Evaluation of Two Different Strategies for Schistosomiasis Screening in High-Risk Groups in a Non-Endemic Setting. A Retrospective Cohort Study.

Background: to evaluate the real-life performance of direct microscopy techniques and ELISA serology for schistosomiasis screening in a high-risk population in a non-endemic setting. Methods: a retrospective cohort study was conducted in two out-patient Tropical Medicine units in Barcelona (Spain) from 2014 to 2017. Asymptomatic adults arriving from the Sub-Saharan region were included. Schistosomiasis screening was conducted according to clinical practice following a different strategy in each setting: A) feces and urine direct examination plus S. mansoni serology if non-explained eosinophilia was present, B) S.mansoni serology plus uroparasitological examination at second-step in case of a positive serology. Demographic, clinical and laboratory features were collected. Schistosomiasis cases, clinical management and 24 months follow-up were recorded for each group. Results: Four-hundred forty individuals were included: 399 in group A, 41 in group B. The majority of the cohort were male (250, 56.4%), mainly from West African countries. No differences in sex distribution (p=0.57), age (p=0.6) and baseline eosinophilia (p=0.17) were found between both groups. Fifty schistosomiasis cases were detected (11.5% group A vs 4 % group B, p=0.733). When both microscopic and serological techniques were performed, discordant results were recorded in 18.2% (16/88)). Eosinophilia was present in 50% cases. Schistosomiasis cases were younger (p<0.001) and presented more frequently eosinophilia and elevated IgE (p<0.001). Pathological ultrasound was described in 33.3% of examinations performed (4 out of 12). Conclusions: Similar results were achieved by serology and microscopy techniques -based strategies. Serology might be prioritized for schistosomiasis screening from regardless peripheric count levels.


Background
Human schistosomiasis is a neglected tropical disease caused by trematodes of the gender Schistosoma sp. Approximately 240 million people worldwide are affected by this condition (1). Meanwhile, individuals with chronic sequelae after infection are estimated in 440 million (2). Endemic areas are located in tropical and subtropical regions of around 70 countries in Asia, America and Africa, although the Sub-Saharan region gathers up to 90% of cases (1). Six species of Schistosoma sp. that affect humans have been identi ed, whose distribution is linked to the fresh-water snail host habitat (3). Schistosoma (S)mansoni, S.japonicum, S. mekongi, S.intercalatum and S. guineensis cause gastrointestinal manifestations through the invasion of mesenteric plexus, while S.haematobium invades the pelvic venus causing genitourinary disease (3). Acute schistosomiasis might evolve to chronic forms, which often remain asymptomatic but may lead to severe consequences in the long-term as result of chronic tissue in ammation. In low-endemic areas, schistosomiasis has gained prominence due to migration streams, increased travelling to endemic regions and variations in climate conditions that affect the intermediate host distribution (5). Prevalence in high-risk groups, such as migrants from endemic areas, could reach up to 15% and even autochthonous outbreaks in Western countries have been documented (6,7). Implementation of different screening strategies in non-endemic settings have been proposed for highrisk groups (8). Direct microscopic observation of eggs in urine or fecal samples by concentration techniques such as Kato-Katz have been considered the gold standard for diagnosis in endemic populations, despite their scarce sensitivity. Direct observation shows a wide intra and inter-individual variability due to circadian changes in the egg load and throughout the different phases of the infection, as well as to the observer's experience (9,10). Antigenic methods such as circulating cationic antigen detection have shown to improve the detection of S. mansoni infections in these high-risk populations (11). On the other hand, antibody detection by commercial non-species-speci c tests (mainly enzymelinked immunosorbent assay, ELISA) is widely used in screening protocols for travelers. However, the clinical interpretation of serological results is often unclear in endemic populations chronically exposed to Schistosoma sp (12). As a result, consensus on recommended screening algorithms using available techniques is lacking. The aim of our study was to describe two different screening strategies for a highrisk group in a non-endemic area, using direct microscopy observation and serological detection by ELISA respectively, in order to determine their accuracy and usefulness in a real-world setting.

De nitions
We use the term migrant based on United Nations Educational, Scienti c and Cultural Organization (UNESCO) de nition as "any person who lives temporarily or permanently in a country where he or she was not born and has acquired some signi cant social ties to this country" (13). Individuals from European origin living in endemic areas were considered emigrant, while subjects arriving from endemic areas were considered immigrants. The term recently arrived migrant was applied to individuals who had arrived to Europe up to 6 months before inclusion. A subgroup of visiting friends and relatives (VFR) was categorized according to CDC de nition as those migrants who had ever returned to their home country since their arrival to Europe (14).

