SEROLOGICAL SCREENING OF PATIENTS WITH CLINICAL SUSPICION OF TRICHINELLOSIS IN BELGRADE FROM 2009 TO 2018 Serological screening of patients with clinical suspicion of trichinellosis in Belgrade from 2009 to 2018. 2019.

Introduction. Trichinellosis is one of the most important foodborne diseases in Serbia. Most patients with which coincides with the months with the most confirmed cases of trichinellosis. A total of 17.4% patients who were tested for trichinellosis had other parasitic infections. Altogether, 223 (14.2%) of tested patients were finally diagnosed with trichinellosis. We detected anti- Trichinella IgG in 68.8% (223) of patients with suspected trichinellosis on admission, which increased to 86.5%, 91.5% and 92.4% after later second, third and fourth testing, respectively. Final diagnoses of toxocariasis, strongyloidiasis, filariasis, and dirofilariasis were made for 2.4%, 0.3%, 0.3% and 0.1% of patients, respectively. Concurrent seropositivity for Trichinella and Toxocara was observed in 18.9% (7/37) of patients with clinical presentation of trichinellosis and who were also tested for toxocariasis. In 3/5 patients with imported filariasis, we found cross-reactivity with Trichinella . Potential cross-reactivity of this ELISA test with antibodies to the autochthonous nematode Toxocara canis demands the introduction of Western blot technology. Trichinellosis must be diagnosed by the combination of clinical, laboratory and epidemiological criteria.


INTRODUCTION
Trichinellosis is one of the most important foodborne diseases in Serbia. Trichinellosis among swine in Serbia had a declining trend from 0.14% to 0.02% between 2001 and 2010. In same period, there were 2257 cases of human trichinellosis with 3 deaths (Sofronic-Milosavljevic et al., 2013). According to the Institute of Public Health of Serbia Dr Milan Jovanovic Batut (IOPHOS-Batut, 2018), the notification rate of trichinellosis in Serbia from 2008-2017 was stable, with annual oscillations. Compared with 2016 (2.68 cases per 100,000 inhabitants), the notification rate decreased in 2017 (0.21 cases per 100,000 inhabitants). A recurring peak occurs in winter. Outbreaks usually occurred among family members after consumption of undercooked pork meat and traditional products thereof, or wild boar and horses and their products in some years (IOPHOS-Batut, 2018). A common cause of outbreaks is consumption of unexamined meat or illegally produced/sold meat products containing Trichinella larvae.
The Clinic for Infectious and Tropical Diseases (CITD) at the Clinical Center of Serbia (CCS) as a University Hospital is a tertiary health care facility that treats patients with trichinellosis from Belgrade who are admitted directly to the Clinic and unclear or complicated cases who cannot be treated in other hospitals in Serbia. The Parasitological Laboratory within the CITD is the clinic diagnostic laboratory for trichinellosis. All patients with clinically suspected trichinellosis who are referred to the CCS for treatment are subjected to serological screening for trichinellosis by this laboratory.

MATERIALS AND METHODS
A retrospective serology study of trichinellosis was carried out from January 2009 to December 2018 and included all outpatients and hospitalised patients from the Clinic for Infectious and Tropical Disease, CCS, who were serologically tested for Trichinella by the Parasitological Laboratory, CCS (n=1,565). A total 1,696 sera of these patients with suspected trichinellosis were tested. Serology testing for Trichinella-specific IgG antibodies was performed for all 1,565 patients upon admission. Testing was repeated in 131 patients with non-reactive initial results and persistent clinical suspicion of trichinellosis. Then, 106 patients (7.7%) were tested for a second time, 20 (1.5%) for a third time and five (0.4%) for a fourth time, repeatedly using the same commercial test. Trichinellosis was diagnosed on the basis of combination of clinical presentation (fever, myalgia, diarrhoea, facial oedema, subconjunctival, subungual and retinal haemorrhages), blood eosinophilia, levels of muscle enzymes, Trichinella-specific antibody response and epidemiological criteria (exposure to contaminated meat or meat products), in accordance with the recent European Union (EU) Trichinella case definition (EU, 2018).
