Laboratory predictors of acute dialysis in hemolytic uremic syndrome.

BACKGROUND

Strict guidelines on use of dialysis in children with post-diarrheal hemolytic uremic syndrome (D + HUS) are lacking. This study investigated laboratory predictors of acute dialysis because they are more objective than clinical features. Added to this, given that urine output is also an objective parameter, its ability to predict dialysis requirements was also investigated.

METHODS

Out of 153 D + HUS children reviewed, 88 received dialysis and 65 did not. Initial laboratory parameters and diuresis between both groups were analyzed.

RESULTS

Dialyzed patients had higher creatinine, urea, alanine aminotransferase, hematocrit and leukocyte count; and lower sodium, bicarbonate, and pH compared to non-dialyzed ones. Serum creatinine was the only independent predictor (P = 0.003) of dialysis; therefore, its ability to predict dialysis was estimated on receiver operating characteristic (ROC) curve analysis and using the Acute Kidney Injury Network (AKIN) staging system. Area under the ROC curve was 0.92 (95% confidence interval [95%CI]: 0.83-1) with a creatinine cut-off of 1.25 mg/dL (sensitivity, 100%; specificity, 76.5%) for children <1 year, and 0.93 (95%CI: 0.88-0.98) with a threshold of 2 mg/dL (sensitivity, 91%; specificity, 87.5%) for older children. AKIN stage 3 at admission predicted dialysis with a sensitivity of 92% and specificity of 84.2%. Urine output had the highest accuracy for dialysis prediction (sensitivity, 100%; specificity, 95.3%).

CONCLUSIONS

Initial serum creatinine concentration was the best laboratory predictor of dialysis, but the first 24 h diuresis was even better for this purpose. But, given that serum creatinine is an immediate available parameter, the cut-offs identified may label D + HUS children who will probably need dialysis, prompting early referral to centers able to provide dialysis.