Effects of Insomnia, Anxiety and Depression on Clinical Outcomes of Frozen Shoulder After Manipulation Under Anesthesia.

Abstract

Background and purpose: Frozen shoulder is a common disorder causing pain, stiffness, disability, and sleep disturbance. While often linked to psychological comorbidities, the impact of preoperative mental health and insomnia on recovery outcomes remains unclear. The purpose of this study was to investigate the effects of preoperative anxiety, depression, and insomnia on postoperative recovery in patients undergoing manipulation under anesthesia (MUA) for frozen shoulder.

Methods: This study was a secondary analysis of a prospective observational cohort study. A total of 197 patients with primary frozen shoulder were enrolled and treated with MUA. Preoperatively, the Insomnia Severity Index (ISI) and Hospital Anxiety and Depression Scale (HADS) classified patients: 33 with clinical insomnia (ISI ≥ 15) and 45 with psychological distress (PD, including anxiety or depression, HADS anxiety ≥ 8 and depression ≥ 8). Shoulder passive range of motion (ROM), pain (visual analog scale, VAS), and functional scores (Constant-Murley Score [CMS], Oxford Shoulder Score [OSS]) were assessed at 3 and 6 months postoperatively. Outcomes were compared between patients with/without insomnia or PD. To control for multiple comparisons, all p-values were adjusted using the False Discovery Rate (FDR) method.

Results: In the insomnia analysis, the insomnia group had a significantly lower proportion of male patients (12.1% vs. 28.7%, p = 0.048) and a lower mean BMI (22.0 ± 2.4 vs. 23.0 ± 2.7, p = 0.033) compared to the group without insomnia. Furthermore, baseline abduction was significantly more restricted in the insomnia group (67.2° ± 21.9 vs. 81.8° ± 24.8, p = 0.001) and OSS was significantly worse (38.5 ± 8.1 vs. 32.6 ± 5.5, p < 0.001). Preoperative PD was more common in right-side cases (60.0% vs. 42.8%, p = 0.042) and was associated with worse baseline external rotation. Following MUA, all patient groups achieved substantial clinical improvement. At 3 and 6 months postoperatively, after FDR adjustment, there were no statistically significant or clinically meaningful differences in ROM, VAS, CMS, or OSS between patients with and without insomnia, nor between patients with and without PD. All observed mean differences between groups were small and fell well below established minimal clinically important difference (MCID) thresholds (e.g., at 6 months, CMS mean difference between PD and no PD was 1.84 [95% CI -0.92 to 4.60]; adjusted p = 0.796).

Discussion: Although preoperative clinical insomnia and PD are associated with worse baseline pain and functional presentation, they were not associated with poorer postoperative recovery following MUA in this cohort. At the 3 and 6 months postoperative follow-ups, there were no statistical differences in pain or ROM between patients with and without these preoperative conditions.