Trichinella-specific IgG antibodies were detected in sera by commercial enzymelinked immunosorbent assay (ELISA) Novagnost® Trichinella spiralis IgG according to the recommendations of the manufacturer (NovaTec Immunodiagnostica GmbH, Dietzenbach, Germany). Before testing, all the sera were allowed to thaw at room temperature and they were mixed thoroughly by vortex. The sera dispensing steps (serum dilution 1+100), subsequent processing of the test (incubation 60 min, washing 3x, adding of 100 µL conjugate, incubation 30 min, wash 3x, addition of 100 µL TMB substrate, incubation 15 min, addition of 100 µL Stop solution), and measurements were performed fully automatically on the Siemens-BEP® 2000 Automatic System. This is a qualitative and semi-quantitative test. Results were reported in Novagnost® Units. Sera were considered positive if the absorbance value was higher than 15% over cut-off or >11.5 Novagnost® Units; those <8.5 were negative and, between 8.5 and 11.5 were intermediate.
Cross-reactivity with other nematodes was also analysed. During the reporting period in this group of 1,565 patients, 601 patients were tested for Toxocara canis, 1,264 for intestinal helminths including 293 for Strongyloides stercoralis and 43 for microfilariae, at the same time as the Trichinella examination was conducted.
The same type of test and protocol was used for detection of Toxocara-specific IgG in sera. Using the Novagnost® ELISA, sera were considered positive if the absorbance value was higher than 15% over the cut-off level or >11.5 Novagnost® Units. Three methods (direct faecal smear, formalin-ether or commercial Parasep® concentration techniques and agar plate culture) were used to examine stool specimens for diagnosis of Strongyloides stercoralis larvae. Blood samples were examined by the Knott method for the presence of microfilariae.
Data was obtained from laboratory protocols and the patient medical records. The statistical analysis of the data was descriptive and analytical. The statistical t-test, Chisquared test and the Mann-Whitney U test were used to test significant difference between two groups. ANOVA was used for testing statistical significance for three or more groups. Distribution of frequency was analysed using the Chi-squared Goodness of Fit test. P<0.05 was considered statistically significant. The analysis was conducted in SPSS version 15.0 (SPSS Inc., Chicago, IL, USA).

RESULTS
Serological detection of Trichinella-specific IgG antibodies was performed in 1,565 patients. The mean age was 43.1±18.1 years (range 6-85); 737 (47.0%) were women and 828 (52.9%) men. The number of serologically tested patients varied greatly per year (χ2=111.67, DF=9, p<0.01). Annually, from 109 (2016 y) to 231 (2014 y) sera were tested, with a mean number of 156.6±44.01 (Figure 1). The monthly variations in Trichinella test requests showed very strong seasonality (χ2=72.27, DF=11, p<0.01). Most patients were tested in December or March (Figure 2). A total of 273 (17.4%) tested patients had parasitic infections. Final diagnoses of trichinellosis were ascribed to 223 (14.2%) tested patients. Final diagnoses of toxocariasis, strongyloidiasis, filariasis, and dirofilariasis were ascribed to 2.4%, 0.3%, 0.3% and 0.1% of tested patients, respectively. Separate cases of amoebiasis and trichuriasis were diagnosed in two patients (0.1%). Depending on clinical presentation and laboratory data, the other patients (n=1294; 82.7 %) were referred for further, differential diagnostic investigations (Table 1). Most of them were referred to the Allergy Clinic and the Clinic for Hematology because of the patients' unclear eosinophilia or to the Clinic of Rheumatology to differentiate causes of polymyositis. In five patients with positive serology for Trichinella, trichinellosis was not confirmed. Three of these patients had imported filariasis. Microfilariae of Loa loa were found in the blood of two patients, while the third patient was amicrofilaraemic with Calabar swellings. One of the five patients had strong positivity for T. canis. These four results were interpreted as cross-reactivity with the Trichinella test, while in the fifth patient, the presence of other parasitic infection was not confirmed. In 12 cases, intermediate results for the Trichinella-specific IgG ELISA were recorded. Due to insufficient clinical and epidemiological suspicion of trichinellosis, testing was repeated. In six cases, the finding was negative, and in the other six cases, no increased antibody levels were observed. Other parasitic infections were not detected in these cases.
