Perioperative complications and readmission rates following total shoulder arthroplasty in patients aged 90 and older

Abstract

Background

Previous literature has cited age as an independent risk factor for surgical outcomes following total shoulder arthroplasty (TSA). The purpose of this study is to determine the short-term outcomes of nonagenarians following primary TSA and compare them to younger cohorts.

Methods

The Nationwide Readmission Database (NRD) was queried from 2010 to 2020 to identify patients who underwent a primary reverse or anatomic TSA (rTSA or aTSA). Cohorts were defined by those aged 70–74 (septuagenarians), 80–84 (octogenarians), or 90 and over (nonagenarians). The septuagenarians and octogenarians were matched 1:1 to the nonagenarians by the Charlson–Deyo Comorbidity Index (CCI), year of procedure, and sex. Length of hospital stay, readmission rate, and postoperative outcomes were collected.

Results

Length of hospital stay, in-hospital death, 180-day mortality, 180-day readmission, and rate of medical complications were highest in the nonagenarians. Rates of periprosthetic fracture were highest in the nonagenarians. The number of total surgical complications was not significantly different between cohorts. The rate of revision within 180 days and discharge to rehabilitation facility was not significantly different between the three groups.

Discussion

The nonagenarians are at an increased risk of medical complications, longer hospital stays, periprosthetic fractures, and death following TSA. Otherwise, nonagenarians have similar rates of total surgical complications to the septuagenarians and octogenarians.

Introduction

In recent years, total shoulder arthroplasty (TSA) has grown in popularity as more surgeons become trained on the technique and indications expand beyond osteoarthritis to include proximal humerus fractures, rotator cuff tears, and aseptic necrosis of the humeral head.^1–3 TSA provides a substantial increase in health-related quality-of-life measures, including physical component summary scores and shoulder-specific measures, according to a meta-analysis by Carter et al. ⁴ By 2013, reverse total shoulder arthroplasty (rTSA) was performed more than anatomic total shoulder arthroplasty (aTSA).^5,6 As of 2017, aTSA and rTSA were performed over 110,000 times per year in the United States, demonstrating a greater than 50% increase between 2012 and 2017.^2,6,7 This increasing incidence is seen across all age groups.^2,6

The American population is aging, working, and staying active for longer periods of time; by 2035, the number of people 85 years and older is expected to double. ⁸ As such, the demand for TSA is increasing to meet the needs of this population. Rabinowitz et al. estimate that the demand for TSA in those over 85 years will increase to 344% by 2040. ⁹ However, many studies report advanced age as an independent risk factor for medical complications, mortality, and increased length of stay (LOS) after TSA.^10–14 Contradictory evidence concludes that TSA in elderly patients is a valid therapeutic option that offers functional improvement and has a low medical complication rate.^15,16

The growing popularity of the procedure combined with the aging population provides an increasing number of patients in their 80s and 90s presenting as candidates for TSA. However, the outcomes of TSA in patients older than 90 years are sparse as there is very little data published. A few studies are published with patients ≥80 that include those over age 90.^15,17,18 Only two previously published studies on TSA outcomes in those ≥80 years old included small cohorts of patients over age 90 (n ≤ 16).^19,20

The purpose of this study is to determine the short-term outcomes of nonagenarians following primary TSA and compare them to younger cohorts. Our hypothesis is that the medical and surgical outcomes of TSA are higher for nonagenarians in comparison to octogenarians and septuagenarians and that they will have a longer LOS, greater readmission rates, and greater revision rates following TSA compared to octogenarians and septuagenarians.

Materials and methods

The Nationwide Readmission Database (NRD) was queried from 2010 to 2020 to identify patients who underwent a primary TSA using the International Classification of Disease (ICD) Clinical Modification (CM) and Procedure Coding System (PCS) codes. The NRD was developed as part of the Healthcare Cost and Utilization Project and contains nationally representative de-identified data of millions of hospital discharges. Patients were included if they underwent either an anatomic (0RRJ0JZ, 0RRK0JZ, 8180) or reverse (0RRJ00Z, 0RRK00Z, 8188) TSA and if they were over 70 years old. The number of TSAs performed in those aged 70–79, 80–89, and over 90 each year between 2010 and 2020 was recorded, and the percent change was calculated.

