Shoulder & Elbow

Serious adverse events and lifetime risk of reoperation after elective shoulder arthroplasty

In a month of headline papers on large joint arthroplasty (see the Bristol group papers in The Lancet), we were delighted to see this paper from the team in Oxford (UK), which looks at the outcomes of shoulder arthroplasty across the UK using large data methods.¹ The number of total shoulder arthroplasties (TSA) performed worldwide has increased significantly over the last two decades, with the number of reverse polarity prostheses, in particular, growing at an increasing rate. As with all relatively new healthcare interventions, a phenomenon described by ‘Scott’s parabola’ will likely occur – and this article is certainly one to pour some water on unfettered expansion. The study was designed to evaluate both the survival of TSA and serious adverse events in patients undergoing TSA. The NHS England hospital episode statistics dataset were interrogated, and 58 000 elective TSAs performed in almost 52 000 adults above the age of 50 years between April 1998 and April 2017 were found. The overall lifetime risk of revision was calculated, and the rates of serious adverse events at 30 and 90 days post-surgery were recorded. Unsurprisingly, the number of TSAs per year increased more than five-fold during the study period. Age and sex stratification showed that the lifetime risk of revision surgery varied from 2.7% in women aged over 85 years, to 23.6% in men aged 55 to 59 years, with the risk of revision highest during the first five years after surgery. The risk of a serious adverse event was 3.5% at 30 days and 4.6% at 90 days and these were unsurprisingly associated with increasing age and comorbidity, but also male sex. The category of serious adverse events included pulmonary embolism, myocardial infarction, lower respiratory tract infection, acute kidney injury, urinary tract infection, and cerebrovascular events. Of these, urinary tract infection formed almost one-third of cases. All-cause death was 0.23% at 30 days and 0.47% at 90 days, with overall mortality lower than the expected rates for the general population. This is a very important piece of work and, with access to large-volume population data from a universal national healthcare system, there is less confounding information from socioeconomic, geographical, or commissioning factors. The data is extremely useful for shared decision-making with patients. All younger patients, especially men, should be aware of the relatively high likelihood of early failure and requirement for revision. Elderly patients with comorbidities form another group where patient selection is very important, meaning that postoperative care pathways should be scrupulously designed to avoid complications.

The terrible triad injury: replacement or reconstruction?

Return to sport after arthroscopic rotator cuff repair

Midshaft clavicle fractures: under meta-analysis

After the cuff: early active motion or sling?

Throughout the history of treatment of rotator cuff tears by surgical means, the pendulum has swung back and forth in regard to the degree of immobilization, or otherwise, which is prescribed postoperatively. In the early days of open repair, a natural evolution from initial caution to more permissive mobilization occurred. Yet, with the introduction of arthroscopic surgery, the pendulum has swung back towards the more cautious end of the spectrum.

With modern repair techniques and anchor technology, the question now is once again being asked if patients are missing out on the potential rehabilitation benefits of early mobilization. As a firm and inescapable endpoint, the radiological outcome of repair failure is sharp in the mind of many surgeons who instinctively err on the side of caution. In the first of two studies addressing this topic in this issue of 360, we review a randomized controlled trial performed in Alberta (Canada) comparing two rehabilitation protocols.⁵

Their intervention group received early active motion, while the control group were treated with sling immobilization after arthroscopic cuff repair. The authors recruited 206 patients with full-thickness tears who were randomized by the team to either early motion with self-weaning off the sling or sling immobilization for six weeks following surgery. Recruited patients had an average age of 56 years and underwent arthroscopic repair. Outcomes were assessed at regular intervals up to a final follow-up of 24 months and included ultrasound assessment of repair integrity at 12 months. Additional outcomes included shoulder range of movement (ROM), pain, and subjective health-related quality of life (HRQOL) scale collected by a blinded assessor. In terms of study retention, 83% of those enrolled completed 24 months of follow-up and, while at six weeks postoperatively the early motion group had significantly better ROM, by the final 24 month follow-up there were no significant differences apparent between the groups. At one year, 25% of patients demonstrated a recurrent full thickness rotator cuff tear on ultrasound examination, but again the distribution was not significantly different between the early motion and sling groups. This does seem to be a relatively small study sample size but does at least seem to demonstrate non-inferiority. Advocates for early mobilization will be pleased with these findings, but we feel this paper does not fully answer the question, although it is a certainly a level of evidence which we would commend others to pursue.

