Removing the sling reduces pain immediately after shoulder surgery

Abstract

Background

A broad arm sling is traditionally applied following shoulder surgery for protection of repair and ‘for comfort’. It has never been proved however that the use of a sling actually leads to reduced pain. The aim of our prospective study was to quantify a patient's immediate post-surgical pain with and without a sling, with the hypothesis being that it is in fact more comfortable not to have it on.

Methods

Following surgery, a sling was applied as usual, and a pain score obtained within the first 4 hours. The sling was then removed and the score immediately repeated. The sling was left off until discharge, then used again as dictated by the nature of the surgery performed.

Results

The mean pain score prior to sling removal was 4.59. The mean score following sling removal was 3.24, a mean reduction in 10-point Numerical Pain Scale of 1.35 points (p < 0.001). Sixty-two out of 85 patients experienced some reduction in pain after removal of the sling. No patient reported an increase in pain.

Discussion

Our study shows that the simple action of allowing the operated arm to rest free of a sling reduces pain in the immediate post-surgical period.

Introduction

Shoulder surgery is a painful procedure and remains so even with the development of arthroscopic techniques that reduce soft tissue trauma and wound size.^1–5 Multi-modal anaesthetic and analgesic regimes have been introduced to improve pain control however post-operative pain levels remain high, especially in rotator cuff repair patients.^6–8 Acute post-operative pain has a wide range of consequences both for the health system and the patient; reducing patient satisfaction, delaying rehabilitation and recovery, necessitating higher analgesic requirements and slowing patient transit from the recovery room. The level of acute post-operative pain has also been associated with the development of chronic pain.^9–11 Any measure that can be taken to reduce acute pain post-operatively may have a profound effect on a patient's recovery and ongoing functional status. Although sling immobilisation in internal rotation (IR) immediately following surgery is traditional, evidence is emerging that this position increases pain post-operatively,^12,13 possibly secondary to abnormal soft tissue balance across the shoulder. ¹⁴ Given that current literature also shows that different positions of longer-term immobilisation (and even no immobilisation of the shoulder) are safe post-operatively without endangering surgical repair,^15–18 we set out to assess the effect on pain of temporarily removing the sling in the acute post-operative period. Our hypothesis was that removing the sling would reduce acute post-operative pain scores.

Methods

This prospective cohort study included consecutive patients presenting for elective shoulder surgery over a period of 16 months from September 2022 to March 2024. Institutional approval was obtained from the Office for Research at our centre and carried out in accordance with the National Statement on Ethical Conduct in Human Research (2007).

Surgical procedures included were arthroscopic or open rotator cuff repair, arthroscopic labral repair, Latarjet procedure, biceps tenodesis, capsular release, subacromial decompression and acromioclavicular joint procedures.

Exclusion criteria were arthroplasty, revision surgery, pain symptoms not related to the shoulder pathology, and patients who were expected to be unable to comply with the pain score assessment after surgery.

All procedures were performed by a single surgeon. Patients consented to participate in the study at their pre-operative appointment and provided written consent. To avoid any suggestion bias, patients are not told at this time what any expected change in pain would be with loosening or removal of their sling. They are simply informed that they will be asked about their pain levels at different times, and that adjustments to their sling may be made in the first 4 hours after their surgery.

All patients had a general anaesthetic with local anaesthetic infiltration to incision sites. A standard cross-body internal rotation sling with no waist strap was applied at the conclusion of the procedure while the patient was still anaesthetised on the operating table and left in position during transfer from operating table to bed. Patients were transferred to the recovery room with the sling on and monitored in the usual way through to wakefulness prior to transit back to the inpatient ward or day surgery centre.

