Letter to the Editors-in-Chief regarding Velmurugan et al.,—Association of co-accumulation of arsenic and organophosphate insecticides with diabetes and atherosclerosis in a rural agricultural community: KMCH-NNCD-I study

Dear Editors-in-Chief,

Velmurugan and colleagues recently reported serum levels of organophosphates in a sample of farmers from the state of Tamil Nadu in South India [1]. Samples were collected in March–April 2015, which is not the time of year pesticides are typically applied in this region.

We were shocked to see the levels of organophosphate (OP) pesticides reported in serum: mean methyl parathion of 45.10 μg/L and chlorpyrifos of 49.57 μg/L, for example. We have extensive experience measuring OP pesticides internationally, in a variety of human matrices in both occupational [2, 3] and non-occupational [4, 5] exposure settings and have never encountered serum levels this high. OP pesticides are inherently unstable in serum, especially if not collected and stored cautiously. Even in properly frozen serum, OP pesticides are unstable over months [6]. In occupational settings with well-documented, extensive use of OP pesticides and serum samples analyzed quickly after collection, the maximum values we have observed were less than 0.5 μg/L [6]. In fact, in forensic applications where death from OP pesticide poisoning occurred, serum levels were typically less than 10 μg/L [7]. In the Velmurugan et al. [1] study, the mean serum level of all OP pesticide concentrations reported was > 10 μg/L. We doubt very much that such high levels could be observed in living study participants.

In addition, we found that the values reported are remarkably similar for each pesticide. In our previous studies of rural populations in Thailand [2] and Bangladesh [4], it was consistently found that para-nitrophenol (a metabolite of methyl and ethyl parathion) and 3,5,6-trichloro-2-pyridinol (a metabolite of chlorpyrifos and chlorpyrifos methyl) are the most commonly detected urinary OP metabolites (detected in 78–100% of samples), whereas other metabolites, such as malathion dicarboxylic acid (a metabolite of malathion), are rarely detected (< 10% of samples).

As written, there are not enough details in the methods to enable proper evaluation of their measurements. In addition, there are several errors. For example, the authors state that they used dispersive liquid–liquid microextraction technique for the OP pesticide assay, but then cite a paper that uses solid-phase microextraction sampling. They also state that they incubate the samples in 5 N HCl at 70 °C for 30 min. (They actually say “70 min for 30 min” but that must be an error.) Even without the elevated temperature, it is unlikely that OP pesticides could withstand that level of acidity. In fact, in our laboratory experiments we were forced to use a “softer” deproteination technique to prevent degradation of the OP pesticides [8]. Furthermore, no information was provided on how samples were collected or stored, or how long they were stored prior to analysis.

We observed other methodological details that were worrisome. Some of the ions monitored are very low in molecular weight, near the chemical noise region of mass spectra, which would make it difficult to have the limits of detection (LOD) they report. Indeed, the LODs reported are unlikely to be achievable with the instrumentation and ions they are monitoring, though it is difficult to know exactly, because the specific instrument used is not mentioned. They also spike the internal standard at ppm levels (probably because they could not detect it at lower concentrations), but the measured values are low ppb values. This can result in a great deal of error when calculating analyte concentrations.

In sum, we are not convinced that what the authors report as OP pesticides are actually this exposure and not simply chemical noise. We kindly request that further details be provided so that the validity of the findings can be evaluated.