Causes of death and early life determinants of survival in homozygous sickle cell disease: The Jamaican cohort study from birth

Abstract

Globally, the majority of persons born with sickle cell disease do not have access to hydroxyurea or more expensive interventions. The objectives were to estimate the survival in homozygous sickle cell disease, unbiased by symptomatic selection and to ascertain the causes of death in a pre-hydroxyurea population. The utility of early life biomarkers and genetically determined phenotypes to predict survival was assessed. A cohort study based on neonatal diagnosis was undertaken at the Sickle Cell Unit, a specialist clinic delivering care to persons with sickle cell disease in Jamaica. Screening of 100,000 deliveries detected 315 babies with homozygous sickle cell disease of whom 311 have been followed from birth for periods up to 43 years. Pneumococcal prophylaxis and teaching mothers splenic palpation were important, inexpensive interventions. Anticipatory guidance, routine care and out-patient acute care were provided. Each participant was classified as alive, dead, or defaulted (usually emigration). Causes of death were ascertained from clinical records and/or post-mortem reports. Survival was assessed using the Kaplan-Meier function. Sex-adjusted Cox semi-parametric proportional hazards and Weibull modelling were used to assess the effects on survival of biomarkers. Survival to 40 years was 55.5% (95% CI 48.7% to 61.7%). Acute Chest Syndrome (n = 31) and septicemia (n = 14) were significant causes of death at all ages. Acute splenic sequestration (n = 12) was the most common cause of early deaths. Survival was significantly shorter in those with lower hemoglobin at 1 year, high total nucleated count at 1 year, and a history of dactylitis ever. In these hydroxyurea naïve patients, survival into midlife was common. Causes of death were often age specific and some may be preventable. Early life biomarkers predictive of decreased survival in SS disease identify a patient group likely to benefit from close clinical supervision and potentially high risk therapies.

Fig 1. Median age at death (interquartile range) for 12 cause-of-death groups (140 causes, 119 deaths) among people with sickle cell disease.

Fig 2

Survival among N = 298 cohort participants stratified by a history of dactylitis at any age (Fig A), modelled effect of dactylitis at any age (Fig B).

Fig 3

Survival among N = 290 cohort participants stratified by early life total nucleated cell count (below median / above median) (Fig A), modelled effect of total nucleated cell count (at 10x10⁹/L, 15 x10⁹/L, 20 x10⁹/L) (Fig B).

Fig 4

Survival among N = 279 cohort participants stratified by early life total haemoglobin level (below median / above median) (Fig A), modelled effect of total haemoglobin (at 6g/dL, 8 g/dL, 10g/dL) (Fig B).