Background: Children with sickle cell anemia (SCA) are highly susceptible to stroke and other manifestations of pediatric cerebral vasculopathy. Inflammation is a risk factor for acute stroke. In a cross-sectional study of Ugandan children, we asked whether inflammatory markers correlated with findings on neurological and neurocognitive testing and magnetic resonance imaging (MRI).

Methods: Stored plasma samples were obtained at a well study visit from 81 children, a subset within a larger sample of study participants with HbSS ages 1-12 years recruited at the Mulago Hospital sickle cell clinic in Kampala. Participants underwent 4 standardized assessments: neurocognitive testing using the Mullen Scales of Early Learning (for ages 1-4 years) or Kaufman Assessment Battery for Children, 2nd edition (for ages 5-12) (abnormal at z-score of -2 or lower), examination for prior stroke (PedsNIHSS) and transcranial Doppler ultrasound (TCD) using criteria for pediatric SCA, plus a non-contrast MRI/MRA on a 1.5 Telsa scanner, analyzed for T1- and T2- weighted images and T2 FLAIR. MR scans were interpreted by clinical and research methods, the latter per SWiTCH protocol (Helton, Blood 2014). Adjudication of differing reads was performed by a blinded third neuroradiologist. Participants undergoing MRI intentionally included 29 without any abnormal findings on the three other evaluations.

Plasma samples were examined for key cytokines and markers of vascular activation by a 20-plex Luminex immunoassay. Immune markers included IL1b, IL1RA, IL2, IL4, IL10, IL18, IL22, IFNg, TNFa, TNFb, sFasL, CXCL1, CXCL8 (IL8), PDGFBB, VEGFA, serpin E1 (Plasminogen Activator Inhibitor (PAI)-1), sICAM1, VCAM1, TGFa, plus assay of hsCRP. Raw immune data were transformed using Box-Cox transformations and standardized; markers with persistently skewed (non-Gaussian) distributions after transformation were dichotomized for analysis. Continuous immune variables were analyzed by repeated measures ANOVA (unadjusted) and ANCOVA (adjusted for baseline covariates of age, sex, hemoglobin and 25-hydroxy-vitamin D3/D2. Dichotomized biomarkers were analyzed using generalized linear mixed-effects models. For these exploratory biomarker discovery-focused analyses, all significance tests were performed 2-tailed without adjustment for multiple comparisons, with a=0.1 for main effects.

Results: Mean age was 6.48 ± 2.75 years, 50.6% male. Mean hemoglobin was 7.26±0.90 g/dl; 16.7% were malnourished using standard international measures established for age and sex. Infarcts on MRI were detected in 42 (52%), including 13 (25%) with no other test abnormalities. Of the 20 inflammatory markers tested, the levels of 13 markers were associated with one or more abnormality. One or more infarcts on MRI as the sole abnormality ("silent" stroke) were associated with levels IL1b, IL2 and TGFb (Table 1). Neurocognitive dysfunction alone was associated only with levels of IL22. Abnormal TCD (primarily "conditional" values) or prior stroke was associated with levels of VEGFA, IL4 and TNFb. In contrast, abnormal TCD or stroke by examination plus neurocognitive dysfunction was associated with nine cytokines levels: CXCL8, CXCL1, IFNg, TNFa, sFasL, IL18, IL22, sICAM1 and VEGFA. In all cases, cytokine levels were negatively associated with each of the abnormal neurological, neurocognitive or imaging results. Adjusting for age, sex, hemoglobin and malnutrition had only modest effects on statistically significant results. No associations were found between the test results and any of the other eight inflammatory markers, including hsCRP.

Conclusions: In this exploratory biomarker analysis in a sample with highly prevalent abnormal test results, the presence of those abnormalities correlated with levels of a majority of the inflammatory markers assayed. The clinically well status at sampling and the finding of lower markers levels associated with existing abnormalities suggest that these children were not currently inflamed. As a cross-sectional analysis, cerebral vasculopathy underlying those abnormalities had previously occurred. Lower levels of inflammation may reflect a survivor status of children affected by a chronic illness with high local mortality. Our findings need to be confirmed in a prospective study children, and may be most interesting at acute events and for subjects receiving disease-modifying therapy.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution