Impact of depressive disorders on primary total shoulder arthroplasties: a matched control analysis of 113,648 Medicare patients

Abstract

Introduction

The purpose of this study was to investigate whether patients with depressive disorders undergoing primary total shoulder arthroplasty have higher rates of (1) in-hospital lengths of stay, (2) readmission rates, (3) medical complications, and (4) implant-related complications.

Methods

A retrospective query was performed using a national claims database. Study group patients were matched to controls in a 1:5 ratio according to age, sex, and medical comorbidities. The query yielded 113,648 patients who were with (n = 18,953) and without (n = 94,695) depressive disorders. Pearson’s χ² analyses were used to compare patient demographics. Logistic regression analyses were used to calculate odds-ratios of complications and readmission rates. Welch’s t-tests were used to test for significance for in-hospital lengths of stay. A p-value less than 0.003 was considered statistically significant.

Results

Study group patients had significantly longer in-hospital lengths of stay (2.7 days versus 2.3 days; p < 0.0001). Patients who have depressive disorders had higher incidences and odds of readmissions (9.4 versus 6.15%; odds-ratio: 1.6, p < 0.0001), medical complications (2.7 versus 0.9%; odds-ratio: 3.0, p < 0.0001), and implant-related complications (6.1 versus 2.4%; odds-ratio: 2.59, p < 0.0001) compared to controls.

Conclusion

Depressive disorder patients have longer in-hospital lengths of stay and increased odds of readmissions and complications following primary total shoulder arthroplasty.

Introduction

The Diagnostic and Statistical Manual of Mental Health Disorders, Fifth Edition (DSM-V) defines depression as having a loss of interest, appetite or weight changes, sleeping difficulties, psychomotor agitation, fatigue or loss of energy lasting greater than two weeks with severe impairment in activities of daily living.¹ Recent studies report the lifetime prevalence of depression within the United States to be 20.6%, which is expected to substantially increase.² In 2004, depression was considered to be the third leading cause of disease or injury worldwide for both sexes, and is projected to be the leading cause of disease burden worldwide by 2030 according to the World Health Organization.^2,3 Studies have shown that patients who have concomitant anxiety symptoms, functional impairment, chronic illness, low mastery, low education, and having no partner were risk factors that best predicted the risk of depression.^2,4,5

Recent studies have demonstrated that patients who have depressive disorders have worse outcomes and are prone to adverse events following orthopedic surgery.^6–10 However, many of these studies combine multiple health disorders into a single cohort, group multiple procedures, or their studies were not pertinent to total shoulder arthroplasty (TSA). With the expected increase in TSA procedures from 10.1 cases per 100,000 to 184.8 cases per 100,000 by 2030 along with the expected increase in prevalence of depression, orthopedic surgeons performing primary TSA will invariably encounter more of these patients in their practice.¹¹

Therefore, the purpose of this study was to investigate the impact on patients who have depressive disorders undergoing primary TSA. Specifically, we evaluated whether patients with depressive disorders are at greater odds of (1) increased in-hospital lengths of stay (LOS), (2) readmission rates, (3) medical complications, and (4) implant-related complications.

Materials and methods

Database

A query from 1 January 2005 to 31 March 2014 was performed using the Medicare Standard Analytical Files from the PearlDiver (PearlDiver Technologies, Fort Wayne, Indiana) supercomputer. The commercially available database has been used extensively for orthopedic-related research.^12–14 The database contains over 100 million patients from the Humana insurance and Medicare insurance claims database and provides information such as diagnoses, procedures, complications, discharge dispositions, in addition to other metrics. Since the database contains deidentified information, our study was exempt from the institution’s Institutional Review Board approvals. Patients and complications were queried using International Classification of Disease, ninth revision (ICD-9) and Current Procedural Terminology (CPT) coding.

