Outcomes of shoulder arthroplasty following axillary lymph node dissection

Abstract

Introduction

Shoulder arthroplasty with previous axillary lymph node dissection historically has unsatisfactory outcomes. We analyzed outcomes of primary shoulder arthroplasty in patients with previous axillary lymph node dissection.

Methods

Thirty-two primary shoulder arthroplasties after prior axillary lymph node dissection were performed. These patients were analyzed for patient-reported outcomes, range of motion, complications, and reoperations.

Results

Average age was 70.8 ± 7.5 years old. There were 19 anatomic total shoulder arthroplasties, four hemiarthroplasties, and nine reverse total shoulder arthroplasties. Eight were performed by a superior approach while 24 were performed by a deltopectoral approach with cephalic vein preservation. There were three complications (one deltoid dehiscence, one axillary nerve palsy, and one postoperative pneumonia). There was one revision (hemiarthroplasty to reverse total shoulder arthroplasty for cuff failure at 91 weeks), two reoperations, and no infections. Patient-reported outcomes were available for 21/26 (80.1%) of the surviving shoulders at 4.8 ± 2.0 years. Average visual analog scale pain score was 7.1 ± 14.5, Simple Shoulder Test score 8.3 ± 2.6 “yes” responses, Single Assessment Numeric Evaluation score 80.2 ± 17.4, and American Shoulder and Elbow Surgeons score 83.6 ± 14.1.

Conclusion

Axillary lymph node dissection is not a contraindication to shoulder arthroplasty. A deltopectoral exposure can be utilized without substantial risk of worsening lymphedema or wound complications. While a superior approach avoids cephalic vein injury, important approach-related complications (deltoid dehiscence and axillary nerve palsy) were observed.

Level of evidence: Level IV—case series.

Introduction

The utilization of shoulder arthroplasty continues to increase.^1–3 Additionally, there is an increasing prevalence of breast cancer with an estimated 3.1 million breast cancer survivors in the United States alone.⁴ These patients commonly undergo axillary lymph node dissection (ALND) as part of both diagnosis and treatment. While primary shoulder arthroplasty has had consistently successful outcomes,^5–7 there is added complexity in patients with prior ALND. These patients often have increased soft tissue scarring, poorer soft tissue healing in the setting of radiation, and both poor healing potential and an increased need for meticulous dissection and preservation of the cephalic vein in the setting of lymphedema.

There is a paucity of guiding literature on this population. An older study by Andrews et al.⁸ evaluated a group of patients who underwent shoulder arthroplasty after prior ALND between the years of 1979 and 1996. The authors reported poor clinical results with an infection rate of 10% and only 4/17 (23.5%) having an excellent clinical result. However, there were major confounding variables that may help explain the poor results. Eight of their 20 patients had some degree of rotator cuff deficiency, but the reverse shoulder arthroplasty (RSA) implant was not yet available in the United States and was not utilized for these patients. Half of the unsatisfactory results were due to limited active elevation in those who had rotator cuff pathology. Additionally, five of the 12 anatomic total shoulder arthroplasties (aTSAs) performed utilized a metal-backed glenoid component. Metal-backed glenoids have since been shown to have higher complication rate and unsatisfactory results.⁹

The purpose of this study was to evaluate outcomes of patients with prior ALND for breast cancer undergoing shoulder arthroplasty. We hypothesized that subjects would have more satisfactory outcomes than was indicated by the review of historical literature. With evolution of both the technique and the implant design (especially the emergence of the reverse implant) in the field of shoulder arthroplasty since publication of these original results, we expect improvement in clinical outcomes.

Methods

After approval from the institutional review board, a retrospective review of all primary shoulder arthroplasty cases performed at a single institution was conducted. An institutional shoulder arthroplasty database was queried by International Classification of Diseases Ninth Revision, Clinical Modification (ICD-9-CM) to identify all primary shoulder arthroplasties. The time period over which these data was collected was January 2005–September 2015. The ICD-9-CM codes utilized were 81.80 (aTSA), 81.81 (partial shoulder arthroplasty (hemiarthroplasty)), and 81.88 (RSA). All female patients entered into our clinical system with a cancer history were identified by one of two ways: prior ICD-9 diagnosis of cancer or self-reported cancer history on our standardized self-reported preoperative medical history form.

