Comparison between clinical and ultrasonographic findings in cases of periportal fibrosis in an endemic area for schistosomiasis mansoni in Brazil

Introduction: Abdominal palpation and ultrasound findings among patients from an endemic area for schistosomiasis in Brazil who had been followed up for 27 years were compared. Methods: In 2004, 411 patients from Brejo do Espírito Santo, in the State of Bahia, were selected for the present investigation after giving their written informed consent. Based on clinical data, they were divided into three groups: 41 patients with evidence of liver fibrosis in 2004 (Group 1); 102 patients with evidence of liver fibrosis in the past (1976-1989) but not in 2004 (Group 2); and 268 patients without evidence of liver fibrosis at any time during the 27-year follow-up (Group 3). All of the patients underwent abdominal ultrasound in which the examiner did not know the result from the clinical examination. The data were stored in a database. Results: The prevalence of periportal fibrosis on ultrasound was 82.9%, 56.9% and 13.4% in Groups 1, 2 and 3, respectively. In the presence of hard, nodular liver or prominent left lobe and a hard palpable spleen, ultrasound revealed periportal fibrosis in 70.9%. However, periportal fibrosis was diagnosed using ultrasound in 25.4% of the patients in the absence of clinical evidence of liver involvement. Thus, ultrasound diagnosed periportal fibrosis 3.1 times more frequently than clinical examination did. Conclusions: Although clinical examination is important in evaluating morbidity due to Manson’s schistosomiasis in endemic areas, ultrasound is more accurate in diagnosing liver involvement and periportal fibrosis. Key-words: Schistosomiasis. Abdominal ultrasound. Clinical examination. Periportal fibrosis. RESUMO Introdução: Neste estudo, se comparou os achados da palpação abdominal e do ultrassom em pacientes de área endêmica de esquistossomose que foram acompanhados por 27 anos no Brasil. Métodos: Em 2004, 411 pacientes de Brejo do Espírito Santo, no estado da Bahia, após consentimento informado e por escrito foram selecionados para o presente estudo. Baseando-se no exame clínico eles foram divididos em 3 grupos: 41 (Grupo 1) com evidência de fibrose hepática no ano de 2004; 102 (Grupo 2) com evidência de fibrose hepática no passado (1976-1989) mas não em 2004; e 268 (Grupo 3) sem evidência de fibrose hepática em 27 anos de seguimento. Todos foram submetidos a exame ultrassonográfico do abdome em que o examinador não sabia o resultado do exame clínico. Os dados foram armazenados em banco de dados. Resultados: A prevalência de fibrose periportal ao ultrassom foi de 82,9%, 56,9% e 13,4% nos Grupos 1, 2 e 3, respectivamente. Na presença de fígado duro, nodular ou lobo esquerdo proeminente e baço palpável duro, o ultra-som revelou fibrose periportal em 70,9%. Porém, fibrose periportal foi diagnosticada através do ultrassom em 25,4% dos pacientes, na ausência de evidência clínica de envolvimento hepático. Assim, o ultrassom diagnosticou fibrose periportal 3,1 vezes mais frequentemente que o exame clínico. Conclusões: O exame clínico tem importância na avaliação da morbidade da esquistossomose mansônica em áreas endêmicas, mas o ultrassom mostra-se mais preciso quando se pretende diagnosticar o envolvimento hepático e a fibrose periportal. Palavras-chaves: Esquistossomose. Ultrassom abdominal. Exame clínico. Fibrose periportal. 1. Graduate Course on Tropical Medicine and Infectology, University of Triângulo Mineiro, Uberaba, Minas Gerais, Brazil. 2. Graduate Course on Health Sciences: Infectology and Tropical Medicine, School of Medicine, Federal University of Minas Gerais, Belo Horizonte, MG, Brazil. Address to: Dra Raiza Ruiz-Guevara. Av San Juan Bosco, Ed. Belveder A-62 Altamira, Caracas 1062, Venezuela. Ph: 58-212 2652559, Fax: 58-212 6930454. e-mail: raizaruiz@yahoo.com Received in 11/09/2009 Accepted in 05/03/2010 Periportal fibrosis, first described by Symmers in 19041, can produce portal hypertension and the formation of gastroesophageal varices. Rupture of these varices causes digestive hemorrhage, the most frequent and feared complication of hepatosplenic Manson’s schistosomiasis. The hepatosplenic form is the usual clinical counterpart of periportal fibrosis. In clinical-epidemiological studies conducted in endemic areas, morbidity used to be evaluated only on the basis of physical examination of the abdomen. However, there is no consensus about the clinical parameters of the liver and spleen that should be considered in the evaluation. Hepatomegaly and splenomegaly have no uniform definition. The limit for determining that an increase in liver volume has occurred remains a debatable issue. Some investigators consider hepatomegaly to be present when the organ exceeds the costal margin by 5cm along the sternal line or by 4cm along the midclavicular line2. Others accept enlargement to be present if the liver is palpable 2cm below the costal margin3 or simply if the liver is palpable4, without mentioning whether this organ is palpable at rest or during inspiration. Yet other investigators attribute more importance to liver characteristics such as hard consistency, presence of nodules on the liver surface or prominence of the left lobe5. Regarding the spleen, palpability would be sufficient for some physicians6-8, whereas others believe that the spleen should be palpable at the costal margin or beyond it9. Thus, the diagnosis of hepatosplenic schistosomiasis varies according to the different criteria mentioned. The routine use of abdominal ultrasound examination has been of great help in studying schistosomiasis because it enables identification of periportal fibrosis10-14. It also provides other important information such as sound evidence of portal hypertension and can measure or better estimate organ sizes15-18.

