- Email: [email protected]
Chronic splenomegaly in Nairobi, Kenya. II. Portal hypertension
107
TRANSACTIONS OF THE ROYAL SOCIETY OF TROPICAL MEDICINE AND HYGIENE (1987) 81, 107-110
Chronic
splenomegaly
in Nairobi,
Kenya.
II. Portal hypertension
KEVIN M. DE COCK’, S. AWADH’, R. S. RAJA’, B. M. WANKYA’, R. A. JUPP~,B. SLAVING,T. SIONGOK’, P. H. REES’, J. BERTRAND AND S. B. LUCAS ‘DDept. of Medicine, ‘Dept. of Diagnostic Radiology, 3Dept. of Human Pathology, Kenyatta National P.O. Box 30588, Nairobi, Kenya; 4Dept. of Chemical Pathology, 6Dept. of Virology, St. Thomas’ London, SE1 7EH; ‘Division of Disease Control and Research, Ministry of Health, Nairobi,
K. ARAP Hospital, Hospital, Kenya
Abstract Eighty-five patients with chronic splenomegaly and proven oesophagealvarices were studied at Kenyatta National Hospital, Nairobi. The major defined groups were hepatosplenic schistosomiasis (24%), cirrhosis (20%) and portal vein occlusion (11%). Hyper-reactive malarial splenomegaly (tropical splenomegaly syndrome) was considered as the cause of oesophagealvarices in only one patient. In 26% of cases liver biopsy was non-diagnostic and the extrahepatic portal vein was demonstrated radiologically to be patent. Such patients were thought to be suffering from idiopathic portal hypertension, not previously described elsewhere in Africa. Hepatitis B surface antigen was detected in 12% of controls and in 58% of patients with cirrhosis (p
Betweenlate 1979andearly 1982,85African patientswith oesophagealvarices were studied at Kenyatta National Hospital. A full history, including geographical and tribal
background, was obtained and physical examination performed in all cases. Varices were confirmed by endoscopy alone using an Olympus GIFK P2 endviewing gastroduodenoscopein 14 patients, by barium swallow alone in 37 patients, and in 34 patients by both investigations. Transsplenic portal venography was performed in standard fashion (SHERLOCK,1981) in 45 cases. In addition, all patients had the following tests performed: full blood count and blood film examination; prothrombin time estimation; measurement of serum bilirubin and albumen; stool examination by the Kato technique; rectal snip examination for ova of
Correspondence too: Dr. Kevin M. De Cock, Special Pathogens Branch, Division of Viral Diseases, Centers for Disease Control, 1600 Clifton Road N.E., Atlanta, GA 30333, USA.
munsoni; and liver biopsy, using either a Menghini or Tru-cut needle. Specimens were processed in standard fashion (SCHEUER,1980), the staining methods used being haematoxylin and eosin, Perl’s prussian blue for iron and Gordon and Sweet’s silver impregnation for reticulin fibres. 56 patients had sera stored at -80°C for subsequent estimation of immunoglobulin levels and examination for markers of hepatitis B virus (HBV) infection. Immunoglobulins were estimated by kinetic turbidimetry (DEVERILL, 1980).Commercially available kits were used for detection of HBV markers (Abbott Laboratories Ltd., Chicago, Illinois, USA). All sera were tested for hepatitis B surface antigen (HBsAg) by solid phase radioimmunoassay (Ausria II). Negative sera were subsequently tested for antibody to HBsAg (anti-HBs) by an analogeas technique (Ausab). Finally, sera negative for HBsAg or anti-HBs were tested for antibody to hepatitis B core antigen (anti-HBc) by an enzyme immunoassay (Corzyme). Sera were available from 82 rural adult controls who showed no clinical evidence of infection or liver disease(DE CGJZKet al., 1987). These were tested in similar fashion for markers of HBV infection.
