Clinical characteristics of Burkitt's lymphoma from three regions in Kenya.

OBJECTIVES To describe the clinical characteristics of Burkitt's lymphoma (BL) from three regions in Kenya at different altitudes with a view towards understanding the contribution of local environmental factors. DESIGN Prospective cross-sectional study. SETTING Kenyatta National Hospital and seven provincial hospitals in Kenya. METHOD Histologically proven cases of Burkitt's lymphoma in patients less than 16 years of age were clinically examined and investigated. MAIN OUTCOME MEASURES For every case the following parameters were documented: chief complaint(s); physical examination, specifically pallor, jaundice, oedema, lymphadenopathy, presence of masses, splenomegaly and hepatomegaly. Reports of evaluation of chest radiograph, abdominal ultrasound/scan, bone marrow aspiration, cerebral spinal fluid cytology, liver and kidney function tests, urinalysis, stool occult blood and full blood count results. Stage of disease was assigned A, B, C or D. Cases of BL from three provinces of Kenya with diverse geographical features were analysed: Central, Coast, and Western. RESULTS This study documented 471 BL cases distributed as follows: Central 61 (males 39 and 22 females), M:F ratio 1.8:1; Coast 169 (111 males and 58 females), M:F ratio 1.9:1; and Western 241 (140 males and 101 females), M:F ratio 1.4:1. The major presenting complaints were: abdominal swelling--Central 36%, Coast 4% and Western 26%; swelling on the face--Central 31%, Coast 81% and Western 64%; and proptosis--Central 3%, Coast 1% and Western 9%. The mean duration of these complaints in weeks were Central 6.9, Coast 6.08, and Western 5.05. The initial physical finding was a tumour mass in 39%, 72% and 54% of cases for Central, Coast and Western respectively. Tumour stage at diagnosis was: stage A--Central 21%, Coast 43% and Western 34%; stage B--Central 10%, Coast 5% and Western 10%; stage C--Central 41%, Coast 34% and Western 30%; and stage D--Central 28%, Coast 17% and Western 26%. For the age and sex matched cases the results show that commonly involved sites were: abdomen--Central 35%, Coast 9% and Western 14%; jaw (mandible)--Central 24%, Coast 22% and Western 31%; maxilla--Central 6%, Coast 24% and Western 11%; and lymph nodes--Central 10%, Coast 4% and Western 8%. The disease stage was A--Central 33%, Coast 44% and Western 36%; stage B--Central 11%, Coast 10% and Western 27%; stage C--Central 39%, Coast 34% and Western 27%; and stage D--Central 21%, Coast 13% and Western 37%. CONCLUSION This study shows that clinical features of childhood BL vary with geographical region. The variations are documented in proportion of jaw, maxilla, abdominal and lymph nodal sites involvement. The differences observed are potentially due to the local environmental factors within these provinces. BL cases from Western province had features, intermediate between endemic and sporadic. Coastal province BL cases were similar to endemic BL, while BL cases from Central province resembled more or less sporadic BL subtypes. Strategies to explain and investigate the local environmental factors associated with the observed differences may certainly contribute towards improved understanding and clinical management of BL.