Sputum IL‐6 level as a potential predictor of acute chest syndrome during vaso‐occlusive crisis in children with sickle cell disease: Exploratory prospective prognostic accuracy study

Acute chest syndrome (ACS), a severe complication of sickle cell disease (SCD), occurs in about 20% of vaso-occlusive crisis (VOC) episodes, but robust predictive markers of ACS are lacking. ACS was found more likely to develop in the context of spine and/or pelvis pain with high reticulocyte and leukocyte or neutrophil counts, decreased platelet count, and increased levels of inflammatory mediators such as secretory phospholipase A2 (sPLA2).1, 2 However, these predictive markers have not been validated in multicenter studies, and their potential clinical role remains uncertain.2 We recently reported a major increase in the levels of interleukin-6 (IL-6) in the sputum of children with SCD during ACS,3 which led us to try tocilizumab in a young child with severe ACS and dramatically high IL-6 levels in endotracheal and pleural fluids, with a very rapidly favorable outcome.4 Also, it was reported in a pediatric cohort of 139 SCD patients at steady state that sputum IL-6 levels were twice as high as in healthy age-matched controls and were positively correlated with the number of past ACS episodes.5 To investigate the prognostic accuracy of sputum IL-6 level to predict ACS in children with SCD hospitalized for VOC, we conducted a prospective exploratory observational study between December 2020 and April 2022 in a French pediatric university hospital SCD reference center. Eligibility criteria were as follows: SCD of all types (i.e., SS, SC, S/β0, and S/β+), age ≥4 years (sputum collection being challenging under this age) and <18 years, and hospitalization for VOC. Exclusion criteria were: other diseases leading to increased pulmonary inflammation (e.g., tuberculosis, pneumonia), use of any immunomodulatory treatment in the last 3 months, and inability to obtain sputum during chest physiotherapy by autogenic drainage, performed within the first 48 h of hospitalization for VOC. Collected sputum samples were immediately frozen and stored at −80°C. All IL-6 measurements were performed at the end of the study, blinded to patient outcomes, using the ELISA technique (human IL-6 ELISA kit, R&D). A blood test, including a hemogram and C-reactive protein (CRP), was performed in all patients within 24 h before sputum collection. Clinical data were extracted from patients' electronic medical records (Table S1). ACS occurrence was defined as the association of fever and/or acute respiratory symptoms with a new pulmonary infiltrate on chest X-ray. Assessors of the ACS outcome were blinded to the results of sputum IL-6. Sputum IL-6 levels measured within the first 48 h of hospitalization were compared between VOC episodes complicated or not by ACS. Because we hypothesized that IL-6 levels in the lungs could increase at an early stage of ACS development without the diagnostic criteria for ACS having yet been met, subgroup analyses were performed in patients with or without respiratory symptoms during VOC (chest pain and/or hypoventilation and/or transient desaturation with SpO2 < 95% in room air, without auscultatory or radiographic abnormality at this stage). Statistical methods are detailed in Supplementary Material. Written informed consent was obtained from all children's parents or legal guardians. The study was approved by a medical ethics committee (GR-Ex/CPP-DC2016-2618/CNIL-MR01). During the study period, 50 consecutive VOC episodes (in 45 unique patients) were included; four VOC episodes (in four patients) were excluded because sputum could not be obtained during chest physiotherapy (Figure S1). None of these four episodes were complicated by ACS. In total, 46 VOC episodes (41 patients) were included in the analysis; 5 children presented two consecutive VOC episodes. Twelve VOC episodes (12 children) were complicated by ACS after a median delay of 1.0 [IQR: 0.5–2.0] days after sputum collection, and 34 VOC episodes (31 children) were not complicated by ACS (Table S1). Two patients presented both a VOC episode complicated by ACS and an episode without. Median sputum IL-6 level measured within the first 48 h of hospitalization was increased in VOC episodes complicated by ACS (86.3 [IQR: 0.8–4178.3] pg/mL) compared to VOC episodes without (1.0 [IQR: 0.0–8.9] pg/mL, p = .019) (Figure 1A). Of note, sputum IL-6 level was not increased for the VOC episodes with documented viral infection (Table S2). With sputum IL-6, the two other clinical and biological parameters associated with ACS occurrence were the number of past ACS episodes (median: 1.5 [IQR: 1.0–3.0] for VOC episodes complicated by ACS versus 0.0 [IQR: 0.0–1.0] for VOC episodes without, p = .005) and CRP level (median: 42.0 [IQR: 8.0–106.5] mg/L for VOC episodes complicated by ACS versus 5.0 [IQR: 2.0–11.0] mg/L for VOC episodes without, p = .014). No statistically significant correlation was found between sputum IL-6 level and any clinical or biological parameter, including the number of past ACS episodes (r = .06, p = .70) and CRP level (r = −.01, p = .94), except for hemoglobin level (r = −.44, p = .002) and platelet count (r = −.32, p = .029), which were negatively correlated with sputum IL-6 level. The area under the receiver operating characteristic (ROC) curve (AUROC) for predicting ACS (Figure 1C,D) was not significantly different between sputum IL-6 (0.72 [95% CI 