Outcomes for type C proximal humerus fractures in the adult population: comparison of nonoperative treatment, locked plate fixation, and reverse shoulder arthroplasty.

Background: This study compares patient-reported outcomes and range of motion (ROM) between adults with an AO Foundation/Orthopaedic Trauma Association type C proximal humerus fracture managed nonoperatively, with open reduction and internal fixation (ORIF), and with reverse shoulder arthroplasty (RSA).
Methods: This is a retrospective cohort study of patients >60 years of age treated with nonoperative management, ORIF, or RSA for AO Foundation/Orthopaedic Trauma Association type 11C proximal humerus fractures from 2015 to 2018. Visual analog scale pain scores, Patient-Reported Outcomes Measurement Information System (PROMIS) scores, ROM values, and complication and reoperation rates were compared using analysis of variance for continuous variables and chi square analysis for categorical variables.
Results: A total of 88 patients were included: 41 nonoperative, 23 ORIF, and 24 RSA. At the 2-week follow-up, ORIF and RSA had lower visual analog scale scores and lower PROMIS pain interference scores (P < .05) than nonoperative treatment. At the 6-week follow-up, ORIF and RSA had lower visual analog scale, PROMIS pain interference, and PF scores and better ROM (P < .05) than nonoperative treatment. At the 3-month follow-up, ORIF and RSA had better ROM and PROMIS pain interference and PF scores (P < .05) than nonoperative treatment. At the 6-month follow-up, ORIF and RSA had better ROM and PROMIS PF scores (P < .05) than nonoperative treatment. There was a significantly higher complication rate in the ORIF group than in the non-operative and RSA groups (P < .05).
Conclusion: The management of AO Foundation/Orthopaedic Trauma Association type 11C proximal humerus fractures in older adults with RSA or ORIF led to early decreased pain and improved physical function and ROM compared to nonoperative management at the expense of a higher complication rate in the ORIF group.

Proximal humerus fractures are treated based on numerous factors, including patient age, fracture type, calcar comminution, osteoporosis, vascular status of the humeral head, rotator cuff function, surgeon expertise, and a patient's health and functional status.,,27, 28, 29,,,,, AO Foundation/Orthopaedic Trauma Association (AO/OTA) type C proximal humerus fractures are challenging fractures to standardize treatment in the adult patient population because of age, complex fracture patterns, osteoporotic bone, high incidence of rotator cuff tears, and variable health and functional status.,,,,,,, As a result, there is a substantial amount of debate regarding the optimal treatment of AO/OTA type C proximal humerus fractures in the adult population.

The mainstay treatments include nonoperative management, open reduction internal fixation (ORIF) with locked plating, hemiarthroplasty (HA), and reverse shoulder arthroplasty (RSA)., The majority of proximal humerus fractures is managed nonoperatively and has been associated with good outcomes for both nondisplaced and displaced fractures.,,, In contrast, some studies suggest nonoperative treatment results in slower return of function, less pain relief, and even higher mortality rate., The literature also supports operative treatment for AO/OTA type C proximal humerus fractures.,,,,,,,,,,,50, 51, 52,, ORIF has been a staple for surgical management of proximal humerus fractures throughout the years; however, the use of RSA in the treatment for proximal humerus fractures has increased as experience with and outcomes of RSA have improved.,,,,,,, Recent studies have compared these surgical options in order to optimize outcomes and assist in surgical decision-making for complex proximal humerus fractures in the adult population.,, However, there is a paucity of evidence directly comparing clinical and patient-reported outcomes (PROs) among nonoperative management, ORIF, and RSA. This information will be helpful to further understand outcomes, guide treatment, and counsel patients with AO/OTA type C proximal humerus fractures

The aim of this study was to compare early outcomes including range of motion (ROM), PROs, complications, and secondary operations of nonoperative management, ORIF with locked plating, or RSA for AO/OTA type C proximal humerus fractures in adult patients older than the age of 60 years.