Schistmom
Results On the basis of liver histology and results of splenoportovenography, patients were allocated to six diagnostic groups; hepatosplenic schistosomiasis (HS); cirrhosis; extrahepatic portal vein occlusion; patients with non-diagnostic biopsies in whom portal venography demonstrated a patent extrahepatic portal vein; patients with non-diagnostic biopsies in whom portal venography could not be performed; and hyper-reactive malarial splenomegaly (for criteria, see DE COCK, 1983, and DE COCKet al., 1987). In one patient the liver biopsy was inadequate for interpretation. The numbers of patients in eachgroup are shown in Table 1. This table also shows the number of Datients in each group with ova of S . mansoni in stool or rectal
snip. 72% of casesof histologically diagnosed HS had ova detected, compared with 20% or less of other groups. Of the 60 patients in the study with nonschistosomal liver disease,eight had positive stool or snip examinations, significantly less than the proportion of positives in the histologically diagnosed cases
108
CHRONIC
SPLENOMEGALY
of HS (x2 = 25.9 on 1 degree of freedom; Yates correction applied; p
No. of man.& in patients stool snip
Diagnosis Hepatosplenic schistosomiasis
25
18
Cirrhosis
17
2
Portal vein occlusion
9
0
Non-diagnostic biopsy: portal vein patent
22
3
Non-diagnostic biopsy: splenoportovenogrsm not performed
10
2
Hyper-reactive malarial splenomegaly
1
0
Inadequate biopsy
1
Table
2-Prevalence
bilirubin
II.
(1.3 mgdl + 0.58) than those in the undi-
agnosedgroup (0.73 mg/dl f 0.25), as well asa lower mean serum albumen, 33.8 g/l + 8.3 compared with
44.5 g/l & 7.39 respectively (p
of HBV markers in major diagnostic
Cirrhosis W)
;;
KENYA.
This study documents two significant findings; firstly, the importance of schistosomiasisas a causeof
85
Total
IN
groups of portal hypertension
Non-diagnostic biopsy portal vein patent (%)
and controls
Controls (%)
HBsAg 4, Anti-HBs &
Total tested
4)
(t!;
3** (25)
10** (67)
(&
0
0
Anti-HBc+ as only marker Total with any marker
(5;;
(76)
(h 2) (I&
(k)
(6)
(&
(I&
+ N.B. Anti-HBc only tested for in sera negative for HBsAg and anti-HBs. *p
(Z)
K. hi. DE cow
portal hypertension in Kenya, and secondly, the occurrence of so-called “idiopathic portal hypertension”. In addition, data are provided concerning HBV infection in patients with liver disease and controls. 32 patients, 37% of the total, had non-diagnostic liver biopsies. In 10 such cases splenic venography was not carried out, so portal vein occlusion could not be strictly excluded, -ad they are not considered further. In the other 22 patients (26%), the portal vein was radiologically demonstrated to be patent. Careful consideration has been given to the possibility that liver biopsy sampling errors led to under-diagnosis of soecific conditions such as HS or cirrhosis. Since the frequency of positive stool or snip examinations was significantly greater in patients with histolagically diagnosed HS than in all other groups, it is unlikely that HS was the cause of portal hypertension in patients with non-diagnostic biopsies. Cirrhosis is also unlikely to have been missed to any great degree. Patients with non-diagnostic biopsies and patent portal veins suffered significantly more episodes of haematemesisand had a significantly lower prevalence of ascites than cirrhotics. In patients with cirrhosis, mean serum bilirubin, IgG and IgA levels were significantly higher. An uncontrolled observation which impressed us was the tolerance shown by some of the undiagnosed patients for repeated variceal bleeding. More complications, including deaths, might have been expected if such patients had been suffering from cirrhosis. Although careful follow-up and autopsy studies are required for completed histological evaluation, we believe that many of these patients with non-diagnostic biopsies were suffering from non-cirrhotic, nonschistosomal, intrahepatic portal hypertension, most often referred to as idiopathic portal hypertension (IPH). It is unlikely that any other specific noncirrhotic conditions, such as nodular regenerative hyperplasia or parual nodular transformation, were responsible. We emphasize that IPH is an entirely distinct condition from hyper-reactive malarial splenomegaly, or so-called tropical splenomegaly. Our patients with non-diagnostic biopsies did not have hepatic sinusoidal lymphocytosis, and their malarial serology and IgM levels did not differ from those of geographically matched controls (De Cock, unpublished). We do not believe that hyper-reactive malarial splenomegaly is an important cause of oesophageal varices in Kenya, although. portal vein presstire may be elevated in this condition (WILLIAMS et al.. 1968). IPH has been described h different parts of the world under a variety of names. Non-cirrhotic portal fibrosis (SAMA et al., 1971), hepatoportal sclerosis (MIKKELSEN et al.. 19651. obliterative wrtal venopathy (NAYAK & ‘RAMALINGASWAMI, i969; IBER, 1969), and benign intrahepatic portal hypertension (KINGHAM et al., 1981; LEVISON et al., 1982) are all terms referring to a non-cirrhotic, non-schistosomal form of intrahepatic portal hypertension that may represent a spectrum of pathology rather than one condition. The importance of the present work is that, apart from our preliminary reports (DE COCK et al., 1982 a,b; 1983), IPH does not appear to have been previously documented in Africa. It is possible that many such caseshave previously been dismissed as examples of obscure splenomegaly. The aetiology of IPH remains uncertain, but the striking preponder-
109
et al.