Assessment of pain using a 10-point Numerical Pain Scale (NPS) was undertaken within the first 4 hours post-operatively. This occurred either in the recovery room or in the inpatient ward or day surgery area. The patient's ability to participate was assessed using a modified Aldrete OoR score, a validated score that measures physiological recovery from anaesthesia and is used to assess readiness for discharge from the recovery room (Figure 1). A score of >9 is considered appropriate for discharge, and we considered that this score constituted adequate wakefulness to proceed. A Numerical Pain Scale rather than a Visual Analogue Scale (VAS) was used whereby a verbal response was taken due to the logistical difficulty in writing. The patient was asked to rate their pain on a scale of 0 to 10, 10 being the worst pain they have ever experienced and 0 being no pain at all. The patient was then told that the sling was going to be adjusted, following which the strap was undone, the elbow extended and the shoulder allowed to gently rotate to neutral with the arm placed by the patient's side in bed (Figures 2 and 3). As soon as the arm was re-positioned as such, the patient was again asked to rate their pain on a scale of 0 to 10. This second score is taken no more than 20 seconds after the first – that is, just as long as it takes to loosen the sling. Unless the score was higher, the sling was removed altogether and the arm allowed to continue to rest by the side. All results were recorded contemporaneously, and the data was entered into an Excel (Microsoft M Redmond, Washington 2018, version 16.16.27) spreadsheet.

Figure 1.

Modified Aldrete score.

Figure 2.

Sling on.

Figure 3.

Sling off.

Upon patient discharge either on the same day or the following morning the sling was used as normal in a fashion dictated by the nature of the surgical procedure.

Based upon a paired samples t-test to detect a difference in NPS between sling and no sling with means equal to half the standard deviation of the difference, power of 0.80, and alpha of 0.05, 2-tailed, a sample size of 34 patients would be required. In order to have sufficient power (0.80), to detect statistically significant (0.05, 2-tailed) Pearson correlations of at least 0.30, a minimum of 85 patients (95 allowing for 10% attrition) would be required, and so the larger number was utilised.

Results

Data was collected for 85 patients from September 2022 to March 2024. After graphically checking data met assumptions such as Normality, the difference in NPS was analysed using a paired t-test. Pearson correlation was used to analyse the association between the initial NPS and the change in NPS after sling removal. Percentages were compared using independent samples binomial tests. Subgroups for rotator cuff repair and stabilisation patients were analysed separately for change in NPS, as well as testing for statistical interaction between subgroup and sling / no sling, that is, ascertaining whether the difference for one subgroup is significantly greater than the difference for the other subgroup. Level of statistical significance (alpha) was set at 0.05, 2-tailed. 95% confidence intervals (CIs) were reported wherever appropriate. Data was analysed using Stata 18 (Stata Corporation, College Station, Texas 2023).

Demographics

There were 18 females and 67 males in the study group.

Change in NPS whole patient cohort

For the whole group, the mean pain score prior to removal of sling was 4.59 (standard deviation (SD) = 2.14), and after removal of the sling was 3.24 (SD = 1.97). The mean reduction in pain score across all cases was 1.35 (SD = 1.10, 95% CI 1.12–1.59), p < 0.001 (Table 1). The range of improvement in pain score was 0 to −4 points (mean 1.35, median 1). Twenty-three subjects had no change in pain score (27%). There were no cases with an increase in pain score after sling removal. Correlation between initial NPS score with the sling on, and change in scores was −0.40, p = 0.0001.

Table 1.

NPS score before and after sling removal for entire cohort.

	Pre-sling removal	Post-sling removal
Min NPS	0	0
Max NPS	10	8
Mean NPS	4.59	3.24
95% CI	4.13–5.05	2.81–3.66
SD	2.14	1.97

CI: confidence interval; NPS: Numerical Pain Scale; SD: standard deviation.

Change in VAS by subgroup

There were 32 rotator cuff repair cases with a mean pain score of 5.25 (SD = 2.50) prior to sling removal and 3.63 (SD 2.24) after sling removal. The mean change in score was an improvement of 1.62 (95% CI = 1.18 to 2.07, p < 0.001). There were 19 stabilisation cases with a mean pain score of 4.26 (SD 1.73) prior to sling removal and 3.11 (SD 1.66) after sling removal. The mean change in pain score was an improvement of 1.16 (95% CI 0.67 to 1.65, p < 0.001). The reduction in pain score appeared to be greater with the rotator cuff repair subgroup than the stabilisation subgroup (1.62 vs. 1.15) however this difference was not statistically significant (p = 0.151).