Periportal fibrosis, first described by Symmers in 1904 1 , can produce portal hypertension and the formation of gastroesophageal varices.Rupture of these varices causes digestive hemorrhage, the most frequent and feared complication of hepatosplenic Manson's schistosomiasis.The hepatosplenic form is the usual clinical counterpart of periportal fibrosis.In clinical-epidemiological studies conducted in endemic areas, morbidity used to be evaluated only on the basis of physical examination of the abdomen.However, there is no consensus about the clinical parameters of the liver and spleen that should be considered in the evaluation.Hepatomegaly and splenomegaly have no uniform definition.The limit for determining that an increase in liver volume has occurred remains a debatable issue.
Some investigators consider hepatomegaly to be present when the organ exceeds the costal margin by 5cm along the sternal line or by 4cm along the midclavicular line 2 .Others accept enlargement to be present if the liver is palpable 2cm below the costal margin 3 or simply if the liver is palpable 4 , without mentioning whether this organ is palpable at rest or during inspiration.Yet other investigators attribute more importance to liver characteristics such as hard consistency, presence of nodules on the liver surface or prominence of the left lobe 5 .Regarding the spleen, palpability would be sufficient for some physicians [6][7][8] , whereas others believe that the spleen should be palpable at the costal margin or beyond it 9 .Thus, the diagnosis of hepatosplenic schistosomiasis varies according to the different criteria mentioned.
The routine use of abdominal ultrasound examination has been of great help in studying schistosomiasis because it enables identification of periportal fibrosis [10][11][12][13][14] .It also provides other important information such as sound evidence of portal hypertension and can measure or better estimate organ sizes [15][16][17][18]

METHODS
The present study was designed to compare clinical and abdominal ultrasound findings in order to determine: a) whether clinical alterations suggestive of liver and/or spleen involvement in schistosomiasis are confirmed by ultrasound; b) the association between periportal fibrosis and other ultrasound findings; and c) the agreement between organ sizes measured by clinical examination and by ultrasound.