Study population

Using Boolean command operations, the inclusion criteria for the study group consisted of all patients undergoing primary TSA who have depressive disorders, whereas patients who were without depressive disorders undergoing TSA served as controls. The database was first queried for all patients undergoing primary TSA using CPT and ICD-9 procedural codes 23472 and 81.80, respectively. Patients who had a diagnosis of depressive disorders were queried using ICD-9 diagnoses codes: major depressive affective disorder, recurrent episode (ICD-9: 296.31–296.34), dysthymic disorder (ICD-9: 300.4), chronic depressive personality disorder (ICD-9: 301.12), adjustment disorder with depressed mood (ICD-9: 309), and depressive disorder, not elsewhere classified (ICD-9: 311). These codes were chosen due to their high specificity in diagnosing patients with depression.¹⁵

Study group patients were randomly matched to controls in a 1:5 ratio according to age, sex, and medical comorbidities—alcohol abuse, chronic obstructive pulmonary disease, general anxiety disorder, hyperlipidemia, hypertension, obesity (BMI >30 kg/m²), and tobacco use.^16–21 The query yielded 113,648 patients with (n = 18,953) and without (n = 94,695) depressive disorders undergoing primary TSA during the study interval. Matching was successful as there were no statistical differences between the cohorts for these matching metrics (Table 1).

Table 1.

Comparison of patient demographics in patients with and without depressive disorders undergoing primary total shoulder arthroplasty.

	Depressive disorders		Controls
Demographics	n	%	n	%	p-value a
Age (years)
<64	1082	5.7	5423	5.7	0.922
65–69	4780	25.2	23,881	25.2	0.996
70–74	4886	25.7	24,405	25.7	0.983
75–79	4312	22.7	21,544	22.7	0.999
80–84	2674	14.1	13,362	14.1	0.994
>85	1006	5.3	5016	5.3	0.951
Unknown	213	1.1	1064	1.1	0.997
Sex					0.369
Female	10,324	54.4	9986	10.5
Male	8416	44.4	42,030	44.3
Unknown	213	1.1	1064	1.1
Comorbidities
Alcohol abuse	619	3.2	3060	3.2	0.823
COPD	310	1.6	1540	1.6	0.951
General anxiety disorder	155	0.8	734	0.7	0.572
Hyperlipidemia	13,703	72.3	68,464	72.3	1.00
Hypertension	16,955	89.4	84,709	89.4	0.987
Obesity (BMI >30 kg/m2)	1863	9.8	9287	9.8	0.944
Tobacco use	2199	11.6	10,967	11.5	0.935

BMI: body mass index; COPD: chronic obstructive pulmonary disease.

^aAssessed by Pearson’s χ².

Outcomes assessed

Primary outcomes assessed included comparing in-hospital LOS, 90-day readmission rates, 90-day medical complications, and two-year implant-related complications. Ninety-day medical complications analyzed and compared included acute post-hemorrhagic anemia, acute kidney failure, cerebrovascular accidents, deep vein thromboses, myocardial infarctions, paralytic ileus, pneumonias, thrombocytopenia, transfusion of blood products, and urinary tract infections. Two-year implant-related complications analyzed and compared included articular bearing surface wear, broken prosthetic joint implants, dislocations of prosthetic joint, mechanical loosening, peri-prosthetic fractures, and peri-prosthetic joint infections.

Data analyses

Patient demographics including age, sex, and medical comorbidities were analyzed and compared using Pearson’s χ² analyses. Welch’s t-tests were used to test for significance for in-hospital LOS between the cohorts. Logistic regression analyses were used to calculate odds-ratios (ORs) and 95% confidence intervals (95%CI) on the impact of depressive disorders on medical complications and readmission rates. Statistical analyses were performed using the open programming language R (R Foundation for Statistical Computation, Vienna, Austria). Due to the number of medical and implant-related complications being assessed, Bonferroni corrections were performed to reduce the probability of type I errors.²² A p-value less than 0.003 was considered statistically significant. This was attained by dividing 0.05 by the number of comparisons being made (n = 16).

