Salicylate toxicity after undetectable serum salicylate concentration: a retrospective cohort study

Background: Salicylates are usually rapidly absorbed and quickly measurable in serum. An undetectable serum salicylate concentration ([ASA]) may occur early after ingestion and may be interpreted as evidence of non-exposure and not repeated. Although cases of delayed salicylate detection are reported rarely, the risk factors associated with this phenomenon are not known. Research question: What factors are associated with an early undetectable [ASA] in salicylate poisoning? Methods: Records from a single regional poison center were searched from 2002 to 2016 for cases of salicylate toxicity treated with bicarbonate and [ASA] > 30 mg/dL. Cases were excluded if initial [ASA] was obtained >4 h after presentation. Case information, serial [ASA], and outcomes were recorded and compared between groups. Results: A total of 313 records met all criteria with 11 initially undetectable [ASA] (3.5%) and 302 detectable [ASA] (96.5%). Time of first [ASA] occurred sooner in the undetectable [ASA] group (89 vs. 137 min, p = 0.011) while time to peak [ASA] was longer (640 vs. 321 min, p <.001). The longest interval between ingestion and undetectable [ASA] was 225 min. Peak [ASA] and reported mean ingested dose were similar in both groups (45 vs. 50 mg/dL, p = NS; 19.7 g vs. 32.9 g, p = NS). Coingestion of agents that delay gastric emptying were similar in both groups (18% [2/11] vs. 25% [76/302], p = NS, chi-square). Hemodialysis was performed in 9% (1/11) of undetectable [ASA] patients and 5.6% (17/302) of detectable [ASA] patients (p = NS, chi-square). A single death occurred in the entire cohort in a patient with an initially detectable [ASA]. Discussion: In this series, a small but significant proportion (3.5%) of patients who developed [ASA] > 30 mg/dL had an initially undetectable [ASA]. Those with an undetectable [ASA] were measured earlier after ingestion with a longer time to peak [ASA]. However, neither coingestion of agents prolonging gastric emptying nor reported dose ingested was different between groups. Formulation was infrequently recorded but one undetectable [ASA] did ingest a non-enteric coated product. Limitations include the small number of patients with undetectable [ASA], use of single poison center data and partial data on co-ingestants and aspirin formulation. Conclusions: [ASA] may be undetectable early after an overdose and need for serial [ASA] in the evaluation of salicylate ingestion should be further explored. Additional research is needed to determine any causative factors and the optimal timing of [ASA] measurements.