Cirrhosis associated with increased complications and healthcare utilization following total shoulder arthroplasty

Abstract

Introduction

Cirrhosis is a known risk factor for morbidity and mortality following surgical procedures and has been associated with increased complications, hospital length of stay (LOS), and cost of admission following total joint arthroplasty. However, a paucity of literature exists evaluating the effect of cirrhosis on postoperative outcomes following total shoulder arthroplasty (TSA). The purpose of this study is to evaluate the short-term outcomes following elective primary TSA in patients with cirrhosis compared to matched controls.

Methods

The Nationwide Readmissions Database was queried from 2016 to 2020 to identify patients who underwent elective primary TSA. Patients with a diagnosis of cirrhosis (n = 627) were matched in a 1:1 proportion to patients who did not have cirrhosis. Bivariate statistical analyses were performed to compare preoperative demographic and comorbidity data, postoperative outcomes, and hospital utilization metrics between the two groups. Following Bonferroni correction, an alpha value of 0.003 defined significance.

Results

Patients with cirrhosis exhibited higher rates of postoperative medical and implant-related complications following primary TSA, including acute renal failure (6.3% vs 1.1%: p < 0.001), urinary tract infection (3.5% vs 0.6%; p < 0.001), transfusions (3.0% vs 0.2%; p < 0.001), acute respiratory distress syndrome (2.9% vs 0.2%: p = 0.002), surgical site infection (2.0% vs 0.2%: p = 0.001), dislocation (2.1% vs 0.0%: p < 0.001), and prosthetic loosening (1.5% vs 0.0%; p = 0.002). These patients also exhibited higher rates of all-cause complications (32% vs 9.2%: p < 0.001) and mortality (1.5% vs 0.0%; p = 0.002) within 180 days of surgery and had an increased cost of admission ($24,633 vs $18,500; p < 0.001) and LOS (2.6 vs 1.5 days; p < 0.001).

Conclusion

Patients with cirrhosis were found to have increased risk of medical and surgical complications, higher costs, and longer LOS following TSA. These findings can assist orthopedic surgeons in developing strategies in the preoperative period to mitigate complications in this at-risk patient group.

Level of evidence

Level III – Retrospective cohort study.

Introduction

Cirrhosis represents the late stage of progressive liver fibrosis or scarring and can be precipitated by numerous acute and chronic medical conditions such as viral hepatitis, alcohol abuse, hemochromatosis, non-alcohol associated fatty liver disease, medication side effect, autoimmune disease, diabetes, smoking, and celiac disease.^1–8 The accumulation of intrahepatic scar tissue leads to decreased overall function of the liver resulting in coagulopathy, impaired immune function, malnutrition, and portal hypertension which are the leading causes of morbidity and mortality.^9–11 Risk factors for development of cirrhosis include amount of alcohol consumed, smoking, and diabetes. Most cases of cirrhosis remain asymptomatic until the liver begins to decompensate, thus it is estimated that one in three people with cirrhosis are unaware of their condition. ¹² With improvements in the overall medical care of cirrhosis, the longevity of these patients has increased, resulting in an increase of cirrhotic patients undergoing elective surgeries such as total shoulder arthroplasty (TSA).^13,14

TSA has proven to be effective in managing a range of degenerative and traumatic shoulder pathologies such as glenohumeral arthritis, rotator cuff tears and arthropathy, and complex humerus fractures. ¹⁵ An estimated 800,000 patients are currently living with a TSA highlighting the common practice and success rate of TSA. ¹⁶ However, the overall altered systemic function seen in cirrhosis can increase complications following TSA, resulting in an increase in perioperative morbidity and mortality. With an increase in overall surgical volume of TSA of nearly 100% from 2011 to 2017, combined with an anticipated increase of an additional 235% by 2025 it is vital to identify preoperative risk factors and comorbidities that may impact patient care and postoperative outcomes.^17–19

Previous studies by Tiberi et al. and Metikala et al. investigating patients with a preoperative diagnosis of cirrhosis undergoing total joint arthroplasty (TJA) have shown that these patients had prolonged length of stay (LOS), higher rates of readmission within 90 days, and a higher overall cost of care roughly $10,000 more than non-cirrhotic patients. These studies also demonstrated that preoperative cirrhosis is associated with higher rates of perioperative complications such as acute renal failure, periprosthetic fracture, prosthetic joint infection, revision, dislocation, urinary tract infection, blood transfusion, as well as increased overall revision and mortality rates.^20,21 Despite these findings, there are few studies investigating the impact of cirrhosis in patients undergoing TSA.