ance of Kamba patients suggeststhat an environmental causeshould be searched for in Machakos District. The importance of HBV infection in Kenya is demonstrated by the finding of some viral marker in 80% of control subjects. The prevalence of HBsAg was significantly higher in patients with cirrhosis than in controls. Previous work in Kenya (PETTIGREW et al., 1976) has demonstrated an association between HBV infection and chronic liver disease. In the present study, 27% of patients with non-diagnostic biopsies and patent portal veins were HBsAg positive, compared with 12% of controls. This could result from inclusion in the non-diagnostic group of some unrecognized cases of cirrhosis. An alternative explanation, however, is that this increased prevalence of HBsAg results from the greater number of blood transfusions which the undiagnosed patients required. Equal proportions of patients with cirrhosis as with non-diagnostic biopsies had some marker of HBV infection, but the prevalence of anti-HBs was significantly greater in the latter. Although equally exposed, therefore, the reaction to HBV was strikingly different in these two groups. Portal or splenic vein occlusion was found in 11%of casesof portal hypertension. We disagree with earlier reports by MILLER (1967; 1972) and BAGSHAWE & CAMERON (1976), that obstruction of the portal vein is the major cause of variceal haemorrhage in Kenya. Nevertheless, there can be little doubt that portal vein occlusion is more common in the tropical developing world than in industrialized countries. The single major cause of portal hypertension identified in this study was HS, accounting for almost one third of cases.Haemorrhage was well tolerated, and biochemical liver tests were more favourable than in cirrhosis. Both cirrhosis and HS were sipnificantlv more frequent in Kamba patients from M>chakos -District than expected from the pattern of general medical admissions. The severity of schistosomal disease is generallv related to intensitv of infection. In Kenva. prevaleice and intensity of
We are grateful to Mrs J. Clarkson,London Schoolof Hygiene and Tropical Medicine, for statistical advice; to Professor Michael Hutt for discussion; to Drs May Ho and
Harrison Spencer for a number of control sera; to the staff of the Walter Reed Institute, Nairobi, for storage of serum; to Drs Sarah Thomas and Virinder Taiwar for help with endoscopy and radiology; fo Rita Musick and Ruth Nao for secretarial assistance; and to the Wellcome Trust for financial support. References Bagshawe, A. F. & Cameron, H. M. (1974). Idiouathic tropical splenomegaly and portal hypertension. In: Health and Disease in Kenya. Vogel, L. C., Muller, A. S.,
CHRONIC
110
SPLENOMEGALY
Odingq, R. S., Onyango, Z. & de Gem, A. (editors). Nairobi: East African Literature Bureau, pp. 445-452. De Cock, K. M., Awadh, S., Raja R. S., Wankya, B. M. & Lucas, S. B., (1982a). Esophageal varices in Nairobi, Kenya: a study of 68 cases.AmericanJournal of Tropical Medicine and Hygiene, 31, 579-588. De Cock, K. M., Awadh, S., Raja, R., Wankya, B. M. & Lucas, S. B. (1982b). Portal hypertension in Nairobi, Kenya. Gut, 23, A446. De Cock, K. M., Awadh, S., Raja, R. S., Wankya, B. M., Lillywhite, J. E., Hodgen, N., Bertrand, J., Jupp, R. & Lucas, S. B. (1983). Further evidence of idiopathic portal hypertension in Kenya: a study of hepatitis B markers, immunoglobulin levels and enzyme-linked immunosorbent assayfor Schiswsoma mansoni in patients and controls. Gut, 23, A483-484. De Cock, K. M., Hodgen, A. N., Lucas, S. B., Jupp, R. A. Slavin, B., Arap Siongok, T. K. & Rees, Ph. H. (1987). Chronic splenomegaly in Nairobi, Kenya. I. Epidemiology, malarial antibody and immunoglobulin levels. Transactions of the Royal Society of Tropical Medicine and Hygiene, 81, 100-106.