Sling versus no sling after cuff repair: take two

To continue on the theme and add to the evidence base, we would also draw 360 readers’ attention to the work of a second group based in Geneva (Switzerland) and their newly reported randomized controlled trial of sling versus no sling for arthroscopic rotator cuff repair.⁶

This group have taken a more pragmatic approach, and only randomized patients who had small or medium rotator cuff repairs performed. Their study, therefore, potentially more closely mirrors clinical practice. Furthermore, the protocol included early passive mobilization exercises in both groups for the first four weeks followed by progressive active mobilization, and the difference between the two was the presence or absence of a sling.

A total of 80 patients were randomized, with 40 in each group. Compared with the group placed in a sling, the no-sling group demonstrated greater external rotation and active elevation at six weeks, and greater internal rotation and active elevation at 12 weeks, postoperatively.

Ultrasound examination at six months after surgery showed no significant differences in terms of tendon thickness, the rates of bursitis, or in repair failure between the sling and no-sling group. The Single Assessment Numeric Evaluation (SANE) score was decreased and pain was increased with sling use. The authors conclude that sling immobilization may not be required following small or medium rotator cuff repairs and, on this evidence, we would tend to agree.

Preoperative shoulder injections are associated with increased risk of revision rotator cuff repair

The injection of non-water-soluble corticosteroids into the shoulder provides an important part of the armamentarium of the shoulder surgeon, but we should always be mindful of the potentially deleterious effects of our treatment. It has previously been recognized that the potent anti-inflammatory effect of these agents may inhibit normal healing processes when we come to repair these tissues. Furthermore, and probably less frequently, there is also a possibility of inoculation of infection into these deep tissues, or increasing infection rates in following secondary surgery. This group from Charleston, South Carolina (USA) have performed a ‘big data’ study to examine the effect of preoperative shoulder injections on the rate of revision rotator cuff repair.⁷ A retrospective investigation of patients undergoing primary rotator cuff repair and the timing of preoperative shoulder injections was undertaken. Data from privately insured subjects was obtained from 2010 and 2014, and multivariable logistic regression models were used to evaluate patients requiring revision surgery after confirming the laterality of injection and the index procedure. Altogether, arthroscopic rotator cuff repair was performed on 4959 patients during the study period, and 392 of these underwent subsequent revision surgery within three years of the index procedure. Patients receiving injections within six months of the primary rotator cuff repair were at a statistically significant increased risk of revision surgery, with adjusted odds ratios of 1.375 when up to three months prior, 1.822 when three to six months prior, and 1.237 when six to 12 months prior, but this last result was not significant. It is unclear if steroid injection was a proxy for the severity of the disease, lack of response to treatment, or other uncontrolled findings; but the results are nevertheless of concern. The risk of revision declines significantly when more than six months have elapsed between injection and ipsilateral primary surgery, and so it makes sense to try to delay intervention for this period of time if it is required. Indeed, the use of injections at all should be considered carefully in patients likely to proceed to rotator cuff repair.

Injections prior to rotator cuff repair are associated with increased rotator cuff revision rates

In a further article in this issue, and on the same topic, this group from California (USA) have also performed a separate ‘big data’ study of shoulder injection prior to rotator cuff repair.⁸ Again, large national insurance databases were interrogated to provide the study cohort and a total of 22 000 patients who received ipsilateral shoulder injections prior to rotator cuff repair were included in the eventual analysis. These patients were then matched by age, sex, body mass index, smoking, and comorbidities to a matched group who underwent rotator cuff repair without prior injections. Revision rotator cuff repair was the endpoint studied. This study again echoes the finding that patients receiving corticosteroids prior to cuff repair are more likely to undergo revision surgery, and this occurred at an odds ratio of 1.52. Furthermore, these findings were time-dependent and patients receiving injections closer to the time of index surgery were more likely to undergo revisions. The effect was also cumulative, in that patients receiving two or more injections had a greater than two-fold increased risk at a combined odds ratio of 2.12. What really is impossible to say is what happens to those patients who have an injection and do not go on to surgery.

If injections are obviating the need for surgery in some groups of patients, which even a few months in shoulder practice will convince the casual observer that they do, then the scale to which this effect might occur becomes important. If there is a significant improvement in these groups, such that subsequent surgery is avoided, then this is potentially a worthwhile pursuit, but if surgery is essentially inevitable for the majority, then these results cast doubt on how appropriate this practice may be. Regardless, surgeons would probably do well to observe the findings here both in terms of the dose effect and the duration of time to surgical intervention where the injection therapy fails.