In contrast, conservative management entailed patient observation for at least 4 hours followed by a repeat chest x-ray. If after the observation period, patients were walking comfortably and without supplemental oxygen, they were discharged. Patients in the observation group underwent an intervention if they met a variety of criteria, including unstable vitals or an enlarging pneumothorax. All patients received standard care with analgesia and supplemental oxygen as needed.
A total of 316 patients were randomized, with 154 assigned to the intervention group and 162 to the observation group. The mean age for all participants was 26. Most patients were male (84.4% in the intervention group and 87.7% in the observation group) and almost half were current smokers (49.3% in the intervention and 42.5% in the observation group). The mean body mass index of participants was 21.4 in the intervention and 21.3 in the observation group. Twenty-five patients (15%) in the observation group underwent interventions for reasons specified in the research protocol (eg, “significant symptoms” such as abnormal physiologic observations and intolerable symptoms, or patient unwillingness to continue in the assigned group), and 10 patients assigned to the intervention group declined treatment.
Using a complete-case analysis, 129 of 131 patients (98.5%) in the intervention group and 118 of 125 patients (94.4%) in the observation group met the primary outcome of radiographic resolution within 8 weeks (risk difference [RD] = –4.1%; 95% CI, –8.6 to 0.5), thereby falling within the prespecified margin for noninferiority of less than 9%.
Per-protocol analysis at 8 weeks also proved observational management noninferior, with 124 of 126 patients (98.4%) in the intervention group and 123 of 130 patients (94.6%) in the observation group achieving lung reexpansion within 8 weeks (RD = –3.8%; 95% CI, –8.3 to 0.7). The time to symptom resolution was similar between groups, with a median time of 15.5 days in the intervention group compared with 14 days in the observation group (hazard ratio = 1.11; 95% CI, 0.88-1.4). A lower risk of serious adverse events (relative risk [RR] = 3.3; 95% CI, 1.37-8.1) and pneumothorax recurrence (absolute RD = 8%; 95% CI, 0.5-15.4) occurred in the observation group vs the intervention group. The average length of hospital stay for patients in the intervention group was 6.1 days, vs 1.6 days in the observation group (RR = 2.8; 95% CI, 1.8-3.6).
Two additional sensitivity analyses were performed because multiple study participants were lost to follow-up or had data collected after 8 weeks. Noninferiority was maintained when data collected after the 8-week visit were included and extended to 63 days (RD = –3.7%: 95% CI, –7.9 to 0.6). However, noninferiority was lost when missing data after 8 weeks were deemed “treatment failure” (RD = –11%; 95% CI, –18.4 to –3.5).
Continue to: WHAT'S NEW
