Autoimmunity is present in patients with stable chronic obstructive pulmonary disease (COPD), playing a role in indirect and direct ways. We aimed to explore whether autoimmunity could play a role in COPD exacerbations and construct autoimmunity-related prediction models. This prospective, longitudinal, observational cohort study enrolled 155 patients with acute COPD exacerbations (AECOPD) followed for at least two years. The laboratory parameters, including complete blood count, serum immunoglobulins G/A/M and complement C3/C4 levels, were collected at enrollment. We studied the demographic characteristics, clinical characteristics and laboratory parameters to identify independent risk factors and build predictive models. The results showed that lower lymphocyte count was associated with noninvasive ventilation (NIV) in patients with AECOPD (the odds ratio [OR] 0.25, the 95% confidence interval [CI]: 0.08–0.81, P = 0.02). Lymphocyte count performed well with an area under the curves (AUC) of 0.75 (P < 0.0001, sensitivity: 78.1%, specificity: 62.3%, cutoff value [Cov] ≤ 1.1). The C index, calibration plot, decision curve analysis (DCA) and bootstrap repetitions indicated that this clinical prediction model based on lymphocyte count for NIV in patients with AECOPD performed well. Having prior home oxygen therapy (OR: 2.82, 95% CI: 1.25–6.36, P = 0.013) and higher COPD Assessment Test (CAT) scores (OR: 1.14, 95% CI: 1.03–1.25, P = 0.011) were associated with the increased risk for respiratory failure. For predicting respiratory failure, CAT scores and home oxygen therapy combined had an AUC-ROC of 0.73 (P < 0.0001). This clinical prediction model based on lymphocyte count may help to assist in treatment decisions for NIV in patients with AECOPD. Lower complement C3 seems to be associated with worse outcomes in patients with AECOPD.
