Secondary Logo

Journal Logo

FEATURED ARTICLES

Do Relaxation Exercises Decrease Pain After Arthroscopic Rotator Cuff Repair? A Randomized Controlled Trial

Weekes, Danielle G. MD; Campbell, Richard E. MD; Wicks, Eric D. MD; Hadley, Christopher J. BS; Chaudhry, Zaira S. MD; Carter, Aaron H. MD; Pepe, Matthew D. MD; Tucker, Bradford S. MD; Freedman, Kevin B. MD; Tjoumakaris, Fotios P. MD

Author Information
Clinical Orthopaedics and Related Research: May 2021 - Volume 479 - Issue 5 - p 870-884
doi: 10.1097/CORR.0000000000001723

Abstract

Introduction

Opioid medications are often the first-line treatment for postoperative pain; however, they are associated with multiple side effects and potential dependence [1]. Because of the potential adverse effects of opioid use, surgeons have implemented multimodal pain management protocols, including other treatments such as preoperative nerve blocks, NSAIDs, and cryotherapy [17]. Although these measures have decreased postoperative pain and narcotic use, the use of cognition-based therapies may further decrease the use of postoperative opioids without subjecting patients to additional costs or substantial risks. Cognitive-based relaxation is one nonpharmacologic method of pain management that warrants consideration. Cognition-based therapies for pain management focus on controlling a patient’s negative thoughts and providing a method to exercise control over how the patient perceives pain. Relaxation exercises can help modulate the perception and interpretation of painful stimuli by focusing the user’s attention away from the painful stimuli, thereby decreasing anxiety and providing a sense of control [2]. Relaxation exercises have been demonstrated to decrease pain after various surgical procedures, including orthopaedic procedures [8, 13, 17]. In multiple studies assessing pain in patients who underwent joint arthroplasty, pain scores measured immediately after patients performed relaxation exercises were lower than preexercise pain scores [5, 9, 10]. There is also evidence that relaxation exercises may decrease postoperative pain for an extended period. After spinal fusion, adolescent patients who were taught relaxation exercises reported less overall pain on the day of discharge, 2 weeks after discharge, and 1 month after discharge than a control group did [3].

However, other studies have failed to identify an analgesic effect of relaxation exercises postoperatively [6, 15]. Although previous evidence supports the incorporation of relaxation techniques into postoperative pain management protocols [3, 8, 9, 14], there are only two randomized controlled trials assessing the efficacy of relaxation exercises for postoperative pain relief after the first postoperative day [3, 6]. Both studies examined patients after spinal surgery. Furthermore, only one study investigated the effects of relaxation exercises after patients were discharged; however, this study also only enrolled adolescents [3]. Pain management during hospitalization differs from the outpatient setting because of the control of medications by nurses, ability for providers to quickly prescribe rescue intravenous and oral pain medications, and the general discomfort of staying in a hospital compared with patients’ homes. At home, patients control the dose and frequency at which they consume pain medications, potentially leading to better or worse pain management compared with the inpatient setting. An increasing number of surgical procedures are being performed in outpatient surgical centers; therefore, the efficacy of relaxation exercises should be studied in a similar setting to allow for greater generalizability [7]. It is also unclear whether the analgesic effects of relaxation exercises translate into a decrease in the consumption of opioid medication.

We therefore asked, (1) Does performing relaxation exercises after arthroscopic rotator cuff repair (ARCR) decrease pain compared with standard pain management medication? (2) Does performing relaxation exercises after ARCR decrease opioid consumption? (3) What proportion of patients who used the relaxation techniques believed they decreased their pain level, and what proportion continued using these techniques at 2 weeks? (4) Does performing relaxation exercises after ARCR affect shoulder function?

Patients and Methods

Study Design and Setting

This was a randomized controlled trial with two treatment groups. This study was performed at one private orthopaedic institution with multiple treatment centers. Patients were enrolled by orthopaedic surgeons and research coordinators between June 2017 and April 2018. The last patient completed follow-up in October 2018.

Participants

Adult patients with full-thickness rotator cuff tears undergoing ARCR were eligible for this trial. Patients included in the study underwent concomitant additional procedures including subacromial decompression, biceps tenodesis, biceps tenotomy, distal clavicle resection, labral debridement, and capsular release. Exclusion criteria included pregnancy, age younger than 18 years, concomitant severe glenohumeral arthritis, concomitant adhesive capsulitis, revision shoulder surgery, concomitant procedures not listed above, workers compensation claim, diagnosed psychological disorders (including depression, schizophrenia, post-traumatic stress disorder), history of alcohol or drug dependence, or preoperative opioid pain management. The diagnosis of severe glenohumeral arthritis was based on radiographs and was subjective. In general, we excluded patients with large humeral or glenoid exostosis, moderate humeral or glenoid sclerosis, or significant glenohumeral joint space narrowing. In addition, we excluded patients already practicing relaxation techniques, including meditation, guided imagery, and breathing exercises. Patients found to have irreparable rotator cuff tears and those undergoing additional surgical procedures not listed above were removed from the trial postoperatively.

