A 2- Year retrospective study on the pattern of Retinobh Storna in Kenya

Abstract

Background: The national epidemiological characteristics of retinoblastoma in Kenya have not been determined. The diagnosis of this cancer is mainly clinical; histology determines tumour extent. Late diagnosis of this otherwise curable malignancy is associated with high mortality. Aim: To determine the incidence and pattern of presentation of retinoblastoma in Kenya. Design: A retrospective case series Setting: All 75 eye care centres in the 8 provinces of Kenya as registered in the Ministry of health eye information system. Methods: With permission, clinical registers at eye care centers were reviewed to identify cases of retinoblastoma that presented from 1sr January 2006 to 31st December 2007. Only centers that reported cases were visited to record patient's clinical and demographic data in a questionnaire. Cross-referred cases were analysed once to avoid double-counting. Results: A total of 206 suspected cases presented to 46 eye care facilities but 58 cases (28.2%) were lost after referral. Of 148 traced cases, 28.4% were self referrals and of the referred cases, most (21.6%) were from central province. Only 63.5% of cases were finally treated at 2 teaching and referral hospitals.After excluding 3 missing files and 13 cases that were ruled out on histology, 132 confirmed cases(166 eyes) were subsequently analysed. The mean delay in first presentation was 6.75 months and delay after referral was 1.69 months. Leukocoria was the most common presenting complaint (91.7% cases) and sign (71.1 % eyes). There were 25.8% bilateral cases and 78.2% unilateral cases with mean ages of 26 and 35.9 months respectively. The male to female ratio was 1.49:1. Only 4.5% had a positive family history. Most (32.6%) cases resided in the Rift valley province. There was no association between ethnicity and bilaterality. The annual incidence of retinoblastoma in 2007 was 1:17,030 live births. Conclusions: A significant proportion of cases were lost after referral. The late presentation was associated with advanced disease. Leukocoria was the most common finding. Most cases resided in the Rift valley province. The incidence of retinoblastoma was similar to most countries but may be an underestimate. Recommendations: Public education ancl screening with the red reflex test by primary health care workers would ensure early detection. Quality control measures in record keeping would ensure accuracy. A retinoblastoma registry would provide accurate estimates through register-based studies. Further research is necessary to investigate the lost cases after referral, delays in presentation and barriers to uptake of services.