Study design
A retrospective longitudinal cohort study was performed in order to compare two different screening strategies for schistosomiasis used in the clinical practice.

Settings
The study was conducted in two out-patient units specialized in tropical medicine, part of the International Health Program of Catalan Institute of Health (PROSICS) in Barcelona, Spain. These units provide free medical care to self-referred individuals and referred cases from general practitioners and other medical specialists, public health services and non-governmental organizations (NGOs). Each clinic performs a different schistosomiasis screening strategy: A) hemogram and feces and urine concentration techniques for direct examination followed by S. mansoni serology if non-explained eosinophilia was present, B) S.mansoni serology and uroparasitological direct examination at a later time in case of a positive serology. Coproparasitological analysis was also performed in group B patients as part of general parasitosis screening, but not speci cally for Schistosoma screening. All samples were analysed at the Parasitology Unit of Vall d'Hebron Microbiology Department.

Study population and data collection.
Asymptomatic adults arriving from the Sub-Saharan African (SSA) region from January 2014 to December 2017 were included. Subjects with previously known either eosinophilia or high levels of immunoglobulin E (IgE), past history of schistosomiasis or macrohematuria were excluded.
Demographic, clinical and laboratory data at baseline were collected in all subjects.
Laboratory tests included hemogram [with absolute eosinophil count (AEC)] and biochemical panel. Mild eosinophilia was de ned as an AEC between 500 and 1000 cells/µl, moderate eosinophilia as AEC between 1000 and 3000 cells/µl and severe eosinophilia as AEC ≥ 3000 cells/µl. The upper normality limit for relative eosinophils count was settled in 4.5%, according to the recommendations of the expert group of the Spanish Society of Tropical Medicine and International Health (SEMTSI) for the diagnosis and treatment of imported eosinophilia (15). Total serum IgE levels were also recorded when available and considered normal below 500KU/L according to the local reference threshold. Routine migrant screening, including hepatitis serology (hepatitis B surface antigen and anti-hepatitis C antibodies), human immunode ciency virus (HIV) antibodies and antigen-p24 detection and syphilis serology (TPHA and RPR), was performed in all subjects. Asymptomatic individuals aged below 35 years were also tested for tuberculosis infection by either tuberculin skin test or tuberculosis blood test (Quantiferon-TB Gold®). Other latent infections, such as malaria infection and lariasis, were screened based on the patient individual risk assessment. Microscopic examination of stool and urine was performed through concentration techniques; stool examination was carried out using the Ritchie's formalin-ether technique, while urine samples were processed by centrifugation. The identi cation of the different Schistosoma species was carried out according to the morphologic characteristics of the eggs. An ELISA for IgG against S.mansoni (Novagnost S. mansoni IgG; Siemens Diagnostics, Marburg, Germany) was performed following manufacturer instructions. The results were expressed by index and considered positive when index ≥1.1, negative when index ≤0.9 and grey zone in values ranging from 0.9 to 1.1.
Complementary examinations [abdominal ultrasound (US) and X-ray] were performed at the discretion of the physician in charge. Treatment regimens and follow-up at month 6, 12 and 24 after therapy were also recorded. Figure 1 shows a owchart of the study design.

Case de nition
Con rmed cases were established after direct observation of Schistosoma eggs in either feces or urine samples. They were classi ed as either intestinal or genitourinary disease according to the identi ed Schistosoma species (16). Probable schistosomiasis case was de ned by a positive serology which was considered true positive by the responsible physician, without parasitological con rmation. The disease location was considered undetermined due to the inability to identify the parasite species in probable cases. Cases considered false positive according to the medical records were also recorded.

Statistical analysis
Univariate analysis of the dataset included measures of distribution, central tendency (median or mean depending on distribution), and dispersion (standard deviation or interquartile range [IQR]). Normally distributed quantitative variables were compared with Student t-test. Non-normally distributed quantitative variables were analyzed with the Mann-Whitney U test. Categorical variables were described in frequency and percentage. The bivariate analysis was carried out using the χ2 test or the Fisher's exact test for frequencies below 5%. Hypothesis testing was conducted with a 5% alpha risk and 95% con dence intervals (CI), and considered statistically signi cant if the two-tailed p-value was below 0.05. Statistical analysis was conducted using IBM SPSS, version 26.0 for Windows (SPSS Inc, Armonk, NY, USA).