Accounting for all Patients

During the study period, 563 patients were eligible for inclusion. However, because of the high surgical volume and limited research staff at the institution, most of these patients were not approached and we were unable to complete the informed consent process. Informed consent was obtained by trained research staff, whose schedule varied each day. Eligible patients were often unable to be informed about the study before their operation. Patients were contacted during the morning, afternoon, and evening. The researchers did not exhibit a preference for patients of a particular demographic or appointment time. One hundred seventy-nine patients were enrolled, but 33 patients were unenrolled because of changes in operative procedure that contradicted the inclusion/exclusion criteria (such as, only debridement was performed) or patient request before any postoperative data could be collected. In total, 146 patients were followed postoperatively. Of the 146 enrolled patients, 117 recorded their pain and postoperative opioid use in a postoperative journal (Fig. 1). Follow-up proportions were similar between the relaxation and control groups (relaxation: 80%, control: 81%; OR 1.1 [95% CI 0.6 to 1.5]; p = 0.90). At the 2-week postoperative follow-up visit, 138 patients provided information on pain and opioid consumption. Follow-up proportions were not different between the relaxation and control groups (relaxation: 95%, control: 94%; OR 1 [95% CI 0.2 to 4.3]; p > 0.99). Eighty-nine percent (66 of 74) of patients in the relaxation group completed the 2-week follow-up survey about the relaxation exercises. Eighty-two percent (119 of 146) of the enrolled patients completed the 6-month postoperative American Shoulder and Elbow Surgeons (ASES) questionnaire. Completion proportions were not different between the relaxation and control groups (relaxation: 78%, control: 85%; p = 0.32).

Fig. 1
Fig. 1:
This flow diagram depicts patient randomization and follow-up throughout the study.

Randomization

Before patient enrollment, the research coordinator used an electronic random-number generator to generate the random sequence for treatment group allocation. Research coordinators assigned patients to treatment groups based on the next available slot in the random allocation sequence. Patients were allocated to treatment groups in a 1:1 fashion. No randomization restrictions such as blocking were used. All surgeons were blinded to group allocation. Because of the study design, patients could not be blinded.

Description of Experiment, Treatment, or Surgery

Multiple sports medicine fellowship–trained surgeons (MDP, BST, KBF, FPT, and three nonauthor physicians) performed the ARCRs, and different physical therapists were involved in patient rehabilitation. Procedures were performed in both the beach chair and lateral decubitus position; however, similar operative techniques and standardized physical therapy protocols were used. All patients were discharged on the same day of surgery, and a consistent perioperative protocol was used for all patients (described below). Patients in the relaxation group received preoperative education on relaxation exercises. Patients in this group watched a 5-minute instructional video detailing how to perform relaxation breathing exercises. Patients also received an educational pamphlet containing the same instructions. Patients were instructed to perform the breathing exercise for approximately 2 minutes every 4 hours before taking pain medications throughout the first 5 days postoperatively. Patients were also encouraged to continue performing the breathing exercise until they stopped using opioid medications. This education did not consist of one-on-one coaching by a trained therapist or psychologist and was not intensive in nature; however, the relaxation breathing technique in our study was adapted from a previous study evaluating the effect of relaxation breathing exercises on pain and opioid consumption after general surgery [4]. The relaxation breathing instructions were as follows: (1) While mentally counting to four (1-2-3-4), inhale slowly and fill up your lungs to maximum capacity. This will correspond to approximately 3 to 5 seconds. (2) Hold your breath for a count of 2 (5-6). (3) Exhale at the same pace as you inhaled for a count of 4 (1-2-3-4). Exhalation and inhalation time should be the same. (4) Breathe in and out in a measured and harmonious manner. Concentrate on your breath while you inhale and exhale. (5) As you are breathing, let the lower jaw drop slightly as though starting a small yawn; keep the tongue quiet and resting in the bottom of the mouth. Breathe slowly and rhythmically with a pattern of inhale, hold, exhale, and rest. (6) Repeat this cycle 10 times. Gradually increase to 20 breathing cycles over time. After reviewing the material, patients were asked if they had any questions and were told to contact the research office if they thought of any additional questions.

Aftercare

All patients in the control and relaxation groups received an ultrasound-guided interscalene nerve block with ropivacaine preoperatively. The nerve block was performed in a similar fashion by board-certified anesthesiologists (nonauthors). General anesthesia was used in all patients. Local anesthesia was not used. The use of postoperative intravenous ketorolac was not standardized. Patients in both groups received 10-mg tablets of oxycodone, 12.5 mg of promethazine hydrochloride, 100 mg of docusate sodium, and a cold therapy shoulder cuff. If oxycodone tablets could not be prescribed, 10-mg hydrocodone tablets were prescribed. Postoperative NSAID and acetaminophen use was not standardized.

Variables, Outcome Measures, Data Sources, and Bias

The prespecified primary outcomes were pain and opioid consumption during the initial 5 days postoperatively. The secondary outcomes included pain and cumulative opioid consumption 2 weeks after ARCR, as well as short-term postoperative shoulder function.

To assess the primary outcome, we instructed patients to complete a pain and meditation journal, beginning on the day of surgery (postoperative day 0) and ending on the fifth day postoperatively. Pain at rest and pain with activity were recorded on a 10-cm VAS twice daily and then averaged. Pain scores on the first day postoperatively were recorded. Daily opioid consumption was also recorded in this journal. Opioid consumption was converted to morphine milligram equivalents (MMEs) based on opioid type and dose [18]. Two weeks after ARCR, patients also completed a follow-up survey assessing pain, opioid consumption, and the continued use of relaxation exercises. Shoulder function was assessed preoperatively and at 2, 6, 13, 18, and 26 weeks postoperatively using the ASES questionnaire.