Results

In-hospital LOS and readmission rates

Patients with depressive disorders undergoing primary TSA were found to have significantly longer mean in-hospital LOS (2.71 days versus 2.34 days; p < 0.0001) compared to matched controls. Additionally, patients with depressive disorders undergoing primary TSA were found to have higher incidence and odds (9.37 versus 6.15%; OR: 1.57, 95%CI: 1.49–1.66, p < 0.0001) of 90-day readmission rates compared to matched controls.

Ninety-day medical complications

Patients with depressive disorders undergoing primary TSA were found to have higher incidence and odds (2.7 versus 0.93%; OR: 3.00, 95%CI: 2.69–3.34, p < 0.0001) of 90-day medical complications compared to matched controls. Study group patients were found to have higher incidence and odds of paralytic ileus (0.07 versus 0.02%; OR: 4.06, 95%CI: 1.95–8.44, p = 0.0002), thrombocytopenia (0.06 versus 0.02%; OR: 3.99, 95%CI: 1.87–8.54, p = 0.0003), acute post-hemorrhagic anemia (0.68 versus 0.18%; OR: 3.78, 95%CI: 3.01–4.76, p < 0.0001), requiring transfusion of blood products (0.27 versus 0.08%; OR: 3.56, 95%CI: 2.49–5.09, p < 0.0001), cerebrovascular accidents (0.10 versus 0.03%; OR: 2.96, 95%CI: 1.68–5.23, p = 0.0002) in addition to other medical complications (Table 2).

Table 2.

Comparison of medical complications in patients with and without depressive disorders undergoing primary total shoulder arthroplasty.

Medical complications	DD (%)	Control (%)	OR	95%CI	p-value
Paralytic ileus	0.07	0.02	4.06	1.95–8.44	0.0002
Thrombocytopenia	0.06	0.02	3.99	1.87–8.54	0.0003
Acute Post-hemorrhagic anemia	0.68	0.18	3.78	3.01–4.76	<0.0001
Transfusion of blood products	0.27	0.08	3.56	2.49–5.09	<0.0001
Cerebrovascular accidents	0.10	0.03	2.96	1.68–5.23	0.0002
Urinary tract infections	0.91	0.32	2.84	2.35–3.43	<0.0001
Acute kidney failure	0.28	0.11	2.52	1.81–3.51	<0.0001
Pneumonia	0.24	0.10	2.42	1.69–3.45	<0.0001
Myocardial infarction	0.06	0.03	1.96	0.97–3.94	0.058
Deep vein thrombosis	0.07	0.04	1.62	0.86–3.03	0.128
Total medical complications	2.74	0.93	3.00	2.69–3.34	<0.0001

CI: 95% confidence interval; DD: depressive disorder; OR: odds-ratio.

Two-year implant-related complication

The study demonstrated patients with depressive disorders undergoing primary TSA had significantly higher incidence and odds (6.07 versus 2.43%; OR: 2.59, 95%CI: 2.41–2.78, p < 0.0001) of implant-related complications compared to matched controls. Patients with depressive disorders were found to have higher incidence and odds of peri-prosthetic fractures (0.60 versus 0.15%; OR: 4.05, 95%CI: 3.16–5.19, p < 0.0001), dislocation of prosthetic joint (2.42 versus 0.92%; OR: 2.65, 95%CI: 2.36–2.97, p < 0.0001), broken prosthetic joint implant (0.49 versus 0.20%; OR: 2.37, 95%CI: 1.85–3.04, p < 0.0001), mechanical loosening (1.35 versus 0.59%; OR: 2.30, 95%CI: 1.98–2.68, p < 0.0001) in addition to other implant-related complications (Table 3).

Table 3.

Comparison of implant-related complications in patients with depressive disorders and controls undergoing primary total shoulder arthroplasty.