Deverill! I. (1980). Kinetic measurement of the immunoprecipitm reaction using the centrifugal analyser. In: Centrifugal Analysers in Clinical Chemisrry.Price, C. P. & Spencer, K. (editors). Eastbourne, UK: Praeger. Hansen, D. P. & Daily, D. S. (1978). A fibre-endoscopic study of acute upper gastrointestinal haemorrhage in Nairobi. Kenya. American 7ournal of T&Cal Medicine . and Hygiene, -27, 197-200.Highton, R. B. (1979). Schistosomiasis. In: Health and Disease in Kenya. Vogel, L. C., Muller, A. S., Odingo, R. S., Onyango, Z. & de Geus, A. (editors). Nairobi: East African Literature Bureau, pp. 347-355. Iber, F. I. (1969). Obliterative portal venopathy of the liver and “idiopathic portal hypertension”. Annals of Internal Medicine,
71, 660-661.
Kingham, J. G. C., Levison, D. A., Stansfield, A. G. & Dawson, A. M. (1981). Non-cirrhotic intrahepatic portal hypertension: a long term follow-up study. Quarter.$ 3ournul of Medicin.?, 50, 259-268. Levison, D. A., Kingham, J. G. C., Dawson, A. M. & Stansfield, A. G. (1982). Slow cirrhosis - or no
IN
KENYA.
II.
cirrhosis? A lesion causing benign intrahepatic portal hypertension. Joumal of Pathology? 137, 233-272. Linsell, C. A. (1967). Cancer incidence m Kenva 1957-1963. British 7oukal
of Cancer. 21. 665-673.
Mikkelseq- W. P.,*Edmondsonj H. A., Peters, R. L., Redeker, A. G. & Reynolds, T. B. (1965). Extra and intrahepatic portal hypertension without cirrhosis (hepatoportal sclerosis). Annals of Surgery, 162, 602-618. Miller, J. R. M. (1967). Portal hypertension in Nairobi. East African Medical 3owna1, 44, 376-386. Miller. I. R. M. (1972). Portal hvoertension. East African Medical Journal,
44, 243-245 . A
Nayak, N. C. & Ramalingaswami, V. (1969). Obliterative portal venopathy of the liver associated with so-called idiopathic portal hypertension or tropical splenomegaly. Archives of Pathology, 87, 359-369.
Pettigrew, N. M., Bagshawe, A. F., Cameron, H. M., Cameron, C. H., Dorman, J. M. St MacSween, R. N. M. (1976). Hepatitis B surface antigenaemia in Kenyans with chronic liver disease. Transactions of the Royal Society of Tropical Medicine and Hygiene, 70, 462-465. Sama, S. K., Bhargava, S., Gopi Nath, N., Talwar, J. R., Nayak? N.C., Tandon, B. N. & Wig, K. L. (1971). Non-cirrhotic portal fibrosis. American Journal of Medicine, 51, 160-169. Scheuer, P. J. (1980). Liver biopsy interpretation (3rd edit.). London: Balliere Tindall. Shah, M. V. (1978). Chronic liver disease in Kenya. M.Med. Thesis. Universitv of Nairobi. Sherlock, S. (1981). i)iseasesof the Liver and Biliary System (6th edit.). Oxford: Blackwell. Thomas. SE.. Mwaungulu. G. S.. Wankva. B. M. & De Cock. K. ’ M. (1983). . Acute- uaoer- aastrointestinal haemorrhage at Kenyatta Nationi kos&al, Kenya: a prospective study. East African Medical .7ournal, 60, 428-43 1.
Williams, R., Parsonson,A., Somers, K. & Hamilton, P. J. S. (1968). Portal hypertension in idiopathic tropical splenomegaly. Lancet, i, 329-333.
