Identification of patients at risk of Clostridioides difficile infection for enrollment in vaccine clinical trials. Issue 3 (15th January 2021)

• Record Type: Journal Article
• Authors: Stevens, Vanessa W.; Russo, Ellyn M.; Young-Xu, Yinong; Leecaster, Molly; Zhang, Yue; Zhang, Chong; Yu, Holly; Cai, Bing; Gonzalez, Elisa N.; Gerding, Dale N.; Lawrence, Jody; Samore, Matthew H.
• Is Part Of: Vaccine. Volume 39:Issue 3(2021)
• Page Start: 536
• Page End: 544
• Publication Date: 2021-01-15
• Subjects: Clostridioides difficile infection -- Vaccine -- Trial enrollment; Vaccines -- Periodicals; 615.372
• DOI: 10.1016/j.vaccine.2020.12.016
• Languages: English
• ISSNs: 0264-410X

Abstract

Highlights: Randomized vaccine trials for healthcare-associated infections are limited by rare events. This algorithm identifies patients with a 20-fold higher risk of CDI than the general population. It can be applied to inpatients, outpatients, or long-term care patients. Abstract: Background: Clostridioides difficile infection (CDI) is an important cause of diarrheal disease associated with increasing morbidity and mortality. Efforts to develop a preventive vaccine are ongoing. The goal of this study was to develop an algorithm to identify patients at high risk of CDI for enrollment in a vaccine efficacy trial. Methods: We conducted a 2-stage retrospective study of patients aged ≥ 50 within the US Department of Veterans Affairs Health system between January 1, 2009 and December 31, 2013. Included patients had at least 1 visit in each of the 2 years prior to the study, with no CDI in the past year. We used multivariable logistic regression with elastic net regularization to identify predictors of CDI in months 2–12 (i.e., days 31 – 365) to allow time for antibodies to develop. Performance was measured using the positive predictive value (PPV) and the area under the curve (AUC). Results: Elements of the predictive algorithm included age, baseline comorbidity score, acute renal failure, recent infections or high-risk antibiotic use, hemodialysis in the last month, race, and measures of recent healthcare utilization. The final algorithm resulted in an AUC of 0.69 and a PPV of 3.4%. Conclusions: We developed a predictive algorithm to identify a patient population with increased risk of CDI over the next 2–12 months. Our algorithm can be used prospectively with clinical and administrative data to facilitate the feasibility of conducting efficacy studies in a timely manner in an appropriate population.