Trends in Practice Among Shoulder Specialists in the Management of Frozen Shoulder: A Consensus Survey

Limitation of Any Particular Movement as a Sign of FS

The specialist and faculty groups did not agree on the restriction of any particular movement as one of the earliest signs of FS. The faculty group gave equal importance to all 3 movement restrictions (external and internal rotations, abduction, and global), and the specialist group favored restricted rotations above other combinations of restrictions. Although the ISAKOS group summarized the clinical findings of FS as “the typical finding in the FS is a global reduction of the range of motion, by definition in 2 or more planes, and equal in passive and active examination,” the group suggested that limitation of external rotation is the earliest sign observed in FS. ²³

Utility of Radiological Investigation in Established FS

Although taking a plain radiograph of the shoulder is essentially normal in patients with FS, both groups strongly agreed (>75%) that one must acquire a plain radiograph of the shoulder to rule out any underlying secondary cause, such as calcific tendinitis, glenohumeral arthritis, or a neoplastic process that classifies a stiff shoulder as a secondary stiff shoulder. ²³ Further, disuse osteopenia of the humeral head is not uncommon in FS, and that should alert the surgeon if MUA is planned. ⁴⁶

Furthermore, both groups strongly agreed that if the plain radiograph is normal, there is no need to proceed with an MRI or USG to confirm the diagnosis of FS unless there is a doubt. Also, both groups preferred MRI over USG to confirm the diagnosis of FS and rule out any underlying pathology, as the former can detect other intra-articular pathologies, has higher diagnostic accuracy, and has a lesser possibility of interobserver variation than USG. ⁶⁹ However, USG is quick, relatively inexpensive, and widely available to patients with reasonable sensitivity and specificity. ¹⁶ Future research is needed to compare USG with MRI for aiding in the diagnosis of FS.

Screening Serological Investigation

While the faculty group remained indecisive, specialist surgeons strongly agreed that HbA1c and RBS assessment should be done together rather than RBS alone to screen for diabetes in a patient who is not known to have it to avoid not identifying those who have prediabetes.

It is widely accepted that FS is common and often severe in patients who have diabetes ^3,71 and thyroid dysfunction. ^11,22,59 Hence, it is essential to rule out diabetes and thyroid dysfunction unknown to the patient. To rule out diabetes, the American Diabetes Association recommends the assessment of fasting blood sugar (FBS, >126 mg%), glycosylated hemoglobin (prediabetes, 5.7%-6.4%; diabetes, ≥6.5%), or RBS (prediabetes, 140-199 mg%; diabetes, >200 mg%) in an outpatient setting to screen for diabetes. ¹ However, in a clinical setting, the patient does not arrive fasting while visiting an orthopaedic surgeon for shoulder pain, which results in the inability to perform FBS. At the same time, RBS could often be <140 mg% in a patient who is prediabetic, resulting in missing the patient’s prediabetic status, and his or her diagnosis would now depend on whether the clinician has asked for glycosylated hemoglobin. HbA1c detects the plasma sugar fluctuation over the past 3 months. Recent literature has discussed the association between those who are prediabetic and FS, which is why the assessment of HbA1c and FBS/RBS could be of interest. ^7,49,51 Several meta-analyses suggest that lone FBS or HbA1c is neither sensitive nor specific enough to detect prediabetes, and a combination of both might help detect prediabetes. ^2,25 In contrast, Safran et al ⁵⁷ concluded that routine diabetes workup is not warranted in patients with FS. However, we believe that both (RBS/FBS and HbA1c) need to be evaluated as a standard screening method through future rigorous prospective studies to confirm their utility in detecting a patient who is prediabetic, compared with mere RBS/FBS.

Screening Thyroid Dysfunction, Especially in Women

The opinion between the 2 groups on this issue was divided. Milgrom et al ⁴¹ reported that the prevalence of thyroid dysfunction as a risk factor for FS in an age-matched population is more common in women than men, with a risk ratio of 7.3 in women and 2.6 in men. Although thyroid dysfunction is more common in women, ⁴³ we suggest that routine TSH screening in all female patients with FS without any clinical features of thyroid dysfunction may not be justified unless we have robust global data justifying the same.

