Potassium levels and egfr do not predict severe hyperkalemia following spironolactone introduction in patients with ckd at high risk of hyperkalemia

Background and Aims Mineralocorticoid receptor antagonists (MRA) reduce blood pressure, albuminuria and the rate of disease progression in patients with chronic kidney disease (CKD) and albuminuria. Despite these apparent benefits, only a very small fraction of patients with CKD are treated with an MRA. This may in part be due to the fear of hyperkalemia (HK), which in the most severe cases can cause life-threatening arrythmias. Indeed, international guidelines and previous studies have excluded patients believed to be at high risk of severe HK from treatment with MRA including patients with pre-existing high serum potassium. To examine if the risk of HK can in fact be predicted by baseline potassium levels or eGFR, we performed a clinical study testing the effect of introducing spironolactone on plasma potassium (P-K) levels in closely monitored, high-risk patients excluded from other studies. Secondly, we analyzed the effect of spironolactone on eGFR and albuminuria. Method We included patients with eGFR 25-60 ml/min/1.73 m2 on maximal tolerated RAAS-blockade (ACEi or ARB) and a history of at least two HK-episodes (P-K > 4.5mmol/l) within 24 months prior to inclusion. Following dietary counselling on avoidance of potassium-rich foods, spironolactone was initiated at 25 mg daily. If tolerated as defined by a decline in eGFR < 30%, a P-K ≤ 5.5mmol/l and the absence of severe hypotension, the dose was increased to 50 mg after two weeks. Total follow-up was four weeks measuring P-K, eGFR, blood pressure and spot urine albumin creatinine ratio. Results from maximal tolerated dose were compared to baseline using paired t-test. In a post-hoc analysis, patients were grouped based on the occurrence of severe HK (P-K > 5.5mmol/l) or not, and baseline characteristics and the change from baseline to maximal dose were compared using unpaired t-test. Linear regression model was used to test the association between baseline P-K vs. P-K at maximal dose of spironolactone and the change in P-K from baseline to maximal dose of spironolactone vs. baseline eGFR. Results Fifty-eight patients were included with a mean age of 65 years. Forty-seven were males and 23 had diabetes. Forty-eight patients reached a spironolactone dose of 50 mg. Following spironolactone introduction, mean eGFR declined from 39 at baseline to 34ml/min/1.73 m2 (p<0.001) and albuminuria was reduced from a median of 1276mg/g to 654 mg/g (49%; 95%CI: 44 – 54%) with no significant change in blood pressure. Mean P-K increased 0.5mmol/l (95% CI 0.3 - 0.7mmol/l) from 4.7mmol/l to 5.2mmol/l. Seventeen patients developed severe HK with a P-K > 5.5mmol/l and four were briefly admitted with a P-K > 6.2mmol/l. Importantly, there was no difference in baseline P-K nor eGFR in patients that developed severe HK when compared with those that did not (4.70 vs 4.67 mmol/L, p = 0.83 and 36.2 vs. 40.1 ml/min/1.73 m2, p = 0,13). Furthermore, baseline P-K did not correlate with P-K at maximum spironolactone dose (Figure 1) and the change in P-K did not significantly correlate with baseline eGFR (Figure 2). Conclusion Short-term treatment with spironolactone in patients with CKD at high risk of HK leads to similar reductions in albuminuria and eGFR when compared with low-risk cohorts. With dietary counseling, 30% of patients will develop severe HK within 4 weeks. Importantly and contrary to common belief, neither baseline P-K levels nor baseline eGFR were associated with the development of severe HK. Thus, excluding patients from MRA treatment based solely on eGFR and P-K levels is not appropriate. Instead, we believe an empirical approach based on dietary counseling and close monitoring of P-K should be used.