|Home | About | Journals | Submit | Contact Us | Français|
Tapered oral dexamethasone for acute chest syndrome (ACS) in sickle cell anaemia was studied using a novel ACS assessment tool and investigational biomarkers. Twelve participants were randomized (mean age 17.3 years) before early study termination. Dexamethasone decreased duration of hospitalization for ACS by 20.8 h compared to placebo (P=0.024). Rebound pain occurred in both groups (3 dexamethasone vs. 1 placebo). Overall, dexamethasone decreased the leucocyte activation biomarker, sL-selectin; however, participants with rebound pain had higher sL-selectin within 24 h of treatment (dexamethasone or placebo). This ACS assessment tool was feasibly applied, and sL-selectin is a promising biomarker of ACS therapy.
Acute chest syndrome (ACS) is a frequent reason for hospitalization and the leading cause of death in sickle cell disease (SCD) (Vichinsky, et al 1997). ACS treatment is empirical and supportive, and more effective treatment is needed. Intravenous pulse-dose dexamethasone decreased the duration and morbidity of ACS in a clinical trial, but it also precipitated “rebound” painful events in a subset of patients (Bernini, et al 1998). The cause of corticosteroid rebound toxicity is unknown but speculated to include the downstream effects of leucocytosis, bone marrow necrosis, and abrupt corticosteroid withdrawal. Efforts to decrease corticosteroid rebound toxicity have included co-treatment with transfusion (Isakoff, et al 2008) and use of prednisone instead of dexamethasone at a lower relative dose (Kumar, et al 2010). However, decreased transfusion burden is one benefit of corticosteroids in ACS (Bernini, et al 1998), and different corticosteroid regimens for ACS have not been tested in clinical trials. Therefore, we designed a randomized, placebo-controlled, double-blind clinical trial to determine whether a tapered regimen of oral dexamethasone could decrease the duration of ACS while minimizing rebound pain. We also studied potential biomarkers of ACS therapy.
This trial (clinicaltrials.gov: NCT00530270) was conducted by the Comprehensive Sickle Cell Centers (CSCC) network. All participants provided informed consent. A central data and safety monitoring board and institutional review boards at each site approved and monitored the study. The trial was terminated early because of slow accrual and closure of the entire CSCC network (Nabel and Shurin 2008). Although the final study sample provided inadequate power to test the primary hypothesis, we explored trends in the efficacy, safety, and biomarker data to provide therapeutic and pathophysiological insights and guide the development of future studies.
The main inclusion criteria were: age ≥5 years; sickle cell anaemia (HbSS) or sickle-β0-thalassaemia; and ACS diagnosed within the preceding 24 h. Exclusions were conditions likely to be exacerbated by corticosteroid therapy. The full inclusion and exclusion criteria are in the Supplemental Methods. ACS was defined as a new lobar or segmental pulmonary infiltrate on a chest radiograph and ≥2 of the following in the 24 h preceding enrollment: temperature ≥ 38.5°C; tachypnea; dyspnea or increased work of breathing; chest wall pain; or peripheral oxygen saturation <90% in room air. ACS was classified as “mild to moderately severe” or “severe” according to pre-specified criteria (Supplemental Methods). The randomized, double-blind study intervention was a tapered regimen of oral dexamethasone or placebo (0.3 mg/kg q12h × 2, 0.3 mg/kg q24h × 2, 0.2 mg/kg q24h × 2, 0.1 mg/kg q24h ×2, then stop). All subjects received standard, protocol-directed supportive care for ACS (Supplemental Methods). A centralized, adaptive randomization schema (Begg and Iglewicz 1980, Pocock and Simon 1975) allocated subjects 1:1 to dexamethasone or placebo, stratified by site and age (<18 or ≥18 years) and severity. Randomization had to occur within 24 h of the diagnostic chest radiograph and study-drug given within 2 h of randomization.