Eight out of 32 cases (25%) in the rotator cuff repair subgroup had no change in NPS after sling removal compared with 6 out of 19 (32%) in the stabilisation subgroup. This difference (7%) is not statistically significant, 95% CI = −18.8% to 32.8%, p = 0.589 (Table 2).

Table 2.

NPS scores: rotator cuff repair group.

	Pre-sling removal	Post-sling removal
Mean NPS	5.25	3.63
SD	2.50	2.24
Median NPS	6	4

NPS: Numerical Pain Scale; SD: standard deviation.

Of the patients who experienced a reduction in score, that is excluding those where there was no change, the mean reduction in NPS was 1.85 (5.09 pre-sling removal vs. 3.24 after sling removal) (Table 3).

Table 3.

NPS scores: arthroscopic stabilisation group.

	Pre-sling removal	Post-sling removal
Mean NPS	4.26	3.11
SD	1.73	1.66
Median NPS	4	3

NPS: Numerical Pain Scale; SD: standard deviation.

Discussion

Our study demonstrated a statistically significant reduction in pain score after removal of the sling in the immediate post-operative period after elective shoulder surgery. This was the case for whole of the cohort studied, as well as for each of the subgroups analysed (rotator cuff repair and stabilisation surgery). Sixty-two out of 85 (73%) of patients reported an improvement in pain with the sling removed, and 23 out of 85 (27%) had no change. No patient experienced an increase in pain after sling removal and no patient asked for the sling to be replaced.

The mean change in pain score was a reduction of 1.35 points (SD = 1.10) on NPS for our whole group. Among those of the group where a change in score occurred, as opposed to those with no change, the reduction was 1.85 (SD = 0.85). These are statistically significant and importantly likely represent clinically significant reductions in pain, in the context of an absence of clearly defined parameters for this. There is no consensus in the literature regarding what constitutes a minimal clinically important difference (MCID) in pain score, and no specific studies able to be identified examining pain scores in elective shoulder patients. A small study looking at trauma patients suggested that a 13 mm change on a 100 mm VAS score, a comparative scale to the NPS used in our study, represented the smallest measurable change in acute pain severity that was clinically important. ¹⁹ Recently a systematic review reported that the absolute MCID quoted ranged widely from 8 to 40 mm, ²⁰ and a large study of 224 acute post-operative patients showed that an absolute change of 10 mm for the 100 mm VAS signifies a clinically important improvement in pain. ²¹ Our mean change of 1.35 in NPS we believe represents a clinically significant reduction in pain. There are numerous potential positive implications of this reduction in pain including magnitude of analgesic use, length of hospital stay and overall patient satisfaction and recovery.

Rotator cuff repair procedures were more painful than stabilisation procedures which is consistent with a general understanding that rotator cuff repair is a particularly painful procedure. Our study showed a mean NPS for rotator cuff repair of 5.25 and a mean NPS for stabilisation of 4.26. Following removal of the sling, NPS reduced in the rotator cuff repair group by 1.62 points, and in the stabilisation group by 1.15 points. There is no statistical significance between the magnitude of these reductions, however the trend towards a greater reduction in NPS in the more painful rotator cuff repair group suggests a potentially greater utility in the intervention of sling removal in this group of patients. Our study was not adequately powered for subgroup analysis however to prove this.

Shoulder surgery is associated with high levels of post-operative pain,^22–24 and severe enough to necessitate opioid use for at least several days afterwards.^2,25,26 This remains the case even with the development of minimally invasive arthroscopic techniques that reduce soft tissue trauma and wound healing requirements. Pain is still severe particularly in the first 48 hours post-operatively^1–5 and initial analgesic requirements are often similar to those following open surgery. ² There is also a widely reported phenomenon of ‘rebound pain’ or ‘pain bounce’ on day 1 post-operatively.^1,4 Rotator cuff repair surgery is associated with particularly high levels of pain.^{1,4,7,14,27–30}

Most arthroscopic surgery is performed as day case surgery, the benefits of which include shorter hospital stay, increased throughput of patients and increased patient satisfaction. ³¹ This makes it even more essential to control pain to allow for planned and timely discharge, and comfort in the home environment. Pain is the most common cause of recovery room delays, affecting 24% of patients overall, ³² as well as being the most common reason for delayed discharge from hospital^33,34 and unanticipated re-admission after outpatient surgery.^33,35–38 Delayed discharge also has financial and logistical implications for both the health system and the patient.

