Paraquat dichloride is a non-selective contact herbicide whose toxicology has been understood in outline since the 1970s. It undergoes redox cycling, generating superoxide at the expense of cellular reducing equivalents, and it is taken up by alveolar type I and type II cells through the polyamine transport system, which concentrates it in exactly the tissue least able to tolerate oxidative injury. The consequence is a characteristic course: oropharyngeal and gastrointestinal corrosion early, renal and hepatic injury over the following days, and in those who survive that phase a progressive pulmonary fibrosis that is usually fatal (Dinis-Oliveira et al., 2008; Bismuth et al., 1990).

Two features make paraquat a particular problem for poison centres. The ingested volume needed to cause death is small, so an accident rather than an act of self-harm can be lethal. And the window in which anything useful can be done is short, measured in hours, while the people affected are often far from a hospital able to measure a plasma concentration. Self-poisoning with pesticides accounts for a large share of deaths in rural areas of low and middle income countries (Gunnell et al., 2007), and comparative work on case fatality has placed paraquat among the most lethal of the agents in common use (Dawson et al., 2010).

Poison centres in affected regions handle such calls, but often without a protocol written for this specific agent. General guidance on poison centre functions exists (WHO, 1997), and the clinical literature on management is substantial (Gawarammana and Buckley, 2011), yet the step of turning that literature into a telephone procedure usable at two in the morning by a duty officer is rarely documented. This article proposes such a procedure. It is a proposal assembled from published sources and has not been implemented or evaluated.

Three bodies of evidence shape the protocol. The first concerns prognosis. Plasma paraquat concentration interpreted against time since ingestion has been used to predict survival since the nomogram published by Proudfoot et al. (1979), refined statistically by Hart et al. (1984) and examined prospectively by Senarathna et al. (2009). Urine concentrations carry prognostic information as well (Scherrmann et al., 1987). The practical value of these tools is that they separate those for whom intensive treatment may help from those for whom palliation is the honest course.

The second concerns treatment. Decontamination has a narrow window and the agents used for it have never been shown in controlled studies to change outcome. Immunosuppressive regimens combining a glucocorticoid with cyclophosphamide have been investigated for decades; the systematic review by Eddleston et al. (2003) found the evidence insufficient, and the later Cochrane review reached a similar conclusion about the quality of the available trials (Li et al., 2021). Supplemental oxygen is the one intervention on which the mechanistic argument is clear in the opposite direction: because the injury is oxidative, oxygen given in the absence of significant hypoxaemia is expected to worsen it, and standard toxicological references advise against routine administration (Hoffman et al., 2019).

The third concerns the circumstances of exposure. Accidental ingestion in agricultural households is associated with storage of concentrate in the home and with decanting into drink bottles and cups. Measures that restrict availability and improve storage have been the main recommendation of the prevention literature for twenty years (Konradsen et al., 2003; Wesseling et al., 2001), and they are reflected in the international code on pesticide management (FAO and WHO, 2014) and in the classification of paraquat under the hazard scheme maintained by the World Health Organization (WHO, 2020).

The protocol divides a call into five tasks. Figure 1 shows how the opening enquiry branches, and Table 1 sets out each task with its purpose and the entry it should leave in the record. The tasks are described in turn below. They assume a centre staffed by trained information officers with pharmacist or clinical supervision available, which is the usual arrangement.

Three things determine whether the protocol is usable. The first is that it fits on a small number of pages with the scripted advice in a fixed place, since a duty officer will not read prose during a call. The second is that the local information, which laboratories measure paraquat, which hospitals can take blood, which products are registered in the country, is filled in locally and reviewed, because that information changes and a protocol with a stale telephone number is worse than none. The third is training, including rehearsal of calls where the product cannot be identified, which is the common case.

The protocol also needs to sit alongside the centre's general procedures for pesticide calls rather than standing apart from them. Organophosphorus compounds remain the commonest agents in many settings and have an entirely different management (Eddleston et al., 2008), and a duty officer who cannot yet identify the product needs a route that keeps both possibilities open.

This is a proposed protocol and has not been implemented or evaluated. What would need to be shown falls into three parts. Whether the identification questions actually distinguish paraquat from other herbicides in calls where the label is absent, which could be examined against centres' existing call records. Whether the advice given is followed, particularly the instruction about oxygen, which requires follow-up with the transferring services. And whether the protocol changes anything that matters, such as the proportion of cases in which a urine test is done and a timed plasma sample taken, which is a realistic process measure even where mortality is not.

A limitation of a different kind should be stated plainly. Better telephone triage cannot substantially change the outcome of a substantial ingestion of concentrate, because no treatment of proven benefit exists. The measures that have reduced deaths where they have been adopted are regulatory: restricting or withdrawing the most hazardous formulations and limiting household storage (Konradsen et al., 2003). A poison centre protocol is a way of handling the consequences of a policy problem, not a solution to it.

Calls about suspected paraquat ingestion give a poison centre little time and few effective options, which is the reason to have the procedure written down in advance rather than assembled during the call. The five tasks set out here, product identification, circumstances, scripted immediate advice, investigations and follow-through, are drawn from the published clinical and prognostic literature and from international guidance on poison centre practice. Their value in service is untested, and the process measures that would test it are within reach of any centre that keeps call records. Prevention through storage and availability remains the part of the problem that matters most.

To be declared by the author(s) before publication.

To be declared by the author(s) before publication.