Description of Study Population

Seventy-four patients were assigned to the relaxation group and 72 patients were assigned to the control group. The mean patient age was 60 ± 10 years. The mean patient BMI was 29 ± 5 kg/m2 (Table 1). A higher percentage of patients in the control group underwent concomitant distal clavicle resection than those in the relaxation group did (14% versus 4%, OR 3.8 [95% CI 1 to 14.5]; p = 0.04), and there was a higher percentage of Black patients in the control group than in the relaxation group (15% versus 4%, OR 4.3 [95% CI 1.1 to 16]; p = 0.03). Thirteen patients had notable preoperative diagnoses (Table 2).

Table 1. - Patient demographics and clinical characteristics
Characteristic Total cohort (n = 146) Relaxation group (n = 74) Control group (n = 72) p value
Age in years 60 ± 10 59 ± 9 60 ± 10 0.75
BMI in kg/m2 29 ± 5 29 ± 5 29 ± 5 0.88
Female 45% (66) 41% (30) 50% (36) 0.25
Race or ethnicity
 White 84% (122) 88% (65) 79% (57) 0.16
 Black 10% (14) 4% (3) 15% (11) 0.03
 Hispanic 4% (6) 4% (3) 4% (3) > 0.99
 Other 3% (4) 4% (3) 1% (1) 0.62
Rotator cuff tendon tear
 Supraspinatus 95% (139) 95% (70) 96% (69) 0.73
 Infraspinatus 17% (25) 19% (14) 15% (11) 0.56
 Subscapularis 15% (22) 14% (10) 17% (12) 0.59
 Teres minor 1% (2) 0% 3% (2) 0.24
 Multitendon tears 24% (35) 20% (15) 28% (20) 0.29
Number of tendons torn 1.3 ± 0.57 1.3 ± 0.58 1.3 ± 0.57 0.71
Repair technique
 Number of anchors 3.1 ± 1.2 2.9 ± 1.1 3.3 ± 1.3 0.10
 Double-row repair 79% (116) 73% (54) 86% (62) 0.07
 Single-row repair 21% (30) 27% (20) 14% (10) -
 Subacromial decompression 47% (69) 43% (32) 51% (37) 0.32
 Biceps procedure 39% (57) 36% (27) 42% (30) 0.52
 Distal clavicle resection 9% (13) 4% (3) 14% (10) 0.04
Preoperative VAS pain score 5.7 ± 2.4 5.5 ± 2.4 5.9 ± 2.4 0.31
Preoperative ASES score 47 ± 19 49 ± 20 45 ± 18 0.36
Data are presented as the mean ± SD or % (n). The sums of percentages may not equal 100% because of rounding approximations.

Table 2. - Notable preoperative diagnoses
Diagnosis Number of patients Relaxation, n Control, n
Healed acromion fracture 1 0 1
Healed clavicle fracture 1 1 0
Gouta 5 2 3
Rheumatic arthritisa 3 3 0
Gout and rheumatic arthritisa 1 0 1
Complex regional pain syndrome affecting bilateral legs and contralateral arm 1 0 1
Fibromyalgia 1 0 1
aIt is unclear which joints were affected.

Ethical Approval

Institutional review board approval was obtained before patient enrollment. This trial was registered at Clinicaltrials.gov (NCT04249089).

Statistical Analysis

An a priori power analysis was performed to determine the sample size needed to assess the primary endpoint: postoperative VAS pain score. The previously published minimum clinically important difference (MCID) for acute postoperative pain measured on a 10-cm VAS is 0.99 cm [12]. Our analysis demonstrated that a two-sided test could generate approximately 80% power with a sample size of 128 (64 patients in each group). To account for anticipated attrition, we increased the sample size.

This study used a per-protocol design, in which all who fit within the inclusion/exclusion criteria were analyzed in their originally assigned group. To assess the primary outcomes, we constructed linear mixed models with repeated measures to analyze the relationship between treatment group and time. We controlled for patient demographic covariates including age, BMI, gender, and injury laterality in the model. Similar repeated-measures models were constructed to analyze the effect of treatment group and time on ASES outcomes and patient subgroups. We used the chi-square and Fisher exact test to compare categorical data. Between-group differences in additional outcomes were analyzed using t-tests as indicated. An α level of 0.05 was used.

Results

Effect of Relaxation Exercises on Pain Scores

There was no difference in postoperative pain scores during the first 5 days postoperatively or at 2 weeks postoperatively between the relaxation and control groups. There was no difference in the resting pain scores, nor a difference in improvement in resting pain scores between the relaxation and control groups during the first 5 days postoperatively (Table 3). There was also no difference in pain with activity, nor a difference in improvement in pain with activity between the relaxation and control groups during the first 5 days postoperatively. There was also no difference in the mean (mean of resting and activity VAS scores) pain scores, and there was no difference in improvement in the mean pain scores between the relaxation and control groups during the first 5 days postoperatively (Fig. 2). At 2 weeks postoperatively, there was no difference in VAS pain scores between the relaxation and control groups (3 ± 3 versus 4 ± 2, mean difference 0 [95% CI -1 to 1]; p = 0.60).

