Galiën, H.T. van der (Hilde) (2014) Predictive value of biomarkers and clinical risk criteria for bacterial infection and sepsis in children with cancer and febrile neutropenia. thesis, Medicine.
![[img]](https://umcg.studenttheses.ub.rug.nl/style/images/fileicons/text.png) |
Text
GalinvanderHT.pdf Restricted to Registered users only Download (1MB) |
Abstract
Objectives: For children with cancer and chemotherapy-induced febrile neutropenia (FN), no reliable risk assessment strategy to discriminate between bacterial infection or sepsis and other causes of fever is available yet. In this study, we aimed at determining clinical variables and biomarkers that may help in risk stratification of these children. We evaluated the predictive value of the biomarkers C-reactive protein (CRP) and Procalcitonin (PCT) and two risk models: one based on low-risk exclusion criteria and the other on a risk score. Nowadays, all children with FN are hospitalized and treated with intravenous antibiotics. The ultimate aim is to reduce the length of hospital stay and use of antibiotics in low-risk patients. Patients and methods: A single-center prospective study was performed in children with cancer and FN. CRP and PCT serum levels were measured on day 1, 3 and 5 of the episode of FN. The two risk models were used only at presentation, based on medical records and clinical observations. All episodes of FN were divided in three groups: group I: patients without any bacterial infection; group II: patients with a local bacterial infection and group III: patients with bacteremia or sepsis. Groups were compared using the appropriate statistical tests. For each biomarker and risk model the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were calculated. Results: During a six-months studyperiod, 23 episodes of FN were enrolled. The predictive value of PCT was superior to that of CRP. Serum levels of PCT on day 1 and 3 were significantly higher in the group with bacteremia or sepsis. Both biomarkers predicted bacteremia best on day 3. PCT > 1.20 μg/L had a sensitivity of 80% and a specificity of 92%. CRP > 65 mg/L had a sensitivity of 100% and a specificity of 64%. Combining CRP and PCT improved the diagnostic performance. The first risk model (sensitivity 85% and specificity 40%) performed better than the second (sensitivity 69% and specificity 30%) at predicting bacterial infection, though both models clearly have a lack of specificity. By combining the risk models a sensitivity of 100% was achieved, but the specificity decreased even more to only 18%. Conclusion: PCT is more useful as a marker for the diagnosis of bacteremia in febrile neutropenia patients than CRP. Combining more than one marker is recommended in the assessment of patients with FN. Monitoring the course of the biomarkers over several days enhances the predictive performance. Using the proposed risk models at presentation is not useful, since they failed at defining patients at low risk for bacteremia or bacterial infection. The first risk model could be simplified and for the second one, reassessment after 24 hours could be useful. These biomarkers and clinical criteria – especially combined – have potential to contribute to the development of more appropriate treatment protocols in low risk patients.
| Item Type: | Thesis (Thesis) |
|---|---|
| Supervisor name: | Tissing, Dr. Wim J.E. |
| Supervisor name: | Ruggiero, Dr. Antonio and “A. Gemelli” Hospital and Università Cattolica del Sacro Cuore and Rome, Italy. |
| Faculty: | Medical Sciences |
| Date Deposited: | 25 Jun 2020 11:04 |
| Last Modified: | 25 Jun 2020 11:04 |
| URI: | https://umcg.studenttheses.ub.rug.nl/id/eprint/2387 |
Actions (login required)
 |
View Item |