Clinicopathological Staging System

Both groups broadly agreed that following the clinicopathological staging system of FS with 3 stages helps them manage FS. ^45,55 It is generally agreed that pain control is the primary aim of the freezing stage by either NSAIDs or steroids. Once the pain of the freezing stage eases, FS moves into the frozen stage, characterized by grossly restricted range of motion (ROM) and mild to moderate pain. Therefore, regular PT to improve ROM and occasional analgesics to relieve pain are essential for treating the frozen stage. With the gradual restoration of ROM and minimal pain, the FS moves into the third stage, thawing, wherein regular, sustained stretching exercises are advocated until symptoms are resolved. If the patient’s ROM does not improve over several months in the frozen stage, surgical treatment can be advocated to hasten the recovery. ⁴⁸ However, not many agree with the staging system in other surveys. ³¹

Timing of Analgesics

To relieve the pain of FS, both groups strongly agreed that NSAIDs must be administered at bedtime. Since the pain in the FS is usually felt at night, disturbing sleep quality, ^44,63 it is prudent to administer analgesics at bedtime to provide adequate pain relief and improve sleep quality. ⁴⁴ Of note, a short course of NSAIDs or any other analgesics does not alter the natural course of FS but enables the patient to carry out one’s activities of daily living more comfortably and perform PT with ease.

Options if NSAIDs Fail to Provide Adequate Pain Relief

Both groups strongly agreed that injecting steroids is the next best option if the pain is not relieved by NSAIDs. Oral steroids were the least preferred option. Multiple studies have confirmed the strong evidence favoring steroid injections in improving pain and ROM compared with placebo in the short term, with moderate evidence in the midterm. ^15,65 Many randomized trials have confirmed that injectable steroids provide superior clinical results to oral steroids. ^37,67 Further, one must not forget the systemic side effects of oral steroids, such as avascular necrosis of the hip, even if administered for a short duration. ⁴⁰ It may be prudent to avoid oral steroids for fear of raising blood sugar levels, as many patients with FS have diabetes.

Even though a few surgeons of the specialist group (13.4%) mentioned using other measures for pain relief (opioids, nerve blocks, pregabalin, etc) as the next step in case NSAIDs fail, the current literature shows no benefit of these measures in isolation. ⁷⁰ These adjunct measures work best along with other standard treatments. ⁷⁰

Steroid Injections

Surgeons in both groups also broadly agreed that the intra-articular site is preferred over the subacromial, and the preferred dose by most surgeons in both groups is 40 mg. A recent meta-analysis by Shang et al ⁶¹ confirmed that improvement in pain and functional scores are better with intra-articular injection than subacromial injection, while blood glucose level is less altered with subacromial injection. Although none of our surgeons preferred a 20-mg dose, recent randomized trials confirmed no difference between 40- and 20-mg doses. ^27,68 Further, the literature rarely supports the use of doses higher than 40 mg. ³⁹

Both groups strongly favored TA (specialist, 76.3%; faculty, 87.6%) over MP. A few studies have found TA slightly superior to MP in providing pain relief and improving ROM, ^10,58 while another author confirmed the superiority of MP. ⁸ It is important to note that steroid injections risk a transient increase in blood glucose levels occurring within 1 to 5 days in patients with diabetes. ⁶⁶ Hence, it seems that a 40-mg dose should be administered until it is proven that higher doses are more efficacious than lower ones.

Almost all surgeons in both groups agreed that a single steroid injection suffices in most situations, and another steroid injection can be given after 3 to 4 weeks if there was no response to the first one. Most surgeons strongly disagreed with the use of multiple injections at repeated intervals. In a study on 1377 patients, Erickson et al ¹⁴ concluded a futility of multiple steroid injections over a single injection in patients with FS.

Most surgeons in both groups preferred the blind technique of steroid injection based on standard landmarks over sonography-guided injection. Although Song et al ⁶² concluded in a systematic review that image-guided injection resulted in significant improvement in ROM, the superiority of image-guided injection over a blind technique needs further evaluation by more prospective comparative studies keeping the logistical and cost challenges of an image-guided procedure in mind.

