Raised lactate levels are known to be a marker of poor outcome in critically ill patients. Although bedside estimation of lactate levels is now possible, it remains a costly investigation and is still not freely available in third world countries. Anion gap (AG) has been suggested as a screening marker for LA. We examined the association of raised AG with hyperlactatemia (HL)/lactic acidosis (LA).
1852 lactate level estimations done on 352 patients in a medical ICU over a 6-month period were analyzed with concurrent AG levels. HL (lactate >2 mmol/L) was seen in 733 samples (39.6%). Samples with LA were identified as those with HL in presence of base deficit >4mmol/L (n=313). Receiver-operator characteristic (ROC) curves were drawn to study the utility of AG in predicting various levels of HL/LA (>2 mmol/L, >5 mmol/L and >10 mmol/L). Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were calculated for best cut-off of AG for predicting various levels of HL/LA.
Overall incidence of raised AG in patients with HL was 84% (n=615) whereas that for LA was 96% (n=300). AG estimation had excellent NPV for HL>5 mmol/L (>95%) as well as for any level of LA (>95%) but, at best, had a PPV of 66%. ROC curves showed AUC values of 0.641, 0.813 & 0.939 for HL and 0.819, 0.873 & 0.952 for LA with lactate levels >2 mmol/L, >5 mmol/L and >10 mmol/L, respectively.CONCLUSIONS: Monitoring of AG estimation is a fairly sensitive technique for screening of HL/LA and even mild HL with or without metabolic acidosis is associated with raised AG levels in 84% samples. Sensitivity and NPV tends to rise with increasing lactate levels and in presence of concomitant acidosis although PPV remains low.
A normal AG level (<15 mmol/L) in presence of metabolic acidosis almost always excludes LA although raised levels may not confirm its presence.
A. Banga, None.