Introduction

Toxocara canis, Toxocara cati and Ascaris suum are worldwide-distributed helminths with zoonotic potential whose final hosts among domestic animals are dogs, cats and pigs, respectively. The adult worms live in the gut of their final hosts and lay eggs that are dispersed into the environment through the hosts’ faeces. Transmission to humans occurs mainly by ingestion of larvated eggs from faecally contaminated soil, water and vegetables that are not properly washed, cooked or peeled. Besides individuals occupationally/recreationally exposed to soil, those exhibiting pica or suffering from mental retardation, as well as those coming into contact with contaminated environments (as it is often the case of children’s playgrounds after indiscriminate defecation by dogs and cats) (Uga and Kataoka 1995) are at high risk of infection (Despommier 2003; Kaplan et al. 2004). People may also acquire Toxocara infection by consuming raw/undercooked meat of potential paratenic hosts like chicken and sheep (Nagakura et al. 1989; Salem and Schantz 1992), and a recent review pointed out that the significance of paratenic hosts as sources of infection for definitive hosts is one of the biggest gaps in our understanding of Toxocara epidemiology (Holland 2015). Although it is theoretically possible that A. suum could be transmitted through consumption of raw/undercooked pork, this has not been proved. Yet, A. suum infection by consumption of raw/undercooked offal from a paratenic host like chicken has been deemed possible based on experiments on pigs (Permin et al. 2000), and the habit of regularly consuming raw bovine or porcine liver was described in some severely ill people infected with A. suum (Izumikawa et al. 2011; Kim et al. 2002).

Since humans are accidental hosts for Toxocara spp. and A. suum, their larvae usually fail to reach the adult stage and undertake aberrant migrations throughout the body. However, it has been reported that A. suum can develop to adult worms in humans (Nejsum et al. 2005), further stimulating the debate of whether A. suum (collected mainly from pigs) and Ascaris lumbricoides (infecting mainly humans) are truly two distinct species (Leles et al. 2012). Although human infection with tissue-dwelling larvae is frequently asymptomatic, they may damage whatever tissue they enter, resulting in a syndrome known as visceral larva migrans (VLM), a serious complication of which is the migration of the larvae in the eye causing ocular larva migrans (OLM) (Smith et al. 2009). Moreover, a positive association between seropositivity to A. suum and wheeze, asthma and food- and aero-allergen sensitization has been found in 4-year-old children in the Netherlands (Pinelli et al. 2009), supporting the hypothesis that low-level or transient infections with such helminths may promote allergic reactions. Several epidemiological and experimental studies reviewed elsewhere (Pinelli and Aranzamendi 2012) suggest that also Toxocara infections contribute to the development of allergic manifestations and exacerbation of airway inflammations.

As detection of Toxocara or Ascaris larvae in biopsies is rare and eggs are hard to find in human faeces in countries like the Netherlands (de Wit et al. 2001a, b), diagnosis of human Toxocara and Ascaris infections relies mainly on serology. We have previously reported that sera of Dutch patients (n = 2838) suspected of VLM/OLM referring to the Netherlands’ National Institute for Public Health and the Environment (RIVM) for routine serodiagnosis of Toxocara and Ascaris infections during 1998–2009 had a seroprevalence (IgG) of 8 % for Toxocara and 33 % for Ascaris (Pinelli et al. 2011). However, these estimates refer to people selected based on clinical grounds. To determine the true magnitude of (the exposure to) these zoonotic helminths and to identify targets for control measures, population-based surveys and risk factor analyses are needed. In the Netherlands, two nationwide cross-sectional serosurveys were conducted, the so-called PIENTER-1 (1995–1996) (De Melker and Conyn-van Spaendonck 1998) and PIENTER-2 (2006–2007) (van der Klis et al. 2009) to establish a large serum bank with accompanying epidemiological information representative of the Dutch population. These serosurveys were primarily aimed not only to perform immunosurveillance to evaluate the Dutch national immunization programme but also to address additional research questions. This provides the unique opportunity to obtain insights into a multitude of infections occurring in the population, including those caused by Toxocara and Ascaris, whose epidemiology in developed countries is still largely unclear. Therefore, the aim of this study was to determine the seroprevalence and factors associated with increased or decreased IgG antibody levels for Toxocara and Ascaris in the general population of a high-income country like the Netherlands, which is characterized by high swine production, freedom from stray dogs and virtual absence of autochthonous A. lumbricoides infections.

