Fever Without a Source in Infants 22 to 90 Days Old: Is the Reliability of Blood Analysis Time-dependent?

To the Editors: Fever without a source (FWS) in infants under 3 months of age is a common reason for visits to pediatric emergency departments, usually caused by self-limiting viral infections.1,2 However, these infants also face a significant risk of invasive bacterial infections (2%–4%) and other potentially serious bacterial infections (PSBIs; up to 20%).3 Advances in biomarkers such as C-reactive protein (CRP) and procalcitonin (PCT) have led to the development of new clinical algorithms that minimize unnecessary treatments and hospitalizations.4,5 The objective of our study was to determine whether the duration of fever at the time of blood analysis (BA) influences the usefulness of analytical results (CRP, PCT and absolute neutrophil count) in ruling out PSBI in patients aged 22 to 90 days with FWS and in a good general condition. We conducted a retrospective observational study at a tertiary hospital, evaluating infants 22 to 90 days of age with FWS, who underwent an initial BA within the first 6 hours of fever onset, from March 1, 2011, to May 31, 2018. Patients in poor general condition or with abnormalities in the urine dipstick were excluded. Participants were divided into 2 groups: patients with an abnormal BA within 6 hours of fever and patients with normal BA who had a second BA between 6 and 24 hours. This approach, involving a second BA when the first is performed within 6 hours of fever onset, is a routine practice in our hospital. Variables such as sex, age, absolute neutrophil count, CRP and PCT values, blood culture results, urine cultures, cerebrospinal fluid cultures, rapid respiratory virus diagnosis, antibiotic administration and final clinical diagnosis were analyzed. Data were managed using Microsoft Excel and analyzed statistically with SAS software, version 9.4. We reviewed 260 medical records, of which 77 (29.6%) met the inclusion criteria. Overall, 56% were male, with a median age of 43 (interquartile range, 27) days. Thirteen patients had an abnormal BA performed within the first 6 hours of fever of which 3/13 presented PSBI: 3 urinary tract infections (1 due to Klebsiella spp and 2 due to Escherichia coli). Two blood analyses were performed in the 64 remaining patients: one within <6 hours and a second BA between 6 and 24 hours of fever. Of the 64/77 patients with normal BA within the first 6 hours of fever, classified as low risk, 13/64 (20%) showed alterations in the BA performed between 6 and 24 hours of fever. Of these, 5/13 (38.4%) had PSBI: 3 bacteremias caused by Streptococcus agalactiae, 1 urinary tract infection caused by P. mirabilis and 1 gastroenteritis caused by Salmonella spp. No patients with normal BA between 6 and 24 hours of fever were later diagnosed with PSBI or invasive bacterial infections. The detection of analytical alterations differed significantly between the first 6 hours of fever and the 6- to 24-hour period (P = 0.007). The objective of this study was to analyze the utility of acute phase reactants, such as PCT and CRP, which are already incorporated into some clinical algorithms,5 while also assessing how the timing of blood sampling about fever duration impacts the reliability of these biomarkers. This study provides evidence that the diagnostic value of blood analyses evolves over time, with samples taken between 6 and 24 hours of fever showing alterations that are pivotal in identifying PSBI, even in cases initially classified as low risk. These findings highlight the importance of considering the duration of fever when interpreting biomarker results to optimize diagnostic strategies and improve patient outcomes.