0.52–0.92]) and serum CRP (0.74 [95% CI 0.56–0.92]; p = .93). Among the 12 VOC episodes complicated by ACS, four (33%) had high sputum IL-6 levels (9116 pg/mL, 3730 pg/mL, 468 pg/mL, and 155 pg/mL respectively) contrasting with CRP levels lower than 10 mg/L (1 mg/L, 1 mg/L, 9 mg/L, and 8 mg/L respectively). Sputum IL-6 and CRP thresholds that provided a specificity of 100% for predicting ACS were 150 pg/mL and 150 mg/L, respectively, for a sensitivity of 50% (95% CI 25–75) and 25% (95% CI 9–53), respectively (p = .50) (Table S3). Combining sputum IL-6 > 150 pg/mL and/or CRP >150 mg/L allowed the detection of 6 additional ACS, and hence provided a higher sensitivity (75% [95% CI 47–91]), for a specificity of 100% (95% CI 90–100), than CRP alone (p = .031). Among the 12 VOC episodes complicated by ACS, the median sputum IL-6 level was higher (5523.0 [IQR: 3730.0–9116.0] pg/mL) in episodes with respiratory symptoms (n = 5) compared to episodes without (n = 7) (0.0 [IQR: 0.0–8.3] pg/mL, p = .003) (Figure 1B). The median delay between sputum collection and ACS diagnosis was 1.0 [IQR: 0.3–1.5] day in VOC episodes with respiratory symptoms and 1.0 [IQR: 0.8–2.5] day in VOC episodes without (p = .38). Among all VOC episodes with respiratory symptoms (n = 16), the median sputum IL-6 level measured within the first 48 h of hospitalization was higher in those complicated by ACS (5523.0 [IQR: 3730.0–9116.0] pg/mL, n = 5) compared to those without (3.0 [IQR: 0.5–6.1] pg/mL, n = 11, p < .001) (Figure S2). In this subgroup of VOC episodes with respiratory symptoms, the sensitivity and specificity of sputum IL-6 level > 150 pg/mL were 100% (95% CI 56–100) and 100% (95% CI 74–100), respectively. Here, we report for the first time the potential interest of measuring sputum IL-6 level during VOC to predict ACS. Increased IL-6 levels in sputum might reflect the recruitment of innate immune cells, including monocytes, in the lungs. Indeed, the main chemokine found in sputum during ACS is monocyte-chemoattractant-protein-1 (MCP1),3 and monocyte stimulation by free HbS was recently shown to induce a predominant increase in IL-6 production mediated by TLR4.6 In our study, hemoglobin level was negatively correlated with sputum IL-6 level, which supports the hypothesis of a role for hemolysis, leading to the release of free HbS and heme, in ACS pathophysiology. Platelet count was also negatively correlated with sputum IL-6 level, which might reflect a role of platelet aggregation and consumption secondary to inflammation in ACS development. In this exploratory prospective prognostic accuracy study, we found that the two biological markers associated with ACS occurrence during VOC were CRP and sputum IL-6 levels, which highlights the importance of inflammation in ACS pathophysiology. Although the AUROC for predicting ACS was not significantly different between sputum IL-6 and CRP, no correlation was observed between these two markers, reflecting pulmonary and systemic inflammation, respectively, and the combination of sputum IL-6 > 150 pg/mL and/or CRP >150 mg/L provided a higher sensitivity for predicting ACS than CRP alone. These findings suggest that pulmonary inflammation may occur independently of systemic inflammation, consistent with our previous observation of 150-fold higher IL-6 levels in sputum compared to plasma during ACS.3 Nevertheless, systemic inflammation, which is classically observed during VOC, may promote ACS as inflammatory cytokines are known to activate endothelial cells and induce increased expression of adhesion molecules. In this study, we did not perform repeated daily measurements of sputum IL-6 levels, although our observation of increased levels, mainly in patients with respiratory symptoms without auscultatory or radiographic abnormality, suggests that sputum IL-6 levels may rise at an early stage of ACS development without the ACS diagnostic criteria having been met yet. Hence, daily monitoring of sputum IL-6 during VOC should be considered in future studies. Because sputum IL-6 levels in patients with respiratory symptoms were found to increase up to 2 days before ACS diagnosis, an early preventive treatment strategy with an anti-human IL-6 receptor monoclonal antibody might be relevant. Our single-center study has several limitations, including small sample size, which impeded multivariable analysis to determine whether sputum IL-6 is independently associated with ACS occurrence, and high patient severity (see more details in Supplementary Material). Furthermore, it would be important to investigate in future studies whether prognostic accuracy could be improved by combining sputum IL-6 with other biomarkers such as sputum IL-8, MCP1, and CCL3.3 In conclusion, sputum IL-6 level appears as a potential predictor of ACS during VOC and may help identify patients who could benefit from preventive targeted anti-inflammatory therapeutics such as tocilizumab. Large multicenter studies are needed to confirm these findings and to develop predictive scores integrating this novel marker of ACS. The authors thank Sandra Manceau and the Laboratory of Excellence GR-Ex for their valuable help. This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. No potential conflict of interest relevant to this work was reported. The data that support the findings of this study are available from the corresponding author upon reasonable request. Data S1: Supporting Information Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.