In the database, exact ages are recorded up until age 90. All patients 90 years and older were recorded as 90. It was presumed that most nonagenarians were in their early 90s. To account for this and represent more uniform differences in age groups, the query was then narrowed to three age cohorts: 70–74, 80–84, and 90 and over.

Demographics recorded for each age group were sex and the Charlson–Deyo Comorbidity Index (CCI). CCI is an age-independent score that weighs 17 comorbidity conditions to predict long-term survival. ²¹ Comorbidities of each age group were recorded, including the presence of alcohol abuse, congestive heart failure (CHF), chronic obstructive pulmonary disease (COPD), diabetes mellitus (DM), DM with chronic complications, hypertension, obesity, peripheral vascular disease (PVD), renal failure, and current smoking status.

Patients in the octogenarian and septuagenarian age cohorts were then matched 1:1 to the nonagenarian cohort by sex, CCI, and year of surgery. Post-operative outcomes recorded were in-hospital death, death within 180 days, readmission within 180 days, revisions, transfer to a rehabilitation facility, medical complications, and surgical complications. Medical complications included acute respiratory distress syndrome (ARDS), acute renal failure (ARF), bleeding, cardiovascular complications, thrombosis, gastrointestinal complications, complications requiring transfusion, UTI, pulmonary embolism, pneumonia, respiratory disorders, sepsis, and shock. Surgical complications included dislocation, fracture, surgical site infection, osteolysis, prosthetic loosening, hardware failure, wound dehiscence, the need for subsequent debridement, and other mechanical failures.

Statistical analysis was conducted using the IBM SPSS^® Statistics software (IBM, Armonk, NY, USA). Continuous variables were compared using analysis of variance, and categorical variables were compared using chi-square tests. Significance was defined by an alpha value of 0.05 for all variables.

Results

The initial database query identified a total of 119,702 patients aged 70 and older. There were 46,383 in the septuagenarian cohort (ages 70–74), 19,800 in the octogenarian cohort (ages 80–84), and 1092 in the nonagenarian cohort (over age 90). The percent of females was highest in nonagenarians (74%) and lowest in septuagenarians (57%). The CCI increased with age. The CCI in nonagenarians (1.1 ± 1.4, p = 0.005) was significantly different from the CCI in octogenarians (1.0 ± 1.4) and septuagenarians (0.9 ± 1.3).

Analyzing the incidence of TSA by age cohort from 2010 to 2020, the nonagenarians had a 2179% increase, representing an increase from 19 cases annually to 433 cases annually. This was greater than the octogenarian’s 1650% increase (from 631 cases annually to 11,044 cases annually) and the septuagenarian's 2060% increase (from 1631 cases annually to 35,233 cases annually) (Figure 1).

Figure 1.

Yearly incidence of TSA from 2010 to 2020.

Septuagenarians and octogenarians were more likely than nonagenarians to have alcohol use disorder (p < 0.001), COPD (p < 0.001), diabetes (p < 0.001), obesity (p < 0.001), and a smoking history (p < 0.001) (Table 1). Nonagenarians were more likely than septuagenarians and octogenarians to have CHF (p < 0.001), hypertension (p < 0.001), PVD (p < 0.001), and renal failure (p < 0.001) (Table 1).

LOS was longer in the nonagenarian cohort (p < 0.001) in comparison to both the septuagenarians and octogenarians. In the nonagenarians, in-hospital death (p = 0.012), death within 180 days (p < 0.001), and readmission within 180 days (p < 0.001) were significantly higher. In-hospital death was reported in 0.6% of nonagenarians, 0% of octogenarians, and 0.1% of septuagenarians. Death within 180 days occurred in 2% of nonagenarians, 0.3% of octogenarians, and 0.1% of septuagenarians. Rates of transfers to a rehabilitation facility (p = 0.285) and rates of revision within 180 days (p = 0.760) were not significantly different (Table 2).

Table 2.

Outcomes.