The purpose of this study is to investigate the association of cirrhosis with postoperative outcomes and healthcare utilization following elective primary TSA. We hypothesize that patients with a preoperative diagnosis of cirrhosis undergoing TSA will have increased rates of complications, readmissions, revisions, and mortality, as well as increased cost of admission, and overall healthcare utilization following TSA compared to those without a preoperative diagnosis of cirrhosis.

Methods

This was a retrospective cohort study using a large national administrative claims database. This study was exempt from Institutional Review Board (IRB) approval, as it utilizes patient data that is publicly available and de-identified. The Nationwide Readmissions Database (NRD) was queried from 2016 to 2020 to identify all patients who underwent elective primary TSA using International Classification of Disease, tenth revision (ICD-10) Procedure Coding System (PCS) codes. ICD-10 PCS codes were used to identify all patients who underwent both anatomic (0RRJ0JZ, 0RRK0JZ) and reverse (0RRJ00Z, 0RRK00Z) TSA. The NRD is a commercially available database of de-identified patient data created by the Healthcare Cost and Utilization Project (HCUP). The NRD draws from HCUP's State Inpatient Databases (SID) to provide a collection of patient data that represents an estimated 61% of United States hospital admissions across 31 states. Each inpatient admission is associated with a unique discharge weight that can be used to extrapolate this data to represent all inpatient admissions in the United States. This database utilizes a verified patient linking variable to allow for tracking of admissions and subsequent readmissions within a given calendar year. The NRD features a large patient sample size, which provides sufficient data for analysis of readmissions for relatively uncommon disorders and procedures. All patients who underwent revision TSA, removal of a TSA component, or shoulder hemiarthroplasty were excluded from the initial patient selection. All patients who underwent any additional procedure within 180 days of the index primary TSA were also excluded to appropriately attribute all postoperative complications to the primary TSA of interest. This included all procedures with an associated ICD-10 PCS code and was not limited to procedures of the shoulder joint.

Patients with cirrhosis were identified using ICD-10 Clinical Modification (CM) codes K74, K74.3, K74.4, K74.5, K74.6, K74.60, and K74.69. This query ultimately identified 627 patients with cirrhosis who underwent primary TSA. To control for the effect of demographic variables and preoperative medical comorbidities other than cirrhosis on postoperative outcomes, patients with cirrhosis were matched 1:1 with patients who underwent primary TSA and did not have a preoperative diagnosis of cirrhosis. These patients were matched based on age, sex, and the presence of preoperative hypertension requiring medication, diabetes mellitus (DM), congestive heart failure (CHF), peripheral vascular disease (PVD), chronic kidney disease (CKD), and chronic lung disease using an optimized matching algorithm. Demographic data, preoperative comorbidities, postoperative medical and surgical complications, and economic and hospital metrics were compared between patients with and without a preoperative diagnosis of cirrhosis.

Preoperative demographic variables included age, sex, active smoking status, presence of alcohol abuse, median household income quartile, and insurance payer. Median household income quartile was computed relative to other households within a given patient's residential ZIP code. Preoperative comorbidities included hypertension requiring medication, DM, CHF, PVD, CKD, chronic lung disease, and Charlson-Deyo Comorbidity Index (CCI). The CCI is a composite index that incorporates 17 common patient comorbidities into a weighted, numeric score that can be used to estimate a patient's risk of mortality within 10 years of surgery.