Materials and methods

Patient population

Following institutional review board approval, patients with a closed proximal humerus fracture treated either nonoperatively, with ORIF, or with RSA at a single, large, tertiary referral center between January 2015 and December 2018 were retrospectively identified using Current Procedural Terminology codes 23600, 23615, and 23472. Definitive treatment was decided by the attending surgeon based on age, fracture morphology, medical comorbidities, activity level, and patient preference. All patients were treated by fellowship-trained orthopedic surgeons from a large academic practice who were experienced in treating proximal humerus fractures. The inclusion criterion was patients aged 60 years or older who sustained an AO/OTA type C (Neer 4-part) proximal humerus fracture as identified on initial injury radiographs and underwent definitive nonoperative management, ORIF with locked plating, or RSA within the first 4 weeks postinjury. Patients were excluded if they were younger than 60 years at the time of injury, sustained an open or pathologic fracture, presented with multiple long bone fractures, and underwent percutaneous pinning, intramedullary nailing, and HA. Patients were also excluded if they did not have a minimum of 1 year of clinical and radiographic follow-up.

Data collection

Patient demographics and baseline characteristics were recorded including age, sex, body mass index, smoking status, and comorbidity burden. Of note, comorbidity burden was assessed using a modified Charlson comorbidity index with the age component excluded in order to evaluate age and comorbidity burden separately. ROM data, including active forward flexion (aFF), passive forward flexion (pFF), and external rotation (ER), were recorded at 2-week, 6-week, 3-month, 6-month, and 1-year clinical follow-up visits. ROM data were collected primarily from physical therapy notes to standardize measurements. PROs were also assessed at follow-up visits, including visual analog scale shoulder pain scores and Patient-Reported Outcomes Measurement Information System (PROMIS) pain interference, physical function, and depression domain scores.

Any complications that occurred up to 1 year after treatment were recorded. Complications included nonunion, avascular necrosis, intraarticular screw cutout, hardware failure, adhesive capsulitis, and periprosthetic fracture. The presence of scapular notching or tuberosity malunion/resorption on postoperative radiographs at any time point was recorded for RSA patients.

Nonunion was defined as a lack of union progression over 3 months or absence of complete union by 9 months., Secondary operations within 1 year of treatment were also recorded including revision ORIF, conversion to arthroplasty, removal of hardware, and lysis of adhesions. Malunion was assessed on postoperative radiographs for the nonoperative and ORIF cohorts and defined as HSA <120° or >150°, or head-shaft translation >5 mm.

Statistical analysis

Statistical analysis was performed using Stata/SE 17.0 for Mac (StataCorp LP, College Station, TX, USA). Baseline patient characteristics were compared among nonoperative, ORIF, and RSA groups using 1-way analysis of variance or chi square analysis for continuous and categorical variables, respectively. Differences in ROM values and PRO scores at follow-up visits between treatment groups were assessed using 1-way analysis of variance. Ross et al previously utilized a threshold of 90° of aFF to indicate satisfactory functional outcome for 3-part and 4-part proximal humerus fractures. This metric was used at the final follow-up to compare satisfactory functional outcomes among nonoperative, ORIF, and RSA cohorts utilizing chi square analysis. Associations between treatment type and malunion, complication, and secondary operation rates were determined using chi square analysis. Statistical significance was set at P < .05.

Results

A total of 149 patients with AO/OTA type 11C proximal humerus fractures were identified. Sixty-one patients were excluded from the study based on the following: pathologic fracture (n = 4), multiple long bone fractures (n = 10), previous shoulder surgery (n = 3), underwent HA (n = 5), inadequate clinical or radiographic follow-up (n = 39). A total of 88 patients were included in the study with the following: 41 were managed nonoperatively, 23 underwent ORIF, and 24 underwent RSA. Average clinical and radiographic follow-up for each one of the groups included 2.0 years for nonoperative, 1.8 years for ORIF, and 2.4 years for RSA (P = .2). On average, patients who underwent ORIF were younger than those managed nonoperatively or with RSA (ORIF = 67 years, nonoperative = 77 years, RSA = 77 years; P < .01). No other significant differences in patient characteristics were identified between groups (P > .05 for all, Table I).