Baseline features
A total of 566 patients were tested for schistosomiasis in both settings during study period. Four-hundred and forty patients met inclusion criteria and were included in the analysis: 399 (90.7%) were screened by urine and feces examination (group A), while 41 (9.3%) were screened by serology (group B). Table 1 summarizes baseline characteristics of the overall cohort and by screening strategy.

Schistosomiasis screening outcomes
Fifty subjects (11.4%) were diagnosed with Schistosoma infection during the study period. Characteristics of schistosomiasis cases and the comparison with subjects without schistosomiasis are presented in Table 2. Absolute and relative eosinophilia were signi cantly more frequent in the schistosomiasis group (50% vs 12.6% in absolute count, 82% vs 31.3% in relative count, both p<0.001). Grades of eosinophilia in schistosomiasis cases are represented in Figure 3. The proportion of schistosomiasis cases did not signi cantly vary between the two strategies (46 cases [11.5%] in group A vs. 4 cases [9.8%], p=0.733). According to the screening strategy: Group A. In group A, based in copro and uroparasitological direct examination, 46 cases (11.5%) of schistosomiasis among 399 screened subjects were detected. Thirty-seven (9.3%) cases were diagnosed by urine and/or feces examination: the vast majority were stool samples (30 cases; 26 subjects presented S. intercalatum eggs, 4 cases with S. mansoni eggs). Uro-parasitological samples yielded positive results in 6 subjects, in which S. haematobium eggs were observed. One case with positive results in both stool (S.intercalatum) and urine (S.haematobium) examination was detected. Serology was performed in a second-step diagnosis in two subjects with con rmed gastrointestinal schistosomiasis by S. intercalatum and S.mansoni respectively, and resulted negative in both cases. Serology was performed in a second time in 45 subjects with previous negative results in urine and stool microscopic observation, as part of the study for eosinophilia and/or hyper IgE in baseline blood test. Eleven subjects tested positive, although a probable diagnosis was assumed in 9 cases (81.8% of the cases with a positive ELISA result). The remaining two cases presented a positive serology for Strongyloides Stercolaris and were diagnosed as such, considering Schistosoma spp. serology as a false positive result due to probable cross-reaction.
Group B. Screening was conducted through baseline Schistosoma ELISA serology in 41 cases. Serology resulted positive in 6 individuals, two of them were not assumed as clinically relevant according to medical records. No cases were detected by stool analysis. Among suspected cases, three subjects underwent uroparasitological examination. Genitourinary schistosomiasis was con rmed in all three cases with the observation of S.haematobium eggs in urine specimen.  *absolute frequency (%). Probable cases were de ned by a positive serology considered true positive by the responsible physician. Con rmed cases were established after direct observation of Schistosoma spp. eggs in either feces or urine samples. Table 4 shows the correlation between serology and coproparasitological or uroparasitological studies when both a direct examination technique and serology were performed.

