Effect of high‐flow oxygen therapy on regional oxygen saturation during vaso‐occlusive pain crisis: An observational study

The vaso-occlusive pain crisis (VOC) is the hallmark manifestation of sickle cell disease (SCD). Hypoxia-induced polymerization of haemoglobin S, red blood cell sickling and microvascular occlusion accounts for the main disease process underlying VOC. Hypoxia seems to be a critical factor inducing the polymerization of deoxygenated sickle haemoglobin, resulting in multicellular adhesion, vascular obstruction and ischaemia–reperfusion injury.1 High-flow nasal oxygen (HFNO) therapy, by inducing arterial hyperoxemia, may mitigate the postcapillary venules desaturation and the associated polymerization. By measuring regional oxygen saturation within a sampled volume mixing arterioles, capillaries and venules, near-infrared spectroscopy (NIRS) allows a non-invasive monitoring of the microcirculation. Patients with SCD exhibit lower values of muscular and cerebral regional oxygen saturation at steady state.2, 3 The effects of oxygen therapy on microcirculation are unknown during VOC. We aimed at describing these effects using NIRS. The study was conducted in an intensive care unit within a university hospital during a two-year period. Adult patients with SCD admitted for VOC were eligible for inclusion if: (1) the painful site of the VOC included at least one limb; (2) the physician in charge of the patient decided to deliver oxygen therapy by using HFNO. This study was approved by an external review board (‘Comité de Protection des Personnes Ile-de-France V’). The requirement for informed consent was waived because of the observational nature of the study, as per French law. All patients received written and oral information about the study. After a 5-min period with standard oxygen therapy (i.e. targeting a pulsed oxygen saturation of 95%), HFNO was delivered by using cannulas (Optiflow™ or Airvo-2™, Fisher-Payckel), during 1 h, at flows of 30–60 L/min and fraction of inspired oxygen (FiO2) of 100%. We recorded baseline characteristics including the type of haemoglobinopathy, the steady-state haemoglobin value and the history and long-term treatment of SCD. Pain and regional oxygen saturation (rSO2) were assessed for each limb site with a categorical pain score (CPS, ranging from 0 to 3 points) and a NIRS monitor (INVOS™ 5100C; Medtronic), respectively.4 Biological data were measured before the HFNO. Three sensors were positioned simultaneously on the forehead (cerebral rSO2) and symmetrically on two limbs (on the flexor digitorum superficialis muscle at the proximal third of the forearm or on the rectus femoris muscle at the middle third of the thigh) above a muscle area (muscular-rSO2). For the peripheral measurement, most painful and less painful limbs referred to the most and the less severe painful side of the limb studied (as assessed by CPS), respectively. rSO2 values were recorded every 30 s and then averaged over 3 min before administering HFNO (T0) and 60 min after HFNO at 100% FiO2 (T60). Sample size calculation and statistical methods are reported in Data S1. Ten patients were included. Patients' characteristics are provided in Table S1. Two patients received red blood cell transfusion before inclusion, and no patient had a cerebral vasculopathy. At baseline, the CPS of the most painful limb was higher as compared to that of the less painful limb: 2.5 (2.0–3.0) vs. 1.0 (0.0–2.0), p = 0.008. Standard oxygen therapy consisted of HFNO at FiO2 between 21% and 40% (nine patients) and low-flow nasal oxygen at 5 L/min (one patient). There were no significant differences in rSO2 measured on the most painful limb, the less painful limb and the forehead at baseline (Figure S1). During HFNO, a within-subject increase for rSO2 was observed, both at cerebral and muscular sites, with no significant change in clinical parameters except for an increase in pulsed oxygen saturation (Table 1 and Figure 1). These results persisted after exclusion of patients receiving transfusion before inclusion (Figure S2). In this exploratory study, we observed an increase in rSO2 in limbs and cerebral areas during HFNO in SCD patients with VOC. This observation could be relevant, as enhanced rSO2 may increase microvascular oxygen tension and favour the process of red blood cell unsickling.5 At baseline, there was a trend towards lower rSO2 in the most painful limb as compared to the less painful limb, but this difference did not reach statistical significance. A previous study found normal sublingual microcirculation during painful crisis in SCD.6 Another study evidenced a 30% reduction in microvascular blood flow during pain crisis as compared to steady state.7 However, no comparison between most painful and less painful limbs was performed in these studies. Values of rSO2 we observed in less painful limbs are in accordance with values of muscle tissue oxygen index previously reported in SCD patients.2 Previous studies evidenced that cerebral rSO2 was significantly lower in patients with SCD as compared to those without,8 but no such difference was evidenced for quadriceps.9 We found slightly lower values of cerebral rSO2 (41%) than previously reported in patients with SCD (48%),3 and there was a trend towards lower values of cerebral rSO2 as compared to limb values. Cerebral and muscle oxygen saturation were reported to be similar amongst patients with SCD2 but not in neonates10 nor adults11 without SCD. Lower values and/or decline of cerebral rSO2 have been associated with cognitive impairment after cardiac surgery.12 In line with our hypothesis, rSO2 increased during HFNO both in limbs and cerebral areas. A previous case report of a child without SCD evidenced improvement in cerebral oxygenation (measured by NIRS) with HFNO.13 The optimal oxygen support for patients with SCD is still controversial: an excessive oxygen supplementation could enhance oxidative stress and inflammation,14 whereas hypoxia could trigger sickling and VOC and enhance pulmonary vasoconstriction and acute chest syndrome.15 Although prolonged continuous oxygen delivery in patients with SCD was associated with a decrease in erythropoietin levels,16 a correlation between erythropoietin levels and haemoglobin is not consistently observed17 and whether hyperoxia could be detrimental in this population is still uncertain. The main strength of our study is that, to the best of our knowledge, it is the first to report rSO2 during HFNO in several patients. The improvement of tissue oxygenation and microcirculation driven by HFNO may improve unsickling during SCD, but further studies are needed to assess the needed value or change in rSO2 to improve clinical outcomes like the pain score. Whether longer exposition to HFNO will be needed to achieve these goals warrants further research. Other limitations of our study include its small sample size and the lack of a direct biological surrogate of sickling. In conclusion, a 1-h session of HFNO at FiO2 1 in SCD patients with vaso-occlusive painful crisis improved regional oxygen saturation both at muscular and cerebral sites but with no change in the pain score. JC, SG, AH, PB, KR, NDP, GC and AMD contributed to the study design, data analysis, interpretation and drafting of the manuscript. JC and FP contributed to data acquisition. All authors revised the article and gave approval for the submitted version. Medtronic provided the monitor for clinical measurements. The authors have no competing interests. The study data (protocol, patients' characteristics and measures, and statistical analysis) are available upon request. Appendix S1. Figure S1. Figure S2. 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.