Thrombocytopenia as a surrogate marker of hepatosplenic schistosomiasis in endemic areas for Schistosomiasis mansoni

Introduction: This study aimed to evaluate whether a low platelet count is a good surrogate marker of hepatosplenic schistosomiasis (HSS) in a rural area of Brazil. A small district in southeastern Brazil, with a population of 1,543 individuals and a 23% prevalence of schistosomiasis, was selected for this investigation. Methods: In July 2012, 384 volunteers were subjected to clinical, ultrasonography (US), and laboratory examinations, including stool sample analysis. The HSS patients were classifi ed into four groups: Group 1 consisted of patients with a spleen >13cm and liver fi brosis; Group 2 consisted of patients with a palpable spleen and spleen>13cm measured by US; Group 3 consisted of patients with a spleen >13cm measured by US; and Group 4 consisted of patients with a palpable spleen. Results: Eight patients were in Group 1 (2.1%), twenty-one were in Group 2 (5.5%), eight were in Group 3 (2.1%), and eighteen were in Group 4 (4.7%). A signifi cant difference in the mean platelet counts was observed between the patients with and without HSS (p<0.01). Based on the receiver operating characteristic (ROC) curve (platelet count <143,000/mm3), the sensitivity was greater than 92% in all groups, and the specifi city varied from 44.4% to 75%. Conclusions: We concluded that in endemic areas, thrombocytopenia demonstrates good sensitivity for detecting HSS and may be used as a screening tool to identify patients with HSS.

It has been estimated that 240 million people worldwide are infected with schistosomiasis, with an additional 700 million at risk of infection.Currently, 2 to 6 million individuals are considered to be infected in Brazil 1 .
Most of the infected individuals are asymptomatic, but in 5% to 10% of infected individuals, liver periportal fi brosis results in severe disease consisting of portal hypertension, splenomegaly, esophageal varices, and recurrent hematemesis 2 .Pathologically, dense bands of fi brosis around the portal tracts are typical of schistosomiasis 3 .
Schistosomal periportal fi brosis is usually assessed using imaging methods, and abdominal ultrasonography (US) has become the imaging technique of choice [4][5][6][7][8] .However, the identifi cation of other non-invasive, inexpensive, and simple routine laboratory tests as surrogate markers is of interest.Such markers would facilitate the selection of patients with hepatosplenic schistosomiasis (HSS) in rural areas for further clinical and laboratory evaluations.
In the hepatosplenic form of schistosomiasis, portal hypertension is the most evident clinical complication, and splenomegaly with platelet sequestration in the spleen is the most accepted explanation for thrombocytopenia 9,10 .
However, thrombocytopenia has been shown to be a good marker of HSS in patients admitted to Brazilian hospitals.A group of Brazilian authors showed that low blood platelet counts were associated with portal hypertension, and others also found an association between low platelet counts and the intensity of liver fi brosis 11,12 .
This study aimed to determine whether low platelet counts in the blood can be used as a screening tool to identify patients with HSS in an area endemic for Schistosomiasis mansoni in Brazil.

Patients
This was a cross-sectional, analytical study.Topázio, a small district in Teófi lo Otoni, State of Minas Gerais, Brazil, with a population of 1,543 individuals and a 23% prevalence of schistosomiasis, was selected for this investigation.A total of 400 volunteers with ages ranging from 16 to 97 years

DISCUSSION
(median=41) were enrolled in the study after signing informed consent forms.During one week in July 2008, a local health agent invited all volunteers to be examined in a local outpatient clinic; three individuals did not respond to our invitation, and 13 (3.3%)were excluded because they did not complete the examinations.Therefore, a total of 384 volunteers were ultimately included in the study.

Clinical examination
The 384 participants were subjected to anamnesis and physical examination by one of the authors (JRL).Particular attention was given to the abdominal examination; specifi cally, the right hepatic lobe was examined along the anterior axillary line, and the left hepatic lobe was examined along a line passing through the xiphoid process.The spleen was palpated and measured under the left costal margin with the patient in the dorsal decubitus position during deep inspiration [13][14][15][16][17][18][19] .

Ultrasound
All were submitted to abdominal US using a portable Medison Sonoace 1500 system with a 3.5-MHz probe (Samsung, Korea) and examined according to the proposed protocol of the World Health Organization (WHO) for US assessment of schistosomiasis-related morbidity 20 .In this study, we used a spleen size cut-off point of >13cm (longitudinal diameter) because spleens tend to be more enlarged in HSS than in other diseases that evolve with splenomegaly, such as cirrhosis and mononucleosis-like syndromes [21][22][23][24][25][26] .