Study population
A cross-sectional study was conducted in Brejo do Espírito Santo in 2004, a rural community in the municipality of Santa Maria da Vitória, Bahia; this area does not present malaria transmission but is considered hyperendemic for schistosomiasis.A total of 3,766 people have been followed up since 1976 by means of clinical and fecal examinations (Kato-Katz technique) 19 , and control measures have been implemented.

Physical examination
The same methodology for clinical evaluation was used on all occasions.The right hepatic lobe was examined on the anterior axillary line and the left lobe on the line through the xiphoid appendix.If it exceeded the costal margin, it was measured in centimeters.When the liver was palpable, it was determined whether its consistency was flaccid (lip consistency), slightly hardened (nose tip consistency) or hard (bone consistency), and whether its surface was smooth or nodular.Prominence of the left lobe was defined when this lobe was clearly larger than the right lobe 5 .The spleen was palpated and measured under the costal margin and its consistency was determined.We defined splenomegaly as present only when the spleen was palpable below the left costal margin.The following clinical forms were described: hepatointestinal (palpable liver), hepatointestinal with advanced hepatic lesions (at least one of the following: hard, nodular or prominent left liver lobe), hepatosplenic (hepatointestinal with advanced liver lesions plus splenomegaly) and postsurgical hepatosplenic (histologically proven periportal fibrosis confirmed during surgery) 9 .
Patients with hepatosplenic or hepatointestinal schistosomiasis with advanced hepatic lesions were assumed to have periportal fibrosis and therefore severe clinical disease 9 .Regarding hepatosplenic patients who underwent surgery, only those with hard and/or nodular liver and/or prominent left lobe were assumed to have liver fibrosis.
Based on clinical examinations performed by one of us (AP) just before the field visit of the present study (2004), the patients who were invited to participate in this study were divided into three groups: 41 patients with clinical evidence of liver disease (Group 1); 102 patients with evidence of liver disease in the past (1976-1989)  but not in 2004 (Group 2); and 268 patients without evidence of liver disease at any time during the 27-year follow-up (Group 3).

Abdominal ultrasound
The abdominal ultrasound examination was performed using a GE LOGIQ 100 instrument with a convex 3.5 MHz transducer.We used indicators of periportal fibrosis proposed by the Niamey Working Group (Niamey-Belo Horizonte classification) 20,21 .In accordance with this protocol, periportal fibrosis was classified in degrees (incipient, possible, probable, definite, advanced and advanced with portal hypertension).The following other parameters were also evaluated: hepatic contours, left lobe hypertrophy, right lobe atrophy, splenomegaly and thickening of the gallbladder wall.The measurements (walls of the intrahepatic portal branches, caliber of the portal vein, left lobe, right lobe and spleen) were adjusted for the patient's height 16 .

Statistical analysis
The data were analyzed using the SPSS program for Windows (Version 10.0; Copyright® SPSS Inc., 1989-1999).After performing descriptive analysis of normality and homogeneity of variance, the data were analyzed using the Student's t, Pearson and McNemar chisquare tests.Sensitivity and specificity were estimated.The kappa (k) concordance index was using as an estimate of reproducibility.In all cases, the null hypothesis was rejected when a was less than 5% (p < 0.05).

Etical
All clinical, laboratory and ultrasound examinations were performed following a masked protocol.The study protocol was approved by the Research Ethics Committees of the University of the Triângulo Mineiro, Federal University of Minas Gerais and the Brazilian National Research Ethics Committee (CONEP).
A total of 411 patients were studied (199 men and 212 women).The mean ages in years for groups 1, 2 and 3 were: 50.3 ± 12.6, 50.5 ± 12.9 and 39.8 ± 20.5, respectively.Group 3 volunteers were significantly younger than the individuals in groups 1 and 2.
Table 1 shows that, from 1976 to 2004, there were marked decreases in the prevalence of S. mansoni infection and in the parasitic load, as well as a significant decrease in the proportion of patients with hepatosplenic schistosomiasis.

