the exclusion of patients at high risk of AKI (those with severe chronic kidney disease or heart failure), or because of the healthy participant effect.^11,12 It is unlikely that the calculated power would have been achieved even if the study had reached the intended sample size. With the incidence of AKI we observed, the required sample size to be able to detect a 40% decrease in the incidence of AKI (from 5% to 3%) would have been approximately 8440 patients. Whether or not such an investment is reasonable given the clinical relevance of a 2% absolute decrease is debatable.

In conclusion, this prematurely ended trial did not generate substantial evidence regarding the question of whether RASi should be continued or discontinued around major surgery. The early termination of this trial highlights the importance of rethinking trial design in perioperative medicine. Alternative designs, such as cluster randomisation or practice-preference randomised consent, might enhance feasibility.

Funding

A grant from the Dutch organization for health research and development ZonMw (project number 848018007).

Declaration of interest

The authors declare no conflicts of interest are present.

Handling Editor: Hugh C Hemmings Jr

Appendix A. Supplementary data

The following are the Supplementary data to this article: