Seasonal ambient temperature and particulate matter as discriminators of hyperemesis gravidarum severity: a retrospective comparative observational study.

INTRODUCTION

Ambient air pollutants and temperature may influence the clinical severity of hyperemesis gravidarum (HG), but their relative contributions remain unclear. This study aimed to determine whether particulate matter (PM₁₀), sulfur dioxide (SO₂), and ambient temperature were independently associated with severe ketonuria in pregnant women with HG.

METHODS

This retrospective observational study included 211 pregnant women diagnosed with HG. Clinical severity was defined by ketonuria category (severe ≥2 + vs. non-severe <2+). Covariates (maternal age, gestational week, BMI, Healthy Eating Index [HEI], gravida, PM₁₀, SO₂, and ambient temperature) were selected a priori based on literature and univariate screening. Multivariable binary logistic regression was used; model discrimination was assessed via ROC analysis.

RESULTS

The autumn/winter group showed significantly higher maternal age, gravida, and PM₁₀ exposure, and lower ambient temperature than the spring/summer group (all p < 0.001). In multivariable analysis, only lower ambient temperature (OR = 0.890, 95% CI: 0.845-0.938, p < 0.001) and higher maternal age (OR = 1.143, 95% CI: 1.057-1.237, p < 0.001) were independently associated with severe ketonuria; PM₁₀, SO₂, HEI, gestational week, BMI, and gravida were not significant (all p > 0.05). The model correctly classified 81.5% of cases (Nagelkerke R 2 = 0.426). ROC analysis confirmed ambient temperature (AUC = 0.799) and PM₁₀ (AUC = 0.754) as significant discriminators for severe versus non-severe ketonuria, while SO₂ was not (AUC = 0.532).

CONCLUSION

Ambient temperature showed the strongest discriminative performance for HG severity among the environmental parameters examined, while the association of PM₁₀ was attenuated after adjustment for temperature. Maternal age was also independently associated with HG severity. These findings should be interpreted as associations rather than causal effects. Further prospective, multi-site studies incorporating individual-level exposure assessments are warranted to confirm these associations and elucidate the underlying mechanisms.