Abdominal ultrasound
Periportal fibrosis was detected in 128 (31.1%) out of the 411 participants.Most of the patients with severe fibrosis were in Group 1 (Figure 1).The mean ages for patients with and without fibrosis were 52 and 34 years respectively (p < 0.01).There was no gender difference.
The liver and spleen sizes of the different groups are presented in Table 3. Ultrasound showed irregular hepatic contours in 11 patients (3%) and gallbladder wall thickening in 114 (29.2%).

Concordance between clinical and abdominal ultrasound examination
There was no relation between palpable right (192/266) or left (153/228) liver lobes and periportal fibrosis on ultrasound.Periportal fibrosis was present in 50 (36.8%) of the 136 patients with a non-palpable right lobe, and in 74 (27.8%) of the 266 with a palpable right liver lobe.Hence, periportal fibrosis was more frequent when the right liver lobe was non-palpable (p = 0.001).
There was a statistically significant association between presence of a palpable left liver lobe and periportal fibrosis (p = 0.001), and Periportal fibrosis was present in 61 (24.2%) out of the 252 patients with a non-palpable spleen, and in 24 (22.9%) of the 105 patients with a spleen that was only palpable during deep inspiration.Interestingly, 16 (72.7%)out of the 22 patients presenting enlarged spleen had periportal fibrosis (p<0.004) and portal hypertension (p<0.004).A hardened spleen was also associated with periportal fibrosis (16 out of 17 patients) and portal hypertension (14 out of 17 patients) (p<0.001).

Comparison between other ultrasound findings and periportal fibrosis
Thirty-seven (66.1%) out of the 56 patients with right liver lobe atrophy had periportal fibrosis, as opposed to 89 (25.5%) out of the 349 patients with normal right liver lobe (p < 0.0001).This was not observed for the left liver lobe.
Twenty-three (60.5%) out of the 38 patients with splenomegaly diagnosed using ultrasound had periportal fibrosis, as opposed to 84 (24.1%) out of the 348 without splenomegaly (p < 0.001).Splenomegaly was also associated with portal hypertension (Table 3).
Irregular hepatic contours and thickening of the gallbladder wall were associated with periportal fibrosis (Table 3).A good agreement was found between gallbladder wall thickening and periportal fibrosis (kappa = 0.8).

Comparison of liver and spleen size determined on clinical examination and ultrasound
Twenty-eight (50.9%) out of the 55 patients who had right liver lobe atrophy on ultrasound had a palpable liver.Conversely, 50 (30.5%)out of the 164 patients with an enlarged left lobe on ultrasound had a non-palpable left liver lobe.
Among the 125 patients with a palpable spleen, ultrasound confirmed the presence of splenomegaly in 31 (24.8%).On the other hand, six (15.8%) out of the 38 patients with splenomegaly on ultrasound had a non-palpable spleen.