Clinical laboratory evaluations were obtained at baseline (before study-drug) and daily until discharge. The ACS assessment tool (Figure 1) was developed for this study, and its’ face validity was established by a team of clinical SCD experts. The ACS assessment tool was applied every 4 h. Biomarkers were obtained immediately before the first dose of study-drug (baseline), 24 h (mid-point of highest dose-level) and 48 h (beginning of taper) after the first dose, and at the 1-week follow-up. The primary endpoint was the duration of ACS, defined as the interval between first dose of study-drug and the time at which all ACS assessment endpoint criteria were met (Figure 1) or the patient was discharged, whichever occurred first. Painful events in the 2 weeks after discharge for ACS were counted as “rebound” events.
High sensitivity C-reactive protein (hsCRP) and the soluble adhesion markers, sVCAM-1, sICAM-1, sP-selectin, sL-selectin and sE-selectin, were assayed by commercially available enzyme-linked immunosorbent assay kits: hsCRP (ALPCO Immunoassays, Salem, NH) (Krishnan, et al 2010); sVCAM-1, sICAM-1, sL-selectin, sP-selectin, and sE-selectin (R&D Systems, Minneapolis, MN) (Setty, et al 2003). For assessment of plasma nitric oxide (NO) metabolites, plasma was ultra-centrifuged, the nitrate present in the ultra-filtrate was reduced using nitrate reductase, and total nitrite measured fluorometrically using 2,3-diaminonaphthalene. The assay was reproducible and linear over the 10-1000 pmol range (Stuart and Setty 1999). Von Willebrand Factor antigen (VWF:Ag) was quantified by an automated immunoturbidometric assay (Instrumentation Laboratories, Bedford, MA), and VWF ristocetin cofactor activity was measured on the Chrono-Log 480 VS Aggregometer (Chrono-Log, Haverton, PA) (Krishnan, et al 2008). Whole blood tissue factor procoagulant activity was measured as previously described (Key, et al 1998). Serum sPLA2 was measured by an enzyme immunometric assay (Cayman Chemical, Ann Arbor, MI).
The intended sample size of 56 per arm was calculated using a two-group Satterthwaite t-test under the hypothesis that dexamethasone shortened the natural log of ACS duration by 3.4 (≈30 h) (Bernini, et al 1998, Griffin, et al 1994). The primary endpoint, duration of ACS, was compared between treatment groups using a Generalized Linear Mixed Model (GLMM) controlling for age group and allowing different variance estimates for each treatment group. Other duration variables used the same model. Total opioid usage was calculated by summing all opioids after conversion to IV morphine equivalents and compared between treatment groups by a Wilcoxon rank-sum test. Rebound painful events were compared between treatment groups using the Fisher’s exact test. GLMMs were used to compare (1) change from baseline in laboratory measures between treatment groups, (2) difference in post-baseline biomarker measurements between treatment groups, and (3) differences in biomarkers between the rebound vs. no rebound pain groups, regardless of treatment.
We enrolled 12 participants with HbSS (9 children, 3 adults; mean age 17.3 years, range 5 – 45) at 5 sites. Twelve were randomized and 11 received study-drug (5 dexamethasone, 6 placebo). One participant did not receive study-drug due to a local pharmacy error and was considered a dropout.
Dexamethasone appeared to reduce duration of hospitalization (41.5 vs. 62.3 h; P=0.024), but not duration of ACS, administration of supplemental oxygen, hypoxaemia, or total opioid usage (Table I). Nevertheless, the direction of treatment effects favoured dexamethasone in all efficacy outcomes (Table I). Study-drug was well tolerated and not prematurely discontinued or dose-modified. There were no statistically significant differences in adverse events or escalation of care between arms (Supplemental Table 1). However, 3 patients treated with dexamethasone had a rebound painful event (1 requiring re-hospitalization), compared to 1 with placebo (0 re-hospitalizations) (P=0.24). No marked leucocytosis (>50 × 109/l) occurred in either group. The convalescent rise in the platelet count (follow-up visit 1) was attenuated by dexamethasone compared to placebo, but other clinical laboratory values were similar between treatment groups (Supplemental Table 2). Three patients in the placebo group received a red blood cell transfusion (0 in the dexamethasone group), and this probably explains the numerically higher post-baseline haemoglobin concentration in the placebo group.
