Background

Giardia is a leading but neglected cause of infectious gastroenteritis worldwide [1] and is treatable. The flagellated protozoan, Giardia lamblia (syn. G. duodenalis and G. intestinalis) comprises eight genetic “assemblages” (A-H) with only A and B affecting humans. Assemblages A and B can also infect pets, livestock and wild animals showing the potential for zoonotic transmission [2]. The reported prevalence of Giardia in human populations is 4–43% and 1–7% in low and high income countries respectively [3, 4]. Prevalence in the UK has been reported as 1.3% in asymptomatic children [5], 1.4% in a general practitioner population and 0.9% in the general population [6]. Between 3000 and 4000 cases are reported annually in England and Wales [7]. There is a substantial burden of undetected Giardia in the UK and for every one case of Giardia reported to national surveillance there are 14 cases in the community [8]. Corresponding ratios for Campylobacter, Cryptosporidium and Salmonella are 9.3, 8.2 and 4.7 respectively. The incidence of Giardia in Northwest England increased four-fold following the introduction of the enzyme linked-immunosorbent assay for the detection of parasite antigens in stools [9]. This test has greater sensitivity than microscopy [10] and was applied universally to all stool samples, i.e. no testing criteria were applied, indicating that the majority of cases did not have the commonly accepted exposures for Giardia cases.

Giardia transmission occurs through the ingestion of the infective cyst stage shed in human or animal faeces. The cyst may be present in faecally contaminated water, food or fomites. The clinical disease (giardiasis) typically includes diarrhoea, flatulence, abdominal pain and bloating [11] and weight loss due to malabsorption [12]. Some infections are relapsing due to re-infection from an ongoing source, possibly an asymptomatic household member, or because they are refractory to metronidazole therapy [13]. The clinical picture is altered in high prevalence countries due to partial immunity from repeated exposure. Chronic infection in children may cause failure to thrive [14].

Estimates of the proportion of Giardia infections that are asymptomatic, but which have the potential for transmission of Giardia cysts, varies considerably from 5 to 15% [15] to 76% [16]. Without treatment infectivity can continue for months, potentially causing household clusters or outbreaks. There is no evidence based guidance on when to treat asymptomatic infection and when it may curtail transmission, and asymptomatic carriage is generally not treated [17, 18]. However expert opinion suggests treatment of asymptomatic Giardia infection in certain circumstances, for example food handlers, day care nurseries and recurrent infection in a household [19, 20].

The aim of this study was to estimate the prevalence of Giardia infection in households of index cases of giardiasis and to identify characteristics of households with more than one case of giardiasis. The study investigated infected rather than diseased cases (defined as symptomatic and therefore more likely to be reported) to provide a more comprehensive insight into transmission dynamics and risk factors within Northwest England households.

Methods

Prevalence survey

An observational study was conducted to estimate the prevalence of Giardia infection in households of confirmed Giardia cases in nine local authorities in Lancashire between June 2014 and June 2015. Giardia cases were identified by the detection of Giardia antigen in a stool specimen using a faecal antigen enzyme immunoassay (EIA) method as previously described [9, 10], by the three participating hospital laboratories in Lancaster, Blackburn and Preston. All positive specimens were further characterised by assemblage, details of the methods are described [21].

Households were invited to participate by an environmental health officer during their routine public health investigation. All household members were asked to provide a stool sample. A household member was defined pragmatically as a person who lived in the same residence as the case for at least two nights per week in the month prior to diagnosis or had household equivalent contact for example in a care home or university residence. All stool samples were tested using the same EIA method as the index cases. Index cases were excluded if they lived in a single person household.

The prevalence of additional Giardia infection was measured by dividing the number of households that had at least one case of Giardia infection in addition to an index case by the total number of households included in the study. The Giardia prevalence amongst household contacts was also estimated by dividing the total number of additional infections detected by the total number of household contacts.