Discussion
Our study analyzes two schistosomiasis screening strategies used in the clinical practice of an International Health Institution in Barcelona, Spain, from 2014 to 2017. All diagnosed cases correspond to immigrants, mostly from Equatorial Guinea. This country is probably over-represented due to historical and cultural ties with Spain and the high degree of awareness of the importance of screening in this group. Similar demographic characteristics have been reported in cohorts within the Spanish territory (17,18).
The rationale for screening schistosomiasis in a population from endemic areas is that the disease is highly prevalent, and that treatment with praziquantel is safe and effective 18, 19 . In studies in endemic populations, the rate of schistosomiasis varies between 20-40% (19,20). In migrants, the prevalence varied between 1% and 18% depending on the screening technique (parasitological vs serological study), according to a recent meta-analysis (21)(22)(23). In our study the overall prevalence, including both screening strategies, was 11.5%.
It should be noted that the effectiveness of the parasitological study depends largely on the experience of the observer, and therefore may vary signi cantly between laboratories; moreover, in contexts with a low egg burden, egg excretion may vary at different times of the day; even in the same sample the eggs may be unevenly distributed (24,25). The use of more objective techniques would allow screening to be expanded and performed in primary care units, and not necessarily in specialized centers.
In this situation, the sensitivity of the technique should be prioritized over speci city, although performing two tests in parallel could be used to increase the accuracy of detection and inform schistosome specy (8). In our study, the percentage of diagnosed cases in our cohort resulted similar in both described strategies. A possible explanation could be the high specialization of the reference laboratory, which optimizes the sensitivity of direct observation techniques. This condition might not be generalized to other contexts with less specialized laboratory personnel.
In this study, the serology used was an ELISA based on crude antigens of S. mansoni. Several studies show that this technique is more sensitive than direct visualization of eggs, mainly in adults in nonendemic areas with light infections (26,27). Nevertheless, the sensitivity of commercial tests for other Schistosoma species could be compromised. Thus, sensitivity varies between 50 and 90% for S. mansoni and 20 and 70% for S. haematobium (28).
Prevalence of imported eosinophilia among travelers and immigrants is reported between 8% and 28.5%, that is consistent with the reported prevalence of 16.8% in our cohort. Etiological diagnosis is often troublesome and, depending on the depth of the study and on the population analyzed, a parasitic cause is identi ed in 17-75.9% of the individuals. Among the di culties encountered to compare studies are the heterogeneity of the studied populations, the study designs and the different diagnostic protocols (15). In schistosomiasis, eosinophilia is not a consistent nding (29,30), mainly in adult migrant patients, with initial infection probably in childhood, and low parasite load. In our study, 50% had eosinophilia.
However, this nding could be in uenced by the usual presence of coinfections in these patients, such as strongyloidiasis, lariasis and soil-transmitted infections. Consistently with other studies (18,31), IgE elevation was higher (68%); however, one third of the patients diagnosed had no abnormality. Therefore, screening for the disease should be based primarily on origin in the case of migrants, and on risk activities in the case of travelers, rather than on the presence of these parameters (21).
Once screening has been performed, the question of what the exact procedure is to follow once individuals are considered positive is not clear. There are few guidelines on the management of these patients, either symptomatic or asymptomatic. Some guidelines recommend performing a parasitological examination in the case of a positive serology, and performing an ultrasound only if eggs are found (32). Although this is essential to differentiate the species, the low sensitivity of the parasitological test and the presence of lesions in patients without egg detection suggest that an ultrasound should be performed in all patients, regardless of the parasitological result and the presence of symptoms (33,34). In our study, in all patients diagnosed, either by parasitological study or by serology, treatment was considered. However, the percentage of ultrasounds performed was very low, re ecting the di culties on an appropriate followup in many contexts. One third of the performed ultrasound showed pathological ndings in relation to the infection.
One limitation of schistosomiasis serology is that its use to differentiate between current and past infection is very limited, since antibody titers vary signi cantly after adequate treatment (35). In our study, only 2.5% of patients underwent control serology, of which 73% were positive. In general, the follow-up of migrant patients is di cult, especially in the screening of asymptomatic population, due to socioeconomic and idiomatic limitations, high mobility and lack of disease perception (36). Mechanisms need to be in place to ensure non-discrimination in health care access, such as access to cultural mediators, improvement of health literacy through targeted health promotion interventions, supranational communication system and effective asylum policies, among others (37,38).
Due to the limitations of both techniques currently used, new tests with higher sensitivity, higher speci city, and capable of differentiating current from past infections, should be developed. Meanwhile, the combination of several serological techniques, the western blot technique, or the immunochromatographic test (ICT) have shown a higher sensitivity (28).
A recent evaluation of several diagnostic tools found that a rapid commercial serological ICT test with 96% sensitivity would have the potential for being used as a single screening test for migrants from Sub-Saharan Africa (39). Molecular techniques (serum or excreta PCR) also have a high performance, improving the sensitivity of serology and allowing con rmation of the Schistosoma species involved. Furthermore, they can be useful for monitoring outcome (40). More efforts should be made to make these tests widely available for their use in daily practice.
Regarding general screening, an important proportion of the cohort was diagnosed with some infectious diseases, which encourages the maintenance of screening programs with the active participation of actors such as NGOs, community organizations and other health care departments. The lower rate of infections in our study compared to similar publications could be probably explained due to the exclusion of individuals with symptoms or known eosinophilia/elevated IgE (17,18). Concerning general migrant health status, it is worthy reported the high prevalence of cardiovascular conditions, even among such a young population as the included in this cohort (41). Speci c, culturally oriented strategies and resources should be available to ensure also detection and management of non-communicable diseases in these vulnerable populations (42).
This study presents important limitations mainly due to its retrospective design. Information bias due to incomplete data in medical records could have in uenced the results concerning serological interpretation. As foresaid, selection bias might not be completely avoided despite tight inclusion criteria. Also, the sample size of both strategies considerably differs, making di cult a signi cant comparison between groups. In the same line, the sample of schistosomiasis cases is too limited to draw robust conclusions.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests. Grades of eosinophilia (by relative frequency) in schistosomiasis cases. AEC absolute eosinophil count in peripheric blood.