Context: Ethylene glycol poisoning manifests as metabolic acidemia, acute kidney injury and death. The diagnosis and treatment depend on history and biochemical tests. Glycolate is a key toxic metabolite that impacts prognosis, but assay results are not widely available in a clinically useful timeframe. We quantitated the impact of serum glycolate concentration for prognostication and evaluated whether more readily available biochemical tests are acceptable surrogates for the glycolate concentration.
Objectives: The objectives of this study are to 1) assess the prognostic value of the initial glycolate concentration on the occurrence of AKI or mortality in patients with ethylene glycol exposure (prognostic study); 2) identify surrogate markers that correlate best with glycolate concentrations (surrogate study).
Methods: A systematic review of the literature was performed using Medline/PubMed, EMBASE, Cochrane library, conference proceedings and reference lists. Human studies reporting measured glycolate concentrations were eligible. Glycolate concentrations were related to categorical clinical outcomes (acute kidney injury, mortality), and correlated with continuous surrogate biochemical measurements (anion gap, base excess, bicarbonate concentration and pH). Receiver operating characteristic curves were constructed to calculate the positive predictive values and the negative predictive values of the threshold glycolate concentrations that predict acute kidney injury and mortality. Further, glycolate concentrations corresponding to 100% negative predictive value for mortality and 95% negative predictive value for acute kidney injury were determined.
Results: Of 1,531 articles identified, 655 were potentially eligible and 32 were included, reflecting 137 cases from 133 patients for the prognostic study and 154 cases from 150 patients for the surrogate study. The median glycolate concentration was 11.2 mmol/L (85.1 mg/dL, range 0-38.0 mmol/L, 0-288.8 mg/dL), 93% of patients were treated with antidotes, 80% received extracorporeal treatments, 49% developed acute kidney injury and 13% died. The glycolate concentration best predicting acute kidney injury was 12.9 mmol/L (98.0 mg/dL, sensitivity 78.5%, specificity 88.1%, positive predictive value 86.4%, negative predictive value 80.9%). The glycolate concentration threshold for a 95% negative predictive value for acute kidney injury was 6.6 mmol/L (50.2 mg/dL, sensitivity 96.9%, specificity 62.7%). The glycolate concentration best predicting mortality was 19.6 mmol/L (149.0 mg/dL, sensitivity 61.1%, specificity 81.4%, positive predictive value 33.3%, negative predictive value 93.2%). The glycolate concentration threshold for a 100% negative predictive value for mortality was 8.3 mmol/L (63.1 mg/dL, sensitivity 100.0%, specificity 35.6%). The glycolate concentration correlated best with the anion gap (R2 = 0.73), followed by bicarbonate (R2 = 0.57), pH (R2 = 0.50) and then base excess (R2 = 0.25), while there was no correlation between the glycolate and ethylene glycol concentration (R2 = 0.00). These data can assist clinicians in planning treatments such as extracorporeal treatments and prognostication. Potentially, they may also provide some reassurance regarding when extracorporeal treatments can be delayed while awaiting the results of further testing in patients in whom ethylene glycol poisoning is suspected but not yet confirmed.
Conclusions: This systematic review demonstrates that the glycolate concentration predicts mortality (unlikely if <8 mmol/L [61 mg/dL]). The anion gap is a reasonable surrogate measurement for glycolate concentration in the context of ethylene glycol poisoning. The findings are mainly based on published retrospective data which have various limitations. Further prospective validation studies are of interest.
Keywords: Ethylene glycol; anion gap; biochemistry; glycolate; overdose; poisoning; prognostication; surrogate.