The transformation of a Cryptosporidium reference microbiology service to tackle the One Health challenge

Abstract

Cryptosporidium has a wide range of hosts and routes of transmission, so public health investigations require Cryptosporidium species identification and discriminatory typing to enhance the microbiological evidence provided by routine diagnostic tests. A pioneering seven-locus genotyping scheme, based on multilocus variable number of tandem repeats analysis, has been validated and implemented for Cryptosporidium parvum by the national Cryptosporidium Reference Unit for England and Wales. In this paper, the journey to implementation as a service to support disease surveillance, epidemiology and outbreak investigations, and the resultant benefits to public health activities are described.

Highlights

• •Multilocus variable number of tandem repeats analysis integrated into Cryptosporidium genotyping workflows.
• •MLVA profiles inform the epidemiological surveillance of Cryptosporidium parvum.
• •In outbreak investigations, MLVA can strengthen links between animals and humans.
• •Genetic clustering by MLVA can lead to outbreak recognition.
• •Automated MLVA cluster analysis is used to guide public health action.

1. Introduction

The protozoan parasite Cryptosporidium is an important cause of gastrointestinal illness (cryptosporidiosis) in humans and livestock globally. Transmission by the faecal-oral route is facilitated by robust oocysts shed in faeces that can survive in the environment and many commonly-used disinfectants (King and Monis, 2007). Thus, food and water are important vehicles of transmission, in addition to fomites and direct faecal-oral ingestion, from animals and humans. Oocysts, shed in stool, also provide the main diagnostic target, for which stained microscopy, enzyme immunoassays, immunochromatographic assays and commercial nucleic acid amplification assays provide presence/absence of the parasite at the genus-level only. Currently, 48 species are recognised, a number that is growing each year, and although some have a broad host-range, for example Cryptosporidium parvum, others such as Cryptosporidium hominis are host-adapted (Ryan et al., 2021). Their identification and further typing can assist in determining sources of infection or contamination, understanding transmission routes and providing further insight on surveillance and epidemiological data. Species identification requires molecular analysis, referred to as genotyping, usually provided by specialist or reference laboratories where subtyping may also be undertaken. However, provision is variable, and these procedures are not widely available globally (Chalmers et al., 2018).

1.1. Provision of Cryptosporidium typing and surveillance

In Europe, human cryptosporidiosis data are reported to The European Surveillance System (TESSy) run by the European Centre for Disease Control (ECDC). The most recently available data are for 2021 when 24 European Union (EU)/European Economic Area countries reported case numbers, seven of which also reported information on the Cryptosporidium species identified (European Centre for Disease Prevention and Control, 2024). In the United Kingdom (UK), where reporting to ECDC ceased in 2020 following withdrawal from the EU, the reporting of Cryptosporidium detected in human clinical diagnostic samples to the statutory authority has been mandatory under health protection regulations since 2010 and forms the basis of national surveillance and public health action.

In England and Wales, all clinical diagnostic laboratories are requested to refer Cryptosporidium positive stools to the national Cryptosporidium Reference Unit (CRU), managed by Public Health Wales (PHW), for genotyping by real-time PCR (Robinson et al., 2025a), a service that is free to users. The results are reported back to the submitting laboratory and are captured from the CRU Laboratory Information Management System (LIMS) by the PHW infectious disease case data system, Tarian (Lingard and Davies, 2017) or extracted for analysis. The Cryptosporidium genotyping data from laboratories in England are integrated from the CRU LIMS into the UK Health Security Agency's Second Generation Surveillance System (SGSS), a database that stores and manages laboratory results for regional and national surveillance of all infectious diseases (UKHSA, 2022). These data are used for monitoring the numbers and epidemiology of Cryptosporidium spp. in England and Wales. The most recently available data for England show that the case numbers (and incidence per 100,000 population) for 2022, 2023 and 2024 were 3739 (6.5), 6837 (11.9) and 5702 (9.9), respectively (UKHSA, 2025a). In Wales the case numbers (and incidence) were 368 (11.6), 657 (20.7) and 635 (20.1), respectively during those years (PHW CDSC data). In 2018, most frontline diagnostic microbiology laboratories in Wales implemented PCR-based kits (EntericBio® Gastro Panel 2, Serosep Ltd.) for the diagnosis of gastro-intestinal pathogens that increase the detection rate of Cryptosporidium spp. relative to more traditional methods (Holliday and Perry, 2022). The increase in England and Wales in 2023 was driven mainly by an increase in C. hominis that was also documented in Europe (Schoeps et al., 2024; Williams et al., 2025). Large outbreaks linked to animal contact events have also been documented recently and contributed to the numbers of cases (Peake et al., 2024; Jones et al., 2025). In 2022 and 2023, totals of 2698/4107 (66 %) and 5504/7494 (73 %) of Cryptosporidium-positive stools, respectively, were referred and successfully genotyped at the CRU, informing national surveillance and exceedance / outbreak investigations (Public Health Wales Microbiology, 2023; Public Health Wales Microbiology, 2025; Peake et al., 2023; Williams et al., 2025).