Following identification of all female shoulder arthroplasty patients with a cancer history, we performed direct radiographic review of all radiographs for vascular clips in the axilla. All patients with this finding were reached by telephone to confirm that they did in fact have a history of ALND for breast cancer. After we identified this population, data were collected on preoperative clinical variables (age, body mass index (BMI), diagnosis for arthroplasty, previous surgery, and age-adjusted Charlson Comorbidity Index (AACCI)), operative variables (procedure, surgical approach, posterior superior cuff integrity, subscapularis management, other concurrent procedures performed, and intraoperative complications), and postoperative variables (postoperative complications, future revisions, reoperations, range of motion (ROM), and patient-reported outcomes).

ROM data were recorded at the final patient follow-up visit. These ROM data were gathered retrospectively from existing clinical records. Active forward elevation (AFE) and active external rotation (AER) were assessed directly by the treating surgeon. Only patients that had at least one year of clinic ROM data were included in the final results for ROM. The patient-reported outcomes were recorded by direct telephone call to the patients by our research personnel. Collection of these patient-reported outcomes was performed prospectively, specifically for this study. The specific outcomes utilized were visual analog scale (VAS) pain score (out of 100), Single Assessment Numeric Evaluation (SANE) score,¹⁰ American Shoulder and Elbow Surgeons (ASES) scores,¹¹ and simple shoulder test (SST) score.¹² Only patients with patient-reported outcomes at over two-year follow-up were included. All statistics were calculated with Microsoft Excel (2013; Redmond, WA).

Results

Retrospective review of our institutional database revealed 30 patients with 32 shoulders that underwent shoulder arthroplasty from 2005 through 2015 in the setting of ipsilateral breast cancer treatment with ALND. Over this same time period, 3317 primary shoulder arthroplasties were performed (incidence of 1.0%). These patients were all female, an average age of 70.8 ± 7.5 years old (range 55.6–85.4), had an average BMI of 28.7 ± 5.7 (18.9–38.2), and had an average AACCI of 4.1 ± 1.0 (2–6). The 32 shoulders in this cohort underwent 19 aTSAs, four hemiarthroplasties, and nine reverse total shoulder arthroplasties (RTSAs) (Table 1). A superior approach was utilized in eight patients (25%) in an effort to prevent cephalic vein injury and theoretic worsening of lymphedema, while 24 patients (75%) underwent a standard deltopectoral approach with protection of the cephalic vein. The decision to perform a superior approach was either a result of a preexisting incision or a result of surgeon preference in an effort to avoid cephalic vein injury in a population at risk for lymphedema. In all deltopectoral cases, the cephalic vein was taken laterally and preserved. There were two intraoperative complications (both in the deltopectoral group): a calcar split that was repaired with dacron tapes and an avulsion of the circumflex humeral vein off the axillary vein that left a small defect in the axillary vein that was suture repaired.

Table 1.

Surgical details for each patient.

Patient	Diagnosis	Previous surgery	Procedure	Surgical approach	Posterior superior cuff	Subscapularis management	Other procedure
1	Postseptic arthritis	Arthroscopic I&D	Hemi	Deltopectoral	SS tear	LTO	Glenoid reaming, RCR
2	OA	Arthroscopic SAD	aTSA	Deltopectoral	Intact	LTO	ACJR
3	OA	None	aTSA	Deltopectoral	SS tear	LTO	RCR, ACJR
4	OA	None	aTSA	Deltopectoral	Intact	LTO	None
5	OA	None	aTSA	Deltopectoral	Intact	LTO	None
6	CTA	None	RTSA	Superior	Intact	Superior approach	None
7	OA	None	aTSA	Deltopectoral	Intact	LTO	None
8	CTA	None	RTSA	Superior	Intact	Superior approach	None
9	Fracture	None	RTSA	Deltopectoral	GT fracture	LTO	GT repair
10	OA	None	aTSA	Deltopectoral	Intact	LTO	None
11	OA	None	aTSA	Deltopectoral	Intact	LTO	Stepped posterior augment glenoid
12	OA	RCR and labral repair	aTSA	Deltopectoral	SS tear	LTO	RCR
13	OA	None	aTSA	Deltopectoral	Intact	LTO	None
14	OA	None	aTSA	Deltopectoral	Intact	LTO	None
15	AVN	None	Hemi	Superior	Intact	LTO	Glenoid reaming
16	Fracture	None	Hemi	Deltopectoral	GT fracture	LTO	GT repair
17	OA	None	Hemi	Deltopectoral	SS tear	LTO	RCR
18	Postseptic arthritis	RCR	RTSA	Superior	Intact	Superior approach	None
19	OA, symptomatic Os acromiale	None	aTSA	Superior	Intact	LTO	Acromion ORIF tension band
20	OA	None	aTSA	Deltopectoral	Intact	LTO	None
21	PTA	None	RTSA	Superior	Intact	Repaired piece	None
22	CTA	None	RTSA	Deltopectoral	Intact	LTO	None
23	OA	None	aTSA	Deltopectoral	SS tear	LTO	RCR
24	CTA	Arthroscopic SAD and RCR (two procedures)	RTSA	Superior	Intact	Superior approach	None
25	OA	Open stabilization	aTSA	Deltopectoral	SS tear	LTO	RCR
26	Fracture	None	RTSA	Superior	GT fracture	Subscap lengthening	Olecranon ORIF; GT repair
27	PTA	Magnuson-Stack	aTSA	Deltopectoral	Intact	Peel	ACJR
28	PTA	Magnuson-Stack	aTSA	Deltopectoral	SS tear	Peel	RCR, ACJR
29	OA	None	aTSA	Deltopectoral	Intact	LTO	None
30	OA	None	aTSA	Deltopectoral	Intact	LTO	None
31	OA	Arthroscopic SAD	aTSA	Deltopectoral	Intact	LTO	None
32	Fracture	None	RTSA	Deltopectoral	GT fracture	Peel	GT repair