The morbidity associated with pain can delay return to normal daily activities and reduce health-related quality of life.^39–43 Inadequate management of pain reduces patient satisfaction with their anaesthesia and overall surgical experience, and increases distress.^39,41,44 This in turn can reduce compliance with follow-up or rehabilitation.

Inadequately treated post-operative pain may lead to persistent post-operative pain,^39,45,46 which is a significant independent predictor of delayed recovery.^32,39 An increase in pain impulses entering the spinal cord results in hyper-excitability in dorsal horn neurons and central sensitisation, leading to reduced pain threshold and amplification of the pain response.^39,47 Studies of orthopaedic patients have demonstrated a link between acute post-operative pain and the development of chronic pain independent of other factors, and patients with a significantly greater severity of post-operative pain are more likely to develop chronic pain.^9,10,11 Even modest reductions in post-operative pain may have profound effects on the patient and their recovery.

Although opioids have long been considered a cornerstone of post-operative analgesia, recent evidence shows that they can compound the risk of developing chronic pain. ⁴⁷ A study in rats suggests that repeated post-operative morphine administration prolongs post-operative pain, ⁴⁸ whilst neuroadaptation in humans can enhance existing pain and facilitate chronic pain development through tolerance and opioid-induced hyperalgesia.^47,49 Opioids themselves have well-documented adverse side-effects including nausea and vomiting, pruritus, sleep disturbance and constipation. ²⁶ More chronically there are risks of tolerance and addiction; with more than 25% of patients on long-term opioid therapy having had them initiated post-surgery, ⁵⁰ 6% of them after relatively minor procedures. ⁵¹ Clearly any strategies to reduce the use of this class of drugs are of great benefit.

Significant work has therefore been done to develop multi-modal regimes to control post-operative pain, including regional anaesthetic blocks, joint infiltration with local anaesthetic, oral medication regimes, and local devices such as cold compression. ⁵² Jones et al. ⁵³ carried out a randomised controlled trial on elective shoulder arthroplasty patients in 2022, demonstrating that a multi-modal analgesic regime reduced opioid consumption and increased patient satisfaction with pain management. Multi-modal strategies are skill-dependent and have their own range of side-effects, and even with the advent of these regimes pain control continues to be difficult, especially after rotator cuff surgery.^7,8 It is incumbent on treating practitioners to utilise any available alternative strategies to control pain where possible, especially if these are non-pharmacological, non-invasive and easy to achieve. We have shown in our study that the profoundly simple intervention of removing a sling reduces post-operative pain. This simple strategy is easy to employ with high compliance and is safe for patients who are monitored in a hospital bed.

The analgesic effect of removing the sling presumably occurs through the change in position of the glenohumeral joint, as the position of immobilisation of the shoulder and arm can influence post-operative pain. ¹² The position of across-body IR has been used to immobilise the shoulder for centuries, even though it is a non-anatomical position with no evidence that it is optimal for its intended purpose, or certainty that it reduces pain. This style of sling was traditionally used for immobilisation of battlefield or other traumatic fractures around the shoulder girdle or arm, not for immediate management after elective shoulder reconstructive surgery. Baumgarten et al. ¹³ suggested that when the glenohumeral joint is placed in a neutral rotation position, the centre of the humeral head and glenoid are better aligned and there is more balanced tension across the anterior and posterior soft tissues than when the joint is in internal rotation. In addition, when neutral rotation position of the shoulder joint is used to manage proximal humerus fractures there is better maintenance of anatomical reduction of fragments and this appears to be a more naturally ‘anatomic’ position for the shoulder. ⁵⁴