Implant-related complications	DD (%)	Control (%)	OR	95%CI	p-value
Peri-prosthetic fractures	0.60	0.15	4.05	3.16–5.19	<0.0001
Dislocation of prosthetic joint	2.42	0.92	2.65	2.36–2.97	<0.0001
Broken prosthetic joint implant	0.49	0.20	2.37	1.85–3.04	<0.0001
Mechanical loosening	1.35	0.59	2.30	1.98–2.68	<0.0001
Peri-prosthetic joint infections	1.11	0.50	2.25	1.91–2.65	<0.0001
Articular bearing surface wear	0.12	0.07	1.61	1.01–2.59	0.044
Total implant complications	6.07	2.43	2.59	2.41–2.78	<0.0001

CI: 95% confidence interval; DD: depressive disorder; OR: odds-ratio.

Discussion

Recent studies have demonstrated that the number of primary TSA procedures performed within the United States will increase over 1700% by 2030 from 10.1 cases per 100,000 patients to nearly 185 cases per 100,000 patients.¹¹ During the same time, the World Health Organization anticipates that increasing incidence of depression will make it the leading medical comorbidity within the United States by 2030.²³ Studies investigating the impact of depressive disorders on patients undergoing primary TSA are limited. The results of the present study demonstrated that patients who have depressive disorders have longer in-hospital LOS in addition to increased odds of complications and readmission rates following primary TSA.

While the large sample size ensures that the study was adequately powered, it was not without limitations, most of which are inherent to the use of an administrative database. The validity of the study is reliant on accurate diagnostic and procedural coding, and it is currently estimated that there are up to 1.3% of coding errors within the Medicare database.²⁴ Additionally, we only analyzed a single insurance database, and the results of the study might not be a true cross-sectional representation on the impact of depressive disorders following primary TSA.²⁵ While our study controlled for patient demographics such as age, sex, and comorbidities associated with depressive disorders, it is possible that other patient-related comorbidities could explain the results of our study. Additionally, while the database provided the mean LOS between the cohorts, the specific range for the cohorts was not available. Furthermore, due to research parameters of the PearlDiver database, we were unable to determine if patients who had depressive disorders had acute or chronic symptoms, which could impact the results of the study, and could be the basis of future studies. Even though we have established that patients with a diagnosis of depression have inferior outcomes following shoulder arthroplasty compared to patients who do not have this diagnosis, it is unclear what the pathophysiologic mechanism is that drives these outcomes. It is possible that future prospective, randomized controlled studies might be able to elicit this mechanism in more detail. Despite these limitations, the study is the first, to the best of the author’s knowledge, to analyze the effects of a growing mental health disorder in a population of over 100,000 patients undergoing primary TSA.

Our results are consistent with the findings in the literature. We found that patients who have depressive disorders had significantly longer in-hospital LOS and odds of readmissions compared to matched controls. In a retrospective study of 224,060 patients, Mollon et al.⁶ analyzed the impact of depression on patients undergoing shoulder arthroplasty by analyzing the Nationwide Inpatient Sample database. The study found patients who have depression had significantly longer in-hospital LOS (2.5 days versus 2.2 days; p < 0.0001) compared to controls. Prina et al.⁷ found similar results when analyzing the LOS in 5411 depressed males older than the age of 69 years by utilizing the Western Australian Data Linkage System and demonstrated male patients with depression had significantly longer in-hospital LOS (5.8 days versus 4.0 days; p = 0.002) compared to male patients without depression. Bot et al.⁸ found that patients who have depressive disorders actually had lower LOS compared to patients without depression (3.0 days versus 3.7 days, p < 0.0001); however, a large limitation to their study was also incorporating the results of patients who underwent partial shoulder arthroplasty without matching the cohorts based on medical comorbidities. Other major limitations to the aforementioned studies included combining patients who underwent reverse shoulder arthroplasty with primary TSA which could potentially overestimate the impact of depressive disorders on patients undergoing TSA.