ACJR: acromioclavicular joint resection; aTSA: anatomic total shoulder arthroplasty; CTA: cuff tear arthropathy; GT: greater tuberosity; Hemi: hemiarthroplasty; I&D: irrigation and debridement; LTO: lesser tuberosity osteotomy; OA: osteoarthritis; ORIF: open reduction, internal fixation; PTA: post-traumatic arthritis; RCR: rotator cuff repair; RTSA: reverse total shoulder arthroplasty; SAD: subacromial decompression; SS: supraspinatus.

There was one patient (1/32; 3.1%) who underwent future revision from hemiarthroplasty to RSA for progressive rotator cuff dysfunction at 1.8 years postoperatively (Table 2). This patient originally had a supraspinatus tear that was repaired at index surgery (done through a deltopectoral approach). Two others also underwent future surgery leaving three total reoperations out of 32 shoulder arthroplasties (9.4%): a deltoid dehiscence requiring repair at 0.67 years and a removal of an acromial tension band (that was placed at the time of index surgery for a symptomatic os acromiale) at 4.8 years postoperatively in a patient who sustained an axillary nerve palsy. These two patients underwent their index surgeries through a superior surgical approach. In addition to the three reoperations, there was one other complication (4/32 total; 12.5%): one patient who developed postoperative pneumonia and secondary heart failure leading to pacemaker placement. Two patients (6.3%) had worsening postoperative lymphedema, one of which resolved to baseline after one month. Both patients with increased postoperative lymphedema underwent a deltopectoral approach. There were no cases of periprosthetic joint infection.

Table 2.

Complications and outcomes.

Patient	Intraoperative complications	Surgical approach	Postoperative lymphedema	Future revision	Reoperation	Postoperative complications
1	Calcar split, dacron tape repair	Deltopectoral	No	No	No	No
2	Axillary vein injury	Deltopectoral	No	No	No	No
3	No	Deltopectoral	No	No	No	No
4	No	Deltopectoral	Yes, resolved after one month	No	No	Upper extremity DVT
5	No	Deltopectoral	Yes	No	No	No
6	No	Superior	No	No	No	No
7	No	Deltopectoral	No	No	No	No
8	No	Superior	No	No	No	Pneumonia/defibrillator placed
9	No	Deltopectoral	No	No	No	No
10	No	Deltopectoral	No	No	No	No
11	No	Deltopectoral	No	No	No	Incisional cellulitis, resolved with keflex
12	No	Deltopectoral	No	No	No	No
13	No	Deltopectoral	No	No	No	No
14	No	Deltopectoral	No	No	No	No
15	No	Superior	No	No	No	Incisional cellulitis, resolved with keflex
16	No	Deltopectoral	No	No	No	0
17	No	Deltopectoral	No	Yes	CTA, revision to RTSA (1.75 years)	0
18	No	Superior	No	No	0	0
19	No	Superior	No	No	ROH (acromial tension band; 4.77 years)	0
20	No	Deltopectoral	No	No	0	0
21	No	Superior	No	No	0	0
22	No	Deltopectoral	No	No	0	0
23	No	Deltopectoral	No	No	0	0
24	No	Superior	No	No	Deltoid dehiscence repair (0.67 years)	0
25	No	Deltopectoral	No	No	0	0
26	No	Superior	No	No	0	0
27	No	Deltopectoral	No	No	0	0
28	No	Deltopectoral	No	No	0	0
29	No	Deltopectoral	No	No	0	0
30	No	Deltopectoral	No	No	0	0
31	No	Deltopectoral	No	No	0	0
32	No	Deltopectoral	No	No	0	0