Methods

Data collection

Data were collected from October 1995 to December 1996 (PIENTER-1) and from February 2006 to June 2007 (PIENTER-2). The design and rationale of both serosurveys are described in detail elsewhere (De Melker and Conyn-van Spaendonck 1998; van der Klis et al. 2009). In brief, a two-stage cluster sampling design, with municipalities nested in regions, was applied: Age-stratified random sampling (<1, 1–4, 5–9, …, 75–79 years) was performed in 48 municipalities within five study-defined geographical regions of approximately equal population size in the Netherlands. In total, 18,217 (PIENTER-1) and 24,147 (PIENTER-2) individuals were invited to participate. In 12 municipalities, an oversampling of the largest non-Western migrant groups in the Netherlands in those years (i.e. from Morocco, Turkey, Suriname and Netherlands Antilles) was performed in PIENTER-2; 2558 people were invited in this extra sample. Each invited individual received an invitation letter, a brochure introducing the study, a questionnaire, an informed consent form and a prescheduled appointment for serum sample donation. The questionnaire contained questions regarding demographic characteristics, medical history, activities and behaviours putatively related to infectious disease transmission (e.g. foreign travel, occupation, eating habits, etc.). Informed consent was obtained for all participants. Information on socio-economic status (SES) and urbanization degree per postcode area was obtained from Statistics Netherlands (www.cbs.nl).

In total, 9948 (PIENTER-1) and 7904 (PIENTER-2) individuals provided a serum sample. However, for practical and budgetary reasons, a selection of the available sera was tested for IgG antibodies against Toxocara and Ascaris. In PIENTER-1, a total of 1159 sera were tested for both Toxocara and Ascaris. In PIENTER-2, 3683 and 3675 sera were tested for Toxocara and Ascaris, respectively. As mentioned before, the PIENTER studies were meant to evaluate the national immunization program, meaning that the priority was to test the collected sera for a number of vaccine-preventable diseases. Therefore, antibodies for other pathogens like Toxocara and Ascaris could be searched for only in those sera that had sufficient material to be analysed further. This means that mainly serum samples from (young) children were underrepresented for Toxocara and Ascaris testing, as they contained relatively smaller quantities of serum than those from adults. Departures of our sample from the underlying population (due to non-random selection, among others) as regard to the variables age, gender, ethnicity and degree of urbanization were accounted for in the analysis using sampling weights (see “Data analysis” section).

Serological analysis

Anti-Toxocara and anti-Ascaris IgG antibodies in the collected sera were detected using an enzyme-linked immunosorbent assay (ELISA) and the excretory/secretory (E/S) antigen derived from cultivated T. canis and A. suum larvae as previously described (Pinelli et al. 2009, 2011). Medium binding ELISA microtiter plates (Nunc, Roskilde, Denmark) were used for the Toxocara ELISA and high binding plates (Greiner, Frickenhausen, Germany) for the Ascaris ELISA. The antibody optical density (OD) units of the tested and reference (cut-off) sera were used to calculate a ratio, with a serum having a ratio ≥1.0 being considered positive. The cut-off value was defined as the mean absorbance of 20 serum samples from healthy blood donors plus three times the standard deviation (Pinelli et al. 2009, 2011). The methods, antigens and controls have not been altered between the two serosurveys.

