Diagnosis of protozoa diarrhoea in Campylobacter patients increases markedly with molecular techniques

Background Cryptosporidium and Giardia are major food and water-borne causes of diarrhoea globally, and two of the most notified infectious diseases in New Zealand. Diagnosis requires laboratory confirmation carried out mostly via antigen or microscopy-based techniques. However, these methods are increasingly being superseded by molecular techniques for diagnostics. Here we investigate the level of protozoa coinfection identified by molecular methods in Campylobacter positive samples missed through use of antigen-based assays and then investigated different molecular testing protocols. Methods We report the findings of two observational studies; the first among 111 people with diarrhoea during a large Campylobacter outbreak in Havelock North, and the second a study during normal surveillance activities among 158 people presenting with diarrhoea and a positive Campylobacter test, but negative Cryptosporidium and/or Giardia antigen-based diagnostic test result. The molecular methods used for comparison with the antigen-based tests were in-house end-point PCR tests targeting the gp60 gene for Cryptosporidium and gdh gene for Giardia. DNA extraction was performed with and without bead-beating and comparisons with commercial real-time quantitative (qPCR) were made using clinical samples diluted down to 10-5 for Cryptosporidium positive samples. Results The coinfection prevalence was 9% (n= 10, 3-15% 95%CI) for Cryptosporidium and 21% (n=23, 12-29% 95%CI) for Giardia in the 111 Campylobacter patients of the Havelock North outbreak. The coinfection prevalence was 40% (n=62, 32-48% 95%CI) for Cryptosporidium and 1.3% (n=2, 0.2-4.5% 95%CI) for Giardia in the 158 routine surveillance samples. Sequencing identified Cryptosporidium hominis, C. parvum, and Giardia intestinalis assemblages A and B among patients. We found no statistical difference in positive test results between samples using end-point PCR with or without bead-beating prior to DNA extraction, or between the in-house end-point PCR and qPCR. The qPCR Ct value was 36 (35-37 95%CI) for 1 oocyst, suggesting a high limit of detection. Discussion In surveillance and outbreak situations we found diagnostic serology testing substantially underdiagnoses Cryptosporidium and Giardia coinfections in Campylobacter patients. These findings suggest that the impact of protozoa infections may be underestimated, through underdiagnosis, but molecular techniques likely improve detection capabilities. Laboratories need to understand clinical, rather than analytical, test sensitivity, to allow clinicians to better understand the disease aetiologies of patients that enable better health advice.

. CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 7, 2023. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 7, 2023. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 7, 2023. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 7, 2023. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 7, 2023. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 7, 2023. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 7, 2023. ; https://doi.org/10.1101/2023.01.05.23284190 doi: medRxiv preprint would be more sensitive than antigen-based assays and a subset (n=19) of the above dilutions were 160 sent to Medlab Central for processing using their new diagnostic Tandem is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 7, 2023.  (Fig 2A) which indicated the Giardia were from 191 assemblages A and B (Fig 2A).   Fig 3A) and . CC-BY 4.0 International license It is made available under a perpetuity.

206
To further confirm the results, we randomly chose 10 positive samples and confirmed the presence Campylobacter data would make cryptosporidiosis the second most notified infectious disease in 213 New Zealand (Table S1). Contrary to our hypothesis we found no significant difference between our ability to detect 222 Cryptosporidium DNA whether we used bead-beating or not ( 2 = 0.75, p-value = 0.39, Fig 4A)  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 7, 2023. ; https://doi. org/10.1101org/10. /2023   is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 7, 2023.  (27). However, we showed in our studies that antigen-based tests were 257 missing samples with levels of oocyst presence easily detectable by microscopy (Fig 3C &3D).

258
Our hypothesis that bead-beating impacted the test sensitivity and limit of detection was not 259 supported. We perform bead-beating to break the (oo)cyst walls, however different Cryptosporidium is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 7, 2023. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 7, 2023. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 7, 2023.

335
Therefore, it will be some time before the impact of changes to the laboratory testing can be seen in 336 New Zealand.

337
Asymptomatic Giardia infection is generally not treated, but treatment may be rational if 338 asymptomatic cases lead to onward transmission. The wider availability of sensitive PCR diagnostic 339 tests may allow a more targeted approach to contact treatment in future. In secondary care when 340 giardiasis is highly suspected but stool results are negative, diagnosis can be made through duodenal 341 aspiration and biopsy, which have been shown to detect infection in the absence of cysts on stool 342 microscopy (43,46,47); improvements in test sensitivity may reduce this need. The symptoms of 343 giardiasis can resemble IBS (11), a condition affecting 1 in 7 people in New Zealand, and IBS or 344 dyspepsia patients may have Giardia (e.g. 6.5% of patients (48), but see (46)). The UK National

345
Institute for Health and Care Excellence recommends that faecal testing for parasites is not routine 346 when confirming IBS in people who meet the diagnostic criteria (43), but that clinicians should be 347 aware of both diagnoses. Improved diagnostics will help clinicians in New Zealand determine when a 348 parasitological examination of a stool sample is required (43).

349
Advice on protozoa for patients and carers will be incorrect with false-negative tests, so our work 350 will help the Ministry of Health, clinicians and community workers reinforce simple health advice, . CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 7, 2023. ; https://doi.org/10.1101/2023.01.05.23284190 doi: medRxiv preprint quantitative molecular diagnostic methods to investigate the effect of enteropathogen infections on 445 linear growth in children in low-resource settings: longitudinal analysis of results from the MAL-ED 446 cohort study. Lancet Glob Health. 2018;6(12):e1319-e28.
. CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 7, 2023. ; https://doi.org/10.1101/2023.01.05.23284190 doi: medRxiv preprint