Pain-Relieving PT in the Freezing Stage of FS

The beneficial role of PT is well-established in the frozen and thawing stages. ^15,19 However, many surgeons explore various adjunct pain-relieving options in the freezing stage and methods to retain and regain movement. While the specialist surgeon group remains divided over the most beneficial pain-relieving modality in the early painful stage of FS, the faculty group strongly agreed (>75%) that one must avoid any such measures in the acute stage.

Although the literature is replete with various modalities of pain-relieving PT, such as laser therapy, short-wave diathermy, ultrasonography, and hot packs, ^15,47 surgeons in another survey were also indecisive about the utility of the same. ³¹ A Cochrane database review suggested very low or low-quality evidence in favor of electrotherapy modalities in isolation for achieving pain relief in FS, except for laser therapy and extracorporeal shock wave therapy. ⁴⁷ Nevertheless, it is essential to understand that the impact of a single modality of PT on the natural course of FS is difficult to determine, as therapeutic modalities are typically applied as additional measures to other options, such as analgesics and steroids, and therapeutic exercises. ²⁶

Further, most surgeons agreed that they were against passive capsular mobilization techniques in the acute freezing stage. The literature suggests that surgeons should avoid aggressive mobilization techniques of the shoulder to avoid exacerbation of pain, ^6,13 as low-grade PT programs show better results than high-intensity programs. ^24,56

Use of HD

Most surgeons (>75%) in both groups did not prefer to perform HD, as they felt that it is more invasive than lone steroid injection, and its efficacy is similar to that of the steroid injection. Although several level 1 evidence studies have not found any difference between HD and steroid injection, ^35,50 few meta-analyses have found HD to be one of the most effective nonoperative measures in treating FS by significantly improving pain and function. ^33,70 We recommend more well-designed prospective trials to establish the utility of invasive HD over a lone steroid injection.

MUA and ACR

Both groups strongly agreed that <10% of patients require any surgical treatment, which is in concordance with the literature that nonoperative treatment of FS is successful in up to 90% of patients. ^54,64 Hence, a sincere effort of nonoperative treatment should suffice in most patients, and only a few would require any surgical intervention.

Further, our surgeons preferred ACR over MUA and had better results with ACR than MUA. However, the literature disagrees with this statement, as most studies have confirmed similar treatment outcomes with both options. ^36,53

Although the 2 groups strongly agreed that either surgical intervention should be done between 4 and 9 months, the groups remained almost equally divided on whether intervention should occur early (4-6 months) or late (7-9 months). Although most of the literature supports surgical treatment between 6 and 9 months, ^52,54 a recent study concluded that the results of surgery before 6 months (mean, 3.8 months) or after 6 months (mean, 11.1 months) are similar. Furthermore, early intervention might shorten the overall duration of symptoms in FS and is not associated with inferior clinical outcomes when compared with late operative intervention. ²¹

Regarding the safety of MUA, almost half the surgeons from both groups did experience some complications associated with MUA. However, they were not asked to specify whether the complications occurred with their patient or someone else’s. The overall complication rate reported in the literature after MUA is 0.4%, ³⁰ including fractures of the humerus, cuff and labral tear, and reflex sympathetic dystrophy. ^38,42,48 In a systematic review of 22 studies, Grant et al ¹⁷ concluded that the complication rate with either procedure (MUA or ACR) is <0.5%. Interestingly, a recently published randomized trial by Rangan et al ⁵³ concluded that ACR is a costlier procedure associated with more serious adverse events (4%) than MUA (1%). ⁵³ Nevertheless, as ACR allows the surgeon to better watch the procedure and the fear of unwanted complications during MUA that might result in legal entanglement have prompted surgeons to shift their practice toward the ACR.

Given similar outcome and complication rates between the 2 procedures, we cannot provide a consensus statement in favor of or against either procedure. Well-designed prospective studies must be conducted to establish the clinical superiority and safety of the 2 procedures, MUA and ACR.

Prognostication of FS

Both groups strongly agreed that patients with diabetes and thyroid dysfunction fare poorly compared with those without diabetes and a euthyroid condition. Although the literature concludes that the recovery of patients with diabetes is poor and prolonged compared with the recovery of those without diabetes, by both nonoperative and operative treatment, ^18,20 more prospective studies are required to confirm the recovery pattern of patients with thyroid dysfunction.