DISCUSSION
In this prospective, case-control study we compared the clinical characteristics and laboratory parameters, including 10 serum proinflammatory cytokines, in SCD children with fever of different etiologies and asymptomatic steady-state controls. We found that IL-6 was a very good biomarker for the diagnosis of SBI in these patients, with similar NPV to other biomarkers and better PPV. To our knowledge, this is the first study describing IL-6 as a marker of SBI in febrile SCD children.
Most of the children included in this study had been diagnosed with SCD by newborn screening, had a complete vaccination status and were receiving penicillin prophylaxis. In the group of SCD patients with fever, we found a low proportion of confirmed SBI (5.1%), similarly to other studies performed in high-income countries1,3–5. However, in more than a half of the cases (52.6%) a virus was detected by multiplex-PCR from respiratory samples, as it has been observed in recent pediatric studies16,17. In the remaining 42.8% of cases an infectious etiology was not confirmed. The only statistically significant differences among the baseline characteristics of study subgroups were a higher percentage of patients with CVC and on hypertransfusional regimen in the group of SBI, suggesting a poorer clinical baseline condition of these patients and also highlighting the presence of long-term CVC as a predisposing factor for bacteremia6.
Despite the low proportion of confirmed SBI in this cohort, almost all patients received at least one dose of antibiotic (96.2%), in agreement to current guidelines18–20. This high use of antibiotics emphasizes the importance of defining laboratory parameters that could early discriminate the etiology of fever in these patients, in order to optimize the antimicrobial prescription. Children with SBI presented more frequently with hemodynamic instability and had higher inflammatory parameters (neutrophils, CRP and procalcitonin) than the rest of the patients, similarly to what was described in other studies1,5–9. We also found that children with SBI had significantly higher IL-6 levels when compared with those with VI, NPI and controls, with the highest value observed in the only patient with confirmed pneumococcal pneumonia. The optimal cut-off value of IL-6 for the diagnosis of SBI was 125 pg/ml, with a very high PPV (significantly better than other biomarkers) and NPV applying to different PR. In a setting with a PR of SBI of 5% in children with SCD (average PR in high-income countries1,4,5), the PPV would be 100% and NPV 98.7%. Therefore, we consider that IL-6 could be a very useful marker for the diagnosis of SCD patients with confirmed SBI.
IL-6 is an important factor in systemic and local immune response. It can be secreted by almost all stromal cells and cells of the immune system at an early stage of inflammation (with rapid elevation in 1-3 hours), and further promote the synthesis and release of CRP by the liver 21,22. IL-6 has been previously described as a marker of different types of SBI (alone or in combination with other biomarkers)23,24, such as early-onset neonatal sepsis25,26, sepsis and septic shock21,23,27, bacteremia in patients with febrile neutropenia28,29 or pneumococcal pneumonia30–33. The early elevation of IL-6 that occurs in the presence of SBI is an important advantage of this cytokine over other biomarkers.
In SCD patients, it has previously been described that IL-6 and other proinflammatory cytokines are elevated in asymptomatic steady-state patients when compared to healthy individuals10–13. However, in our study all controls (steady-state SCD patients) had undetectable levels of IL-6. A possible explanation for this could be that almost all children included in this study were receiving vitamin D supplements. According to the study performed by Adegoke et al., children with SCD had a maintained vitamin D deficiency with a high level of proinflammatory cytokines (IL-6, IL-8 and IL-18), which normalizes after vitamin D supplementation34. Other previous studies also found higher levels of IL-6 in patients with VOC14,15, a finding that we did not observe in our study.
The strengths of this study were its prospective design, the homogeneity of the patients and that it was carried out in a reference center for patients with SCD in Spain; therefore, this cohort may be quite representative of the pediatric population with SCD in high-income countries. However, it has also several limitations. Most importantly, the sample size was relatively small, with few cases of confirmed bacterial infection, and the categorization in several study groups might have decreased the statistical power. Blood samples were collected regardless the previous duration of fever and serial samples were not collected, so it was not possible to evaluate the trend over time of the different biomarkers during the febrile episode. Finally, only confirmed SBI were included in the study to avoid bias in the analysis, but maybe some not-confirmed bacterial infections (e. g. pneumonia) were included in the NPI group.
Conclusion
In this prospective study, we found that IL-6 (with an optimal cut-off of 125 pg/ml) was a good marker of SBI with a high PPV and NPV. Therefore, given its rapid elevation, IL-6 may be useful (alone or in combination with other biomarkers) to early discriminate SCD children at risk of SBI, in order to guide their management. In settings with a low PR of SBI, its use could help to reduce unnecessary antibiotic treatments and hospital admissions. These findings should be confirmed in larger cohorts and multicenter studies.