Table I

Baseline characteristics

	Nonoperative (N = 41)	ORIF (N = 23)	RSA (N = 24)	P value∗
Age (mean ± SD, yr)	77.4 ± 10.1	67.1 ± 5.5	77.3 ± 9.5	<.01
Sex (n)				.15
Female	90.2% (37)	78.3% (18)	95.8% (23)
Male	9.8% (4)	21.7% (5)	4.2% (1)
Body mass index (n)				.87
Nonobese (<30 kg/m2)	57.5% (23)	63.6% (14)	62.5% (15)
Obese (≥30 kg/m2)	42.5% (17)	36.4% (8)	37.5% (9)
Smoking status (n)				.80
Nonsmoker	95.1% (39)	91.3% (21)	91.7% (22)
Smoker	4.9% (2)	8.7% (2)	8.3% (2)
Ageless CCI (mean ± SD)	2.2 ± 2.0	1.4 ± 1.7	1.6 ± 1.8	.19

CCI, Charlson comorbidity index; SD, standard deviation; ORIF, open reduction internal fixation; RSA, reverse shoulder arthroplasty.

Boldface indicates statistical significance.

P values calculated using analysis of variance and chi square analysis for continuous and categorical variables, respectively.

Table II and Figure 1, A–C detail the ROM data for the nonoperative, ORIF, and RSA cohorts. At 2 weeks, patients among all 3 groups demonstrated 0° of aFF but had statistically significant differences in pFF (RSA = 47°, ORIF = 10°, nonoperative = 0°; P < .01) and ER (RSA = 3°, ORIF = 1°, nonoperative = 0°; P < .01), which reflects the early activity limitations ascribed for each treatment option. At 6 weeks, RSA patients had greater aFF (57° vs 34° vs 18°; P < .01), pFF (114° vs 58° vs 51°; P < .01), and ER (23° vs 15° vs 14°; P = .04) than ORIF and nonoperative patients. Similarly, at 3 months, the RSA cohort demonstrated significantly greater aFF (125° vs 85° vs 59°; P < .01) and pFF (138° vs 108° vs 70°; P < .01) than the ORIF and nonoperative cohorts. At 6 months, significantly greater aFF (RSA = 133°, ORIF = 106°, nonoperative = 93°; P < .01) and pFF (RSA = 148°, ORIF = 125°, nonoperative = 116°; P < .01) were observed for RSA patients, but ER was similar between groups (P = .56). At the final follow-up, achievement of a satisfactory functional outcome (aFF >90 degrees) was significantly higher in the RSA (91.7%, n = 22) and ORIF (65.2%, n = 15) cohorts than in the nonoperative (31.7%, n = 13) cohort (P < .01).

Table II

Range of motion outcomes

	Nonoperative (N = 41)	ORIF (N = 23)	RSA (N = 24)	P value∗
Active forward flexion (mean ± SD, °)
2-week follow-up	0 ± 0	0 ± 0	0 ± 0	--
6-week follow-up	18.4 ± 29.2	34.1 ± 31.7	57.1 ± 50.8	<.01
3-month follow-up	58.9 ± 47.3	84.5 ± 37.9	125.1 ± 20.4	<.01
6-month follow-up	93.0 ± 31.2	105.6 ± 28.3	133.0 ± 22.5	<.01
Passive forward flexion (mean ± SD, °)
2-week follow-up	0 ± 0	9.6 ± 27.6	46.5 ± 47.7	<.01
6-week follow-up	50.5 ± 41.4	58.3 ± 41.7	114.1 ± 31.5	<.01
3-month follow-up	70.0 ± 52.6	108.2 ± 37.9	137.5 ± 21.2	<.01
6-month follow-up	116.0 ± 23.1	125.4 ± 14.8	147.9 ± 15.1	<.01
External rotation (mean ± SD, °)
2-week follow-up	0 ± 0	0.5 ± 1.5	3.3 ± 6.1	<.01
6-week follow-up	13.5 ± 15.1	14.7 ± 14.9	23.4 ± 13.8	.04
3-month follow-up	23.8 ± 24.0	30.8 ± 21.2	37.3 ± 21.3	.09
6-month follow-up	37.9 ± 20.5	38.1 ± 19.5	44.1 ± 19.4	.56

ORIF, open reduction internal fixation; RSA, reverse shoulder arthroplasty; SD, standard deviation.

