Stopping RAS Inhibitors to Minimize AKI

Drugs that inhibit the renin-angiotensin system, in particular angiotensin-converting enzyme inhibitors (ACEIs) and angiotensin receptor blockers (ARBs), are cornerstones of nephrology care. In particular, they slow progression of proteinuric CKD and markedly improve prognosis for patients with heart failure with reduced ejection fraction (HFrEF). However, it is widely thought that they cause or exacerbate AKI, particularly in patients who are hypovolemic or septic. The supposed causal contribution of ACEI/ARB to AKI frequently leads to practice recommendations that such drugs are stopped during intercurrent illness. What is the evidence to support this practice? Randomized clinical trials are the gold standard for defining the risk of drug-related adverse effects. Higher rates of AKI on ACEI/ARB were not reported in any of the landmark trials that demonstrated the benefits of these drugs in slowing progression of diabetic or non-diabetic kidney disease—but this could partly be due to the fact that these trials were conducted in an era before agreement on a uniform biochemical definition of AKI and partly be because the studies were largely short term, excluding patients with comorbidities and those at risk of intercurrent illness, and required very close monitoring. By contrast, in high-quality recent clinical trials, such as the VA Diabetes In Nephropathy Study (VA-NEPHRON-D) study, dual blockade with combined use of ACEI and ARB was strongly associated with increased risk of AKI. However, combination therapy was associated with better prognosis after AKI, with greater recovery of kidney function, lower 30-day mortality, and lower hazard for the primary end point of death, ESKD, or decline in kidney function (1). This raises the possibility that ACEI/ARB use reduces GFR, leading to an increase in biochemically defined AKI but without causing genuine kidney injury. Similarly, recent publications from the Action to Control Cardiovascular Risk in Type 2 Diabetes (ACCORD) and the Randomized Trial of Intensive versus Standard Blood Pressure (SPRINT) suggest that a fall in GFR associated with intensive BP reduction is not associated with increased markers of tubular injury, although long-term outcome data are awaited. In a systematic review (2) of the evidence on elective cessation of ACEI/ARB to prevent AKI, randomized trials of cessation before endovascular or radiologic procedures involving contrast use did not show reduced postprocedure AKI; a large observational study showed that cessation of ACEI/ARB before cardiac surgery was associated with reduction in biochemical AKI but that there was no reduction in AKI defined by urinary biomarkers. In the absence of randomized study data, information on the link between ACEI/ARB and AKI comes from numerous observational studies, often undertaken among hospital inpatients. These are limited by the difficulty in accounting for confounding by indication—that people who are prescribed ACEI/ARB have conditions that place them at higher risk of developing AKI, regardless of the medication—and the difficulty of identifying an appropriate comparison group. There may also be confounded by contraindication: clinicians may elect to discontinue ACEI/ARB in sicker patients. ACEI/ARB may also be stopped more readily in those with preexisting CKD or because of hyperkalemia. The most robust observational evidence comes from population-based cohort studies where baseline comorbidities and other drug use can be accounted for. Recent studies of this type have shown weak or no association between ACEI/ARB use and AKI (3,4). However, patients with comorbidities for which ACEIs/ARBs are indicated (e.g., diabetes, HFrEF, and CKD) are at high risk of AKI, with rates increasing as the list of comorbidities grows (3). Because these vulnerable patients are frequently admitted to the hospital, the result is that physicians, particularly nephrologists, routinely see patients prescribed ACEI/ARB admitted with AKI, and as a result, they form potentially erroneous conclusions about the causal pathway. This repeated reinforcement of the idea of a link with AKI is likely to contribute to the fact that ACEI/ARB drugs have come to be widely but wrongly seen as “nephrotoxic” (5). The concept that ACEIs/ARBs cause AKI is cemented in multiple clinical guidelines. For instance, the Kidney Disease Improving Global Outcomes guideline on CKD advises clinicians to review or stop the medications in patients with or at risk of AKI, and it includes ACEI and ARB under the umbrella term “potentially nephrotoxic drugs” (guideline 4.4.3)—although the AKI guideline avoids the topic altogether other than to recommend more research on whether ACEI/ARB should be discontinued before radiologic contrast. The United Kingdom National Institute for Health and Care Excellence guidelines make a similar recommendation and use similar terminology. The concept that ACEIs/ARBs cause AKI during intercurrent illness is also embedded in the idea of “sick-day rules.” These are