Clinical characteristics and indicators associated with poor in-hospital prognosis among children with severe pertussis in Shandong Province, China (2022–2024)

Pertussis is a life-threatening respiratory infection among young infants. This retrospective study enrolled only children already diagnosed with severe pertussis. We aimed to describe clinical characteristics and identify indicators associated with poor in‑hospital prognosis within this severe cohort, rather than exploring predictors for progression from mild to severe pertussis. A retrospective observational study was conducted among 80 children diagnosed with severe pertussis who were hospitalized between January 2022 and December 2024. Clinical data, laboratory results, treatment, and outcomes were collected. Multivariate logistic regression and receiver operating characteristic (ROC) curve analysis were applied to screen indicators associated with composite poor in-hospital prognosis and evaluate their prognostic performance.Composite poor in-hospital prognosis was defined as new in-hospital adverse events including invasive mechanical ventilation, pertussis encephalopathy with persistent neurological sequelae, or all-cause in-hospital death; 6 out of 80 participants met this composite endpoint. White blood cell count was included as a continuous variable in multivariate regression. The optimal cutoff of > 30 × 10⁹/L was derived from ROC analysis for descriptive and potential bedside reference only and was not used for regression modelling. The median age was 4 months and 19 days (range, 1 month 14 days to 5 years); 65.0% (52/80) of patients were infants ≤ 3 months who could not receive routine diphtheria, pertussis, and tetanus (DPT) vaccine. All patients presented paroxysmal cough, and 95% had inspiratory whoop. Leukocytosis (WBC > 20 × 10⁹/L) was observed in 45.0% of children; 75.0% had patent foramen ovale (PFO), and 18.80% developed secondary cardiac abnormalities. Erythromycin combination therapy carried a significantly higher risk of liver dysfunction than monotherapy (36.40% vs. 8.60%, P < 0.05). Multivariate logistic regression identified three independent correlates of composite poor prognosis: unvaccinated status (OR = 4.2, 95%CI 1.8–9.6, P = 0.001), mixed bacterial-viral co-infection (OR = 3.5, 95%CI 1.2–10.1, P = 0.023), and elevated continuous white blood cell count (OR = 5.1, 95%CI 2.3–11.4, P < 0.001). Given the limited number of poor-prognosis events ( n = 6), these adjusted associations should be interpreted cautiously as exploratory findings.ROC analysis using continuous WBC count yielded an AUC of 0.678, whereas the combined prognostic model derived from multivariable logistic regression achieved an AUC of 0.689. Infants aged ≤ 3 months constituted the majority of children hospitalized with severe pertussis. After adjusting for age-related vaccination eligibility, unvaccinated status, mixed bacterial-viral co-infection, and higher white blood cell count were independently associated with poor in-hospital prognosis. ROC analysis demonstrated only modest discriminative performance of these indicators; at present, they serve merely as potential references and require external validation before any clinical application. Not applicable