Boldface indicates statistical significance.

P values calculated using analysis of variance.

Figure 1

Range of motion data including (A) active forward flexion, (B) passive forward flexion, and (C) external rotation among nonoperative, ORIF, and RSA cohorts at 2 weeks, 6 weeks, 3 months, and 6 months. ∗∗ = statistically significant difference between groups. ORIF, open reduction internal fixation; RSA, reverse shoulder arthroplasty.

PROs were also compared among the groups (Table III and Fig. 2, A–D). Significant differences in visual analog scale pain scores were observed among the groups at 2 weeks (RSA = 2.7, ORIF = 2.4, nonoperative = 6.2; P < .01), 6 weeks (RSA = 1.1, ORIF = 2.1, nonoperative = 3.5; P < .01), and 1 year (P < .01) but not at 3 months (P = .33) or 6 months (P = .24). In addition, PROMIS pain interference scores were significantly different among the groups at 2 weeks (RSA = 63.8, ORIF = 64.1, nonoperative = 68.9; P = .02), 6 weeks (RSA = 57.4, ORIF = 57.3, nonoperative = 61.0; P = .04), and 3 months (RSA = 50.0, ORIF = 55.2, nonoperative = 59.6; P < .01) but not at 6 months (P = .10) or 1 year (P = .07). PROMIS physical function scores were significantly different among the groups at 6 weeks (RSA = 33.8, ORIF = 35.9, nonoperative = 30.7; P < .01), 3 months (RSA = 39.0, ORIF = 40.8, nonoperative = 35.4; P = .045), 6 months (RSA = 42.2, ORIF = 44.8, nonoperative = 37.4; P = .01), and 1 year (RSA = 46.0, ORIF = 48.1, nonoperative = 37.9; P = .01) but not at 2 weeks (P = .23). No differences in PROMIS depression scores were identified at any follow-up time point (P > .05 for all).

Table III

Patient-reported outcomes

	Nonoperative (N = 41)	ORIF (N = 23)	RSA (N = 24)	P value∗
VAS pain score (mean ± SD)
2-week follow-up	6.2 ± 3.3	2.4 ± 2.2	2.7 ± 2.6	<.01
6-week follow-up	3.5 ± 2.9	2.1 ± 2.5	1.1 ± 1.7	<.01
3-month follow-up	1.9 ± 2.1	1.5 ± 1.7	1.1 ± 2.0	.33
6-month follow-up	1.7 ± 2.3	1.1 ± 1.5	0.8 ± 1.6	.24
1-yr follow-up	0.6 ± 1.3	1.7 ± 1.4	0.1 ± 0.3	<.01
PROMIS pain interference (mean ± SD)
2-week follow-up	68.9 ± 5.8	64.1 ± 7.2	63.8 ± 6.9	.02
6-week follow-up	61.0 ± 5.9	57.3 ± 4.4	57.4 ± 5.7	.04
3-month follow-up	59.6 ± 6.1	55.2 ± 6.4	50.0 ± 7.3	<.01
6-month follow-up	57.9 ± 8.5	52.7 ± 7.8	53.3 ± 6.0	.10
1-yr follow-up	58.4 ± 9.7	50.3 ± 7.5	51.2 ± 5.9	.07
PROMIS physical function (mean ± SD)
2-week follow-up	28.4 ± 6.9	31.3 ± 6.0	28.3 ± 4.8	.23
6-week follow-up	30.7 ± 4.9	35.9 ± 5.4	33.8 ± 4.4	<.01
3-month follow-up	35.4 ± 8.7	40.8 ± 5.5	39.0 ± 4.3	.045
6-month follow-up	37.4 ± 8.5	44.8 ± 5.6	42.2 ± 5.2	.01
1-yr follow-up	37.9 ± 7.7	48.1 ± 8.2	46.0 ± 4.7	.01
PROMIS depression (mean ± SD)
2-week follow-up	58.1 ± 9.1	53.0 ± 8.9	55.8 ± 7.0	.18
6-week follow-up	52.3 ± 9.2	52.0 ± 7.7	52.8 ± 7.3	.96
3-month follow-up	51.0 ± 8.5	47.7 ± 7.1	49.3 ± 10.2	.47
6-month follow-up	51.3 ± 9.9	45.9 ± 10.6	46.8 ± 10.6	.27
1-yr follow-up	49.7 ± 7.7	43.4 ± 8.0	47.1 ± 8.6	.24

ORIF, open reduction internal fixation; RSA, reverse shoulder arthroplasty; PROMIS, Patient-Reported Outcomes Measurement Information System; SD, standard deviation; VAS, visual analog scale.