Postoperative outcomes included both medical and surgical complications. Medical complications included bleeding complications, cardiovascular complications, central nervous system (CNS) complications, gastrointestinal (GI) complications, acute renal failure (ARF), urinary tract infection (UTI), peripheral vascular complications, postoperative shock, respiratory complications, complications requiring transfusion, cellulitis, pulmonary embolism (PE), pneumonia, acute respiratory distress syndrome (ARDS), sepsis, and thrombosis. Surgical complications included surgical site infection (SSI), wound dehiscence, hardware failure, dislocation, prosthetic loosening, osteolysis, and fracture. Additionally, rates of all-cause complications, hospital readmission, revision TSA, and mortality within 180 days of the index procedure were generated and compared between groups.

Statistical analysis was conducted using IBM SPSS^® Statistics software, version 28 (IBM, Armonk, NY, USA). Categorical variables were compared using Pearson chi-square test, while continuous variables were compared using Student t-test. For all comparisons of postoperative outcomes that achieved statistical significance, odds ratios (OR) with respective 95% confidence intervals (95% CI) were computed. OR could not be computed for comparisons in which the incidence of a given complication for a patient group was zero. Following implementation of post-hoc Bonferroni correction, an alpha value of 0.003 defined statistical significance for all comparisons.

Results

The results of a comparison of demographic characteristics between primary TSA patients with and without cirrhosis are shown in Table 1 . Patients with a preoperative diagnosis of cirrhosis were found to have a higher mean CCI (6.1 vs 1.0; p < 0.001) and exhibited higher rates of alcohol abuse (6.5% vs 1.8%; p < 0.001). These patients were also more likely to use Medicare (75% vs 67%; p < 0.001) and Medicaid (7.7% vs 5.4%; p < 0.001) insurance and less likely to use private insurance (13% vs 19%; p < 0.001).

Table 1.

Demographic data for patients with and without cirrhosis undergoing primary total shoulder arthroplasty.

Demographic	Cirrhosis (n = 627)	No cirrhosis (n = 627)	p-value
Age (avg. years ± std. dev.)	68 ± 8.4	68 ± 8.4	0.884
Sex (percent female)	61%	61%	1.000
CCI	6.1 ± 1.8	1.0 ± 1.2	* < 0.001
Active smoker	30%	24%	† 0.013
Alcohol abuse	6.5%	1.8%	* < 0.001
Median household income			0.371
1st quartile	28%	28%
2nd quartile	28%	28%
3rd quartile	25%	28%
4th quartile	19%	16%
Insurance			* < 0.001
Medicare	75%	67%
Medicaid	7.7%	5.4%
Private	13%	19%
Self-pay	0.6%	0.3%

^*Statistically significant.

^†Not significant following Bonferroni correction.

Comparisons of postoperative medical and surgical complications are detailed in Tables 2 and 3, respectively. The results of bivariate analysis found patients with cirrhosis to exhibit increased rates of postoperative ARF (6.3% vs 1.1%; OR 6.3 [95% CI 2.8–14]; p < 0.001), UTI (3.5% vs 0.6%; OR 5.9 [95% CI 2.0–17]; p < 0.001), complications requiring transfusion (3.0% vs 0.2%; OR 21 [95% CI 2.8–154]; p < 0.001), ARDS (2.9% vs 0.2%; OR 4.9 [95% CI 1.6–14]; p = 0.002), SSI (2.0% vs 0.2%; OR 13 [95% CI 1.7–102]; p = 0.001), dislocation (2.1% vs 0.0%; p < 0.001), and prosthetic loosening (1.5% vs 0.0%; p < 0.001) following primary TSA. Additionally, patients with cirrhosis exhibited higher rates of all-cause complications (32% vs 9.2%; OR 4.6 [95% CI 3.4–6.3]; p < 0.001) and mortality (1.5% vs 0.0%; p = 0.002) within 180 days of TSA.

Table 2.

Rates of medical complications and outcomes for patients with and without cirrhosis undergoing primary total shoulder arthroplasty.