Table 3. - Initial patient pain scores postoperatively
Pain at rest VAS scores Pain with activity VAS scores Mean (rest and activity) VAS pain scores Opioid use (MME)
Day after surgery Relaxation group Control group Mean difference (95% CI) Relaxation group Control group Mean difference (95% CI) Relaxation group Control group Mean difference (95% CI) Relaxation group Control group Mean difference (95% CI)
POD 0 3 ± 2 4 ± 2 -1 (-2 to 0) 5 ± 3 6 ± 3 -1 (-2 to 0) 4 ± 2 5 ± 2 -1 (-2 to 0) 19 ± 20 27 ± 37 -8 (-20 to 4)
POD 1 5 ± 3 6 ± 3 0 (-2 to 1) 6 ± 3 6 ± 3 -1 (-2 to 1) 5 ± 3 6 ± 3 -1 (-2 to 1) 43 ± 36 56 ± 45 -12 (-28 to 4)
POD 2 5 ± 3 5 ± 3 0 (-1 to 1) 6 ± 3 6 ± 3 0 (-1 to 1) 6 ± 3 6 ± 3 0 (-1 to 1) 45 ± 43 41 ± 37 3 (-12 to 19)
POD 3 4 ± 3 4 ± 3 0 (-1 to 1) 5 ± 3 5 ± 3 0 (-1 to 1) 5 ± 3 5 ± 3 0 (-1 to 1) 29 ± 34 32 ± 32 -3 (-16 to 10)
POD 4 4 ± 3 4 ± 3 -1 (2-1) 4 ± 3 5 ± 3 -1 (-2 to 0) 4 ± 3 5 ± 3 -1 (-2 to 0) 18 ± 26 20 ± 25 -2 (-12 to 8)
POD 5 3 ± 3 3 ± 3 0 (-1 to 1) 4 ± 3 4 ± 3 0 (-1 to 1) 4 ± 3 4 ± 3 0 (-1 to 1) 13 ±22 15 ± 23 -2 (-11 to 7)
p value p value p value p value
Relaxation versus control 0.46 0.22 0.29 0.25
Group x time interaction 0.64 0.62 0.55 0.45
Data are presented as the mean ± SD. p values were calculated from a longitudinal linear mixed model; MME = morphine milligram equivalents; POD = postoperative day.

Fig. 2
Fig. 2:
This line chart depicts the mean daily pain VAS scores during the first 5 days postoperatively. The error bars represent the 95% CI.

A subgroup analysis of patients who underwent concomitant biceps tenodesis demonstrated similar results; no difference in postoperative pain scores during the first 5 days postoperatively or at 2 weeks postoperatively between the relaxation and control groups. There was no difference in the mean (mean of resting and activity) pain scores, and there was no difference in improvement in the mean pain scores between the relaxation and control groups during the first 5 days postoperatively (Table 4). At 2 weeks postoperatively, there was no difference in VAS pain scores were between the relaxation and control groups (3 ± 2 versus 4 ± 2, mean difference: -1 [95% CI -2 to 0]; p = 0.12).

Table 4. - Biceps tenodesis subgroup analysis
Mean (resting and active) VAS pain scores Opioid use in MMEs
Day after surgery Relaxation group Control group Mean difference (95% CI) Relaxation group Control group Mean difference (95% CI)
POD 0 4 ± 2 5 ± 2 -0.9 (-2.3 to 0.5) 13 ± 21 23 ± 24 -10 (-25 to 5)
POD 1 5 ± 3 6 ± 3 -1.0 (-2.9 to 0.8) 43 ± 35 58 ± 47 -15 (-43 to 13)
POD 2 6 ± 2 6 ± 2 -0.5 (-2.2 to 1.1) 53 ± 42 43 ± 36 9 (-16 to 35)
POD 3 5 ± 3 6 ± 2 -0.8 (-2.5 to 1.0) 40 ± 38 36 ± 26 3 (-20 to 26)
POD 4 4 ± 2 5 ± 2 -0.9 (-2.5 to 0.8) 24 ± 32 24 ± 25 -1 (-21 to 19)
POD 5 4 ± 2 5 ± 3 -0.2 (-1.7 to 1.4) 17 ± 30 15 ± 26 2 (-20 to 18)
p value p value
Relaxation versus control 0.24 0.26
Group x time interaction 0.65 0.35
Data are presented as the mean ± SD. p values were calculated from a longitudinal linear mixed model; MME = morphine milligram equivalents; POD = postoperative day.

Relaxation Exercises and Opioid Consumption

Although there were no differences in postoperative pain scores during the first 5 days postoperatively, the relaxation group consumed fewer opioid medications than the control groups at 2 weeks postoperatively (309 ± 241 MMEs versus 442 ± 307 MMEs, mean difference -133 [95% CI -225 to -4]; p < 0.01). There was no difference in opioid consumption between the relaxation and control groups during the first 5 days postoperatively (Table 3). There was no difference in improvement in opioid consumption between the relaxation and control groups during the first 5 days postoperatively (Fig. 3). At 2 weeks postoperatively, patients in the relaxation group consumed fewer MMEs than those in the control group did (309 ± 241 MMEs versus 442 ± 307 MMEs, mean difference -133 [95% CI -225 to -42]; p < 0.01). Twenty percent (14 of 70) of patients in the relaxation group and 27% (19 of 70) of patients in the control group reported continued use of opioid medication 2 weeks postoperatively (OR 0.7 [95% CI 0.3 to 1.4]; p = 0.32). At 2 weeks postoperatively, there was no difference in percentage of patients reporting use of nonopioid pain medications (NSAIDs or acetaminophen) between patients in the relaxation group (57% [40 of 70]) and those in the control group (60% [42 of 70]) (OR 0.9 [95% CI 0.5 to 1.7]; p = 0.73).