Boldface indicates statistical significance.

P values calculated using analysis of variance.

Figure 2

Patient Reported Outcomes including (A) PROMIS pain interference, (B) physical function, (C) depression, and (D) VAS pain scores among nonoperative, ORIF, and RSA cohorts at 2 weeks, 6 weeks, 3 months, 6 months, and 1 year. ∗∗ = statistically significant difference between groups. VAS, visual analog scale; ORIF, open reduction internal fixation; RSA, reverse shoulder arthroplasty.

Malunion, complication, and secondary operation rates are described in Table IV. Nonoperative management was associated with a higher malunion rate than ORIF (77.5% vs 38.9%, P < .01). The complication rate was significantly higher in the ORIF group (nonoperative = 9.8%, ORIF = 34.8%, RSA = 4.2%; P < .01). Complications in the nonoperative group included nonunion (n = 1) and avascular necrosis (n = 3). Complications in the ORIF group included avascular necrosis (n = 3), hardware failure (n = 2), screw cutout (n = 1), and adhesive capsulitis (n = 2). Complications in the RSA group included a nondisplaced periprosthetic fracture (n = 1). Scapular notching was identified on 20% of RSA patients' postoperative radiographs (n = 5). ORIF patients also experienced a significantly higher rate of secondary operations (nonoperative = 4.9%, ORIF = 34.8%, RSA = 0.0%; P < .01). The most common secondary operation was conversion to RSA (nonoperative = 2, ORIF = 4). Other secondary operations for the ORIF group included revision fixation (n = 1), removal of hardware (n = 1), and lysis of adhesions (n = 2).

Table IV

Complications

	Nonoperative (N = 41)	ORIF (N = 23)	RSA (N = 24)	P value∗
Malunion (n)	77.5% (31)	38.9% (7)	--	<.01
Complications (n)†	9.8% (4)	34.8% (8)	4.2% (1)	<.01
Secondary operation (n)‡	4.9% (2)	34.8% (8)	0.0% (0)	<.01

ORIF, open reduction internal fixation; RSA, reverse shoulder arthroplasty.

Boldface indicates statistical significance.

P values calculated using chi square analysis.

Complications: Nonoperative – nonunion (1), avascular necrosis (3), ORIF, avascular necrosis (3), screw cutout (1), hardware failure (2), adhesive capsulitis (2), RSA, periprosthetic fracture (1).

Secondary operations: Nonoperative – conversion to arthroplasty (2); ORIF, conversion to arthroplasty (4), revision fixation (1), removal of hardware (1), lysis of adhesions (2).

Discussion

The optimal treatment of AO/OTA proximal humerus fractures in the older adult population remains highly debated, with literature to support nonoperative management, ORIF, and RSA.,,37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53 To our knowledge, the presented study is the first one to directly compare functional outcomes among nonoperative, ORIF, and RSA for the management of AO/OTA type C proximal humerus fractures in adult patients older than the age of 60 years. In our cohort, surgical treatment was associated with an early reduction in pain and improved physical function when compared to nonoperative management. Pain appears to normalize between groups by the 1-year follow-up visit; however, the improved physical function is sustained in the surgical groups at 1 year. In addition, early increases in ROM were demonstrated by the surgical cohorts compared to nonoperative treatment, with RSA leading to the greatest amount of motion, which was sustained at the 1-year follow-up. ORIF in this study led to higher complications rates than the nonoperative and RSA cohorts. ORIF had higher secondary operations within the first 2 years after treatment, whereas no secondary operations occurred among the RSA group.