Outcome	Nonagenarian (n = 1092) n (%)	Octogenarian (n = 1091) n (%)	Septuagenarian (n = 1092) n (%)	p-Value
Length of stay (days)	3.1 ± 3.2	1.9 ± 2.0	1.4 ± 1.6	<0.001*
Death	6 (0.6)	0 (0)	1 (0.1)	0.012*
Death within 180 days	22 (2)	3 (0.3)	2 (0.2)	<0.001*
Readmission within 180 days	163 (15)	105 (9.6)	94 (8.6)	<0.001*
Transfer to rehab facility	10 (0.9)	5 (0.5)	5 (0.5)	0.285
Revision within 180 days	13 (1.2)	10 (0.9)	10 (0.9)	0.760

*Indicates statistical significance.

Table 1.

Comorbidities of unmatched cohorts.

Comorbidity	Nonagenarian (n = 1092) n (%)	Octogenarian (n = 19,800) n (%)	Septuagenarian (n = 46,383) n (%)	p-Value
Alcohol use	2 (0.2)	106 (0.5)	433 (0.9)	<0.001*
Congestive heart failure	133 (12)	1292 (6.5)	1829 (3.9)	<0.001*
Chronic obstructive pulmonary disease	142 (13)	3439 (17)	8748 (19)	<0.001*
Diabetes mellitus	83 (7.6)	2511 (13)	7298 (16)	<0.001*
Hypertension	852 (78)	15,051 (76)	32,553 (70)	<0.001*
Obesity	51 (4.7)	2464 (12)	9506 (21)	<0.001*
Peripheral vascular disease	61 (5.6)	948 (4.8)	1460 (3.2)	<0.001*
Renal failure	189 (17)	2310 (12)	3176 (6.9)	<0.001*
Smoker	221 (20)	5038 (25)	12,705 (27)	<0.001*

*Indicates statistical significance.

The total number of patients who experienced a medical complication was greater in nonagenarians in comparison to octogenarians and septuagenarians (15%, 4.4%, 2.7%, p < 0.001). Nonagenarians were more likely to experience ARDS (p = 0.011), ARF (p < 0.001), cardiovascular complications (p = 0.002), transfusion (p < 0.001), UTI (p < 0.001), pneumonia (p = 0.002), and sepsis (p = 0.045). There were no significant differences between the cohorts for bleeding, thrombosis, gastrointestinal complications, pulmonary embolism, respiratory disorders, or shock (Table 3).

Table 3.

Medical complications.

Outcome	Nonagenarian (n = 1092) n (%)	Octogenarian (n = 1091) n (%)	Septuagenarian (n = 1092) n (%)	p-Value
Total patients with any medical complication	158 (15)	48 (4.4)	30 (2.7)	<0.001*
ARDS	18 (1.7)	5 (0.5)	8 (0.7)	0.011
ARF	72 (6.6)	15 (1.4)	6 (0.6)	<0.001*
Bleeding	2 (0.2)	1 (0.1)	0 (0)	0.368
Cardiovascular	10 (0.9)	0 (0)	3 (0.3)	0.002
Thrombosis	4 (0.4)	1 (0.1)	1 (0.1)	0.223
Gastrointestinal complication	0 (0)	1 (0.1)	1 (0.1)	0.606
Transfusion	36 (3.3)	14 (1.3)	4 (0.4)	<0.001*
Urinary tract infection	53 (4.9)	16 (1.5)	11 (1.0)	<0.001*
Pulmonary embolism	1 (0.1)	1 (0.1)	2 (0.2)	0.779
Pneumonia	13 (1.2)	2 (0.2)	0 (0)	0.002*
Respiratory disorders	1 (0.1)	1 (0.1)	2 (0.2)	0.779
Sepsis	7 (0.6)	1 (0.1)	2 (0.2)	0.045*
Shock	0 (0)	1 (0.1)	0 (0)	0.369

*Indicates statistical significance.

Overall, 2.9% of nonagenarians experienced any surgical complication in comparison to 1.9% of octogenarians and 1.9% of septuagenarians, which was not significantly different (p = 0.188). However, the risk of fracture was significantly higher in nonagenarians (p = 0.019). In nonagenarians, the odds of experiencing a fracture postoperatively were 2.8 times greater than octogenarians and 3.5 times greater than septuagenarians. There was no statistically significant difference in surgical site infection, osteolysis, prosthetic loosening, hardware failure, wound dehiscence, subsequent debridement, or other mechanical failure (Table 4).

Table 4.

Surgical complications.