1.2. Investigating links

From an animal-health perspective, C. parvum is a major cause of diarrhoeal disease in neonatal livestock with economic impacts for farmers (Innes et al., 2020). In the UK, Cryptosporidium in neonatal diarrhoea is diagnosed by private veterinary surgeons or laboratories using stained microscopy, enzyme immunoassays, immunochromatographic assays and commercial nucleic acid amplification kits, or by the Animal and Plant Health Agency (APHA) using modified Ziehl-Neelsen stained microscopy. In a One Health approach, animal faeces from healthy or sick animals may be sampled during the investigation of human outbreaks linked to animal contact. These are tested by the APHA using immunofluorescence microscopy and Cryptosporidium-positive samples are referred to the CRU for genotyping and subtyping, which has provided useful linkage information in outbreaks (Smith et al., 2021).

Further subtyping for outbreak-related specimens is undertaken at the CRU by sequencing part of the gp60 gene (Chalmers et al., 2019, Robinson et al., 2025b) and latterly for C. parvum by multilocus variable number of tandem repeats (VNTR) analysis (MLVA) (Risby et al., 2025). The need for a standardised multilocus genotyping (MLG) scheme prioritising C. parvum had been agreed at an international expert workshop in 2016, funded by EU COST Action FA1408 “A European Network for Foodborne Parasites: Euro-FBP” (http://www.euro-fbp.org) (Chalmers and Caccio). Clinical scientists and microbiologists, animal scientists and veterinarians, and epidemiologists participated in the meeting. Investigations at the CRU under the Aquavalens project (European Union Seventh Framework Programme) led to the discovery of new VNTR loci, some of which were identified as suitable for a One Health-focussed MLG scheme (Pérez-Cordón et al., 2016; Chalmers et al., 2018) and were included in the subsequent evaluation and validation of seven selected loci (Robinson et al., 2022). The work came to fruition in 2021 when the CRU began reporting results derived from the MLVA scheme in outbreaks, the first being linked to the consumption of milk from a producer-processor farm gate vending machine (Gopfert et al., 2022). That investigation of human cases of illness required a One Health approach from the outset.

A further development has been the application of MLVA to all C. parvum specimens from laboratories in Wales and the North West region of England. This was funded initially as a 4-month pilot study in 2022 by the National Institutes of Health Research Health Protection Research Unit for Gastrointestinal Infections to evaluate the use of MLVA for case-cluster identification for public health action. The demonstrable value in identifying outbreaks and strengthening microbiological evidence has led to the continuation of this work to provide longer term, year-round understanding and public health benefits (Risby et al., 2023).

Here, we describe the implementation of the MLVA scheme as a clinical service, the development and automation of case cluster detection, and action for public health investigations. We provide an overview of plans to maximize the utility and provide the public health benefits of MLVA for the investigation of all C. parvum cases in England and Wales. The detailed laboratory methods for MLVA can be found in Risby et al., 2025 and in the updated protocol on our website phw.nhs.wales/services-and-teams/cryptosporidium-reference-unit/mlva-assay-protocol/.

2. Ensuring an efficient workflow for the identification and reporting of MLVA profiles

Implementation of MLVA has required little change to the upstream laboratory processes at the CRU as suitable C. parvum DNA is already extracted from referred stools for identification of Cryptosporidium species by real-time PCR (Fig. 1). DNA from C. parvum samples is usually tested weekly by MLVA; up to 44 samples per run, outbreaks may necessitate additional test runs. Two multiplex PCRs (a three-plex and a four-plex) with bespoke fluorophores are used to amplify the seven loci and the PCR products are sized by capillary electrophoresis using a SeqStudio Genetic Analyzer (Thermo Fisher Scientific). The number of tandem repeats at each locus are inferred from the fragment sizes using BioNumerics software (bioMérieux), that allocates the capillary electropherogram peaks to bins determined from sequenced reference standards.

Fig. 1.