Accepted for publication
28th November,
1985
TRANSACTIONS OF THE ROYAL SOCIETY OF TROPICAL MEDICINE AND HYGIENE (1987) 81, 107-110
Chronic
splenomegaly
in Nairobi,
Kenya.
II. Portal hypertension
KEVIN M. DE COCK’, S. AWADH’, R. S. RAJA’, B. M. WANKYA’, R. A. JUPP~,B. SLAVING,T. SIONGOK’, P. H. REES’, J. BERTRAND AND S. B. LUCAS ‘DDept. of Medicine, ‘Dept. of Diagnostic Radiology, 3Dept. of Human Pathology, Kenyatta National P.O. Box 30588, Nairobi, Kenya; 4Dept. of Chemical Pathology, 6Dept. of Virology, St. Thomas’ London, SE1 7EH; ‘Division of Disease Control and Research, Ministry of Health, Nairobi,
K. ARAP Hospital, Hospital, Kenya
Abstract Eighty-five patients with chronic splenomegaly and proven oesophagealvarices were studied at Kenyatta National Hospital, Nairobi. The major defined groups were hepatosplenic schistosomiasis (24%), cirrhosis (20%) and portal vein occlusion (11%). Hyper-reactive malarial splenomegaly (tropical splenomegaly syndrome) was considered as the cause of oesophagealvarices in only one patient. In 26% of cases liver biopsy was non-diagnostic and the extrahepatic portal vein was demonstrated radiologically to be patent. Such patients were thought to be suffering from idiopathic portal hypertension, not previously described elsewhere in Africa. Hepatitis B surface antigen was detected in 12% of controls and in 58% of patients with cirrhosis (p
Betweenlate 1979andearly 1982,85African patientswith oesophagealvarices were studied at Kenyatta National Hospital. A full history, including geographical and tribal
background, was obtained and physical examination performed in all cases. Varices were confirmed by endoscopy alone using an Olympus GIFK P2 endviewing gastroduodenoscopein 14 patients, by barium swallow alone in 37 patients, and in 34 patients by both investigations. Transsplenic portal venography was performed in standard fashion (SHERLOCK,1981) in 45 cases. In addition, all patients had the following tests performed: full blood count and blood film examination; prothrombin time estimation; measurement of serum bilirubin and albumen; stool examination by the Kato technique; rectal snip examination for ova of
Correspondence too: Dr. Kevin M. De Cock, Special Pathogens Branch, Division of Viral Diseases, Centers for Disease Control, 1600 Clifton Road N.E., Atlanta, GA 30333, USA.
munsoni; and liver biopsy, using either a Menghini or Tru-cut needle. Specimens were processed in standard fashion (SCHEUER,1980), the staining methods used being haematoxylin and eosin, Perl’s prussian blue for iron and Gordon and Sweet’s silver impregnation for reticulin fibres. 56 patients had sera stored at -80°C for subsequent estimation of immunoglobulin levels and examination for markers of hepatitis B virus (HBV) infection. Immunoglobulins were estimated by kinetic turbidimetry (DEVERILL, 1980).Commercially available kits were used for detection of HBV markers (Abbott Laboratories Ltd., Chicago, Illinois, USA). All sera were tested for hepatitis B surface antigen (HBsAg) by solid phase radioimmunoassay (Ausria II). Negative sera were subsequently tested for antibody to HBsAg (anti-HBs) by an analogeas technique (Ausab). Finally, sera negative for HBsAg or anti-HBs were tested for antibody to hepatitis B core antigen (anti-HBc) by an enzyme immunoassay (Corzyme). Sera were available from 82 rural adult controls who showed no clinical evidence of infection or liver disease(DE CGJZKet al., 1987). These were tested in similar fashion for markers of HBV infection.