recommendations for patient-led drug cessation (including ACEIs/ARBs) if they become unwell with diarrhea and vomiting or with features suggestive of sepsis. Sick-day rules have been widely introduced in Europe and North America. However, a recent systematic review found that no research had been conducted investigating the harms and benefits of sick-day rules to reduce harm from AKI in the community setting (2). In fact, advice on temporary withdrawal of drugs during acute illness has been shown to cause confusion and a reluctance to restart treatment after the illness has resolved as well as require substantial ongoing support and resources (see reference 14 in the reference 5 cited here) (5). Despite the absence of evidence of benefit, it is frequently argued that temporary cessation of ACEI/ARB therapy is unlikely to cause harm. However, in addition to the concern that patients (and their primary care physicians) may be reluctant to restart drugs termed “nephrotoxic,” there are other potential harms. In particular, there is a risk of acute hemodynamic deterioration and worsening of circulatory congestion among patients with HFrEF previously stabilized on ACEI or ARB. Acute pulmonary edema is the leading cause of readmission after a hospital admission complicated by AKI (6). Although it is conceivable that this is due to (yet to be clarified) deleterious effects of AKI on cardiac function, an alternative explanation is the inappropriate reflex cessation of ACEI or ARB (or diuretics) in response to a rise in serum creatinine concentration. Observational findings comparing patients who continue ACEI/ARB drugs after AKI with those who do not, although again limited by unmeasured confounding, also show possible evidence of benefit. Discharge from an intensive care unit on ACEI/ARB after an episode of AKI is associated with lower 1-year mortality than discharge of otherwise similar patients (matched using propensity scores) discharged without such treatment (7). Prescription of ACEI or ARB after a hospital admission that included an episode of AKI is associated with lower 2-year mortality, although it was also associated with a higher risk of hospitalization for a kidney-related cause (8). An acute reduction in GFR after initiation of an ACEI or ARB is commonly seen in proteinuric CKD and HFrEF, and it is due to altered glomerular hemodynamics caused by relaxation of the efferent arteriole. This reduction in GFR is rapidly reversible on withdrawal of ACEI/ARB therapy. The resulting acute fall in serum creatinine after drug withdrawal during acute illness complicated by AKI may also convince clinicians that the drug was causing the AKI. Indeed, this hemodynamically mediated increase in GFR may well be responsible for the apparent reductions in incidence or severity of AKI reported in many recent studies incorporating a “bundle” of interventions, because these measures often include withdrawal of ACEI and ARB. Although efferent arteriolar vasodilation may indeed cause a further fall in GFR during hypotension, this also maintains blood flow to the kidney tubules, a region highly sensitive to hypoxia and where sustained hypoperfusion leads to acute tubular necrosis. This plus anti-inflammatory effects of Renin/Angiotensin/Aldosterone System inhibition may actually protect against genuine AKI (9). In the United Kingdom, the national “Think Kidneys” AKI program has provided pragmatic advice to primary care clinicians on “sick-day guidance,” when to restart medications stopped during a hospital admission with AKI, and how to respond to changes in kidney function and serum potassium during treatment with ACEI, ARB, and diuretic treatment (www.thinkkidneys.nhs.uk/aki/). This advice is in concordance with a professional consensus paper drawn up by the Renal Association and the British Society for Heart Failure (10). Despite the strong weight of clinical belief and practice, we argue that there is equipoise about the relationship between ACEI/ARB and AKI. Clearly, better evidence is needed to guide practice. There are clinical situations in which equipoise is not present: no clinician would continue ACEI/ARB in the face of severe systemic hypotension or hyperkalemia. However, the point at which the deleterious effects of systemic hypotension outweigh the possible protective effects of ACEI/ARB against tubular injury and inflammation has not been defined; we urgently need randomized, controlled trials of cessation versus continuation in patients who are normokalemic with reasonably preserved systemic BP at high risk of or who have developed AKI. Such trials will also need to distinguish between “hemodynamic” and “structural” AKI with the use of biomarkers and long-term follow-up. Until such trials have been performed, clinicians will have to use clinical reasoning rather than guideline-based reflexes to decide whether ACEI/ARB should be stopped in response to a small change in serum creatinine concentration in a patient for whom there is a strong evidence-based indication for long-term use of the drug. Disclosures None.