Serum lactate: a poor predictor of postoperative mortality in paediatric cardiac surgery

Introduction

Initial serum lactate > 4.5 mmol/l has been found to predict mortality in children following cardiopulmonary bypass for open heart surgery [1]. Since in our experience initial lactate varies widely in both survivors (S) and non-survivors (NS), we hypothesized that persistent hyperlactataemia, or hyperlactataemia in combination with base deficit, might better predict postoperative mortality.

Design

One hundred and ninety-three children, median age 7 months (inter-quartile range 0.5-63 months) were operated on for congenital heart disease over a 1 year period. One hundred and forty-four patients were selected on the basis of surgical complexity. Sequential serum lactates (mmol/l) were measured prospectively at 0, 6, 12, 18 and 24 h. Mean blood pressure, base deficit, bypass time, complications, length of ICU stay and outcome were recorded. Complicated postoperative course was defined as the presence of: seizures; INR/APTT > 3 or AST > 4× normal; and peritoneal or haemodialysis. Data were analysed on 124 patients by Spearman rank correlation, Mann-Whitney and Fisher's exact test. In 19 patients there were insufficient data.

Results

Nine patients died (7.3%) and the postoperative course was complicated in 30 patients (NS and S) by seizures (n = 4), liver dysfunction (n = 16) and dialysis (n = 16).

There was considerable overlap in initial lactate values between the S and NS groups. Initial lactate was significantly (P = 0.0002) elevated in non-survivors (median 8.66, range 1.9-17.6 mmol/l) compared to survivors (median 2.17, range 0.55-13.6 mmol/l), as was lactate at 6, 12, 18 and 24 h. Median 6 h lactate was 5.59 mmol/l in non-survivors (range 2-17.1 mmol/l) and 1.13 mmol/l in survivors (range 0.41-7.42 mmol/l). Twenty-two patients (17.7%) with initial lactate > 4.5 mmol/l survived to discharge. Using ROC analysis, an initial lactate level of 6 mmol/l had the best predictive value for mortality.

Sequential lactate fell significantly only in survivors (non-parametric ANOVA, P < 0.001) but persistent hyperlactatatemia, ie initial and 6 h lactate > 6 mmol/l, predicted mortality with a positive predictive value (PPV) of only 50%. Postoperative lactate correlated poorly with total bypass time (r = 0.23) and initial base deficit did not correlate significantly with either bypass time or initial lactate. Indeed, the presence of hyperlactataemia with base deficit, ie initial lactate > 6 mmol/l plus base deficit > 4 mmol/l, predicted mortality with PPV of only 38%.

Conclusion

Initial lactate levels vary widely in both survivors and non-survivors following complex paediatric cardiac surgery and have low predictive value for mortality. Serum lactate determination may be useful in predicting postoperative complications.

Table Predictive value of initial postoperative serum lactate > 6 mmol/l