Diagnosis of protozoa diarrhoea inCampylobacterpatients increases markedly with molecular techniques

AbstractBackgroundCryptosporidiumandGiardiaare 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 However, these methods are increasingly being superseded by molecular techniques for diagnostics.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 inCampylobacterpositive samples missed through use of antigen-based assays and then investigated different molecular testing protocols.MethodsWe report the findings of two observational studies; the first among 111 people with diarrhoea during a largeCampylobacteroutbreak in Havelock North, and the second a study during normal surveillance activities among 158 people presenting with diarrhoea and a positiveCampylobactertest, but negativeCryptosporidiumand/orGiardiaantigen-based diagnostic test result. The molecular methods used for comparison with the antigen-based tests were in-house end-point PCR tests targeting thegp60gene forCryptosporidiumandgdhgene forGiardia. 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 10−5forCryptosporidiumpositive samples.ResultsThe coinfection prevalence was 9% (n= 10, 3–15% 95%CI) forCryptosporidiumand 21% (n=23, 12– 29% 95%CI) forGiardiain the 111Campylobacterpatients of the Havelock North outbreak. The coinfection prevalence was 40% (n=62, 32-48% 95%CI) forCryptosporidiumand 1.3% (n=2, 0.2-4.5% 95%CI) forGiardiain the 158 routine surveillance samples. Sequencing identifiedCryptosporidium hominis, C. parvum, andGiardia intestinalisassemblages 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.DiscussionIn surveillance and outbreak situations we found diagnostic serology testing substantially underdiagnosesCryptosporidiumandGiardiacoinfections inCampylobacterpatients. 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.