The sample size was based on an estimated additional Giardia household prevalence of 10% (local surveillance data showed that a second symptomatic case of Giardia was reported in 7% of households plus an additional 3% estimated for asymptomatic individuals). The significance level used was 0.05 (corresponding to 95% confidence intervals) with a confidence width of 0.05 with an estimated a response rate of 80% the sample size was 130 households.

Cross-sectional analysis

We compared households with and without secondary infection to determine the characteristics of households with secondary infections. A “case” household was defined as a household with at least one additional Giardia infection in a household member. A “control” household had no additional Giardia infections. We compared household characteristics between “case” and “control” households using univariable analyses to calculate odds ratios (OR) and p values using Fisher’s exact test. The characteristics were identified from the index case study questionnaire and additional assemblage information. We used multivariable logistic regression to identify independent risk factors for additional Giardia infections. All risk factors that had a p value less than 0.2 were considered in a multivariable analysis. The final model included risk factors that were significantly associated with additional Giardia infections i.e. p value <0.05.

Questionnaire information was entered using EpiData and data was analysed using Stata v12 (StatCorps). The date of onset for cases was estimated from the in-house Public Health England case management system if not available from the questionnaire. The dataset was checked for accuracy using Stata to clean the data and check for impossible values and missing data items were sought via environmental health officers or the in-house case management system.

Results

Prevalence survey

The number of index Giardia cases identified was 186. Of these 17 were excluded as a single person household and 26 could not be contacted, resulting in 143 eligible index cases (Fig. 1). Of these, 91 households participated giving a response rate of 64%. There were no significant differences between households that participated and those that didn’t for household size, age or sex of the index case. Within the 91 study households there were 237 household members of whom 212 (89%) provided a stool sample (Fig. 1).

Fig. 1
figure 1

Numbers of participants included in the study including details of positivity and further typing

Full size image

At least one additional case of Giardia infection was detected in 27 households, giving a household prevalence of 30% (27/91) (95% confidence intervals (CI): 20–39%). Giardia was detected in 41/212 (19%) of household members of whom 37 (90%) were asymptomatic, giving a prevalence of asymptomatic infection of 17% (37/212) (95% CI: 12–23%) and a prevalence of undetected symptomatic infection of 2% (4/212) (95% CI: 0.06–3.7%). The age groups most affected by asymptomatic infection were 0–4 and 5–9, with 51% and 35% of household members being affected, respectively (Table 1). The number of additional Giardia positive individuals in a single household ranged from one to five.

Table 1 Number of household members by asymptomatic infection status and percentage of asymptomatic infection in household members
Full size table

Cross-sectional analysis

The univariable analyses for risk factors when comparing households with additional Giardia infection against those without additional Giardia infection can be seen in Table 2.

Table 2 Univariable analyses comparing households with additional Giardia infection to those without
Full size table

The highest association with additional household Giardia infection was with having a child under 5 years old in the household (OR 29; 95% CI 8–114). There were other risk factors linked to having children; children in nappies in the household (OR 4; 95% CI 1–18), being involved in changing nappies in the household (OR 3; 95% CI 0.8–16), being an index case under 5 years old (OR 3; 95% CI 0.7–11) and children attending nursery in the household (OR 2; 95% CI 0.6–10). Other characteristics associated with additional infection were related to the number of people and ratio to facilities; having less than one bedroom per person in the household (OR 4; 95% CI 1–12), having less than one toilet per person in the household (OR 3; 95% CI 1–14) and having four or more people in the household (OR 3; 95%CI 1–8). No association was found between having additional Giardia infection in the household and owning a cat (OR 0.3; 95% CI 0.5–1) and/or dog (OR 0.6: 95% CI 0.4–2). Of the four index cases with previous Giardia infection (ranging from 3 to 26 months prior to the current infection), none had any additional infection in the household.

In the multivariable analysis two risk factors remained significantly associated with additional Giardia infection in the household; having children under 5 years in the household (OR 42.35; 95% CI 10.09–177.69) and having anyone with gastrointestinal symptoms in the household in the 3 weeks before the index case (OR 8.55; 95% CI 1.51–48.28).

