Stroke in Pediatric Sickle Cell Disease in Sub-Saharan Africa: Systematic Review and Estimated Prevalence

Background: Most people with sickle cell disease (SCD) live in Sub-Saharan Africa (SSA); many do not survive to adulthood (Grosse et al. Am J Prev Med 2011). Stroke is a common complication of SCD in the region (Ohene-Frempong et al.Blood 1998), yet the region's burden of pediatric SCD stroke is unknown. This systematic review aims to describe what is known, and to estimate the number of children affected by SCD stroke in SSA.

Methods: Systematic reviews using Pubmed, Embase, and Web of Science were performed for all studies reporting pediatric SCD stroke prevalence data. Keyword search strategy included these or related terms: "stroke" or "brain" and "sickle cell" and "Africa" or each country in SSA available in English or French. Publication dates were selected between 1980 and December 13, 2016. Using pre-selected inclusion criteria, titles and abstracts were screened and full-text articles were reviewed. For publications meeting inclusion criteria, extracted data included: study and participant characteristics, number of patients with stroke, diagnostic criteria, imaging, management, recurrence. The number of children affected by SCD stroke in SSA was estimated based on current estimates of national population and S allele frequency (Piel et al. Lancet 2013) and the stroke prevalence in the extracted studies. The stroke prevalence in SSA was estimated based on both fixed-effect analysis and random-effect analysis.

Results: After duplicates were removed, 133 individual articles remained. Ten full text articles met review selection criteria, including one in French. In all, papers reported data from seven countries in the region, including four reports from Nigeria. Cross-sectional SCD clinic-based data, with a sample size of 6,289 patients, reported pediatric stroke prevalence of 2.9-16.9%. Diagnostic criteria of stroke varied by site and were limited by access to brain imaging, which was reported in only half of the included studies. Treatment for stroke included packed red blood cell transfusion, exchange transfusion, or hyperhydration. Secondary stroke prevention with chronic blood transfusion or hydroxyurea was offered at two of 10 sites. Recurrence data was limited by variable study follow-up and limited reporting times. Using available sickle gene frequencies by country, estimated births affected by SCD per country in SSA, and a linear- or random-effects model, pediatric stroke prevalence in SSA was projected as 279,908 (95% CI: 240,335-319,481) and 561,746 (347,472-775,056), respectively. Estimates do not include stroke- or disease-associated mortality.

Conclusion: This meta-analysis estimates prevalence of pediatric SCD stroke in SSA, underscoring the region's high stroke burden. Multiple clinic-based cross-sectional studies reported stroke prevalence comparable to data from high income countries prior to implementation of standardized prevention strategies. Accuracy is limited by different approaches to stroke diagnosis, paucity of population-based data on disease prevalence and mortality, regional variation in disease prevalence and severity, and overall high rates of early death in children with SCD. Available data would skew towards surviving patients, leading to possible underreporting of people at risk for SCD stroke. Improved diagnostic specificity with standardized criteria and imaging would be useful for identification of strokes. Understanding the large numbers affected by SCD stroke is critical for focusing public health efforts to introduce preventive measures to reduce stroke risk and to implement rehabilitative strategies. Adopting standardized primary and secondary stroke prevention measures in SSA may decrease the burden of SCD-associated stroke.

This work was supported by funding from the National Institutes of Health (5R21HD089791, PIs Idro, Green).