With regards to the rotator cuff specifically, shoulder position has been shown to influence the amount of tension on rotator cuff tendons.^55,56 Tensile forces in supraspinatus are increased with shoulder adduction^57,58 and internal rotation. ⁵⁹ Adduction may reduce its vascularity. ⁶⁰ Biomechanical studies have suggested the optimal position to reduce supraspinatus tension is elevation and ER, ⁶¹ or abduction at an average of 30 degrees in the scapular plane. ⁵⁹ The degree of shoulder rotation in particular has been shown to greatly affect rotator cuff repair tension. Howe et al. ⁶² showed a two-fold increase in load between the anterior and posterior suture when the arm was placed in IR or ER, respectively. Kulwicki et al. ⁶³ have shown that placing the arm in 45 degrees ER or IR will increase load by 125% on anterior or posterior anchors, respectively.

These biomechanical observations have led to recent studies directly comparing the effects on post-operative pain of a traditional IR sling with an alternative positional sling, in a range of different shoulder pathologies. Baumgarten et al. ¹³ found that a group in a neutral rotation sling had significantly less night pain at 2 weeks post-operatively and better pain relief overall at 12 months than the IR sling group. Conti et al. found that patients with a brace in 15 degrees of external rotation had significantly lower VAS scores than those in an IR sling during the early post-operative time and up to 6 months after surgery. ¹² Our study had a more focused time frame, the immediate 4 hours post-operatively only, however with results again consistent with published outcomes in showing that changing the shoulder position from one of IR to one of neutral rotation resulted in a reduction in pain scores in 73% of our patients.

In a pertinent study by Gumina et al. normal subjects in a workplace with no identified shoulder pathology were placed into either a standard internal rotation sling or an abduction brace at 15 degrees with a more neutral rotation, and the two groups compared. Those subjects wearing the standard sling reported severe or very severe discomfort three times more often than those who used the abduction brace. ⁶⁴ And 91% of participants subjectively preferred the abduction brace. It is clear that maintaining the shoulder (whether normal or pathological) in a position of internal rotation is painful. One could hypothesise this relates to compression or other adverse positioning of inflamed anterior tissues in the region of the richly innervated rotator interval. However, causation remains ill-understood. In a practical sense however, unless there is a clinical or mechanical reason to keep a shoulder internally rotated (e.g. a tight repair of the subscapularis) then it should not be forced into this position in a sling, but rather allowed to rest in a neutral position, with or without an alternate brace.

It is also worth considering the other benefits of not being unnecessarily immobilised in a sling apart from improvement in pain, the effect on the elbow and wrist in particular. Elbow immobilisation is highly disabling for simple tasks such as eating and self-care. If carried on for the longer-term stiffness and elbow pain can develop. As wearing an ankle immobiliser boot immediately after knee surgery would be unnecessary and counterintuitive to recovery, so is immobilising the elbow after shoulder surgery. Those of our patients who did not achieve a quantitative reduction in pain with sling removal still benefitted from the free use of their elbow and hand, optimising their hospital experience and basic functional ability in the early post-operative recovery phase.

Concerns regarding the safety of adopting a neutral shoulder position out of a sling in the short-term post-operatively can be allayed by the many longer-term follow-up studies showing no increase in repair failures with different sling positions and usage protocols.^15–18 We believe that the immediate post-operative period, when the patient is in bed with a nurse to help them, is actually one of the lowest risk time periods during their recovery. Even after discharge, when the risk of inadvertent erroneous movement or injury increases, long-term studies have shown no increase in failure of the repair procedure.

Limitations of our study include the relatively small number of subjects and the multiple different pathologies involved which made subgroup analysis by pathology type difficult, other than for rotator cuff and stabilisation patients. The results may not be generalisable to arthroplasty procedures. The population studied however represents a typical shoulder practice in a large centre, and as such the results are generalisable to equivalent settings. A strength of the study is that the patient is their own control, negating a plethora of confounders in assessing pain control modalities.

Conclusion

Our study showed a significant reduction in pain with temporary removal of a standard internal rotation sling in the acute post-operative period after shoulder surgery. No patient experienced an increase in pain upon sling removal. Rotator cuff repair patients demonstrated the largest reduction in pain. Further work is planned to formally quantify any reduction in analgesic requirements and other flow-on effects of reduced post-surgical pain.