Data analysis

Seroprevalence rates for Toxocara and Ascaris were estimated either for 1995/1996 (PIENTER-1) or 2005/2006 (PIENTER-2). Generalized linear models with gamma family and log link were used to identify factors associated with increased or decreased OD ratio units for Toxocara and Ascaris in the tested sera, since these were continuous, positive and right-skew outcome variables with constant variance on the log scale. Models were built in stepwise fashion; 21 (PIENTER-1) and 52 (PIENTER-2) variables were tested for association with the outcome (Tables 5 and 6). After univariate selection (p < 0.10) of candidate predictors to be assessed multivariately, non-significant (p > 0.05) variables were dropped one by one from the multivariable models after having evaluated each partial effect; variables causing a change of >10 % in the coefficients of the other covariates were retained in the models. Age group, gender, ethnicity, urbanization degree, SES group and education level (see Tables 1, 2, 3 and 4 for categorization of these variables) were always included in the models to account for potential confounding effects. Biologically plausible interactions between independent variables were also assessed. All analyses accounted for the survey design, including the geographical regions as strata, the municipalities as clusters (principal sampling units) and a weighting adjustment for age, gender, ethnicity and urbanization degree to the corresponding population from which the samples were drawn in order to account for deviations of the sample distribution from the general population in the Netherlands. Model residuals were inspected to confirm absence of any remaining structure not accounted for by the models. Statistical analyses were performed using STATA 13 (StataCorp, College Station, USA).

Table 1 Factors associated with increased or decreased OD ratio units for Toxocara in the first nationwide serosurvey (PIENTER-1) in the Netherlands, 1995/1996
Full size table
Table 2 Factors associated with increased or decreased OD ratio units for Toxocara in the second nationwide serosurvey (PIENTER-2) in the Netherlands, 2006/2007
Full size table
Table 3 Factors associated with increased or decreased OD ratio units for Ascaris in the first nationwide serosurvey (PIENTER-1) in the Netherlands, 1995/1996
Full size table
Table 4 Factors associated with increased or decreased OD ratio units for Ascaris in the second nationwide serosurvey (PIENTER-2) in the Netherlands, 2006/2007
Full size table

Results

Seroprevalence

Overall Toxocara seroprevalence was estimated at 10.7 % (95 % confidence interval [95 % CI] 8.9–12.8 %) in PIENTER-1 and 8.0 % (95 % CI 6.7–9.5 %) in PIENTER-2. For Ascaris, seroprevalence was 30.4 % (95 % CI 27.7–33.3 %) in PIENTER-1 and 41.6 % (95 % CI 39.6–43.5 %) in PIENTER-2. Looking at the differences in these seroprevalence rates revealed that while Toxocara seroprevalence had decreased significantly from PIENTER-1 to PIENTER-2 (z = −2.67, p = 0.008), that of Ascaris was significantly higher in the second serosurvey compared to the first one (z = 6.37, p < 0.0001). In general, seroprevalence of both Toxocara and Ascaris rose with increasing age (Figs. 1 and 2) and decreased with increasing education level (Tables 1, 2, 3 and 4). Toxocara seroprevalence was higher in males and in non-autochthonous Dutch individuals, particularly in first-generation migrants from Surinam/Netherlands Antilles (Figs. 1, 2 and 3; Tables 1 and 2). Also Ascaris seroprevalence was higher in non-autochthonous Dutch individuals, particularly those of non-Western origin (Fig. 3, Tables 3 and 4).

Fig. 1
figure 1

Toxocara and Ascaris seroprevalence estimates per age group in the first (PIENTER-1) and in the second (PIENTER-2) nationwide serosurvey in the Netherlands. Error bars represent 95 % confidence intervals. Toxocara and Ascaris prevalence estimates from PIENTER-1 and PIENTER-2 are adjusted for the variables shown in Tables 1, 2, 3 and 4, respectively

Full size image
Fig. 2
figure 2

Toxocara and Ascaris seroprevalence estimates per gender in the first (PIENTER-1) and in the second (PIENTER-2) nationwide serosurvey in the Netherlands. Error bars represent 95 % confidence intervals. Toxocara and Ascaris prevalence estimates from PIENTER-1 and PIENTER-2 are adjusted for the variables shown in Tables 1, 2, 3 and 4, respectively

Full size image
Fig. 3
figure 3

Toxocara and Ascaris seroprevalence per ethnic group in the first (PIENTER-1) and in the second (PIENTER-2) nationwide serosurvey in the Netherlands. Error bars represent 95 % confidence intervals. Toxocara and Ascaris prevalence estimates from PIENTER-1 and PIENTER-2 are adjusted for the variables shown in Tables 1, 2, 3 and 4, respectively

