Table 4.
Prediction of Extrapulmonary Complications
| Predictora | OR (95% CI) |
|---|---|
| Age ≥80 y | 2.27 (.49–9.92) |
| PMH: heart disease | 2.13 (.63–7.41) |
| PMH: HTN | 1.74 (.81–3.69) |
| PMH: dementia | 3.60 (.58–25.33) |
| PMH: CKD | 4.56 (1.36–17.27) |
| Smoker | 1.88 (.80–4.42) |
| Obesity | 2.64 (1.23–5.84) |
| ED: RR high | 1.63 (.79–3.35) |
| ED: SpO2 low | 1.34 (.60–2.93) |
| RNAemia | 2.81 (1.26–6.36) |
Abbreviations: CI, confidence interval; CKD, chronic kidney disease; ED, emergency department; HTN, hypertension; OR, odds ratio; PMH, past medical history; RR, respiratory rate; SpO2, oxygen saturation.
aPotential predictors of extrapulmonary complications (EPCs) included demographic features (age ≥60 or ≥80 years; sex), past medical history features (lung disease, cancer, diabetes, immunosuppression, heart disease, HTN, angiotensin-converting enzyme inhibitor or angiotensin receptor blocker use, stroke, dementia, deep venous thrombosis or pulmonary embolus, CKD, tobacco smoking, and obesity), binary indicators of abnormal ED vital signs (low or high mean arterial pressure, heart rate, respiratory rate, and temperature; low SpO2), pneumonia on initial chest radiography or computed tomography, and patient-reported symptoms at enrollment, excluding those constitutive of extrapulmonary diagnosis (fever, chills, cough, sore throat, congestion, shortness of breath, chest pain, and myalgias). Laboratory values were not included, because many were constitutive of extrapulmonary diagnoses.
To prevent overfitting, predictors were selected via elastic net regression of EPCs (1 if a patient had ≥1 EPC; 0 if a patient had none) on these features with 10-fold cross-validation, selecting the regularization parameter λ minimizing mean cross-validated error, and yielding the features listed in Table 4. In a logistic model regressing EPCs on these features, significant predictors of EPC included CKD, obesity (body mass index >30 [calculated as weight in kilograms divided by height in meters squared]), and RNAemia. RNAemia was associated with 2.8 times the odds of EPC, comparable in magnitude to the association between obesity and development of EPCs. The mean cross-validated area under the receiver operating characteristic curve of the model in predicting EPC was 0.73. The Akaike Information Criterion (AIC) was 222.25.