Fig. 3
Fig. 3:
This line chart depicts the mean opioid consumption (morphine milligram equivalents) during the first 5 days postoperatively. The error bars represent the 95% CI.

A subgroup analysis of patients who underwent concomitant biceps tenodesis demonstrated similar results; there was no difference in postoperative opioid consumption between the relaxation and control groups during the first 5 days postoperatively or at 2 weeks postoperatively. There was no difference in opioid consumption, and no difference in improvement in opioid consumption (group × time interaction; p = 0.35) between the relaxation and control groups during the first 5 days postoperatively (Table 4). At 2 weeks postoperatively, patients in the relaxation group consumed fewer MMEs than those in the control group did (307 ± 269 MMEs versus 499 ± 358 MMEs, mean difference -192 (-373 to -10); p < 0.38).

Patients’ Perceptions of Analgesic Efficacy of Relaxation Exercises

Sixty-two percent (41 of 66) of patients believed the relaxation exercises decreased their pain levels. The other 38% (25 of 66) believed the exercises had no effect on their pain level. No patients believed the exercises increased pain levels. Additionally, 52% (34 of 66) of patients were still performing the exercises at 2 weeks postoperatively. In the relaxation group, there was no difference in VAS pain scores between patients who continued to perform the relaxation exercises at 2 weeks postoperatively and those who did not (4 ± 3 versus 3 ± 2, mean difference 1 [95% CI 0 to 2]; p = 0.19). There was no difference in MME consumption between patients who continued to perform the relaxation exercise after 2 weeks postoperatively and patients who did not (361 ± 313 MMEs versus 289 ± 232 MMEs, mean difference 72 [95% CI -59 to 204]; p = 0.28).

Relaxation Exercises and Shoulder Function After ARCR

There was no difference in postoperative shoulder function between the relaxation and control groups. There was no difference in ASES scores, nor was there an improvement in ASES scores (group × time interaction; p = 0.64) between the relaxation and control groups during the first 6 months postoperatively (Table 5). Higher 2-week postoperative pain scores were associated with lower 6-month ASES scores (r = -0.22; p = 0.2). There was no relationship between 6-month ASES scores and total 2-week opioid consumption (r = 0.16; p = 0.06).

Table 5. - Shoulder functional outcome measures over time
Timepoint ASES
Relaxation group Control group Mean difference (95% CI)
Preoperative 49 ± 20 46 ± 19 3 (-3 to 9)
2 weeks postoperatively 41 ± 13 39 ± 14 2 (-2 to 7)
6 weeks postoperatively 50 ± 19 49 ± 18 3 (-6 to 7)
3 months postoperatively 67 ± 19 62 ± 17 3 (-1 to 11)
4.5 months postoperatively 78 ± 18 74 ± 18 3 (-2 to 11)
6 months postoperatively 83 ± 17 78 ± 20 3 (-2 to 11)
p value
Relaxation versus control 0.73
Group x time interaction 0.646
Data are presented as the mean ± SD. p values were calculated from a longitudinal lineal mixed model.

Discussion

Because of the adverse effects of opioids, there is considerable interest in the use of nonopioid pain management strategies. Ideally, these strategies would be effective, inexpensive, and easy to use. There is increasing evidence that relaxation exercises can be used as an inexpensive supplemental pain management strategy [3, 5, 8-10, 14]; however, prior studies often used in-depth education or interventions requiring trained staff or occurring in an inpatient setting. The effects of these exercises have not been studied in a practical, easily reproducible protocol after arthroscopic rotator cuff repair. Our results demonstrate that the use of our easily and quickly taught relaxation exercises did not decrease pain scores or opioid consumption during the first 5 days postoperatively, but despite pain scores being no different, by 2 weeks after ARCR, the use of relaxation exercises did decrease patients’ overall narcotic consumption. This suggests that patients were able to use lower opioid dosages without experiencing increased pain levels.