Overview of the laboratory workflow for genotyping and subtyping at the Cryptosporidium Reference Unit.

A DNA desiccator (Concentrator plus, Eppendorf) is used to concentrate selected DNA samples and improve typability. Typability is defined here as the proportion of samples tested resulting in alleles at all seven genetic loci; this includes specimens with a null allele at locus 1 (cgd1) as sequence mis-matches have been identified in the PCR primer regions (Risby et al., 2023). Initially, desiccation was done when samples produced only a partial MLVA profile, which were subsequently repeated (Risby et al., 2023). However, analysis of 901 specimens from the first six months of 2024 indicated that the cycle threshold (Ct) value from the PCR that identified C. parvum (Robinson et al., 2025a) was predictive of the need for concentration (Fig. 2). For that dataset, initial typability was 698/901 (78 %) but following concentration of the DNA typability increased to 822/901 (91 %). Concentration is now done prospectively on all DNA samples where the Ct value is ≥34, approximately 25 % of specimens. Overall typability for all samples tested by MLVA between March 2022 and May 2025 is 94 %. This prediction avoids repeating tests and has a beneficial effect on workflow, costs and turnaround time (TAT).

Fig. 2.

Comparison of C. parvum PCR Ct value and MLVA typability categorised as full or partial MLVA profiles in 901 unconcentrated DNA samples. Samples that are not fully typable by MLVA are candidates for concentration and increased in number above Ct 34. Note: full typability is defined as alleles identified at loci 2 to 7; locus 1 may have a null return due to primer sequence mismatches.

Samples with partial MLVA profiles are concentrated and repeated by MLVA PCR. By concentrating any sample with a Ct value 34 or higher, most samples that would be likely to demonstrate a partial profile do not need repeating, thereby saving time and resources.

The PCR products are subject to capillary electrophoresis, and peaks indicating amplicon size are identified in the electropherogram, determined by relative fluorescence units (RFU) above background noise; although ideally ≥150 RFU, we have found true peaks at ≥50 RFU. True peaks are differentiated from: stutter by proximity to a true peak size; artefacts that give omni-present, wide and low peaks with an apparent size unique to each dye; and bleed-through from intense peaks. Amplicon sizes are determined by allocation of each peak to a pre-determined bin and the number of tandem repeats inferred using BioNumerics software (BioNumerics 7.6, Applied Maths). MLVA profiles are allocated to each sample by expressing the VNTR number for each locus separated by a hyphen in the order they appear in the genome (Fig. 3). When a new fragment size is identified, the allele is confirmed by re-amplifying the individual locus and Sanger sequencing the fragment. The new allele bin is then added to the BioNumerics database.

Fig. 3.

The expression and reporting of MLVA profiles from the seven-locus C. parvum MLVA scheme.

The target TAT for MLVA results is five to ten working days from specimen receipt, but reports can be issued in as few as two working days. Since 2021, MLVA profiles in outbreak investigations have been reported directly to the investigating Health Protection Team by secure email and from January 2023 routinely entered in LIMS. Other outputs that may be of value to help communicate the context of MLVA results include minimum spanning trees, which can provide additional insight into the clustering of samples by providing a visual representation. This can help put outbreak samples into context of other cases or assist hypothesis generation by colour coding samples by demographic or epidemiological variables (Risby et al., 2023; Morris et al., submitted).

3. Quality controlling the MLVA process in the laboratory

The CRU is an accredited medical laboratory by the United Kingdom Accreditation Service (UKAS) to ISO 15189:2022 and the scope of accreditation was extended to include MLVA in 2023. A structured procedure ensures quality control, including: a standard operating procedure and batch record templates published as controlled documents; documented staff training and competency assessment; internal quality control, provided by the inclusion of a verified C. parvum DNA sample and a molecular grade water sample in every test run; internal quality assessment, provided by re-testing of ∼1 % of the annual number of samples tested; internal fragment size standards, added to every DNA sample; a library of sequenced reference standards, that calibrate binning software to infer number of repeats from fragment sizes; separate sequencing primers, that enable confirmation of the number of repeats from new fragment sizes encountered during testing; scripts, to ensure standardised data analysis processes.

