Ascariasis, trichuriasis and hookworm infection together affect a substantial fraction of the world population and remain concentrated where sanitation is poor (Bethony et al., 2006). Control rests on periodic mass administration of benzimidazoles to school-age children and other at-risk groups, a strategy set out in the expert committee report that still frames national programmes (WHO, 2002). Monitoring is done by counting eggs in stool, almost always with the Kato-Katz thick smear, which fixes a known mass of faeces under cellophane soaked in glycerol and malachite green and allows eggs to be counted under low power (Katz et al., 1972).

The method is well matched to its original purpose, which was to measure intensity in heavily infected populations. It is poorly matched to the situation that successful programmes create. As intensity falls, a single thick smear misses a growing proportion of infections, and the loss is worst for hookworm, whose eggs clear rapidly from the preparation. A meta-analysis that estimated test sensitivity in the absence of a reference standard quantified the extent of the problem across the three parasite groups (Nikolay et al., 2014). Alternatives based on flotation improve recovery and allow larger volumes to be examined (Cringoli et al., 2010), but they remain microscopy, with its dependence on technician skill and on prompt examination.

Molecular detection answers that limitation and introduces others. Quantitative PCR for soil-transmitted helminths is well established in reference laboratories (Verweij and Stensvold, 2014), yet its interpretation in programme settings is debated, because the relationship between a cycle threshold and the egg burden that drives transmission is not straightforward (Papaiakovou et al., 2019). Loop-mediated isothermal amplification sits between these options. It requires only a heat block, and the chemistry tolerates crude templates. This article proposes a workflow for using it on stool in the field. The workflow is a proposal: it has not been applied to field specimens in preparing this article, and no performance results are reported here.

The reaction uses a strand-displacing polymerase and four primers recognising six regions of the target, which generates concatemeric products with stem-loop structures at a single temperature (Notomi et al., 2000). Two additional loop primers shorten the reaction considerably (Nagamine et al., 2002). The practical consequences are that no thermal cycler is needed, that the six-region recognition gives high specificity, and that the very large mass of DNA produced permits readout without electrophoresis. Products can be seen as magnesium pyrophosphate turbidity (Mori et al., 2001), or with metal-indicator dyes such as hydroxynaphthol blue (Goto et al., 2009), or through the pH shift caused by proton release during synthesis, which drives a colour change in a weakly buffered mixture (Tanner et al., 2015). A general protocol for setting up and reading these reactions is given by Tomita et al. (2008).

The same properties create the principal hazard. A reaction that yields very large amounts of product and is read by opening the tube is an efficient source of amplicon contamination, and a laboratory that has once contaminated its workspace will see positives everywhere. Physical separation of preparation and readout areas, unidirectional workflow, and readout formats that keep tubes closed are therefore not refinements but requirements. Table 1 compares the proposed workflow with the methods a programme already uses.

The workflow has five stages. Figure 1 pairs each with the failure mode it must control, which is the organising idea of the design: the amplification itself is robust, and almost all the loss of performance occurs before the tube is closed.

Three practical constraints shaped these choices. The first is staffing: the whole workflow is intended for a technician already competent with pipetting and microscopy, trained over a few days, rather than for a molecular biologist. The second is power and cold chain, which is why lyophilised reagents, battery-capable heat blocks and ambient-temperature preservatives appear throughout. The third is the purpose to which the result will be put. For mapping and for deciding whether to continue mass administration, a qualitative result at the level of the community may be sufficient, and pooling specimens before extraction becomes attractive; for individual diagnosis, pooling is not acceptable. That distinction should be settled before the first specimen is collected, because it determines the sampling design.

The workflow has not been validated. An evaluation would need to establish four things. The first is analytical sensitivity, expressed as eggs per gram of stool, determined on spiked negative stool across the three parasite groups. The second is diagnostic accuracy against a composite reference standard combining repeated Kato-Katz examinations and quantitative PCR, reported in the manner that the STARD statement sets out (Bossuyt et al., 2015), with reagent lots and operating conditions documented as fully as the MIQE guidelines require of quantitative PCR studies (Bustin et al., 2009). The third is reproducibility between operators and sites, since a visual readout is the weakest link. The fourth is programmatic utility: whether a more sensitive test changes the decisions a programme actually takes, which is a separate question from whether it detects more infections.

Two cautions remain. A positive result indicates target DNA, not a viable infection capable of contributing to transmission, and DNA from ingested or recently cleared material can give a signal of no epidemiological consequence. And the published work applying this chemistry to helminths is still limited; studies such as the hookworm assay described by Mugambi et al. (2015) show feasibility rather than readiness for programme use. The proposal here should be read as a design to be tested, not as a method to be adopted.

Isothermal amplification is a reasonable candidate for field detection of soil-transmitted helminth eggs because it removes the thermal cycler and tolerates imperfect templates. The obstacles lie in egg disruption, inhibitor removal and contamination control rather than in the amplification itself, and the workflow proposed here is organised around those three problems. Whether it performs well enough to supplement or replace the thick smear in programme monitoring is an empirical question, and the evaluation outlined above is the work that would answer it.

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

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