Limitations

This study has several limitations. First, our trial only assessed relaxed breathing exercise taught mainly through the use of a 5-minute video. Although these exercises were not taught by a formally trained professional, they were based on prior literature [4]. Other studies have used more in-depth interventions [3, 10]; however, our goal was to create and test an easy-to-implement intervention. We believed that patients would be more receptive to a 5-minute video than a 30-minute session. We also recognize that many clinicians cannot hire a trained therapist solely to teach relaxation exercises. We aimed to assess a real-world intervention. However, more personalized and in-depth instruction may have also yielded different results. Furthermore, the use of different techniques or combinations of techniques may have also changed our results. Further research investigating these differences is necessary to clarify the effects of relaxation techniques on postoperative pain. Only 32% of potential patients were approached for enrollment, which creates a concern for selection bias. Patients were approached based on the research coordinators’ daily schedules, which varied each day. Patients were approached for enrollment throughout the morning, afternoon, and evening. Because this research was conducted at a high-volume surgical practice, it would be very difficult for all eligible patients to be approached. It is possible that patients with very brief preoperative visits may have been selected against; however, it is unlikely that this would affect the internal or external validity of this study. If selection bias was present, it would have affected both study groups. A per-protocol analysis was used, which can increase the risk of Type 1 error [11]. However, this method was used to exclude patients who were enrolled but ultimately did not undergo ARCR or who underwent a procedure included in the exclusion criteria. Inclusion of patients with only debridement or excluded procedures would not allow for fair comparison of postoperative pain. Our methods also included unblinded patients recording pain and opioid use in a journal, introducing a risk for bias. However, unblinded studies tend to demonstrate increased differences between groups and larger intervention effects, and most of our results demonstrated no difference between groups. Therefore, the chance of this bias changing our results is small. The collection of data at 2 weeks postoperative also has potential for recall bias. Patients were asked to think about how many pills they had left, as well as how many they used, to improve the accuracy of their responses. Future studies should maintain daily logs until patients are no longer taking opioid medications.

Although patients were randomly assigned to each group, there were more patients in the control group who underwent concomitant distal clavicle resection than in the relaxation group, which may have biased the control group toward worse pain scores and more opioid use. Similarly, patients were included with preoperative diagnoses (Table 2) that may have affected postoperative pain. However, the number of patients with distal clavicle resection and conditions such as prior fractures, gout, or rheumatic arthritis was relatively small and likely did not have an effect on our results. Because only a small number of patients underwent distal clavicle excision and multitendon repair, we were unable to perform subgroup analysis of these patients. These patient groups may be at higher risk for increased postoperative pain. The use of relaxation exercises in these patients may yield different results and should be investigated as well. The effects of slight variations in surgeon technique and physical therapists may have also biased the results. Patients were randomized to each group, decreasing the effect these variables might have on the results.

The interscalene nerve blocks were performed by multiple anesthesiologists. Variations in nerve block procedures may have resulted in variations in efficacy; however, interscalene nerve blocks are relatively standardized procedures. Furthermore, variations in the length of analgesia are common after nerve blocks. It is unlikely that the use of multiple anesthesiologists affected our results. Similarly, the use of intravenous ketorolac was not standardized. The effects of a single dose of ketorolac would be minimal after the first 8 to 12 postoperative hours. The use of ketorolac may have biased postoperative day 0 results but not subsequent results. Randomization also decreases the potential of bias due to these inconsistencies. Daily NSAID and acetaminophen use was also not quantified and may have differed between groups. There was no difference in the proportion of patients who used nonopioid pain medications at 2 weeks postoperative, and patients in the relaxation group were not instructed to change their use of NSAIDs or acetaminophen; it is unlikely that the teaching of relaxation exercises led to a change in the usage proportion of these medications. NSAID use should be standardized or recorded in future studies.

This study may have been underpowered to identify a statistical significant difference; however, it was powered based on the MCID, a proxy for clinical significance. Although statistical significance is important, a clinical difference is needed to change protocols. Additionally, less than 80% of patients in the relaxation group completed the ASES questionnaire at 6 months postoperative; however, more than 80% of patients completed the ASES at all other time points. The increase in attrition for the 6-month follow-up likely did not have an effect on our results. Finally, only 52% of patients were performing relaxation exercises at 2 weeks postoperative. This represents a relatively large percentage of patients included in the analysis who were not per protocol. Even though this decreases our ability to comment directly on the efficacy of relaxation exercises, our study used a practical intervention and demonstrated how this intervention would affect patients in the real world.

Effect of Relaxation Exercises on Pain Scores

Our results demonstrate that the use of relaxation exercises did not decrease pain levels during the first 5 days postoperatively or at the 2-week follow-up interval. These results did not differ with subgroup analysis of patients with concomitant biceps tenodesis. Multiple studies have assessed the analgesic effects of relaxation exercises, including controlled breathing, guided imagery, and muscle relaxation after abdominal and orthopaedic procedures (Table 6) [5, 9, 10, 13, 14, 16]. There is consistent evidence that the use of these exercises can decrease perception of pain immediately after the exercises are performed [5, 9, 10]. However, the reduction in pain is only beneficial if it lasts longer than immediately after performing the exercises. We measured pain scores throughout the first 5 postoperative days to assess the effects of relaxation exercises throughout each day. Prior studies have yielded mixed results when pain is measured throughout the postoperative period. In contrast to our results, three studies demonstrated that relaxation exercises can decrease pain during the postoperative period [2, 3, 8], with an additional study demonstrating a greater improvement in postoperative pain between the first and second postoperative day [6]. It is important to note that the study performed by Charette et al. [3] was performed with pediatric patients undergoing spinal fusion for scoliosis and may not generalize well to adult patients undergoing ARCR. The interventions described in these studies were more in-depth than to our intervention, with multiple 20- to 60-minute sessions. One prior study had similar results to our study, with no effect of relaxation exercises on postoperative pain [15]. The intervention used in that study was less intense; patients listened to a relaxation audio file during the preoperative and postoperative period. Given our results in the context of prior studies, simple, easily implemented relaxation interventions are not effective in decreasing postoperative pain. If providers want to incorporate relaxation techniques into postoperative pain management protocols, these interventions will likely require one-on-one personal interventions lasting more than 5 minutes. This may be unrealistic for healthcare providers and institutions with limited resources. For example, one study found relaxation exercises decreased pain at 1 month postoperative, but a nurse taught patients exercises preoperatively, at discharge, and called 2 weeks later to reinforce relaxation exercises [3]. Further studies evaluating the efficacy of easier-to-implement relaxation interventions after various surgeries are needed to identify practical interventions that are both time and cost-effective.

