Comparison of Diagnostic Accuracies of qSOFA, NEWS, and MEWS to Identify Sepsis in Older Inpatients With Suspected Infection.

OBJECTIVES

To determine and compare the accuracies of the quick Sequential (Sepsis-related) Organ Failure Assessment (qSOFA) and Modified and National Early Warning Scores (NEWS and MEWS) to identify sepsis in older inpatients with suspected infection.

DESIGN

Prospective diagnostic accuracy study.

SETTING AND

PARTICIPANTS

Patients admitted to an acute geriatric unit of an Italian University Hospital with at least one sepsis risk factor and suspected infection defined as antibiotic prescription and associated culture test during hospital stay.

METHODS

Sepsis diagnosis was defined as the presence on discharge documents of International Classification of Diseases, Ninth revision, Clinical Modification codes for severe sepsis, septic shock, or for infection and acute organ disfunction. For each patient, clinical parameters were evaluated at least twice daily throughout hospital stay; qSOFA, NEWS, and MEWS were derived, and worst scores recorded. Positive cutoffs were set at ≥2, ≥7, and ≥5, respectively. Sensitivity, specificity, positive and negative predictive values (PPV and NPV, respectively), and positive and negative likelihood ratios, as well as areas under the receiver operating characteristic curve (AUROCs) were calculated.

RESULTS

Among 230 geriatric patients with suspected infection at risk for sepsis (median age 86 years, 49% women), 30.9% had a sepsis diagnosis. A qSOFA ≥2 was recorded in 111 (48.3%) patients, a MEWS ≥5 in 65 (28.3%), and a NEWS ≥7 in 115 (50.0%). The qSOFA showed the highest sensitivity [81.7%, 95% confidence interval (CI) 71.7%-89.5%], but low specificity (66.7%, 95% CI 59.1%-73.7%), resulting in a high NPV (89.1%; 95% CI 82.7%-93.8%) and poor PPV (52.3%, 95% CI 43.0%-61.4%). The AUROC for qSOFA was 0.76 (95% CI 0.69-0.83), comparable with that of NEWS (0.74, 95% CI 0.67-0.81, P = .44), but significantly higher than that of MEWS (0.70, 95% CI 0.63-0.77, P = .04).

CONCLUSIONS AND IMPLICATIONS: Repeated qSOFA determinations are useful to rule out sepsis in geriatric inpatients with suspected infection, but poorly support its diagnosis due to low specificity. More complex MEWS and NEWS do not perform better. Implementation of clinical scores to reliably identify sepsis in older patients is urgently needed.