CTA: cuff tear arthropathy; DVT: deep venous thrombosis; ROH: removal of hardware; RTSA: reverse total shoulder arthroplasty.

Patient-reported outcomes were available at over two-year follow-up in 21 of 26 surviving shoulders (80.1%, three reoperations, three deceased). The average follow-up was 4.8 ± 2.0 (2.0–8.6) years. The average VAS (out of 100) was 7.1 ± 14.5 (0–50), SST was 8.3 ± 2.6 (3–12) yes responses, SANE was 80.2 ± 17.4 (40–100), and ASES was 83.6 ± 14.1 (48.3–98.3) (Table 3).

Table 3.

Patient-reported outcomes by procedure.

	Follow-up (years)	VAS	SST	SANE	ASES
Overall
Mean	4.9	11.9	8.0	76.5	79.4
Median	5.1	0.0	9.0	80.0	85.0
SD	2.1	23.5	2.8	21.0	21.2
Range	2.0–8.6	0.0–99.0	2.0–12.0	25.2–100	6.7–98.3
Hemi
Mean	6.9	5.0	8.5	65.0	79.2
Median	6.9	5.0	8.5	65.0	79.2
SD	0.3	7.1	3.5	21.2	20.0
Range	6.6–7.1	0.0–10.0	6.0–11.0	50.0–80.0	65.0–93.3
aTSA
Mean	4.7	10.7	8.9	82.1	82.9
Median	4.8	0.0	9.0	85.0	88.3
SD	2.1	26.7	2.3	20.6	23.1
Range	2.0–8.6	0.0–99.0	3.0–12.0	25.2–100.0	6.7–98.3
RTSA
Mean	4.9	18.6	4.8	63.0	68.3
Median	4.8	25.0	6.0	70.0	73.3
SD	2.3	16.8	2.2	17.2	12.7
Range	2.2–8.5	0.0–40.0	2.0–7.0	40.0–80.0	48.3–78.3

ASES: American Shoulder and Elbow Surgeons; aTSA: anatomic total shoulder arthroplasty; Hemi: hemiarthroplasty; RTSA: reverse total shoulder arthroplasty; SST: simple shoulder test; SANE: Single Assessment Numeric Evaluation; VAS: visual analog scale.

ROM data at over one-year follow-up were available for 23 of 26 surviving shoulders (88.4%). The average physical examination follow-up was 2.6 ± 1.7 (1.0–7.2) years. Average AFE was 147 ± 18 (95–160) and AER (for 22 patients) was 38 ± 9 (15–45) (Table 4).

Table 4.

Range of motion by procedure.

	Follow-up (years)	AFE (deg)	AER (deg)
Overall
Mean	2.9	142.8	37.6
Median	2.3	150.0	40.0
SD	2.1	21.4	9.3
Range	1.0–7.8	95–160	15–45
Hemi
Mean	2.0	135.0	40.0
Median	2.0	135.0	40.0
SD	NA	NA	NA
Range	NA	NA	NA
aTSA
Mean	2.8	155.4	41.2
Median	2.3	160.0	40.0
SD	1.9	8.8	4.2
Range	1.0–7.6	130–160	30–45
RTSA
Mean	3.3	116.7	28.0
Median	1.8	122.5	20.0
SD	2.8	17.5	13.5
Range	1.2–7.8	95–135	15–45

AER: active external rotation; AFE: active forward elevation; aTSA: anatomic total shoulder arthroplasty; deg: degrees; Hemi: hemiarthroplasty; RTSA: reverse total shoulder arthroplasty; NA: not applicable.