Among the 223 patients with clinically confirmed trichinellosis, initial detection of Trichinella-specific IgG antibodies was successful in 68.6% (n=153) of patients. In the second test conducted on 70 patients, seroconversion was present in 40 patients. Repeated serological testing (third time) was performed in 19 patients and showed seroconversion in 11 of these. Repeated serological testing (fourth time) was performed in two patients and showed seroconversion in both of them. Thus, a total of 206/223 (92.4%) patients were serologically positive for trichinellosis (Table 2). In the remaining 17 of the 223 patients, the testing was not repeated because they were involved in family outbreaks and had clear clinical presentation and/or the test was not available at the time of diagnosis. Antibodies against T. canis were detected in 18.9% (7/37) of tested patients, who also had trichinellosis at the same time (Table 3). Three patients with history of travel to tropical areas were tested for the presence of microfilariae, but test results were negative. Coproparasitological testing did not show the presence of strongyloidiasis (Table 3). Trichinellosis is a potentially fatal disease (Sofronic-Milosavljevic et al., 2013). Clinically manifest human trichinellosis starts with non-specific symptoms, so many other diseases should be considered during differential diagnosis (Dupuy-Camet & Bruschi, 2007;Gottstein et al., 2009). This unspecific and wide range of clinical manifestations was a main reason why so many of our patients were subjected to testing for trichinellosis. Despite the non-specific clinical manifestation of trichinellosis, new patients are usually registered in outbreaks, which makes clinical suspicion easier, but identification of initial cases can be difficult. For this reason, all recommended criteria in the EU Trichinella case definition should be included in analysis and carefully combined (EU, 2018).
Moreover, sporadic cases of trichinellosis, especially those with an atypical course, can be very difficult to recognise, and for this reason, trichinellosis should be excluded from various other diseases with similar clinical presentation. Non-specific symptoms were the reason three of our patients were first admitted to other specialist Clinics. Typically, trichinellosis in Serbia occurs in the form of outbreaks, most family outbreaks (Sofronic-Milosavljevic et al., 2013). The appearance of cases within outbreaks accelerated diagnosis, and was why no wider testing for other parasitic infections was performed for 223 patients with trichinellosis.
In our study, most patients presented for testing in December and March, which coincided with the months with the greatest numbers of confirmed trichinellosis cases. Traditionally in Serbia, swine are slaughtered during November and December, and undercooked meat is often consumed before veterinary inspection. Traditionally prepared meat products containing Trichinella larvae are usually consumed within the family circle, leading to family outbreaks. This explains the peak in testing and confirmed cases of trichinellosis in March, when meat products prepared from pigs slaughtered in November/December are ready for consumption. Our previous study showed the strong seasonality of trichinellosis in Serbia during the eight studied years (Ofori-Belić et al., 2010).
According to the case definition of trichinellosis that is suggested by the EU, a combination of a defined number of clinical criteria (fever, muscle soreness or pain, diarrhoea, facial oedema, eosinophilia, subconjunctival, subungual and retinal haemorrhages), laboratory criteria (demonstration of Trichinella larvae in tissue obtained by muscle biopsy or Trichinella-specific antibody response obtained by immunofluorescence assay (IFA), ELISA or Western blot (WB)) and epidemiological criteria (exposure to contaminated food or exposure to a common source) defines probable (clinical plus epidemiological criteria) and confirmed (clinical plus laboratory criteria) cases (EU, 2018). Although a muscle biopsy is included in the laboratory criteria of EU Trichinella case definition, there was no need for its use in our patients.
In the previously investigated period, in two patients with presumed trichinellosis who were not associated with outbreaks, muscle biopsy, although an invasive method, had to be performed for a definite diagnosis (Ofori-Belić et al., 2010). Even though other immunological specific assays (IIF, WB) tests are commercially available in Serbia, they were not performed in our laboratory due to technical reasons. Confirmatory testing is provided by the National Reference Laboratory for trichinellosis in Serbia, which should resolve all unclear cases, but we used this only sporadically, since financing for all NRLs in Serbia has not yet been regulated. Resolving the cross-reactivity between Toxocara and Trichinella in the ELISA test we used has not yet been performed for technical reasons.
Seroconversion, i.e. production of IgG antibodies, usually occurs between 12 to 60 days after infection (Turk et al., 2006), and the level of the specific IgG response correlates with the number of infective muscle larvae ingested, Trichinella species and the host immunity (Pozio et al., 1993;Yang et al., 2016). The window period with a high rate of false negative results during the early stage of infection is the main disadvantage of tests detecting anti-Trichinella IgG . Repeated serological testing increased the number of serologically confirmed cases of trichinellosis. We detected anti-Trichinella-IgG in 68.8% of confirmed cases on admission, which increased to 86.5%, 91.5% and 92.4% after second, third and fourth testing, respectively. In hospitalised patients, seropositivity increased from 58.5% on admission to 86.1%, at second testing (after 10 days on average). There was a wide range from 7 to 22 days between the first and second tests, which depended on individual clinician's decisions. Turk et al. (2006) reported detection of anti-Trichinella IgG in 42.2% of confirmed cases on admission, which increased to 67.8% and 73% after 15 and 30 days, respectively. Our recent survey between 2001 and 2008 on 50 hospitalised patients with trichinellosis showed Trichinella-specific IgG antibodies were detected (by the same ELISA used in the current study) in 13/25 (52%) patients on admission (Ofori-Belić et al., 2010). The mean time between onset of symptoms and admission was 9 days. Repeated serological testing two weeks later performed in four patients showed seroconversion in all (Ofori-Belić et al., 2010).