Complications	Cirrhosis (n = 627)	No cirrhosis (n = 627)	p-value	OR (95% CI)
Any complication within 180 days	32%	9.2%	* < 0.001	4.6 (3.4–6.3)
Bleeding	0.2%	0.2%	1.000	-
Cardiovascular, including MI	0.3%	0.3%	1.000	-
CNS, including stroke	0.0%	0.0%	1.000	-
GI	0.0%	0.0%	1.000	-
ARF	6.3%	1.1%	* < 0.001	6.3 (2.8–14)
UTI	3.5%	0.6%	*   < 0.001	5.9 (2.0–17)
Peripheral vascular	0.0%	0.0%	1.000	-
Post-operative shock	0.2%	0.1%	0.317	-
Respiratory	0.0%	0.2%	0.317	-
Transfusion necessary	3.0%	0.2%	* < 0.001	21 (2.8–154)
Cellulitis	0.5%	0.2%	0.317	-
PE	0.2%	0.2%	1.000	-
Pneumonia	0.9%	0.3%	0.156	-
ARDS	2.9%	0.2%	* 0.002	4.9 (1.6–14)
Sepsis	0.6%	0.0%	† 0.045	-
Thrombosis	0.6%	0.0%	† 0.045	-
Readmission within 180 days	17%	13%	† 0.025	-
Revision within 180 days	2.9%	1.2%	† 0.032	-
Death within 180 days	1.5%	0.0%	* 0.002	-

^*Statistically significant.

^†Not significant following Bonferroni correction.

Table 3.

Rates of surgical complications for patients with and without cirrhosis undergoing primary total shoulder arthroplasty.

Complications	Cirrhosis (n = 627)	No cirrhosis (n = 627)	p-value	OR (95% CI)
SSI	2.0%	0.2%	* 0.001	13 (1.7–102)
Wound dehiscence	0.0%	0.0%	1.000	-
Hardware failure	0.3%	0.0%	0.157	-
Dislocation	2.1%	0.0%	*   < 0.001	-
Prosthetic loosening	1.5%	0.0%	* 0.002	-
Osteolysis	0.0%	0.0%	1.000	-
Fracture	1.7%	0.2%	† 0.004	-

^*Statistically significant.

^†Not significant following Bonferroni correction.

The results of an analysis of economic and hospital metrics are shown in Table 4 . Patients with cirrhosis were found to have a significantly higher cost of admission ($24,633 vs $18,500; p < 0.001) for primary TSA, as well as a significantly longer total hospital LOS (2.6 vs 1.5 days; p < 0.001) compared to patients without cirrhosis. These patients were also more likely to be discharged to another care facility (13% vs 6.2%; p < 0.001) or home with home healthcare (25% vs 20%; p < 0.001) and were less likely to be discharged routinely (61% vs 73%; p < 0.001) than patients who did not have a preoperative diagnosis of cirrhosis.

Table 4.

Average and relative costs for patients with and without cirrhosis undergoing primary total shoulder arthroplasty.

Resource	Cirrhosis (n = 627)	No cirrhosis (n = 627)	p-value
Total cost (2020 USD)	24,633 ± 17,418	18,500 ± 6756	* < 0.001
Total hospital length of stay (days)	2.6 ± 3.6	1.5 ± 1.1	* < 0.001
Disposition			* < 0.001
Routine ‡	61%	73%
Short-term hospital	0.0%	0.3%
Another type of facility	13%	6.2%
Home Health Care (HHC)	25%	20%

^*Statistically significant.

^‡Discharge home or self-care.

Discussion

The results of this study showed that patients with a preoperative diagnosis of cirrhosis have higher rates of medical and surgical complications compared to non-cirrhotic patients following primary TSA. Patients with cirrhosis also exhibited higher rates of all-cause complications and mortality within 180 days of surgery compared to patients without preoperative cirrhosis. They were also more likely to be discharged non-routinely, had longer hospital LOS, and incurred a higher cost of care. These findings largely supported our initial hypothesis.