In the absence of a formal External Quality Assessment (EQA) scheme for Cryptosporidium genotyping or subtyping, we have established an informal scheme acceptable for accreditation by UKAS. This includes species determination, gp60 subtyping, and MLVA, which has been included since 2023. We can participate in the scheme as the staff that administer the scheme, distribute the blind-coded DNA samples and manage the results are different from those that undertake the testing. Of the current eight participating laboratories from six nations in Europe (England, Scotland, Wales, The Netherlands, France and Sweden), three subscribe to the MLVA element. Assessment of the results has helped establish where the pitfalls lie for users, especially in interpretation of bins and reporting of profiles, and led to improvement in instructions for these processes that are captured in the updated assay protocol on our website https://phw.nhs.wales/services-and-teams/cryptosporidium-reference-unit/mlva-assay-protocol/

4. Automated and integrated analysis of MLVA data for public health interventions has been pioneered in Wales

Case specimens that have identical MLVA profiles are considered likely to have been exposed to a common source of contamination or infection, thus signifying potential outbreaks. This is based on validation studies and the pilot study in 2022, that indicated that cases with identical MLVA profiles may be epidemiologically linked and share a common exposure or source of infection (Robinson et al., 2022; Risby et al., 2023).

In Wales, a process has been developed to automate the assessment of MLVA profiles and clusters for public health purposes, involving integration with epidemiological data and application of cross-disciplinary working encompassing microbiological, bioinformatics and epidemiological expertise. This is enabled by the extraction of MLVA profiles from LIMS and automation of cluster identification, reporting and management through integrated analysis by CDSC. Their capture has been enabled by a script in R with an Excel output providing weekly analysis to determine if the MLVA profiles fall into new or pre-existing clusters (Fig. 4). A MLVA cluster is defined as two or more cases with matching simple or complex MLVA profiles as described in Fig. 4. The demographics, size, timeframe and growth rate of the clusters are used to gain an understanding of the basic person, time and place of the cluster. In Wales exposure information, collected through the administration of a questionnaire, is available for approximately 85 % of cases of Cryptosporidium who are routinely followed-up by Environmental Health Officers from the relevant Local Authority. Cross checking the cluster cases with this exposure information and case notes identifies whether there are any potential transmission events or common exposures between the cases in each MLVA cluster and helps determine escalation for public health action (Fig. 5). This is a dynamic process as new information may become available and previous decisions are reviewed.

Fig. 4.

Overview of the weekly script for MLVA cluster identification in Wales.

Fig. 5.

Escalation of MLVA clusters for public health action in Wales.

5. Public health impact of MLVA in England and Wales

In the pilot study, one significant outbreak was newly identified by MLVA profiling and the microbiological evidence in other outbreaks that were already under investigation was strengthened by linking cases (Risby et al., 2023).

In the first outbreak investigated in real-time by MLVA, which was prior to the pilot study, test results from suspected sources strengthened the microbiological evidence when calves from a dairy herd had the same MLVA profile as cases that drank milk from an on-farm processor-retailer vending machine (Gopfert et al., 2022).

In 2023 in England, there were 21 Cryptosporidium outbreaks reported to the UKHSA of which eight were C. parvum (UKHSA, 2025b). Of these, six were linked to animal contact and two to swimming pools. All were investigated by MLVA; two were identified de novo by MLVA clustering of cases from the North West region of England. In Wales, there were four Cryptosporidium outbreaks reported to PHW; three were C. parvum and all were investigated by MLVA. Two were linked to animal contact and one to consumption of milk/milkshakes from an on-farm processor-retailer vending machine. All were investigated by MLVA and once again the same MLVA profile was found in the dairy herd as the cases of illness.

Since the introduction of the automated MLVA clustering process in Wales in autumn 2024 to the middle of May 2025, 51 unique MLVA clusters have been assessed by the CDSC, Wales. Of these, nine met the threshold for escalation to the relevant Health Protection Team. These alerts prompted further investigations including the reinterviewing of cases by Environmental Health Officers. During this time, one large outbreak was identified; the clustering process proved vital at the early stages of the investigation in confirming that cases were likely linked by a common exposure.

6. Steps towards full service provision for England and Wales

The intention now is to continue the application of MLVA for all C. parvum cases from Wales and the North West region of England and extend provision in a step-wise fashion to the other regions of England. This needs to be supported by cluster identification, reporting and management through integrated analysis as has been demonstrated in Wales. Currently, the MLVA clusters identified as an extension of the pilot study in the North West are assessed manually in the CRU and relayed to the UKHSA's regional field epidemiology service. They have their own processes for alerting Health Protection Teams locally. Automation of the process for England is dependent on integrating MLVA in SGSS which is a work in progress. Alternatively, a data sharing platform hosting case and cluster information could be considered.