Table 6. - Prior studies assessing analgesic effects of relaxation exercises
Author Year Study type Surgical procedure Intervention Outcomes Outcome assessment time period Result of intervention
Ceccio [2] 1984 RCT Hip fracture ORIF Preoperative and postoperative instruction of the Jacobson relaxation technique: tongue and jaw relaxation, rhythmic breathing, lack of attention to thoughts, words, and speech Pain and distress while being turned, and postoperative opioid medication use First postoperative 24 hours Decreased pain, distress, and opioid use compared to control group
Charette et al. [3] 2015 RCT Pediatric spinal arthrodesis Preoperatively 30 minutes regarding the surgery, analgesic drugs, turning in bed, physical therapy exercises, and the Jacobson relaxation method; a nurse instructor was present; repeat instructions at discharge and 2 weeks postdischarge Pain, anxiety, and coping ability Throughout the first postoperative month Decreased pain and anxiety with no effect on ability to cope compared to control group
Forward & Greuter [5] 2015 RCT Total knee or hip replacement Group 1: 18 to 20 minutes of structured touch therapy was administered by a trained therapist Group 2: guided imagery instruction with 2 guided imagery audio files on MP3 player Pain, anxiety, pain medication use, and patient satisfaction Postoperatively immediately after intervention Decreased pain and anxiety compared to control group, with no effect on opioid medication use
Gavin et al. [6] 2006 RCT Lumbar and cervical surgery 20- to 30-minute preoperative meeting reviewing relaxation protocol focused on deep breathing and guided imagery and an instructional brochure Pain and opioid use First 2 postoperative days No difference in daily pain scores, but greater reduction of pain between first and second day compared to control group; increased opioid medication consumption compared to control group
Lawlis et al. [8] 1985 Retrospective Spinal surgery Patients attended a 1-hour relaxation instruction session the evening before their surgery Length of stay, complications, pain complaints, pain medication use During postoperative hospitalization Decreased amount of pain complaints, length of stay, oral opioid, and nonopioid pain medications compared with matched cohort
Lim et al. [9] 2014 Quasi-experimental Total knee replacement Three separate 1-hour individual sessions with theoretical and practical instruction focused on emotional tension, postoperative pain, relaxation breathing exercises, and guided imagery Pain, subjective and objective stress, anxiety, and self-efficacy Postoperatively immediately after intervention Decreased pain, stress, and anxiety, as well as increased self-efficacy compared with preintervention
Lin [10] 2012 Quasi-experimental Total knee or hip replacement 20-minute daily breathing relaxation and guided imagery tape from the day before surgery to postoperative day three Pain, anxiety, blood pressure, heart rate Postoperatively immediately after intervention Greater improvement in pain and anxiety compared with control, with no effect on mean blood pressure or heart rate
Seers et al. [14] 2008 RCT Total knee or hip replacement Group 1: total body relaxation exercises Pain and anxiety Four hours after intervention Greater improvement in pain (all three groups) compared with usual care control group, with no effect on anxiety
Group 2: jaw relaxation exercises
Group 3: The attention control group was asked to describe what they do, feel, and think when in pain; patients were instructed by researcher with extensive experience of teaching relaxation and attention control techniques
Thomas & Sethares [15] 2010 Quasi-experimental Total joint arthroplasty Audio CD describing methods to develop a sense of relaxation and harmony twice daily for 5 preoperative and 2 postoperative days Pain and anxiety First 3 postoperative days No effect on pain or anxiety level compared with control group
RCT = randomized controlled trial; ORIF = open reduction internal fixation.

Relaxation Exercises and Opioid Consumption

Although relaxation exercises did not decrease participants’ opioid consumption during the first 5 days postoperatively, they did decrease the total narcotics consumption at the 2-week follow-up visit. These results did not differ with subgroup analysis of patients with concomitant biceps tenodesis. The reason for this discrepancy is not clear. It is plausible that the use of relaxation exercises allowed patients in the relaxation group to obtain postoperative pain control similar to patients in the control group with the use of less opioid medication. Prior studies investigating the effect of relaxation exercises on narcotic use have also yielded mixed results, with one study demonstrating a decrease in opioid and nonopioid pain medication use during the first postoperative day and another study demonstrating an increase in opioid medication use in patients performing relaxation exercises [2, 6]. Our results, in the context of prior studies, do not strongly support the use of relaxation exercises to decrease postoperative use of opioid medications. The prior study demonstrating an increase in opioid consumption by patients performing relaxation exercises used a 20- to 30-minute instructional intervention [6], indicating that the intensity of relaxation training may not mediate the effects of relaxation exercises on opioid use. Further studies, with more rigorous protocols evaluating the effect of relaxation exercises on opioid and nonopioid pain medication use, are needed to draw stronger conclusions. Ideally, these studies would monitor pain medication use daily until patients stop using opioid medications. These studies should also provide more in-depth relaxation exercise training, use a combination of relaxation techniques, and reinforce the training at multiple points.