Schistmom
Results On the basis of liver histology and results of splenoportovenography, patients were allocated to six diagnostic groups; hepatosplenic schistosomiasis (HS); cirrhosis; extrahepatic portal vein occlusion; patients with non-diagnostic biopsies in whom portal venography demonstrated a patent extrahepatic portal vein; patients with non-diagnostic biopsies in whom portal venography could not be performed; and hyper-reactive malarial splenomegaly (for criteria, see DE COCK, 1983, and DE COCKet al., 1987). In one patient the liver biopsy was inadequate for interpretation. The numbers of patients in eachgroup are shown in Table 1. This table also shows the number of Datients in each group with ova of S . mansoni in stool or rectal
snip. 72% of casesof histologically diagnosed HS had ova detected, compared with 20% or less of other groups. Of the 60 patients in the study with nonschistosomal liver disease,eight had positive stool or snip examinations, significantly less than the proportion of positives in the histologically diagnosed cases
108
CHRONIC
SPLENOMEGALY
of HS (x2 = 25.9 on 1 degree of freedom; Yates correction applied; p
No. of man.& in patients stool snip
Diagnosis Hepatosplenic schistosomiasis
25
18
Cirrhosis
17
2
Portal vein occlusion
9
0
Non-diagnostic biopsy: portal vein patent
22
3
Non-diagnostic biopsy: splenoportovenogrsm not performed
10
2
Hyper-reactive malarial splenomegaly
1
0
Inadequate biopsy
1
Table
2-Prevalence
bilirubin
II.
(1.3 mgdl + 0.58) than those in the undi-
agnosedgroup (0.73 mg/dl f 0.25), as well asa lower mean serum albumen, 33.8 g/l + 8.3 compared with
44.5 g/l & 7.39 respectively (p
of HBV markers in major diagnostic
Cirrhosis W)
;;
KENYA.
This study documents two significant findings; firstly, the importance of schistosomiasisas a causeof
85
Total
IN
groups of portal hypertension
Non-diagnostic biopsy portal vein patent (%)
and controls
Controls (%)
HBsAg 4, Anti-HBs &
Total tested
4)
(t!;
3** (25)
10** (67)
(&
0
0
Anti-HBc+ as only marker Total with any marker
(5;;
(76)
(h 2) (I&
(k)
(6)
(&
(I&
+ N.B. Anti-HBc only tested for in sera negative for HBsAg and anti-HBs. *p
(Z)
K. hi. DE cow
portal hypertension in Kenya, and secondly, the occurrence of so-called “idiopathic portal hypertension”. In addition, data are provided concerning HBV infection in patients with liver disease and controls. 32 patients, 37% of the total, had non-diagnostic liver biopsies. In 10 such cases splenic venography was not carried out, so portal vein occlusion could not be strictly excluded, -ad they are not considered further. In the other 22 patients (26%), the portal vein was radiologically demonstrated to be patent. Careful consideration has been given to the possibility that liver biopsy sampling errors led to under-diagnosis of soecific conditions such as HS or cirrhosis. Since the frequency of positive stool or snip examinations was significantly greater in patients with histolagically diagnosed HS than in all other groups, it is unlikely that HS was the cause of portal hypertension in patients with non-diagnostic biopsies. Cirrhosis is also unlikely to have been missed to any great degree. Patients with non-diagnostic biopsies and patent portal veins suffered significantly more episodes of haematemesisand had a significantly lower prevalence of ascites than cirrhotics. In patients with cirrhosis, mean serum bilirubin, IgG and IgA levels were significantly higher. An uncontrolled observation which impressed us was the tolerance shown by some of the undiagnosed patients for repeated variceal bleeding. More complications, including deaths, might have been expected if such patients had been suffering from cirrhosis. Although careful follow-up and autopsy studies are required for completed histological evaluation, we believe that many of these patients with non-diagnostic biopsies were suffering from non-cirrhotic, nonschistosomal, intrahepatic portal hypertension, most often referred to as idiopathic portal hypertension (IPH). It is unlikely that any other specific noncirrhotic conditions, such as nodular regenerative hyperplasia or parual nodular transformation, were responsible. We emphasize that IPH is an entirely distinct condition from hyper-reactive malarial splenomegaly, or so-called tropical splenomegaly. Our patients with non-diagnostic biopsies did not have hepatic sinusoidal lymphocytosis, and their malarial serology and IgM levels did not differ from those of geographically matched controls (De Cock, unpublished). We do not believe that hyper-reactive malarial splenomegaly is an important cause of oesophageal varices in Kenya, although. portal vein presstire may be elevated in this condition (WILLIAMS et al.. 1968). IPH has been described h different parts of the world under a variety of names. Non-cirrhotic portal fibrosis (SAMA et al., 1971), hepatoportal sclerosis (MIKKELSEN et al.. 19651. obliterative wrtal venopathy (NAYAK & ‘RAMALINGASWAMI, i969; IBER, 1969), and benign intrahepatic portal hypertension (KINGHAM et al., 1981; LEVISON et al., 1982) are all terms referring to a non-cirrhotic, non-schistosomal form of intrahepatic portal hypertension that may represent a spectrum of pathology rather than one condition. The importance of the present work is that, apart from our preliminary reports (DE COCK et al., 1982 a,b; 1983), IPH does not appear to have been previously documented in Africa. It is possible that many such caseshave previously been dismissed as examples of obscure splenomegaly. The aetiology of IPH remains uncertain, but the striking preponder-
109
et al.