ELISA is the most commonly used approach for the detection of Trichinella infection in humans. However, numerous cross-reactions are possible if patients are affected by other diseases, as more often occurs in developing countries where parasitic infections are more present (Gómez-Morales et al., 2010). The sensitivity and specificity of different ELISA tests depends on the antigens used. In this study, reliable sensitivity and specificity could not be calculated because a confirmatory serological test like WB is unavailable in CCS. Sun et al. (2015) showed that in ELISA, adult worm excretorysecretory (AW ES) antigens were superior to those of the AW crude antigens and of T. spiralis muscle larvae (ML) ES antigens. Several studies showed that ML ES antigens have a risk of cross-reactivity with the sera of patients with other helminthiases (paragonimiasis, schistosomiasis, clornorchiasis, cysticercosis and anisakiasis) (Dea-Ayuela et al., 2001;Ciu et al., 2015;Yera et al., 2003).
According to the user manual of the commercial ELISA we used, both the sensitivity and specificity are >95%. However, the test manufacturer notes that cross-reactivity with antibodies against T. canis is possible. The simultaneous detection of antibodies for Trichinella and Toxocara in 18.9% of cases in our patients could be a result of crossreaction or could indicate simultaneous infections by both autochthonous nematodes. The actual situations for individual patients would likely be clarified and more accurate if these sera were further analysed by a confirmatory WB test, which was not available to us. Using an ELISA, Gabrielli et al. (2017) reported 23.5% Toxocara seroprevalence, which was confirmed by confirmatory WB tests in 13.0% of the examined population in different areas of Serbia. Cross-reactivity between Toxocara and Trichinella has been documented before (Yera et al., 2003).
The basic limitation of this study is the unavailability of confirmatory WB tests for Trichinella and Toxocara in CCS. For reasons stated above, diagnosis of trichinellosis should not be established only on the basis of a single test result. Reliable diagnosis takes into consideration epidemiological data, clinical presentation and serological data.
In our hospitalised patients, the mean LOS was 8 days, which was half the LOS we determined in our previous study (Ofori-Belić et al., 2010). This could be a result of more prompt epidemic alerts and earlier patient arrival times at our Clinic. Eleven (15.1%) hospitalised patients developed complications, while in our previous study, 16% developed complications (Ofori-Belić et al., 2010). The cardiological and neurological complications of trichinellosis that predominated in our current study are common (Dupuy-Camet & Bruschi, 2007). Gastrointestinal manifestations of trichinellosis are not pathognomonic and are usually moderate (Turk et al., 2006). However, two patients in this study developed a chronic diarrhoeal syndrome, probably due to massive ingestion of infectious muscle larvae. Albendazole was continuously administered to these patients and was well tolerated.
Our results show that trichinellosis remains a major public health problem in Serbia. Taking into account that trichinellosis not only endangers human health and causes significant economic costs, but also endangers the traditional way of life and eating habits, multidisciplinary preventive work is required to significantly reduce the threat of human trichinellosis and animal Trichinella infections. Strict respect of prescribed veterinary control measures and a permanent public health education program are essential.

CONCLUSION
The diagnostic disadvantages of the ELISA used to detect Trichinella-specific IgG antibodies in patients means results must be interpreted carefully and in correlation with patients' clinical presentation. Particularly sensitive categories of patients are those not associated with trichinellosis outbreaks and those with history of travel to tropical areas where numerous nematode species, including loiasis, are present. Potential crossreactivity of this ELISA test with antibodies to the autochthonous nematode T. canis demands the use of WB in our CCS laboratory. The availability of more types of tests, including WB, would enhance interpretation of test results, speed up trichinellosis diagnoses and reduce diagnostic errors. All diagnostically complicated cases should be analysed by the NRL. Finally, the continuous presence of human trichinellosis in Serbia requires a more repressive approach to illegal meat product marketing.

Authors contributions
ZD designed the study, analyzed data and wrote the manuscript in consultation with all authors; NM contributed to the designed the study, performed and analyzed statistical data and wrote the manuscript in consultation with all authors; SJ, BM, IM, GS, NN, AM, AU, JM, BB analyzed data, discussed the results and contributed to the final manuscript; MK involved in planning, designing, analyzing data and supervised of this work.