Full size image

Factors associated with anti-Toxocara IgG levels

Factors independently associated with increased OD ratio units in the final multivariable model for Toxocara were increasing age, male gender, having gardened and/or had contact with soil and/or sand (including that of sandpit playgrounds) with bare hands in the last 12 months and having owned cattle, pigs or cats in the last 5 years. These factors were significant in both serosurveys (Tables 1 and 2). In PIENTER-2, other factors associated with increased anti-Toxocara OD ratio units were increasing monthly income, having hay fever and being a first-generation migrant from Surinam/Netherlands Antilles or from other non-Western countries (other than Morocco and Turkey) as compared to being autochthonous Dutch (Table 2). Conversely, a very high education (i.e. from university-level institutions) and living in an area with an intermediate degree of urbanization (PIENTER-1) or in moderately urbanized areas (PIENTER-2) were independently associated with decreased OD ratio units for Toxocara (Tables 1 and 2).

Factors associated with anti-Ascaris IgG levels

In both PIENTER-1 and PIENTER-2, increasing age and having owned pigs in the last 5 years were independently associated with increased OD ratio units in the final multivariable model for Ascaris (Tables 3 and 4). A significant interaction was found between geophagia and (preschool) age, with ≤5-year-old children displaying geophagia having increased anti-Ascaris antibodies as compared to non-geophagic children (Table 4). Other factors associated with increased anti-Ascaris antibodies in PIENTER-2 were being a first-generation migrant from either Morocco/Turkey or from a Western country (in Europe, North America, Australia and New Zealand) other than the Netherlands or being a first- or second-generation migrant from either Surinam/Netherlands Antilles or from other non-Western countries, vs. being an autochthonous Dutch. Factors independently associated with decreased OD ratio units for Ascaris were increasing education level and having owned rabbits or poultry in the last 5 years (Tables 3 and 4).

Discussion

While the seroprevalence of Toxocara in the general Dutch population was found to decrease from 10.7 % in 1995/1996 to 8.0 % in 2006/2007, the one of Ascaris increased from 30.4 to 41.6 % during the same period. These trends agree with our previous findings on Toxocara and Ascaris seropositivity among patients suspected of VLM/OLM during 1998–2009 in the Netherlands (Pinelli et al. 2011), as in these patients, Toxocara seropositivity decreased significantly over time, whereas Ascaris seropositivity remained unchanged. Campaigns promoting regular deworming of dogs and cats have existed for many years in the Netherlands (Overgaauw and Boersema 1996). Although this might explain the observed decrease in Toxocara seroprevalence, it is difficult to ignore that regular deworming of cats is infrequent (Overgaauw et al. 2009) and that Toxocara prevalence in dogs has remained almost unchanged (at relatively low levels of ∼5 %) over the last two decades in the Netherlands (Nijsse et al. 2015b). A recent modelling paper (Nijsse et al. 2015a) also suggested that cats, rather than dogs, are responsible for most of the environmental contamination with Toxocara eggs, accounting for 46 % of the overall Toxocara egg output of >6-month-old hosts in the Netherlands, followed by dogs (39 %) and foxes (15 %). The same study also reported that Toxocara egg output in urban areas is dominated by free-ranging cats (81 %), as the Netherlands is a country free of stray dogs. Moreover, simulated intervention scenarios indicated that the currently advocated four-times-a-year deworming advice for adult dogs would have little impact on environmental contamination with Toxocara eggs unless its compliance is very high (>90 %) (Nijsse et al. 2015a), which is not realistically enforceable. Altogether, these findings suggest that nowadays dogs might not be the most problematic source of human toxocariasis in a country like the Netherlands, so deworming is unlikely to explain the observed decrease in Toxocara seroprevalence. Rather, such decrease is likely to reflect the interplay of several factors entailing hygiene enhancements between 1995/1996 and 2006/2007, e.g. more frequent enforcement of clean-up/disposal of dog faeces as well as prohibition of dogs being kept off-leash in most places, establishment of dedicated (confined) walking areas for dogs in urban areas, increased use of commercial pet food as opposed to (raw) kitchen scraps, increased practice of covering sandpit playgrounds when not in use, periodic replacement of sand in public sandpits, etc.