DISCUSSION
Palpable right liver lobe and palpable spleen are not good indicators of periportal fibrosis in field-based work.Among the cases of a prominent, hard nodular palpable liver, ultrasound revealed periportal fibrosis in 70.9%.It is worth noting that periportal fibrosis occurred in 25.4% of the patients without clinical evidence of liver involvement.In addition, ultrasound diagnosed periportal fibrosis 3.1 times more frequently than clinical examination would suggest.Even in the cases with severe periportal fibrosis on ultrasound, with or without portal hypertension, ultrasound diagnosed 2.1 times more cases than clinical examination did.
Regarding the size of the right liver lobe, there was more fibrosis when this lobe was non-palpable and even more when the lobe was found to be atrophic by ultrasound, as has been previously reported 2,11 .
With reference to the size of the left hepatic lobe, Houston et al 22 observed that hepatomegaly showed high specificity (94%), but low (28%) sensitivity as a fibrosis marker.It is common knowledge that the liver size decreases with increasing degree of fibrosis 15,[22][23][24] .Doehring-Schwedtsfeger et al 25 found no relationship between liver size and periportal fibrosis on ultrasound.In the present study, the association between left liver lobe beyond the costal margin and periportal fibrosis was important only for Group 1, when the left lobe was also prominent.Thus, we do not think that increased left lobe size alone is a good clinical marker of hepatic fibrosis.When analyzing left lobe hypertrophy on ultrasound, Cerri et al 11 and Kardorff et al 2 detected correlation with portal fibrosis, a finding that was not confirmed in our study.
A distinction should be made between small and large spleen sizes.Only in individuals with large spleens has a relationship with liver fibrosis been found.Thus, in endemic areas, the importance of small palpable spleens is uncertain; it may or may not be related to schistosomiasis (initial stages of the disease).Other causes of splenomegaly should be investigated, as reported by Lambertucci et al 26 .
In our study, we also detected a relationship between clinical splenomegaly and the WHO fibrosis patterns 25,[27][28][29][30][31][32] .In addition, patients with severe periportal fibrosis without splenomegaly were also found 30,33 .Splenomegaly diagnosed using ultrasound was related both to the presence of the WHO patterns and of periportal fibrosis, as already described 11,20 .
The liver was found by ultrasound to have irregular contours in individuals who had current or past clinical evidence suggestive of periportal fibrosis.However, the absence of irregular contours did not rule out periportal fibrosis.Their presence, as well as gallbladder wall thickening, correlated with periportal fibrosis but not with portal hypertension.This finding has been reported before 2,14,20,25 .In one study, the finding of gallbladder wall thickening was also suggestive of non-reversal of liver fibrosis 31 .
No agreement was found between ultrasound and clinical examination regarding organ size, as already reported 34 .One advantage of the Niamey-Belo Horizonte protocol is the adjustment of the measurements to the patients' height 21 .Nonetheless, we occasionally still detected periportal thickening or dilatation of the portal vein in individuals with no qualitative signs of fibrosis on ultrasound (WHO patterns).
Although ultrasound is superior to clinical examination for detection of periportal fibrosis, it does not allow evaluation of important data such as liver and spleen consistency.On the other hand, the subjectivity of clinical examination was unexpectedly demonstrated in the present study: the prevalence of hepatointestinal schistosomiasis with advanced liver disease increased from 4% in 1976 to 14.8% in 1980 and that of hepatosplenic schistosomiasis from 8% to 10.7% (Table 1).After a careful investigation, it was noted that a different physician was responsible for the clinical examination in the study area in 1980.
Brejo do Espírito Santo has not been a hyperendemic area for schistosomiasis since the implementation of control measures, which started in 1976.The current clinical and ultrasound findings, many of them slight and some in a regression phase, represent the sequelae from lesions that previously were severe.These were established in the past when S. mansoni transmission was high.The implementation of effective measures has changed the disease profile in this area, along with the emergence of hepatosplenic schistosomiasis.However, it is important to emphasize that severe disease persists even after achieving a good control of transmission 31 .
In conclusion, although clinical examination is still important in evaluating morbidity due to Manson's schistosomiasis in endemic areas, ultrasound is more accurate in diagnosing liver involvement and periportal thickening.

Prata A et al -
Clinical and ultrasonograpic diagnosis of schistosomiasis mansoni in Brazil

TABLE 3 -Results from clinical examination and abdominal ultrasound and their relationship with periportal fibrosis and portal hypertension, Brejo do Espírito Santo, State of Bahia, Brazil, 2004.
fibrosis was also found in seven (87.5%) out of the eight patients with nodular liver and in 78 (27.6%) out of the 283 patients with smooth liver surface.Periportal fibrosis was detected in 19 (70.4%) out of the 27 patients with a prominent left lobe and in 107 (28.5%) out of the 375 patients without a prominent left lobe.All these differences were statistically significant (p=0.001).These associations were not observed with portal hypertension.