Response to “Hyperammonemia following glufosinate-containing herbicide poisoning: A potential marker of severe neurotoxicity” by Yan-Chido Mao et al., Clin Toxicol (Phila) 2011; 49:48–52

To the Editor:

We are grateful to Yeon et al. for their valuable comments on our paper.¹ In their letter data, 13 additional cases of GLA poisoning are reported, of whom 8 had their serum ammonia measured.² Based on the data of the 8 cases, they suggested that “rising or maintained patterns of hyperammonemia within 24 hours post-ingestion may be more predictable (of severe neurotoxicity) rather than absolute value of hyperammonemia.”

While it may be true that rising or persistently elevated serum ammonia levels better predict the development of severe neurotoxicity following GLA poisoning, their case series and our observations show that more data in this area are required before firm conclusions can be drawn. A confounder may be the additional toxic causes of hyperammonemia, including liver failure and complicated poisoning. In the data reported by Yeon, it is unclear whether any patient had pre-existing liver failure or other diseases that might lead to hyperammonemia. The potential impacts of other causes of neurotoxicity, such as cerebral ischemia or hypoxia, should also be evaluated. Finally, in the assessment of the relation between hyperammonemia and severe neurotoxicity following GLA poisoning, one may need to differentiate between earlier onset neurotoxicity, such as stupor, coma, delirium, and seizures, and delayed-onset neurotoxicity, namely, prolonged amnesia.

In our report,¹ the serum ammonia level was generally measured only after the development of neurotoxicity. Thus, we did not have adequate data to comment on the predictive value of specific time courses of hyperammonemia on earlier onset neurotoxicity, and we agree this is a major limitation.

In a recent retrospective analysis of 115 GLA-poisoned patients reported to the Taiwan National Poison Control Center, we found that 6 patients with severe/fatal effects and available serum ammonia levels all presented with hyperammonemia (median serum ammonia level 122.5 μg/dL; reference value 5–69 μg/dL), whereas 5 mildly poisoned patients who had their serum ammonia level measured did not have hyperammonemia (unpublished data). Thus, we believe that hyperammonemia, either with a value above a certain “threshold” or with the presentation of specific time courses, could have a role in the prediction of severe neurotoxicity following GLA poisoning. We agree that the value of hyperammonemia nevertheless awaits further larger-scale prospective studies.