The use of larval ES antigens is the most common approach for serodiagnosis of Toxocara and Ascaris infections in humans (Smith et al. 2009; van Knapen et al. 1992). It is worth mentioning that due to cross-reactivity, the Toxocara ELISA does not differentiate between T. canis and T. cati and the Ascaris ELISA does not differentiate between A. suum and A. lumbricoides. Moreover, we cannot rule out that cross-reactivity with other helminths like Toxocara leonina might occur, as this has not yet been assessed. However, autochthonous A. lumbricoides infections are virtually absent in the Netherlands as indicated by at least two large population-based studies on enteropathogens in stool samples screened for different parasites and in which no helminth eggs were found (de Wit et al. 2001a, b). Indeed, A. lumbricoides has a worldwide distribution, but it occurs primarily in developing countries where conditions of poor sanitization exist, whereas it is extremely rare in developed countries (Bethony et al. 2006; Umetsu et al. 2014). Moreover, people infected with A. lumbricoides do not always develop circulating antibodies, as these are not usually elicited by the presence of adult worms in the intestine but rather by the extra-intestinal larval migrations in proportion to the number of migrating larvae. Therefore, small numbers of migrating larvae and short-lasting migrations may not suffice to prompt a detectable humoural immune response (Haswell-Elkins et al. 1992). Importantly, the Netherlands is one of the largest pig producers in Europe, and the farm-level prevalence of A. suum in Dutch swine farms is as high as 50 % in free-range farms, 73 % in organic farms and 11 % in conventional farms, with fattening pigs within conventional farms showing the highest A. suum prevalence (55 %) (Eijck and Borgsteede 2005). Therefore, our results about Ascaris more likely refer to A. suum than to A. lumbricoides. Yet, in both PIENTER-1 and PIENTER-2, Ascaris seroprevalence was higher in non-Dutch participants, particularly those from non-Western countries, where A. lumbricoides infection might still be frequent.

Although Toxocara and Ascaris have common antigens that can lead to cross-reactivity (Lozano et al. 2004), this has been reported to be below the detection limit of the respective assays (Pinelli et al. 2009). Moreover, we found different trends in the seroprevalences of Toxocara and Ascaris, and Ascaris seroprevalence was much higher than the one of Toxocara, making any cross-reactivity unlikely to have affected the results. Ascaris seroprevalence was, however, surprisingly high, as it seems rather unclear how the general population is exposed to A. suum at a growing rate. Pig manure, which has widely been used as fertilizer for arable lands and compost soil during the 10 years between the two serosurveys, is often contaminated with A. suum eggs (van Knapen et al. 1992). Thus, contaminated soil has long been deemed a likely source of human A. suum infections in the Netherlands (Pinelli et al. 2009, 2011) and in other European countries (Schneider and Auer 2015), whereas the exposure to A. suum through food remains to be investigated. A recent Austrian study examined 4481 sera from patients with suspected VLM during 2012–2014 using immunoblot and found an A. suum seropositivity of 13.2 %, which is lower than any A. suum seropositivity recorded in the Netherlands. As the Austrian study pointed out, besides differences in pig production and possibly the amount of pig manure contaminated with A. suum eggs used as fertilizer, alternative explanations for the observed difference might be the overrepresentation of the urban population with a putative lower risk of infection in the Austrian study, the diagnostic assay used (immunoblot vs. ELISA) and possible reactions with other antigens in the Dutch sera (Schneider and Auer 2015). Therefore, fine tuning of the cut-off to increase specificity are warranted in future studies to further address this matter.

Antibody levels for both Toxocara and Ascaris increased with age, suggesting continuous exposure to these helminths throughout life. While no significant gender effects were found for Ascaris, males showed significantly higher Toxocara OD ratio units than females. This has been reported previously (Pinelli et al. 2011) and might be related to traditionally male-oriented behaviours and activities, including the occupation. A study on risk factors for Toxocara infections in different occupational groups found that farmers have the highest Toxocara seroprevalence (44 %), with free-roaming farm cats and dogs being indicated as the main sources (Deutz et al. 2005). This agrees with our other finding that having owned cattle or pigs in the previous 5 years was associated with increased levels of anti-Toxocara antibodies, as these animals are highly unlikely to play a role in the direct transmission of these helminths to humans but rather act as proxies for rural lifestyle. A direct causal relationship may instead be derived from the positive associations we found between cat ownership and increased anti-Toxocara antibodies and pig ownership and increased anti-Ascaris antibodies.