7. Further developments

From a laboratory point of view, we have streamlined the MLVA process but the bioinformatic element of the scheme will ultimately require an alternative to BioNumerics, as this software has ceased to be provided or supported by bioMerieux. While there are a few options that can provide some of the features and utility of the BioNumerics MLVA software, we are yet to identify an alternative that encompasses all.

From an epidemiological perspective, the MLVA scheme relies on specimen referral and on the collection of epidemiological data. In some areas, questionnaire response rates with useable data can be low (Chandra et al., in preparation). Both the pilot study and its continuation, and the automated clustering in Wales, has highlighted the importance of case follow-up, but also how the extant questionnaire does not currently sufficiently target exposures of relevance to Cryptosporidium. Accordingly, the questionnaire used in Wales is being revised. It is hoped that improvements in epidemiological data quality will facilitate the assessment process of cases linked by the MLVA clustering process and lead to meaningful health protection action.

The anonymised sharing of microbiological findings is important to a One Health approach. We intend to make C. parvum MLVA profiles and (where relevant during validation of new alleles) DNA sequences publicly available through the development of a Cryptosporidium organism database in PubMLST, an online platform that hosts public databases for molecular typing and microbial genome diversity (www.pubMLST.org). The purpose of putting MLVA profiles in PubMLST is to share those defined MLVA profiles with public and animal health professionals and allow for data comparisons of trends and between countries. It is not intended as a comparison of individuals or for management / investigation of outbreaks. The database will include metadata including the month, year, country, isolation source (e.g., human, animal, food), gp60 subtype and MLVA profile. This resource will become increasingly important as more countries adopt the MLVA scheme, which has been successfully used to investigate outbreaks in Finland and Sweden (Suominen et al., 2025) and in France (Loic Favennec, personal communication). Through sharing of MLVA profiles causing clusters in France, cases with the same MLVA profiles in England have been epidemiologically linked through questionnaire data. This demonstrates the utility of cross-border communication and sharing of information about MLVA profiles.

The MLVA loci are reviewed for suitability and utility, but this is an ongoing process. Analysis of the four-month pilot study data indicated that two loci (cgd1_470_1429 and cgd6_4290_9811) might be candidates for review (Risby et al., 2023), but data gathered over a longer period indicates that seasonality in MLVA profiles that merits reconsideration (CRU unpublished data). Such reviews will necessitate international input, for which a user group will be established that also links in with the PubMLST initiative.

In an era of whole genome sequencing that is now used for many bacterial gastrointestinal pathogens for surveillance and outbreak detection in real-time, the question might reasonably be asked, why study just seven loci when the whole genome could be analysed? For Cryptosporidium, challenges remain in the preparation of parasite DNA without lengthy purification and concentration of oocysts from stools (Wang et al., 2024). New approaches to enrichment by bait capture are promising (Bayona Vasquez et al., 2024; Khan et al., 2024), but a cost-benefit analysis may indicate a combined approach with MLVA for the greater number of samples and WGS for resolving as necessary. It may well be that for C. hominis these new approaches may be more applicable, as the current MLVA scheme was not developed, and is sub-optimal for this species. However, at this stage, pursuing MLVA for the fine discrimination of C. parvum is a valuable One Health approach to public health microbiology and investigations, with tangible benefits for both the identification and investigation of outbreaks.

8. Conclusion

The MLVA scheme has improved differentiation between C. parvum subtypes, enabling improved outbreak investigations and the ability to detect previously unrecognised outbreaks, as well as providing a greater understanding of C. parvum epidemiology. It has been established as a clinical service at the national reference unit and supports outbreak investigations in England and Wales. It provides additional information to supplement epidemiological investigations and has been integrated into routine surveillance in Wales. Thus, a model has been provided for application more widely, to define clusters of cases and alert to emerging outbreak situations.

CRediT authorship contribution statement

Rachel Chalmers: Writing – review & editing, Writing – original draft, Visualization, Supervision, Resources, Project administration, Methodology, Funding acquisition, Formal analysis, Data curation, Conceptualization. Guy Robinson: Writing – review & editing, Validation, Supervision, Methodology, Investigation, Formal analysis, Data curation. Harriet Risby: Writing – original draft, Visualization, Methodology, Investigation, Formal analysis, Data curation. Kristin Elwin: Writing – review & editing, Investigation, Data curation. Rebecca Howarth: Writing – review & editing, Software, Methodology, Formal analysis, Data curation. Felicity Simkin: Writing – review & editing, Software, Methodology, Formal analysis, Data curation. Andrew Nelson: Writing – review & editing, Supervision, Resources, Methodology, Formal analysis, Data curation.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.