Discussion

ALND in the setting of breast cancer treatment can theoretically increase the risk of lymphedema on the ipsilateral side. When evaluating results of shoulder arthroplasty in patients with prior ALND, we observed successful postoperative results with low complication and reoperation rates. Eight patients underwent a superior approach in an effort to limit lymphedema exacerbation by preserving the cephalic vein. Importantly, two of these patients (25%) sustained an approach-related complication (deltoid dehiscence requiring a return to the operating room and an acromial tension band failure with subsequent hardware removal in setting of an axillary nerve palsy). Both axillary nerve palsy and deltoid dehiscence have been described with the utilization of the superior approach for shoulder arthroplasty.¹³ The low rate of permanent postoperative lymphedema with the deltopectoral approach raises the question of whether a superior approach offers a true advantage in this clinical setting. Our current approach is to use a deltopectoral approach in all patients.

Given the increase in both shoulder arthroplasty utilization^1–3 and breast cancer survival in the United States⁴ knowledge of the efficacy of shoulder arthroplasty in this population becomes increasingly important. While primary shoulder arthroplasty has shown successful outcomes,^5–7 breast cancer patients may be at increased risk for poor outcomes given the prevalence of chest wall surgery, radiation therapy, and ALND. Persistent lymphedema is a well-described complication after ALND.^14–16 In the most recent analysis, a rate of 3.5% after a sentinel lymph node biopsy and 19.1% after complete removal of all axillary lymph nodes was reported.¹⁷ The only previous literature on the topic by Andrews et al.⁸ reported results of patients undergoing shoulder arthroplasty after prior breast cancer treatment. Between 1979 and 1996, the authors identified 17 patients who had undergone mastectomy for breast cancer prior to shoulder arthroplasty. At a mean of 4.8 years after surgery, their analysis found a higher infection rate (10%), a higher revision rate (10%), and a greater occurrence of lymphedema in the postoperative setting (41%) compared to our reported results. Only four patients in their series had results that rated as excellent (7 satisfactory and 6 unsatisfactory) by the methodology described by Neer et al.¹⁸ and Cofield.¹⁹ All of these cases were performed through a deltopectoral approach. Notably, there were major confounding variables that may help explain the poor results. The RSA implant was not yet available, which may explain less robust results in their patients with rotator cuff deficiency. Additionally, five of their aTSAs utilized metal-backed glenoids—a technique that has since been shown to have a higher complication rate and unsatisfactory results.⁹ The results of our current analysis suggest that as shoulder arthroplasty techniques and breast cancer treatment have evolved, the outcomes in this population may be more promising than previously described.

The results of this study must be carefully considered in the context of the limitations. First, this is a purely retrospective analysis and therefore it is subject to the same limitations as all retrospective studies. Second, the complete details of the patients’ full treatment course for breast cancer were not widely known. Third, preoperative patient-reported outcomes were not widely available. Therefore, knowing how much improvement each patient achieved from their preoperative baseline is unknown. Fourth, validated examination of patients’ lymphedema was not performed either before or after surgery. We relied on clinic notes and the patients’ recall of whether or not their lymphedema had worsened after surgery. Recall bias may have underestimated or overestimated the true lymphedema experience in the postoperative setting. Finally, swelling of the fingers, hand, and forearm after shoulder arthroplasty is not uncommon and can be seen in patients without prior ALND. Given that swelling is not routinely or reliably documented in chart notes, we were unable to generate a suitable control group with which to compare our cohort. Despite these limitations, this study identified more encouraging results in patients undergoing shoulder arthroplasty after prior ALND than have been previously reported.⁸ Our results suggest that patients who elect to undergo shoulder arthroplasty after ALND may expect a consistent outcome with a lower complication and revision rate than previously reported.

Conclusion

Based on the results of this analysis, ALND does not represent a contraindication to shoulder arthroplasty. A deltopectoral exposure can be utilized without a substantial risk of worsening lymphedema or wound complications. Permanent worsening lymphedema only occurred in 1 patient out of 32. While a superior approach avoids cephalic vein injury, important approach-related complications (one deltoid dehiscence and one axillary nerve palsy) were observed in our cohort, albeit in a small number of patients.

Authors’ note

This investigation was performed at the Rothman Institute of Orthopaedics, Thomas Jefferson University Hospitals, Philadelphia, Pennsylvania.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Ethical Review and Patient Consent

This study was approved by the Thomas Jefferson University Hospital Institutional Review Board. The approval number was 45 CFR 46.110 Control #17D.037.