Outcome	Nonagenarian (n = 1092) n (%)	Octogenarian (n = 1091) n (%)	Septuagenarian (n = 1092) n (%)	p-Value
Total # of patients with any surgical complication	32 (2.9)	21 (1.9)	21 (1.9)	0.188
Dislocation	11 (1.0)	3 (0.3)	6 (0.6)	0.085
Fracture	14 (1.3)	5 (0.5)	4 (0.4)	0.019*
Osteolysis	0 (0)	0 (0)	0 (0)	—
Prosthetic loosening	3 (0.3)	6 (0.6)	6 (0.6)	0.547
Hardware failure	1 (0.1)	1 (0.1)	0 (0)	0.606
Wound dehiscence	0 (0)	1 (0.1)	0 (0)	0.367
Surgical site infection	1 (0.1)	3 (0.3)	0 (0)	0.173
Debridement	1 (0.1)	0 (0)	0 (0)	0.368
Other mechanical failure	2 (0.2)	4 (0.4)	7 (0.6)	0.231

*Indicates statistical significance.

Discussion

This study found that following TSA, patients over age 90 are at increased risk of medical complications, in-hospital death, and death within 180 days. The total number of patients who experienced at least one surgical complication was similar among all cohorts, but the risk of fracture was significantly higher among the nonagenarians. Patients over age 90 also had a longer LOS and higher rate of readmission within 180 days in comparison to patients in their 70s or 80s.

The nonagenarians had a higher CCI in comparison to septuagenarians, a score that is designed to predict mortality risk. The nonagenarian cohort more commonly had comorbidities of CHF and renal failure compared to the younger patients, but the comorbidities of COPD, diabetes, obesity, and smoking were less common among the nonagenarians in comparison to the septuagenarians. This is likely due to the nonagenarians' robustness that allowed survival into their 90s, while those with these comorbidities died at a younger age.

The incidence of TSA grew in all three cohorts. In 2020, the incidence of TSA in nonagenarians was 22 times higher than 2010; this was the largest growth of all three of the cohorts. Wagner et al. analyzed the data from the National Inpatient Sample database and demonstrated that the largest incidence of rTSA was in men and women 75 years of age and older and the largest incidence of aTSA was in men and women 65 to 74 years old. ⁷

Nonagenarians experienced a LOS that averaged 1.2 days longer than the octogenarians and 1.67 days longer than the septuagenarians. This result is consistent with previous reports that age is an independent risk factor for extended LOS following TSA.^22–24 Dunn et al. found that those over age 80 had a 2-fold risk of extended LOS in comparison to those under the age of 60. ²³ Menendez et al. identified age as an independent risk factor for a LOS greater than the 75th percentile for TSA. ²² Notably, they also saw that LOS varied between aTSA and rTSA. Those over age 85 had a 4-fold increase in LOS after aTSA but only a 2-fold increase after rTSA. This may be explained by the higher mean age of patients undergoing rTSA, resulting in an increased mean LOS. ²²

Nonagenarians experienced higher rates of medical complications following TSA in comparison to octogenarians and septuagenarians (15%, 4.4%, and 2.7%, respectively). The oldest cohort was more likely to have complications including ARDS, ARF, pneumonia, and sepsis, require a transfusion, or acquire a UTI. However, the oldest cohort was at a risk similar to the younger two cohorts for other medical complications including bleeding, thrombosis, gastrointestinal complications, pulmonary embolism, respiratory complications, and shock.

The results of this study support the hypothesis that patients over age 90 have a similar risk of all surgical complications in comparison to septuagenarians and octogenarians. When considering the nonagenarian patients, 2.9% of the 1092 patients experienced any surgical complication. This was not significantly different than the septuagenarian cohort rate of 1.9% and the octogenarian cohort rate of 1.9%. However, when surgical complications were considered individually, nonagenarians were 2.8 times more likely than octogenarians and 3.5 times more likely septuagenarians to experience a fracture, likely due to their diminished bone quality. ²⁵ The nonagenarians did not have an increased risk of other surgical complications, including dislocation, surgical site infections, osteolysis, prosthetic loosening, hardware failure, wound dehiscence, or other mechanical failure.