While studies comparing readmission rates in patients with depressive disorders undergoing TSA were limited, our results were consistent with those findings in other orthopedic surgical procedures. Gold et al.⁹ found patients with depression increased the odds of 90-day hospital readmissions in patients undergoing either primary total knee (OR: 1.21, 95%CI: 1.13–1.29, p < 0.001) or total hip (OR: 1.24, 95%CI: 1.13–1.35, p < 0.001) arthroplasty. These results are further supported by Ali et al.²⁶ who assessed for patient-related risk factors for readmissions following primary total hip arthroplasty in 514,455 patients utilizing the United Kingdom National Health Service and found depression significantly increased the odds (OR: 1.52, 95%CI: 1.31–1.76, p < 0.001) of 30-day readmissions following primary total hip arthroplasty. In addition to longer LOS and higher odds of readmissions, our study demonstrated that patients who have depressive disorders had higher incidence and odds of medical and implant-related complications.

Mollon et al.⁶ found that patients who have depression had higher odds of requiring transfusion of blood products (OR: 1.30, p < 0.05), acute renal failure (OR: 1.37, p = 0.012), and postoperative anemias (OR: 2.09, p = 0.045). However, a major limitation was that they included patients with reverse shoulder arthroplasty within their population, potentially overestimating the impact of depressive disorders in patients undergoing shoulder arthroplasty.⁶ Davydow et al.²⁷ analyzed the risk of hospitalization for pneumonia in patients greater than the age of 50 years and found that depression significantly increased the odds of pneumonia (OR: 1.28, p < 0.05) compared to controls. After adjusting for baseline cognitional and functional impairments, multivariate analyses demonstrated that depression continued to increase the odds of pneumonia (OR: 1.24, p < 0.05).²⁷ Utilizing the National Hospital Discharge Survey, Bot et al.⁸ analyzed the impact of depression on patients undergoing primary TSA and partial shoulder arthroplasty and found that patients who have depression had significantly higher incidences of postoperative anemia (10 versus 7.7%; p < 0.0001), pulmonary insufficiency (1.2 versus 0.7%; p < 0.0001), and requiring transfusions of blood products (10 versus 6.2%; p < 0.0001) compared to patients without depression. On multivariate logistic regression analyses for predictors of adverse events, Bot et al.⁸ found depression significantly increased the odds (OR: 1.70, p < 0.001) of adverse events following shoulder arthroplasty.

In addition to greater medical complications, our study demonstrated patients with depressive symptoms had higher incidence and odds of implant-related complications. Bozic et al.²⁸ identified depression to be an independent risk factor for peri-prosthetic joint infections (hazard ratio: 1.28) in patients undergoing primary total knee arthroplasty. While the pathophysiology has not been adequately established, proposed theories include malnutrition and subsequent need for allogenic blood transfusions which are separate risk factors for developing peri-prosthetic joint infections.^28,29 In addition to PJIs, our study found patients with depressive disorders were at increased odds of peri-prosthetic fractures. In a population-based, cross-sectional study of 1566 Medicare patients above the age of 65 with depression, Robbins et al.³⁰ found patients with depression had significantly lower bone mineral density compared to controls after adjusting for age, body mass index, estrogen use, gender, race, smoking, and drinking. These metabolic changes could potentially explain the increased incidence of peri-prosthetic fractures following primary TSA.

Conclusion

After matching for age, sex, and medical comorbidities, this analysis of over 100,000 patients demonstrated that patients who have depressive disorders undergoing primary TSA have increased in-hospital LOS, in addition to greater odds of readmissions, medical complications, and implant-related complications. Future studies should study the incidence of adverse events in patients who have acute or chronic depressive disorders, as varying durations of depressive disorders could potentially have different impact on outcomes following primary TSA. Nevertheless, with the growing prevalence of patients who have depression, the present study is important in allowing orthopedic surgeons to adequately counsel and educate patients of the potential adverse events which may occur following their procedures.