ance of Kamba patients suggeststhat an environmental causeshould be searched for in Machakos District. The importance of HBV infection in Kenya is demonstrated by the finding of some viral marker in 80% of control subjects. The prevalence of HBsAg was significantly higher in patients with cirrhosis than in controls. Previous work in Kenya (PETTIGREW et al., 1976) has demonstrated an association between HBV infection and chronic liver disease. In the present study, 27% of patients with non-diagnostic biopsies and patent portal veins were HBsAg positive, compared with 12% of controls. This could result from inclusion in the non-diagnostic group of some unrecognized cases of cirrhosis. An alternative explanation, however, is that this increased prevalence of HBsAg results from the greater number of blood transfusions which the undiagnosed patients required. Equal proportions of patients with cirrhosis as with non-diagnostic biopsies had some marker of HBV infection, but the prevalence of anti-HBs was significantly greater in the latter. Although equally exposed, therefore, the reaction to HBV was strikingly different in these two groups. Portal or splenic vein occlusion was found in 11%of casesof portal hypertension. We disagree with earlier reports by MILLER (1967; 1972) and BAGSHAWE & CAMERON (1976), that obstruction of the portal vein is the major cause of variceal haemorrhage in Kenya. Nevertheless, there can be little doubt that portal vein occlusion is more common in the tropical developing world than in industrialized countries. The single major cause of portal hypertension identified in this study was HS, accounting for almost one third of cases.Haemorrhage was well tolerated, and biochemical liver tests were more favourable than in cirrhosis. Both cirrhosis and HS were sipnificantlv more frequent in Kamba patients from M>chakos -District than expected from the pattern of general medical admissions. The severity of schistosomal disease is generallv related to intensitv of infection. In Kenva. prevaleice and intensity of
We are grateful to Mrs J. Clarkson,London Schoolof Hygiene and Tropical Medicine, for statistical advice; to Professor Michael Hutt for discussion; to Drs May Ho and
Harrison Spencer for a number of control sera; to the staff of the Walter Reed Institute, Nairobi, for storage of serum; to Drs Sarah Thomas and Virinder Taiwar for help with endoscopy and radiology; fo Rita Musick and Ruth Nao for secretarial assistance; and to the Wellcome Trust for financial support. References Bagshawe, A. F. & Cameron, H. M. (1974). Idiouathic tropical splenomegaly and portal hypertension. In: Health and Disease in Kenya. Vogel, L. C., Muller, A. S.,
CHRONIC
110
SPLENOMEGALY
Odingq, R. S., Onyango, Z. & de Gem, A. (editors). Nairobi: East African Literature Bureau, pp. 445-452. De Cock, K. M., Awadh, S., Raja R. S., Wankya, B. M. & Lucas, S. B., (1982a). Esophageal varices in Nairobi, Kenya: a study of 68 cases.AmericanJournal of Tropical Medicine and Hygiene, 31, 579-588. De Cock, K. M., Awadh, S., Raja, R., Wankya, B. M. & Lucas, S. B. (1982b). Portal hypertension in Nairobi, Kenya. Gut, 23, A446. De Cock, K. M., Awadh, S., Raja, R. S., Wankya, B. M., Lillywhite, J. E., Hodgen, N., Bertrand, J., Jupp, R. & Lucas, S. B. (1983). Further evidence of idiopathic portal hypertension in Kenya: a study of hepatitis B markers, immunoglobulin levels and enzyme-linked immunosorbent assayfor Schiswsoma mansoni in patients and controls. Gut, 23, A483-484. De Cock, K. M., Hodgen, A. N., Lucas, S. B., Jupp, R. A. Slavin, B., Arap Siongok, T. K. & Rees, Ph. H. (1987). Chronic splenomegaly in Nairobi, Kenya. I. Epidemiology, malarial antibody and immunoglobulin levels. Transactions of the Royal Society of Tropical Medicine and Hygiene, 81, 100-106.