Surgical treatment of AO/OTA type C fractures was associated with a significant reduction in pain early in recovery compared to nonoperative treatment. The most dramatic differences between the groups were observed at the 2-week follow-up visit, although these differences converged over time. Calvaro et al similarly reported early increased pain, decreased functional capacity, and reduction in patient-perceived health associated with nonoperative management. Sixty-seven percent of the patients reported moderate-extreme pain or discomfort early in recovery for nondisplaced proximal humerus fractures. Similar to our study, meta-analysis and randomized studies report equivalent pain scores at 6 months, 1 year, and 2 years after injury between nonoperative and surgical treatments., However, these studies did not record outcomes prior to 6 months, which limits their assessment of short-term pain improvement. There was a slight increase in the visual analog scale pain score for the ORIF group at the 1-year mark compared to the nonoperative and RSA groups; however, this was not reflected in the PROMIS pain interference scores. This may be related to the limited sample size of our study or differences between these outcome measures. Our study suggests that surgical treatment significantly reduces pain in the early recovery of AO/OTA type C proximal humerus fractures compared to non-operative treatment, which may be an important consideration to discuss with patients as this may lead to reduced narcotic use and earlier return to activities of daily living. It is noteworthy that RSA had quicker improvement in pain postoperatively. However, it remains unclear if the statistical values associated with pain relief are of clinical significance, and further investigation is warranted.

AO/OTA type C proximal humerus fractures managed surgically demonstrated improved early ROM compared to nonoperative management, with RSA appearing to outperform both ORIF and nonoperative management. RSA had a significantly better pFF at 2 weeks than the ORIF and nonoperative cohorts likely secondary to difference in rehabilitation protocols among the groups. Notably, at the 2-week follow-up, all patients demonstrated similar aFF and ER because active ROM is not part of the rehabilitation protocol at our institution early in recovery for all 3 treatment groups. As a result, ROM data at this time point are of limited clinical utility. PROMIS physical function scores were also significantly better in the surgical cohorts than the nonoperative group at the 6-week, 6-month, and 1-year follow-up visits. Similar studies suggest that surgical treatment significantly improves ROM and physical function in the early recovery of AO/OTA type C proximal humerus fractures.,,,, Two studies directly comparing ORIF, HA, and RSA also support good function outcomes associated with surgical treatment of proximal humerus fractures., Similar to our study, others have demonstrated superior early ROM and shoulder function after RSA compared to nonoperative management and ORIF; however, none have directly compared all 3., Improved ROM metrics were seen at the final follow-up for the RSA and ORIF groups as highlighted by significantly higher achievement of satisfactory functional outcomes (aFF >90°) than the nonoperative group. Our study suggests that surgical treatment significantly improves ROM and physical function in the early recovery of AO/OTA type C proximal humerus fractures compared to nonoperative treatment in adult patients older than the age of 60 years.

Conversely, Li et al performed a meta-analysis on nonoperative management compared to ORIF, which did not show a difference in functional outcome, but did show a significantly higher complication rate in the ORIF group. Additional studies further support good outcomes for AO/OTA type C fractures with nonoperative management compared to ORIF.,, Differences between our results and these prior investigations may be related to differences in age, patient demographics, and fracture morphology categorization. While early results for RSA are promising, some studies express caution when using RSA for proximal humerus fractures. Additional literature suggests that RSA should be utilized exclusively for very disabling shoulder pathology and in patients older than 70 years of age. However, improved techniques along with good clinical outcomes for RSA provide support for broadening the indications for RSA usage in treating proximal humerus fractures in adults older than the age of 60 years.,,,,,, The results of this short-term study provide a foundation for directly comparing nonoperative to surgical treatment for complex proximal humerus fractures in adult patients. Furthermore, mid-term and long-term studies are needed to evaluate whether these short-term benefits are sustained and clinically significant for activities of daily living and to determine the rate of medium- and long-term complications of RSA specifically for the treatment of proximal humerus fractures.

