A polygenic score for height identifies an unmeasured genetic predisposition among pediatric patients with idiopathic short stature.

BACKGROUND

A subset of children with short stature do not have an identified clinical explanation and are assigned a diagnosis of idiopathic short stature (ISS). We hypothesized that a polygenic score for height (PGSheight) could identify children with ISS who have an unrecognized heritable predisposition to shorter height.

METHODS

We examined 534 pediatric participants in an EHR-linked DNA biobank (BioVU) who had undergone an evaluation for short stature by an endocrinologist. We used a previously validated PGSheight and standardized it to a standard deviation (SDS) of 1. PGSheight differences between short stature subtypes was estimated using Tukey's HSD. The PGSheight and mid-parental height (MPH) were then used to predict adult heights for each participant and these predictions were compared using Cohen's d stratifying by short stature subtype. The ability of the PGSheight to discriminate between ISS and short stature due to underlying disease was evaluated using logistic regression models with area under the ROC curve (AUC) analyses and testing the incremental benefit (ΔAUC) of adding the PGSheight to prediction models.

RESULTS

Among the 534 participants, 22.1% had ISS (median [IQR] PGSheight SDS = -1.31 [-2.15 to -0.47]), 6.6% had familial (genetic) short stature (FSS) (-1.62 [-2.13 to -0.54]), and 45.1% had short stature due to underlying pathology (-0.74 [-1.23 to -0.19]). Children with ISS had similar PGSheight values as those with FSS (ΔPGSheight [95% CI] = 0.19 [-0.31 to 0.70], p = 0.75), but predicted heights generated by the PGSheight were lower than the MPH estimate for children with ISS (d = -0.64; p = 4.0×10-18) but not FSS (d = 0.05; p = 0.46), suggesting that MPH underestimates height in the ISS group. Children with ISS had lower PGSheight values than children with pathology (ΔPGSheight = -0.60 SDS [-0.89 to -0.31], p < 0.001), suggesting children with ISS have a larger predisposition to shorter height. In addition, the PGSheight improved model discrimination between ISS and pathologic short stature (ΔAUC, + 0.07 [95% CI, 0.01 to 0.11]).

CONCLUSIONS

Some children with ISS have a clinically unrecognized polygenic predisposition to shorter height that is comparable to children with FSS and larger than those with underlying pathology. A PGSheight could help clinicians identify children who have a benign predisposition to shorter height.