Biomarkers that showed statistically significant changes were sL-selectin, NO metabolites, and VWF:Ag (Supplemental Table 3). sL-selectin was lower at 1-week follow-up with dexamethasone compared to placebo (573.8 vs. 742.8 ng/ml, change from baseline −121.2 vs. 57.2, P<0.001). However, patients who had rebound pain (N=4), regardless of treatment group, had a higher sL-selectin concentration after 24 h of study-drug (746.0 vs. 568.7 ng/ml, change from baseline P=0.042). NO metabolites were lower after 24 and 48 h of dexamethasone compared to placebo, adjusting for covariates including baseline in the model (24 h: 6.5 vs. 9.1 nmol/ml, change from baseline −2.7 vs. 4.4, P=0.008; 48 h: 6.8 vs. 6.5, change from baseline −3.8 vs. 1.5, P=0.01). VWF:Ag decreased less from baseline at 1-week follow-up with dexamethasone compared to placebo (2.4 vs. 1.8, change from baseline −0.6 vs. −1.0, P=0.012).
This clinical trial was terminated early because of slow accrual. Episodes of ACS that met our stringent definition occurred less frequently than we anticipated, especially because we excluded mild and rapidly resolving ACS episodes that frequently occur in young children. Moreover, the unpredictable presentation of acutely ill, potential study participants was a great logistical challenge for study sites.
Even in this small sample, dexamethasone appeared to shorten duration of hospitalization by 20 h, consistent with prior studies and suggesting a large treatment effect size of corticosteroids in SCD (Bernini, et al 1998, Griffin, et al 1994). No other benefits were obvious, although the direction of treatment effects favoured dexamethasone in all efficacy outcomes (Table I). Rebound pain occurred in both groups, with numerically more rebound pain episodes with dexamethasone (P=0.24). A larger, significant difference may have emerged in the much larger intended sample size. Simply adding a taper may not have been sufficient to prevent rebound toxicity, but the magnitude of benefit of the taper, if any, cannot be determined from this study. Finally, although we attempted to standardize supportive care for all participants, differences in supportive care between treatment groups might have accounted for some of the observed treatment effect or adverse effects, especially in the small sample.
The leucocyte activation marker, sL-selectin, decreased with dexamethasone. Given that activated leucocytes contribute to vaso-occlusion (Okpala 2004), this might represent a laboratory correlate of response to corticosteroids for ACS and pain (Bernini, et al 1998, Griffin, et al 1994). Patients who had rebound pain (dexamethasone or placebo group) had higher sL-selectin after 24 h of study-drug, further suggesting that suppression of leucocyte activation might prevent rebound pain and ameliorate ACS. Dexamethasone also decreased levels of nitric oxide metabolites, probably explained by suppression of inducible nitric oxide synthase (Walker, et al 1997). As such, dexamethasone has several different effects that individually may or may not be beneficial for ACS.
Randomized, controlled acute intervention studies are lacking but critically needed for patients with SCD. Most of the care for ACS, and other acute vaso-occlusive complications, is based on anecdotal and uncontrolled clinical evidence. A multi-centre, randomized trial of tapered dexamethasone is feasible but, based on our study, may require as many as 25 clinical sites and 5 years of patient accrual. We conclude that our novel ACS assessment tool was feasible to apply and should be further studied to establish its construct validity, and that sL-selectin is a promising biomarker of ACS therapy. Future acute intervention studies for ACS should incorporate objective assessment tools and biomarkers of therapy.
Supported by grants from the National Institutes of Health (U54-HL70588, U54-HL070585).
clinicaltrials.gov identifier: NCT00530270