Divergence from Hardy Weinberg Equilibrium in Newborn Screening Cohorts for Sickle Hemoglobin

Abstract

Newborn screening (NBS) for hemoglobinopathies was first introduced in the 1970’s to prevent infectious complications of homozygous sickle cell disease (hemoglobin [Hb] SS). The frequency of HbSS in newborns can be estimated using carrier rates (heterogygotes), assuming Hardy Weinberg Equilibrium (HWE). The current study examined whether NBS cohorts matched expected HWE distribution, and possible reasons for divergence. For the purpose of this study, all variants other than HbS were considered to be HbA. A total of 82 studies (1956 to 2013) were retrieved, of which 7 were excluded (carrier parents selected or screening sickle test used). Distribution of Hb SS, AS and AA was examined in 75 studies, from the following regions: Sub Saharan Africa n=26, Brazil n=12, USA n=12, EU n=10; Middle East/North Africa n=7, Caribbean n=6, India n=1, Canada n=1. The following screening laboratory methods were used: cellulose acetate electrophoresis (n=22), isoelectric focusing (n=21), high performance liquid chromatography (n=16), filter paper or micro column (n=7), other (n=9). A total of16,377,450 newborns or infants were included. In mixed random models, relative risk (RR) of detecting HbSS compared to expected HWE distribution was 2.06 (95% confidence interval [CI] 1.86, 3.66), p<0.001. In cohorts where HbAS carrier rates was >= 15%, RR was 1.44 (95% CI 1.23, 1.70), p=0.005; if carrier rates were between >=5% and <15%, RR was 1.62 (95% CI 1.20, 2.19), p=0.004; if carrier rates were <5%, RR was 5.74 (95% CI 3.4, 9.78), p<0.001. Thirteen studies provided data by subgroup: place of origin, ethnicity or race. Affected individuals were more likely to cluster in studies with lower HbS gene frequency (figure 1). Conversely, HbSS RR observed compared to expected HWE distribution was highest in populations with lowest overall gene frequency (figure 2). Overall RR for detecting HbAS compared to expected was 0.96 (95% CI 95, 97) p<0.001. Controlling for gene frequency, the Caribbean region was least likely to deviate from the expected HWE distribution, compared to other regions (RR 1.05, 95% CI 1.01, 1.05, P=0.007); cellulose acetate electrophoresis based methods were most likely to deviate from expected HWE distribution, compared to other methods (RR 0.97, 95%CI 0.95, 0.100, p=0.038). Deviation from HWE is commonly observed in NBS cohorts. It is an expected finding when smaller populations (e.g.migration) carrying a gene, exist within a larger population less affected with the same gene. However, even when controlling for gene frequency, regional variations were noted. Reasons may include further variations in population homogeneity, non-random mating within smaller groups (consanguinity), variations in hemoglobin variant distribution (alpha, beta thalassemia, fetal hemoglobin expression). Lab methods may also affect agreement with HWE distribution, as correct classification of non-affected individuals may vary between methods (normal homozygotes, variants and carriers). Divergence from HWE, especially when observed HbSS occurrence is less than expected, or when population Hb S gene distribution is known, may serve as a rapid and inexpensive quality control measure. Further examination of possible reasons for HWE deviation and caution when estimating burden of disease using carrier rates derived from NBS cohorts appear warranted. [Display omitted] [Display omitted] DisclosuresNo relevant conflicts of interest to declare.