Deverill! I. (1980). Kinetic measurement of the immunoprecipitm reaction using the centrifugal analyser. In: Centrifugal Analysers in Clinical Chemisrry.Price, C. P. & Spencer, K. (editors). Eastbourne, UK: Praeger. Hansen, D. P. & Daily, D. S. (1978). A fibre-endoscopic study of acute upper gastrointestinal haemorrhage in Nairobi. Kenya. American 7ournal of T&Cal Medicine . and Hygiene, -27, 197-200.Highton, R. B. (1979). Schistosomiasis. In: Health and Disease in Kenya. Vogel, L. C., Muller, A. S., Odingo, R. S., Onyango, Z. & de Geus, A. (editors). Nairobi: East African Literature Bureau, pp. 347-355. Iber, F. I. (1969). Obliterative portal venopathy of the liver and “idiopathic portal hypertension”. Annals of Internal Medicine,
71, 660-661.
Kingham, J. G. C., Levison, D. A., Stansfield, A. G. & Dawson, A. M. (1981). Non-cirrhotic intrahepatic portal hypertension: a long term follow-up study. Quarter.$ 3ournul of Medicin.?, 50, 259-268. Levison, D. A., Kingham, J. G. C., Dawson, A. M. & Stansfield, A. G. (1982). Slow cirrhosis - or no
IN
KENYA.
II.
cirrhosis? A lesion causing benign intrahepatic portal hypertension. Joumal of Pathology? 137, 233-272. Linsell, C. A. (1967). Cancer incidence m Kenva 1957-1963. British 7oukal
of Cancer. 21. 665-673.
Mikkelseq- W. P.,*Edmondsonj H. A., Peters, R. L., Redeker, A. G. & Reynolds, T. B. (1965). Extra and intrahepatic portal hypertension without cirrhosis (hepatoportal sclerosis). Annals of Surgery, 162, 602-618. Miller, J. R. M. (1967). Portal hypertension in Nairobi. East African Medical 3owna1, 44, 376-386. Miller. I. R. M. (1972). Portal hvoertension. East African Medical Journal,
44, 243-245 . A
Nayak, N. C. & Ramalingaswami, V. (1969). Obliterative portal venopathy of the liver associated with so-called idiopathic portal hypertension or tropical splenomegaly. Archives of Pathology, 87, 359-369.
Pettigrew, N. M., Bagshawe, A. F., Cameron, H. M., Cameron, C. H., Dorman, J. M. St MacSween, R. N. M. (1976). Hepatitis B surface antigenaemia in Kenyans with chronic liver disease. Transactions of the Royal Society of Tropical Medicine and Hygiene, 70, 462-465. Sama, S. K., Bhargava, S., Gopi Nath, N., Talwar, J. R., Nayak? N.C., Tandon, B. N. & Wig, K. L. (1971). Non-cirrhotic portal fibrosis. American Journal of Medicine, 51, 160-169. Scheuer, P. J. (1980). Liver biopsy interpretation (3rd edit.). London: Balliere Tindall. Shah, M. V. (1978). Chronic liver disease in Kenya. M.Med. Thesis. Universitv of Nairobi. Sherlock, S. (1981). i)iseasesof the Liver and Biliary System (6th edit.). Oxford: Blackwell. Thomas. SE.. Mwaungulu. G. S.. Wankva. B. M. & De Cock. K. ’ M. (1983). . Acute- uaoer- aastrointestinal haemorrhage at Kenyatta Nationi kos&al, Kenya: a prospective study. East African Medical .7ournal, 60, 428-43 1.
Williams, R., Parsonson,A., Somers, K. & Hamilton, P. J. S. (1968). Portal hypertension in idiopathic tropical splenomegaly. Lancet, i, 329-333.
Accepted for publication
28th November,
1985