When deciding how to treat AO/OTA type C fractures, complications and reoperation rate among treatment options play an important role in the decision-making process. Overall complication rates were higher in the ORIF group than in the nonoperative and RSA groups. The secondary operation rate was also higher in the ORIF group than in the nonoperative and RSA groups. Prior studies also report higher complication and reoperation rates with ORIF.,,,,, The most common complications following RSA include instability (4.7%), infection (4%), scapular notching (35%), acromion/scapula fractures (1.5%), periprosthetic fracture, tuberosity resorption, and tuberosity malunion/nonunion.,,, This is consistent with the results of our study. Zumstein et al conducted a systematic review for complications, reoperations, and revisions after RSA for rotator cuff arthropathy and reported rates of 24%, 3.5%, and 10%, respectively. The time frame for associated glenoid loosening from scapular notching in this review was 114 months. Other studies have previously reported mid-term survivorship for RSA as 76% to 91% at 10 years, with early loosening occurring within the first 3 years postoperatively. Additional studies report reoperation rates of 14.4% and revision rates of 11.2% in patients younger than 65 years treated with RSA., The observed differences between our study's complication and reoperation rate and prior studies are likely explained by the short-term design of our study, limited sample size, and variable definitions of complications. Scapular notching is important for the long-term follow-up for possible bone resorption, glenoid loosening, and possible revision surgery; however, we did not include it as a complication in the present study because our goal was to evaluate short-term outcomes, of which scapular notching has less relevance in the short term. Our scapular notching rate is consistent with prior studies mentioned previously. Mid- and long-term follow-up is needed to assess the true risk of complication and reoperation rates in the RSA cohort specifically for the treatment of proximal humerus fractures. Nonetheless, RSA had a lower complication and reoperation rate than ORIF treatment, which is supported by prior studies.,, This suggests RSA is a safe treatment option for proximal humerus fractures in the short term in adults older than 60 years. Mid-term studies have been performed on RSA for rotator cuff arthropathy, which show stability to decreased shoulder outcome scores at 10 to 15 years, which correlate with implant survivorship and retention.,, However, further dedicated studies are needed to elucidate the mid-term and long-term survivorship of RSA specifically for complex proximal humerus fractures in the adult population compared to ORIF or conservative management.

This study is subject to several limitations. Our study introduces an element of selection bias due to the lack of randomization or algorithmic selection. Treatment decisions were made at the surgeon's discretion and likely influenced by a myriad of factors including surgeon and patient preferences, patient age, comorbidities, functional status, and fracture characteristics. In addition, there was a relatively small sample size, which may have limited the statistical power and increases the possibility of type II error. An additional limitation is ROM data documentation at the 1-year time point. While the reasons for this are likely multifactorial, patients were often no longer participating in physical therapy at this time point, and therefore, ROM measurements were less consistent. As such, ROM analysis was only performed up to the 6-month follow-up time point. Furthermore, patients may have developed complications or required a secondary operation outside of 1 year that were not captured in the results of this study, reflecting the limitations of the short-term follow-up from the study design. This is particularly important regarding the RSA group. Another limitation of this study is that the use of PROMIS is not disease specific and could be influenced by other pathology producing pain or interfering with physical function; however, PROMIS has been shown to correlate strongly with conventional measures of upper extremity function (Constant score, DASH, SMFA, ASES). Patient's age, health, function, compliance with rehabilitation, and postoperative restrictions should all be taken into account when considering nonoperative management, ORIF, and RSA treatments. Despite these limitations, this is the largest sample to date directly comparing the nonoperative, ORIF, and RSA cohorts for type C proximal humerus fractures in adults older than 60 years and serves as an important foundation for future mid- and long-term studies.

Conclusion

The results of this study indicate that management of geriatric AO/OTA type C proximal humerus fractures with RSA is associated with the greatest postoperative ROM, followed by ORIF, then nonoperative management. Additionally, operative treatment with either RSA or ORIF leads to earlier improvements in pain and physical function compared to nonoperative management, despite higher complication and secondary operation rates after ORIF. This information will be important when counseling patients regarding treatment options and recovery expectations. The results from this study create a foundation for which future mid- and long-term prospective trials to further create an algorithmic approach and delineate the optimal treatment for AO/OTA type C proximal humerus fractures in the adult population.