Defi nition of hepatosplenic schistosomiasis
HSS was classifi ed into four groups: Group 1 consisted of patients with a spleen >13cm and liver periportal fi brosis, as confi rmed by US; Group 2 consisted of patients with a palpable spleen and spleen >13cm, as measured by US; Group 3 consisted of patients with a spleen >13cm using US; and Group 4 consisted of patients with a palpable spleen.

Laboratory studies
Stool samples from 1,543 individuals (2 slides each) were examined using the Kato-Katz technique.A 10-ml sample of blood was collected from a peripheral vein in a tube containing ethylenediamine tetraacetic acid (EDTA) as an anticoagulant, and a hemogram with platelet counts was performed (Sysmex, E 2100 D).Individuals positive for schistosomiasis were treated with a single oral dose of praziquantel (50mg/kg body weight) 27 .

Statistical analysis
The questionnaire data were transferred to a databank using the EpiData 3.1 program (EpiData Association, Odense, Denmark) and analyzed using Statistical Package for Social Sciences (SPSS) 18.0 software (SPSS, IBM Company, Chicago, IL).Tables were constructed, and the percentages, median values, and mean values were calculated.Categorical variables were compared using the χ2 test; mean values were compared using Student's t-test or analysis of variance (ANOVA); and median values were compared using the Kruskal-Wallis test.Additionally, Spearman correlation analysis was performed for continuous variables, and receiver operating characteristic (ROC) curves were used to defi ne the cut-off value of platelet counts to aid in identifying the best combination of sensitivity and specifi city.The areas under the ROC curves were calculated to defi ne the accuracy of the platelet counts in the identifi cation of patients with HSS.Differences between the groups were evaluated using the Mann-Whitney test.Statistical signifi cance was accepted at the 5% level.