Patients’ Perceptions of Analgesic Efficacy of Relaxation Exercises

Although our intervention was abbreviated compared with previous studies, most of the patients in the relaxation group continued to perform these techniques 2 weeks after ARCR, and they believed the exercises decreased their pain. However, patients’ perceptions contrast with the objective results of this study. Perhaps patients experienced an improvement in pain immediately after performing the exercises, as supported by prior studies [5, 9, 10], but this analgesia rapidly dissipated, resulting in no improvement in daily VAS scores.

Relaxation Exercises and Shoulder Function After ARCR

There was no difference in short-term postoperative shoulder functional outcomes between patients in the relaxation group and those in the control group. Theoretically, improvements in postoperative pain secondary to the use of relaxation exercises may lead to improved participation with physical therapy and performance of home stretches and exercises, which could improve short-term postoperative shoulder function. This is supported by the small correlation between lower postoperative pain and increased ASES scores. However, the interventions used in this study did not lower pain scores; therefore, it is unsurprising that we did not observe a difference in postoperative shoulder function. To our knowledge, no prior study has assessed the effects of postoperative relaxation exercises on shoulder function or physical therapy participation. Further research assessing the efficacy of different relaxation protocols should also collect data regarding postoperative function to assess whether improvements in pain throughout the postoperative period translate into improved functional outcomes.

Conclusion

The preoperative administration of quick, basic relaxation exercises allowed patients to use appreciably lower opioid analgesic doses over the first 2 weeks after ARCR, without any worsening of pain scores. Most patients in the relaxation group reported a perceived benefit from the use of relaxation exercises, and more than half of the patients were still using the exercises 2 weeks after ARCR. Although our results demonstrate a potential indication for relaxation exercises, further research using more in-depth relaxation exercise training and rigorous protocols is needed to better understand the potential effects of relaxation exercises after rotator cuff repair.

References

1. Brummett CM, Waljee JF, Goesling J, et al. New persistent opioid use after minor and major surgical procedures in US adults. JAMA Surg. 2017;152:e170504.
2. Ceccio C. Postoperative pain relief through relaxation in elderly patients with fractured hips. Orthop Nurs. 1984;3:11-19.
3. Charette S, Fiola JL, Charest M-C, et al. Guided imagery for adolescent post-spinal fusion pain management: a pilot study. Pain Manag Nurs. 2015;16:211-220.
4. Flaherty GG, Fitzpatrick JJ. Relaxation technique to increase comfort level of postoperative patients: a preliminary study. Nurs Res. 1978;27:352-355.
5. Forward JB, Greuter NE. Effect of structured touch and guided imagery for pain and anxiety in elective joint replacement patients—a randomized controlled trial: M-TIJRP. Perm J. 2015;19:18-28.
6. Gavin M, Litt M, Khan A, Onyiuke H, Kozol R. A prospective, randomized trial of cognitive intervention for postoperative pain. Am Surg. 2006;72:414-418.
7. Hollenbeck BK, Dunn RL, Suskind AM, Zhang Y, Hollingsworth JM, Birkmeyer JD. Ambulatory surgery centers and outpatient procedure use among medicare beneficiaries. Med Care. 2014;52:926-931.
8. Lawlis GF, Selby D, Hinnant D, McCoy CE. Reduction of postoperative pain parameters by presurgical relaxation instructions for spinal pain patients. Spine (Phila. Pa 1976). 1985;10:649-651.
9. Lim YC, Yobas P, Chen HC. Efficacy of relaxation intervention on pain, self-efficacy, and stress-related variables in patients following total knee replacement surgery. Pain Manag Nurs. 2014;15:888-896.
10. Lin PC. An evaluation of the effectiveness of relaxation therapy for patients receiving joint replacement surgery. J Clin Nurs. 2012;21:601-608.
11. Mccoy CE. Understanding the Intention-to-treat principle in randomized controlled trials. West J Emerg Med. 2017;18:1075-1078.
12. Myles PS, Myles DB, Galagher W, et al. Measuring acute postoperative pain using the visual analog scale: the minimal clinically important difference and patient acceptable symptom state. Br J Anaesth. 2017;118:424-429.
13. Roykulcharoen V, Good M. Systematic relaxation to relieve postoperative pain. J Adv Nurs. 2004;48:140-148.
14. Seers K, Crichton N, Tutton L, Smith L, Saunders T. Effectiveness of relaxation for postoperative pain and anxiety: randomized controlled trial. J Adv Nurs. 2008;62:681-688.
15. Thomas KM, Sethares KA. Is guided imagery effective in reducing pain and anxiety in the postoperative total joint arthroplasty patient? Orthop Nurs. 2010;29:393-399.
16. Topcu SY, Findik UY. Effect of relaxation exercises on controlling postoperative pain. Pain Manag Nurs. 2012;13:11-17.
17. Uquillas CA, Capogna BM, Rossy WH, Mahure SA, Rokito AS. Postoperative pain control after arthroscopic rotator cuff repair. J Shoulder Elbow Surg. 2016;25:1204-1213.
18. Von Korff M, Saunders K, Ray GT, et al. De facto long-term opioid therapy for noncancer pain. Clin J Pain. 2008;24:521-7.
    © 2021 by the Association of Bone and Joint Surgeons