Optimal cut-off value of procalcitonin and procalcitonin/albumin ratio for predicting bacteremia among patients living with cancer: a test-negative case-control study

Abstract

Objective

The occurrence of bacteremia is critically important for the survival of cancer patients. Therefore, our study aims to evaluate the efficacy of procalcitonin (PCT) and the procalcitonin to albumin ratio (PAR) in predicting bacteremia among this population.

Methods

In this retrospective test-negative case-control study, we included 903 hospitalized cancer patients, divided into two groups: the bacteremia-positive group (BSI group, n = 384) and the bacteremia-negative group (non-BSI group, n = 519). We assessed the diagnostic significance of PCT and PAR through receiver operating characteristic (ROC) analysis and determined the optimal cut-off values using Youden’s index.

Results

Both the duration of hospital stay and the 30-day mortality rate were significantly higher in the BSI group. The areas under the curve (AUC) for PAR and PCT were 0.749 (95% CI: 0.715-0.782) and 0.742 (95% CI: 0.708-0.776), respectively, indicating higher levels in the BSI group. The optimal cut-off values for predicting bacteremia were 0.72 for PAR and 1.32 for PCT. PAR showed the highest specificity (92.7%) and positive predictive value (PPV = 83.4%), while PCT demonstrated the highest sensitivity (51.3%) and negative predictive value (NPV = 71.6%).

Discussion

This study is the first in the literature to suggest that PAR and PCT are valuable biomarkers for diagnosing bacteremia in cancer patients. The identified cut-off values offer practical thresholds for bacteremia diagnosis.

GRAPHİCAL ABSTRACT

Keywords:

• Procalcitonin
• cancer
• bacteremia
• procalcitonin/albumin ratio
• cut-off value

Transparency

Declaration of funding

There is no sponsorship/funding.

Declaration of financial/other relationships

The authors of this paper declare no conflicts of interest related to the research, funding, or any other aspect that could potentially influence the results or interpretation of this study. We have received no financial support or grants from any organizations or companies. Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Author contributions

Y. Nadir was involved in the manuscript’s conception, design, and writing. P. Kiran carried out the analysis, SS. Karabulut helped with data interpretation, and HS. Barut, SS. Senger, and M. Degirmenci critically revised the manuscript. All authors agree to be accountable for all aspects of the work.

Acknowledgements

No assistance in the preparation of this article is to be declared.