Ethical considerations
This study was approved by the Human Research Ethical Board at the Federal University of Minas Gerais School of Medicine.
The comprehensive fl ow chart shown in Figure 1 summarizes the results of the present study.Demographic data and physical examination fi ndings in the 384 individuals of this study are presented in Table 1.Eight patients in the group with HSS and 136 in the group without HSS (p=0.92) were male.Overall, the patients with HSS were older than the individuals without HSS (p<0.00).
Using the ROC curve, we found that a platelet count <143,000/mm 3 yielded the highest sensitivity and specifi city for each HSS defi nition (Figure 1).
In Table 2, the area under the ROC curves (AUC), confi dence intervals (CIs), and sensitivity and specifi city of the ROC curves for a platelet cut-off point of <143,000/mm 3 are presented for each defi nition of hepatosplenic schistosomiasis.This report is the fi rst fi eldwork investigation demonstrating that thrombocytopenia can be used as a surrogate marker of HSS in areas highly endemic for schistosomiasis.Souza et al. 11 showed that low blood platelet counts were related to portal hypertension in schistosomiasis.In addition, Lambertucci et al. 12 also found that blood platelet counts less than 140,000/mm 3 identifi ed patients with HSS (a relationship was observed between the spleen size and liver periportal fi brosis grading).Both studies were performed in patients who were admitted to University Hospitals in large cities.
If the successful use of this simple and inexpensive marker for HSS in endemic areas is confi rmed by others, then this marker would hasten the identifi cation of morbidity in poor areas of the country.Patients selected based on a thrombocytopenia assessment would then have their diagnosis confi rmed by other tests, such as US.
In the present study, the patients with HSS were found to be older than the individuals without HSS (p<0.00).A possible explanation for this observation is the initiation of a mass chemotherapy program for schistosomiasis in the late 1970s  in Brazil.As the program was directed especially toward school children, it can be seen that the mean age of patients with HSS has shifted to an older population (age-cohort effect).
The median platelet counts were significantly lower in individuals with HSS (p<0.01), and platelet counts <143,000/mm 3 identified patients with HSS with a sensitivity >92%.The prevalence of the disease was 23% in the study area.As expected, with a platelet count <143,000/mm 3 , sensitivities and specifi cities for the diagnosis of HSS varied depending on the defi nition of HSS.Based on the HSS defi nition of a palpable spleen and a spleen >13cm measured by US, 21 individuals (5.5% of the people enrolled in the study) had HSS, with a sensitivity of 93.4% and specifi city of 47.6%.Some individuals with enlarged spleens most likely had other diseases, which would explain the low specifi city.In addition, errors that occur during abdominal palpation or US examination cannot be excluded.When the HSS defi nition consisted of a spleen >13cm that was associated with periportal fi brosis as confi rmed by US, 8 (2.1%) patients had HSS, with a sensitivity of 92.3% and specifi city of 75%.This defi nition appears to be more accurate because it includes the characteristic schistosomal liver periportal fi brosis.
A group of 20 individuals with periportal fi brosis confi rmed by US but with normal spleens was identified in our study, and the platelet counts were normal in this group.Symmer's periportal fibrosis without splenomegaly has been reported previously 2,13,28 .This finding demonstrates the importance of enlarged spleens in schistosomal thrombocytopenia, and this condition has been referred to as the hepatic form of the disease.Knowing whether such patients have esophageal varices would be interesting.It is of interest to know whether these individuals bleed similarly to patients with HSS.Severe upper digestive bleeding in patients without esophageal varices but with duodenal varices has been reported 29,30 .An upper digestive endoscopy may partially address this issue in such patients.If this hypothesis is confi rmed, this clinical form of schistosomiasis may be classifi ed as hepatic with or without esophageal varices.
To confi rm the fi eld diagnosis, we brought 22 individuals with thrombocytopenia to our University Hospital in Belo Horizonte to undergo examination.Every week, two patients were admitted to the hospital for fi ve days.During the fi rst two weeks, we found one patient who presented a variable number of platelet counts: 35,000, 27,000, 93,000, 37,000, and zero.In addition, platelet clumping was observed in a peripheral blood smear.After changing the anticoagulant (EDTA to heparin), the platelet count increased to 160,000/mm 3 .The patient was diagnosed with pseudothrombocytopenia (an in vitro error) and was reallocated to the group with normal platelet counts.Pseudothrombocytopenia has been reported to occur in as many as 17% of patients who visit hematology units to seek a diagnosis for their thrombocytopenia [31][32][33][34][35][36] .We are planning a new study in rural areas of Brazil that are endemic for schistosomiasis to assess the frequency of pseudothrombocytopenia.Thrombocytopenia is a nuisance in the life of patients with HSS.Minor surgical, dental, and diagnostic procedures have been refused to these patients by misinformed healthcare professionals due to the possibility of bleeding because of low platelet counts.A percentage of such patients may have pseudothrombocytopenia, an artifact that can be easily avoided.
The major mechanisms associated with reduced platelet counts are the following: 1) decreased production (e.g., due to diseases or toxic drugs that affect the bone marrow); 2) increased destruction (caused by cytotoxic agents and other drugs, viral infections, and vitamin B12 or folic acid defi ciency); 3) dilutional thrombocytopenia (caused by blood transfusions with packed red blood cells); 4) distributional thrombocytopenia (caused by splenomegaly); 5) increased platelet destruction [caused by autoimmune diseases, use of heparin, and diseases such as infectious mononucleosis, cytomegalovirus, and acquired immune defi ciency syndrome (AIDS)].Therefore, all these situations should be kept in mind and considered when analyzing thrombocytopenia in areas endemic for schistosomiasis.
We concluded that in areas endemic for schistosomiasis with a prevalence of 23%, platelet counts <143,000/mm 3 identifi ed HSS patients with good sensitivity and moderate specifi city and can be used as a screening tool to identify severe disease in endemic areas.However, one question that remains to be answered is whether similar results will occur in areas with different schistosomiasis prevalences (5%, 10%, or 40%).

TABLE 1 -
Demographic and physical examination data of 384 individuals with and without hepatosplenic schistosomiasis in the district of Teófi lo Otoni, State of Minas Gerais, Brazil, who were examined during a week in July 2012.
FIGURE 1 -Flow chart of the study.Patients with spleens >13cm and liver periportal fi brosis were evaluated using ultrasound.HSS: hepatosplenic schistosomiasis.

TABLE 2 -
Areas under the ROC curves (AUCs), confi dence intervals, and sensitivity and specifi city of the ROC curves for a platelet cut-off point of <143,000/mm 3 for each defi nition of hepatosplenic schistosomiasis.