In comparison, Clark et al. saw just 2 (14%) surgical complications among 14 patients, both of which were delayed wound healing, and zero medical complications. ¹⁹ Almasri studied 16 nonagenarians and reported higher complication rates than the current study: 38% (6/16) had medical complications and 13% (2/16) had surgical complications. ²⁰ Five TSAs in nonagenarians by Churchill had zero blood transfusions, but the numbers are too small to have any meaning. ²⁶ In a systematic review of TSA in patients over age 80, DeBernardis et al. found a wide range of complication rates (3.5–29%) and reoperation rates (0–13%) among aTSA and rTSA. ²⁷ The most common late complications were scapular notching and acromial or scapular spine fracture. ²⁷

The nonagenarians also had higher readmission rates within 180 days. Villacis et al. however reported contradictory results, with patients over age 85 having the lowest risk of readmission 1 year after TSA. ²⁸ Nonagenarians were not at increased risk for transfers to rehabilitation facilities postoperatively nor revisions within 180 days. Revision rates were 1.2% in nonagenarians, 0.9% in octogenarians, and 0.9% in septuagenarians—higher than the 0/14 reported by Clark et al. and 0/5 reported by Churchill.^19,26 Almasri et al. found no difference in reoperation rates between those 85–89 and those over 90. ²⁰

In-hospital deaths were found to be low across all cohorts, but significantly different for the nonagenarians (p = 0.012). One hundred and eighty-day mortality was highest in the oldest cohort: 2% in nonagenarians, 0.3% in octogenarians, and 0.2% in septuagenarians (p < 0.001). These rates were higher than Clark et al.'s small study of 14 nonagenarians, which reported a 0% 90-day mortality rate. ¹⁹ In their most recent follow-up, 29% (4/14) of patients had died from causes unrelated to TSA at a median of 3.2 years. ¹⁹ Almasri et al. found no difference in 30-day deaths between those 85–89 over 90, and 56% (9/16) of patients had died at a median of 3.3 years at the most recent follow-up. ²⁰

Older patients have much to benefit from TSA, and outcomes exceed minimally clinically important differences between preoperative and postoperative functions. ²⁰ Among Churchill’s five TSAs in nonagenarians, all had improved functional outcomes including mental health, vitality, social function, and comfort. ²⁶ DeBernardis’ systematic review concluded that patients over age 80 have favorable clinical outcomes. ²⁷ Boettcher et al. found that both 369 patients over age 80 and 1764 patients between ages 60 and 79 experienced improvements in pain scores, active range of motion, and patient-reported outcomes; satisfaction rates were similar with 92% and 93% reported being “much better” or “better,” respectively. ¹⁷ Friedman et al. reported that older patients had better outcome scores than younger patients: for every 1-year increase in age, ASES score improved by 0.19 points. ²⁹

Orthopedic surgeons must decide how old is “too old” for a TSA. Nonagenarians should not be excluded purely based on their age. Patient selection is critical for favorable outcomes in nonagenarians. The element of selection bias of these older patients who are determined fit for TSA is apparent from the decreased prevalence of comorbidities. Modifiable risk factors must be optimized prior to surgery, and patients must be counseled on an increased risk of medical and surgical complications. Even in the oldest but medically appropriate patient, TSA can provide significantly improved function with acceptable risks.

This study has limitations. The NRD is a multi-institution national database that relies upon accurate data entry and is subject to clinical coding error. The database does not capture complications, revisions, or readmissions that occur during a different calendar year, in a different state, or at a non-participating hospital. It also does not capture complications that are addressed in clinic and do not require readmission or revision. It is possible that these complications are addressed conservatively due to the patient’s age. The database does not provide information about indications for surgery or pre-operative assessments, and thus, patients were not matched based on indication for TSA, which may impact complication profile. Analysis of the database does not elucidate if death or readmission was related to TSA nor does it account for variations in surgeon skill. The limitations are mitigated by the database’s large sample size and diverse geographical representation across 31 states. Previous studies that have reported on TSA outcomes in nonagenarians all have small cohorts that do not allow for any meaningful conclusions.^19,20,26

Conclusion

Nonagenarians are at an increased risk of medical complications, longer hospital stays, and periprosthetic fractures following TSA. Otherwise, nonagenarians have similar rates of surgical complications to septuagenarians and octogenarians. TSA can be performed in patients over the age 90, but there are increased medical and surgical risks that should be taken into consideration. However, this should not preclude them from undergoing this procedure.