Genetic analysis was able to assign an assemblage to 108/132 (82%) specimens. Of these 62 (57%) were assemblage B, 24 (20%) assemblage A and 22 (20%) were mixed A and B. Three households did not have assemblage typing completed. If the index case in the household had assemblage A, the household was significantly less likely to have any additional infection (OR 0.1; 95% CI 0–0.8). The assemblage of the index case and the assemblage of the household members were concordant in 92% of households (22/24). This included both single assemblage concordance i.e. A and A or B and B (54%; 13/24) and where there was mixed infection and single assemblages present e.g. A/B and A (38%; 9/24).

Discussion

To our knowledge this is the first study of Giardia prevalence amongst household contacts of sporadic Giardia cases. During an outbreak of giardiasis in 1977 no additional cases of Giardia were found in 23 household contacts of eight index cases [15]. In contrast we found a high prevalence of asymptomatic Giardia infection with an additional Giardia infection detected in 30% of households and 17% of all household contacts were found to have asymptomatic infection. This high prevalence may fuel transmission in the household and in the community to a greater extent than is currently recognised. Only four additional symptomatic infections were identified however this was a point prevalence and the true rate of symptomatic infection amongst household contacts will be higher and requires evaluation by a longitudinal study of Giardia infected households.

Limitations of this study included the small sample size and the single geographical study area. Further studies would determine whether the findings were generalisable more widely in a low Giardia prevalence setting taking account of Giardia diagnostic methods and environmental and socioeconomic factors, for example household size and type of housing. Although a small study, the sample size was sufficient to provide an estimate of the true prevalence of additional asymptomatic infections. The response rate of 64% was lower than anticipated although it seems unlikely that the Giardia prevalence in non-responding households would be substantially different to responding households due to the similarity in demographics and risk factors. A potential limitation of the study - poor compliance with stool sampling by asymptomatic household members – did not occur, with a compliance rate of 89% adding to the strength of the findings.

We found a strong association between the presence of additional Giardia cases and having children aged under 5 years in the household. The transmission of Giardia has been reported frequently in day care or nursery settings previously [5, 22] and has been shown to be associated with changing nappies [23]. This result adds to the body of evidence that close contact with children, even those without symptoms, can play an important role in the transmission of Giardia.

Developing a ‘high risk household’ definition may be useful for the communication of risk and advising households on ways of reducing this risk. Household factors such as a greater concentration of individuals and fewer facilities also increased the risk of transmission. This may be related to a higher risk of transmission from a contaminated toilet area in households with proportionately fewer toilet facilities per household member. A larger study powered to investigate this association is required to confirm this finding.

Assemblage typing found that the majority of infections were assemblage B, which has previously been associated with person to person transmission [21, 22]. There were a high percentage of mixed assemblage infections (20%). The concordance rate of index and secondary household cases has not been previously reported. Two households had discordant assemblages, possibly due to differing sources or a missed mixed infection in the household.

Conclusions

Our finding of a high household prevalence of asymptomatic infection has raised the public health question of whether treatment of asymptomatic household contacts may be justified in preventing Giardia re-infection of the index case or in preventing household clusters. This is particularly pertinent in households containing children under 5 years in which 50% of household members under 5 years had asymptomatic Giardia infection. Currently, asymptomatic carriage is generally not treated due to lack of evidence, but treatment seems rational in failed treatment of a case or where there is a household cluster. Evidence of the effectiveness of treating asymptomatic infection in curtailing transmission could lead to the offer of routine testing of household contacts or a pragmatic alternative of offering blind treatment to all household contacts. Wider availability of sensitive PCR diagnostic tests may aid a more targeted approach to contact treatment in the future. As this was a prevalence study it cannot provide evidence on the impact of treating asymptomatic infection. A longitudinal study on a larger population is required before recommending any change in current practice.