Previous studies by Newman et al. have found that patients with cirrhosis undergoing lower extremity TJA had increased odds of developing acute postoperative anemia, thrombocytopenia, hematoma/hemorrhage, ARF, device-related complications, infection, and complications of medical care. ²² Additionally, Tiberi et al. reported that patients with cirrhosis undergoing lower extremity TJA had significantly higher rates of blood transfusions, renal failure, infections, dislocations, revision surgery, and mortality. ²⁰ These findings corroborate the results of this study in which we found that patients diagnosed with cirrhosis had increased rates of all-cause complications and mortality within 180 days of TSA.

The results of this study also demonstrate increases in postoperative ARF, complications requiring transfusions, ARDS, SSI, dislocation, and prosthetic loosening in patients with cirrhosis following TSA. These complications may be due to disruption of the protein synthesis in the coagulation cascade by the liver, resulting in an increased risk of bleeding and thrombotic complications. The liver also plays a large role in the clearance of pathogens, cytokine production, and wound healing, so in the case of fibrosis within the liver these processes may be compromised leading to an increased risk of infection and poor wound healing.^23–25 ARF in the postoperative setting in these patients may be due to both perioperative blood loss secondary to poor coagulation and portal hypertension leading to splenchnic vasodilation and decreased arterial blood volume, resulting in activation of the renin-angiotensin-aldosterone system and ultimately decreasing renal perfusion.^26,27

Previous literature has shown that patients with an increased preoperative comorbidity are at increased risk for increased overall medical cost, which is largely driven by an increase in hospital LOS. ¹² Prior literature has found the presence of cirrhosis to be associated with a mean increase in total cost of care of $12,745 per patient annually. Newman et al. reported that patients with a preoperative diagnosis of cirrhosis undergoing TJA had statistically significant difference in additional cost and LOS, with an additional cost of $4516 and 2.2 days longer LOS in those who underwent total hip arthroplasty, and an additional cost of $1532 and 0.6 days longer LOS in those who underwent total knee arthroplasty. ²² Similarly, the results of this study showed that patients with cirrhosis were found to have a higher cost of admission and a longer total hospital LOS following primary TSA compared to patients without a diagnosis of cirrhosis. Additionally, patients with a preoperative diagnosis of cirrhosis were more likely to be discharged to another care facility or home with home healthcare and were overall less likely to be discharged routinely compared to patients without a diagnosis of cirrhosis undergoing primary TSA.

While the effects of cirrhosis have been demonstrated previously in THA and TKA, few studies have investigated the role of cirrhosis in TSA. The results of this study demonstrate that, like THA and TKA, cirrhosis is a major obstacle in the planning and execution of TSA and is associated with increased risks of acute medical and implant related complications, all of which can result in an increase in healthcare utilization.

Strengths of this retrospective cohort study include a large patient sample size, which allowed for high powered statistical analysis for conditions with lower incidence rates that would likely not be possible with a retrospective review at a single institution. A national database was used allowing for geographically dispersed patients across institutions of different sizes and surgical volumes increases generalizability of the results. The follow-up of 180-days is also longer than other national databases. There are several limitations in this study that should be considered. First, analyzing data from an extensive administrative database may limit the availability of other relevant perioperative information. This was a retrospective study relying on proper input and ICD-10 coding thus incorrect coding and entry remain possible. Plus, the database does not provide any additional information about the severity of cirrhosis. Additional complications may also occur outside of the 180-day capture period used by this database, thus additional investigation with longer postoperative follow-up should be performed to elucidate additional long-term risks in this patient population. Further investigation should also be performed to quantify the levels of preoperative laboratory values, such as liver function enzymes and coagulation measures, that render patients with cirrhosis susceptible to increased risk of adverse outcomes to further guide appropriate preoperative risk stratification.

Conclusion

A preoperative diagnosis of cirrhosis is associated with increased risk of several medical and surgical complications and mortality following primary TSA. Patients with cirrhosis are also more likely to incur higher costs of case and increased healthcare utilization compared to patients who do not have a preoperative diagnosis of cirrhosis. These results may help guide clinical decision making when considering TSA in complex cirrhotic patients, while also providing insight into possible risks and complications of TSA in this at-risk patient population.