LAUSR

Buckley et al. retrospectively assessed the need for haemodialysis based on the EXtracorporeal TReatments In Poisoning (EXTRIP) workgroup criteria in a cohort of 361 lithium-poisoned patients. We would like to congratulate the authors for their attempt to improve EXTRIP criteria and discuss their results in the light of our recently published study assessing EXTRIP criteria similarly in a cohort of 128 lithium-poisoned patients. It is noteworthy that we used more stringent criteria to define severe lithium toxicity and that our patients were more severely poisoned, that is, including patients more susceptible to benefit from haemodialysis to limit the risk of fatal outcome or neurological sequelae development (Table 1). These two studies point out the difficulties in interpreting some criteria loosely defined in the EXTRIP guideline. The ‘decreased level of consciousness’ criterion was interpreted as a Glasgow coma score (GCS) of <15 in Buckley's study but as GCS of <12 in ours. The ‘confusion’ criterion was construed as onset of confusion in Buckley's study but as GCS of 12–13 in ours. The ‘serum lithium concentration expected to be >1 mmol/L at 36 h with optimal management’ criterion was determined in Buckley's study using the Ct=Co e−0:161 eGFR+6:47 55 t equation but in ours as serum lithium concentration > 2.5 mmol/L measured 24 h after admission, taking into account the lithium half-life of 24 h in the patient with normal kidney function. Such discrepancies between two teams of toxicologists in translating EXTRIP criteria clearly underline the difficulties expected when non-specialist physicians will use these criteria at the bedside to target lithium-poisoned patients who could benefit from haemodialysis. Interpretation of EXTRIP criteria was more stringent in our approach than in Buckley's one. Interestingly, the use of Buckley's equation in our cohort to interpret the EXTRIP kinetic criteria would have led to perform haemodialysis in 28 additional patients in our cohort. The development of a nomogram to predict serum lithium concentration at 36 h is an interesting proposal that clearly facilitates EXTRIP criteria application. However, we would like to raise some concerns regarding the determination and validation of the provided nomogram. The authors used the Cockroft and Gault equation with the standard male and female weights derived from the Australian population medians to estimate creatinine clearance in their patients to build the nomogram while they used the CKD-EPI formula to determine the glomerular filtration rate in the same patients to validate it. They built their equation assuming a normal volume of distribution while acute-on-chronically lithium-poisoned patients may frequently be dehydrated on admission due to vomiting and diarrhoea, and chronically lithium-poisoned patients may be either dehydrated for