When the diagnosis of toxic dystonia

When the diagnosis of toxic dystonia, and not meningitis, was suspected, the strategy of investigation changed to a search for a toxic agent causing extrapyramidal reactions. The management of patients with dystonic reactions was modified, lumbar punctures were only performed for those with clear evidence of meningitis, use of empirical GS-9620 was reduced, and biperiden was used to alleviate dystonia. Toxicological analysis was performed on urine samples from nine patients with dystonia (February and March, 2015) and 39 medicines available in pharmacies or procured at a government health centre (March and June, 2015) at the Pharmacology-Toxicology Service, Hôpital Raymond Poincaré, Garches, France. Liquid chromatography coupled to diode array, tandem mass spectrometry, gas chromatography coupled to mass spectrometry, and Q-Exactive hybrid quadrupole-Orbitrap mass spectrometry were conducted on all urine and drug samples (). Haloperidol was detected in urine from all nine patients with dystonia (median urinary concentration 4 μg/L [range <1 to 49 μg/L]). Haloperidol was also detected in all of nine yellow tablets imprinted with the letters “AGOG” and labelled as diazepam (). The median haloperidol dose per tablet was 13·1 mg (range 9·5–19·9). 15 mg of haloperidol is 20–25 times the usual daily maximal recommended dose for a 5 kg child. No diazepam or other toxic substances were detected. No haloperidol was detected in the other 30 medicine samples, which all contained the appropriate active pharmaceutical ingredients. Six of the nine AGOG-marked tablets were obtained from patients with dystonia who had ingested them, and three tablets were bought in a pharmacy or procured at a government health centre.
Martyn Plummer and colleagues (September, 2016) analyse the global burden of cancers attributable to infections in 2012. Despite the relatively low attributable proportion compared with other infectious agents, liver flukes, including and , were estimated to have caused 1300 new cases of cholangiocarcinoma in 2012. However, this number might underestimate the real situation. As described in Plummer and colleagues\' appendix, parameters applied in estimation included the baseline incidence of cholangiocarcinoma from cancer registry data in Japan, the relative risk (RR) of cholangiocarcinoma in those infected with liver flukes, and the population with liver fluke infections. However, Fürst and colleagues used a different approach and reached an annual figure for new cholangiocarcinoma cases attributable to liver flukes of nearly 7000 (). In that study, odds ratios (ORs) of 4·39 for and 6·12 for were combined with the cholangiocarcinoma incidence of a population highly affected by in northeast Thailand. In our earlier study, we also attempted to analyse the cholangiocarcinoma cases attributable to , using data from South Korea. After extrapolation to the global infected population, we estimate that more than 4700 cholangiocarcinoma cases were attributable to this organism, which approaches the figure of 5591 estimated by Fürst and colleagues (). The figure for the infected population in Plummer and colleagues\' analysis is similar to those of the two other studies, and the OR is higher. Thus, their substantially lower estimation of the number of cholangiocarcinoma cases attributable to liver flukes arises from the introduction of baseline cholangiocarcinoma data from Japan. The raw data from case-control studies used to calculate combined ORs were done in liver-fluke-endemic areas. Therefore, the lower cholangiocarcinoma incidence in Japan could not represent that in the control populations in these studies. Because of inadequate studies and insufficient data, and the complex and chronic characteristics of cancers, it is challenging to capture an accurate estimate of the number of cholangiocarcinoma cases attributable to liver flukes. Thus, more research is needed, which will have the added benefit of raising awareness of the importance of liver flukes in east Asia and promote control and elimination.
We thank Men-Bao Qian and Xiao-Nong Zhou for highlighting their concerns. Although our analysis focused on the four main infectious agents (human papillomavirus, hepatitis B virus, hepatitis C virus, and ) responsible for 92% of the global burden of infection-attributable cancer worldwide, we welcome any improvements for the rarer infections such as and , which are important causes of mortality and morbidity in endemic populations. According to Fürst and colleagues, parasitic diseases are “characterised by highly focal spatial occurrence and scarce, scattered, and patchy information”. The extent of this disease burden is therefore extremely difficult to estimate. As noted by Qian and Zhou, our estimates of infection prevalence and corresponding relative risk (RR) for cholangiocarcinoma are largely similar to other estimates because they are mainly based on the same sources. Differences in the estimates of the burden of cancer must therefore be due to other methodological issues. We would like to highlight two of these.