Summary

Multivariable analyses of evaluation items (demographic data, preoperative range of motion, and preoperative clinical scores) between the postoperative good results group and the poor results group at the follow up for more than 2 years after manipulation under anesthesia for frozen shoulder showed that diabetes mellitus and smoking as the risk factors.

Abstract

Introduction

Manipulation under anesthesia (MUA) for frozen shoulder (FS) is a useful treatment that can perform in an outpatient. Many good results have been reported regarding the postoperative results of MUA. However, we often experience poor results such as persistent pain and limited range of motion. This study aimed to evaluate the preoperative prognostic factors influencing of postoperative results of MUA for FS.

Methods

From 2014 to 2018, MUA was performed for FS, and 57 shoulders who were able to be followed up for more than 2 years after MUA were examined. Patients with = 135 degrees of passive flexion without rotator cuff tear, traumatic contracture, or any other known disorders that could result in contracture as a complication were defined as FS. There were 31 shoulders in 28 males and 26 shoulders in 25 females, and the average age at MUA was 57.5 ± 10.0 years (42-83 years). MUA was performed by shoulder joint mobilization with cervical 5,6 nerve blocks under the ultrasound guidance. Patients demographic data (age, gender, disease period, diabetes mellitus, hypertension, hyperlipidemia, thyroid disease, steroid use, and smoking), preoperative shoulder range of motion (active flexion, passive flexion, active external rotation, active internal rotation), preoperative clinical scores {Constant shoulder score, University of California at Los Angeles Shoulder score, Numerical Rating Scale (NRS)} were evaluated. These items were divided into good results group and poor results group. The poor results group was defined as those who underwent multiple MUA and those whose NRS at the last evaluation did not improve to less than half of that before MUA. Univariable analysis was performed between 2 groups using ?² test and Fisher exact test on the categorical variables and Mann-Whitney U test on the numeric variables. Evaluation items with P value < 0.05 were added to multivariable analyses. Finally, evaluation items with P value < 0.05 in multivariable analyses were determined as the independent preoperative prognostic factors influencing of postoperative results of MUA for FS.

Results

The good results group had 42 shoulders (73.7%), and the poor results group had 15 shoulders (26.3%). The poor results group consisted of 11 shoulders those who underwent multiple MUA and 4 shoulders those whose NRS at the last evaluation did not improve to less than half of that before MUA. Disease period, diabetes mellitus, and smoking were significantly higher in the poor results group at univariable analysis (P = 0.02, 0.01, and 0.01, respectively). Multivariable analyses showed that diabetes mellitus and smoking were significantly higher in the poor results group (P = 0.01 and 0.02, respectively).

Conclusions

We evaluated the preoperative prognostic factors influencing of postoperative results of MUA for FS. Diabetes mellitus and smoking were significantly higher in the poor results group than the good results group by multivariable analysis. These factors may be the factors for predicting postoperative results of MUA for FS.