We found that exposure to Toxocara was more likely to occur in people having direct contact with soil, like those gardening with bare hands or playing in a sandpit, and that preschool children exhibiting geophagia have higher levels of anti-Ascaris antibodies. This agrees with the positive association between having a high income and increased anti-Toxocara antibodies, as wealthy people in the Netherlands are more likely to have houses with gardens. However, the lack of a significant gender effect for Ascaris is somewhat suggestive that also food plays a role, as hypothesized previously (Pinelli et al. 2011; Schneider and Auer 2015), and that the association between high education and decreased anti-Ascaris or anti-Toxocara antibody levels may essentially mirror the effect of (the knowledge of) hygiene measures in addition to occupation. We also found that suffering from hay fever was associated with increased anti-Toxocara antibody levels. As mentioned in the introduction, there is evidence indicating that migration of Toxocara larvae through the lungs may result in hyper-reactivity of the airways, contributing to the development of allergic manifestations (Pinelli and Aranzamendi 2012). Our findings about the “protective” effects of living in moderately urbanized areas, as well as owning rabbits or poultry, are difficult to grasp and might act as proxies of other (hitherto unknown) factors entailing a lower risk of encountering Toxocara and/or Ascaris.

One of the most interesting results was the effect of ethnic background on increased levels of anti-Toxocara and anti-Ascaris antibodies. This concerned specifically the first-generation migrants from Surinam/Netherlands Antilles and from other non-Western countries (excluding Turkey and Morocco) for Toxocara and either the first- or second-generation migrants from Surinam/Netherlands Antilles and from other non-Western countries, as well as the first-generation migrants from Turkey or Morocco, for Ascaris. Ethnicity as a risk factor for Toxocara infection has been reported in the USA for Hispanic children of Puerto Rican descendent (Sharghi et al. 2001). However, it is largely unclear whether these associations are due to specific lifestyle characteristics posing these ethnic groups at increased risk of helminth infections or whether this is due to the high frequency of travel to the countries of origin to visit relatives and friends. Surely, free-ranging dogs and cats are widespread in many non-Western countries, including the Caribbean basin (Georges and Adesiyun 2008; Krecek et al. 2010; Thompson et al. 1986), where Toxocara seroprevalence in humans is generally much higher than that in the Netherlands, especially among children (e.g. 40–86 % in Caribbean children vs. 2–6 % in those of the present study) (Baboolal and Rawlins 2002; Kanobana et al. 2013; Lynch et al. 1993; Thompson et al. 1986). This applies to Ascaris as well (Lynch et al. 1993), whereas countries with predominantly Muslim population are hardly exposed to pigs (either via food or the environment) given their religious restrictions on the consumption of pork and the very limited pig farming therein. Thus, the significantly higher levels of anti-Ascaris antibodies we found among the first-generation migrants from Turkey/Morocco were quite unexpected. However, there are reports of relatively high rates of A. lumbricoides infection in these countries (El Guamri et al. 2011; Yentur Doni et al. 2015), so cross-reactivity might be an explanation. It is nonetheless interesting to note that the levels of anti-Ascaris antibodies among the second-generation migrants from Turkey/Morocco did not differ significantly from those of the autochthonous Dutch population. This suggests that people who are born and live in a virtually A. lumbricoides-free country like the Netherlands and refrain from consuming pork would show similar levels of antibodies against Ascaris than those of the native (and largely pork-consuming) Dutch population, thereby supporting the hypothesis of predominant exposure to Ascaris through the environment.

In conclusion, we reported different trends for Toxocara and Ascaris seroprevalences in the Netherlands, possibly reflecting improvements in hygiene management of pets and increased exposure to soil contaminated with pig manure. Future research should focus on the sources of A. suum contamination in the environment, not only in the Netherlands but also in other swine-reach regions. Besides identifying several factors associated with increased anti-Toxocara and anti-Ascaris antibodies, allowing for the identification of specific (age, ethnic, socio-economic, etc.) groups of the general population at high risk of Toxocara or Ascaris infections, our results suggest that these infections mainly occur through environmental, rather than foodborne, routes, with direct contact with soil and ownership of definitive hosts like cats and pigs being potentially modifiable exposures.