Translate this page into:
Perioperative complications and length of stay in patients with diabetes mellitus after aseptic revision total shoulder arthroplasty
, Sarah I. Goldfarba, Jake DePalob, Henry Maxwell Foxa, Ansh Kishanc, Umasuthan Srikumarana,
⁎Corresponding author: Umasuthan Srikumaran. us@jhmi.edu
-
Received: ,
Accepted: ,
This article was originally published by Reed Elsevier India Pvt. Ltd. and was migrated to Scientific Scholar after the change of Publisher.
Abstract
Abstract
Total shoulder arthroplasty (TSA) is increasingly performed, but complications and revisions are rising concerns, especially in the presence of comorbidities such as diabetes mellitus (DM). Our study aimed to assess the association between DM and perioperative complications and length of hospital stay after aseptic revision TSA, which has been underexplored.
Using the National Surgical Quality Improvement Program (NSQIP) database, we analyzed 1979 patients undergoing revision TSA from 2006 to 2021. Patients were categorized as with DM (n = 404) or without DM (n = 1575). We examined patient demographics, comorbidities, and surgical characteristics. Multivariate logistic regression was used to identify risk factors for major complications and prolonged hospital stays.
Patients with DM showed similar rates of major and minor complications and hospital stays compared to those without DM. Hypertension emerged as a significant risk factor for major complications (odds ratio [OR] 4.0, 95 % confidence interval [CI] 1.3, 17.1), whereas American Society of Anesthesiologists classification >2 (OR 3.6, 95 % CI 1.4, 12.6) and preoperative hematocrit <35 % (OR 4.3, 95 % CI 2.1, 8.6) were associated with prolonged hospital stays. Multivariate analysis did not find DM to be an independent predictor of prolonged hospital stays.
In patients undergoing aseptic revision TSA, 30-day complication rates and lengths of hospital stays did not differ significantly between patients with and without DM. These findings challenge previous assumptions about DM's impact on TSA revisions. The present study emphasizes the importance of individualized risk assessment, focusing on comorbidities such as hypertension and anemia, which appear to play a more critical role in outcomes. Further research is needed to explore the heterogeneity within the diabetic population and to refine bundled-care models to optimize risk stratification and resource allocation in TSA revisions.
Abstract
Highlights
•DM did not increase 30-day complication rates after revision TSA.•Hypertension had higher complication risk than other comorbidities after revision TSA.•ASA >2 and hematocrit <35 % predicted prolonged hospital stays.•Findings suggested reevaluating DM's role in bundled-care models.
Keywords
Total shoulder arthroplasty
Aseptic revision
Diabetes mellitus
Postoperative complications
Healthcare costs
Bundled care model
1 Introduction
Total shoulder arthroplasty (TSA) is a surgical procedure used to relieve symptoms of multiple diseases, including osteoarthritis and rotator cuff pathology.1 TSA procedures and subsequent revisions have become increasingly common since the 1990s. More than 10,000 TSA revisions were performed in the United States in 2017, with a cost to the U.S. healthcare system of about $205 million.2 With the incidence of this costly procedure on the rise, it is important for physicians to understand which patient populations are most likely to need revision TSA and to explore ways to minimize patients’ risk of needing a second surgery.
According to the Centers for Disease Control and Prevention, about 1 in 10 Americans (or approximately 37.3 million people) has diabetes mellitus (DM).3 Diabetes is well known to increase the risk of postoperative complications after orthopaedic surgery, including poor wound healing, hyperglycemia, ketoacidosis, and hyperglycemic hyperosmolar syndrome.4,5 In addition, studies have shown that poor glycemic control in the preoperative stage is associated with an increased rate of postoperative complications.6,7 Currently, however, there is limited research on the effects of diabetes on complications in the setting of TSA revisions.
Our study examined the incidence of postoperative complications in diabetic versus non-diabetic patients who underwent aseptic revision TSA. We hypothesized that patients with diabetes mellitus would be more likely to experience postoperative complications and longer hospital lengths of stay after aseptic revision TSA than those without diabetes.
2 Materials and methods
2.1 Study design and setting
We conducted a retrospective analysis using data from the National Surgical Quality Improvement Program (NSQIP) database spanning January 2006 through December 2021. NSQIP is a surgical outcomes database that contains deidentified patient medical chart data from patients across the United States.Access to the NSQIP database can be obtained through application to the American College of Surgeons.8
2.2 Patient selection
We used Current Procedural Terminology (CPT) codes 23473 and 23474 to identify patients who underwent 1-component or 2-component revision shoulder arthroplasty between January 2006 and December 2021. Of the 2435 patients initially identified, 1979 (81 %) were included in this study (Fig. 1). Exclusion criteria included incomplete height or weight data (n = 19), lack of American Society of Anesthesiologists (ASA) classification (n = 2), incomplete preoperative hematocrit data (n = 244), surgery lasting fewer than 15 min (n = 6), International Classification of Diseases (ICD) diagnosis description that did not match the stated CPT code (n = 3), ICD diagnosis involving periprosthetic infection and/or sepsis (n = 175), incomplete readmission data (n = 5), and incomplete length of stay data (n = 2). Infection and sepsis cases were excluded to reduce variability between groups, because we sought to compare the impact of DM on aseptic revision outcomes. Cases with short operation times were excluded because of the implausibility of completing a revision shoulder arthroplasty within such timeframes; we selected 15 min as a threshold to effectively eliminate erroneous inputs while still encompassing all valid surgeries.
The septic versus aseptic revision status of a patient was determined using postoperative diagnosis defined by ICD 9 and ICD 10 codes (see Supplemental Table A1). Specifically, the ICD code affiliated with each included patient was noted and manually classified as a septic or aseptic revision; any diagnosis that was not explicitly characterized as septic was categorized as aseptic.
2.3 Variables and primary study outcomes
We collected data on patient age, body mass index (BMI), sex, ASA classification, number of components in the revision (1 or 2), comorbidities (hypertension, current smoker, history of severe chronic obstructive pulmonary disease, congestive heart failure, bleeding disorder, and hematocrit <35 %), and total operation time. We also collected information on postoperative data such as major complications (sepsis, septic shock, acute renal failure, pulmonary embolism, ventilator for >48 h, unplanned intubation, myocardial infarction, cardiac arrest requiring CPR, stroke, and mortality), minor complications (urinary tract infection, pneumonia, superficial surgical site infection, deep wound disruption, deep wound infection, and deep vein thrombosis), readmission, and length of hospital stay. Patients were divided into 2 groups, diabetic or non-diabetic, and variables were compared between the groups. Primary outcomes were major complications and prolonged hospital stay.
2.4 Ethical approval
Approval from an ethical committee was not required for this study because the study used data from a national, deidentified database.
2.5 Statistical analysis
We used descriptive statistics such as median, interquartile range, mean, standard deviation, and range to summarize our findings. Comparisons of demographic data and data on complications between the 2 groups were performed using Wilcoxon rank sum test, Fisher's exact test, one-way analysis of variance, and Pearson's chi-squared test, as indicated. Specifically, Wilcoxon rank sum test was used to compare ages, as all ages ≥89 are placed in a ≥89 years category by the NSQIP; Fisher's exact test was used for binary variables; one-way analysis of variance was used for continuous variables; and Pearson's chi-squared test was used for categorical variables. Statistical significance was set at P < .05. Additionally, multivariable logistic regression was used to identify risk factors for both major complications and prolonged hospital stay (defined as > 7 days).
Using univariate analyses, we found similar rates of both major and minor complications in the diabetic and non-diabetic cohorts (Table 2). Additionally, there was no significant difference between the groups in the rate of readmission. Notably, using univariate analysis alone, patients with diabetes had longer hospital stays (1.98 days versus 1.72 days, P = .03 by Pearson's chi-squared test) and a higher rate of prolonged hospital stays (3.5 % versus 1.5 %, P = .02 by Fisher's exact test).
| N (%) | ||||
| Characteristic | Total N = 1979 | No Diabetes N = 1575 | Diabetes N = 404 | P |
| Age (years) | 69 (62, 75)a | 69 (62, 75)a | 69 (63, 74)a | .99b |
| BMI | 31 ± 6.8c | 30 ± 6.6c | 34 ± 6.9c | <.001d |
| Sex | ||||
| Male | 905 (45.7) | 700 (44.4) | 205 (50.7) | .03e |
| Female | 1074 (54.4) | 875 (55.6) | 199 (49.3) | |
| ASA classification | ||||
| ASA I (normal health) | 23 (1.2) | 22 (1.4) | 1 (0.3) | .22f |
| ASA II (mild systemic disease) | 712 (36.0) | 638 (40.5) | 74 (18.3) | |
| ASA III (severe systemic disease) | 1164 (58.8) | 869 (55.2) | 295 (73.0) | |
| ASA IV (life-threatening) | 80 (4.04) | 46 (2.9) | 34 (8.4) | |
| Number of components replaced | ||||
| 1 | 736 (37.2) | 562 (35.7) | 174 (43.1) | .007e |
| 2 | 1243 (62.8) | 1013 (64.3) | 230 (56.9) | |
| Comorbidities | ||||
| Hypertension | 1310 (66.2) | 958 (60.8) | 352 (87.1) | <.001e |
| Current smoker | 235 (11.9) | 189 (12.0) | 46 (11.4) | .80e |
| History of severe COPD | 155 (7.8) | 114 (7.2) | 41 (10.2) | .06e |
| Congestive heart failure | 22 (1.1) | 15 (1.0) | 7 (1.7) | .19e |
| Bleeding disorder | 61 (3.1) | 41 (2.6) | 20 (4.95) | .02e |
| Hematocrit <35 % | 269 (13.6) | 190 (12.1) | 79 (19.6) | <.001e |
| Total operation time (minutes) | 126.1 ± 62.4c | 126.1 ± 62.4c | 125.8 ± 62.4c | .93d |
| Postoperative condition | Total N = 1979 | No Diabetes N = 1575 | Diabetes N = 404 | P |
| Major complications | 32 | 25 | 7 | .83b |
| Sepsis | 11 | 9 | 2 | >.99b |
| Septic shock | 1 | 1 | 0 | >.99b |
| Acute renal failure | 2 | 1 | 1 | .37b |
| Pulmonary embolism | 11 | 8 | 3 | .48b |
| Ventilator for >48 h | 1 | 0 | 1 | .20b |
| Unplanned intubation | 2 | 2 | 0 | >.99b |
| Myocardial infarction | 7 | 7 | 0 | .36b |
| Cardiac arrest requiring CPR | 3 | 2 | 1 | .50b |
| Stroke | 1 | 1 | 0 | >.99b |
| Mortality | 2 | 1 | 1 | .37b |
| Minor complications | 58 | 51 | 7 | .14b |
| Urinary tract infection | 16 | 13 | 3 | >.99b |
| Pneumonia | 6 | 6 | 0 | .61b |
| Superficial surgical site infection | 13 | 12 | 1 | .49b |
| Deep wound disruption | 2 | 2 | 0 | >.99b |
| Deep wound infection | 9 | 7 | 2 | >.99b |
| Deep vein thrombosis | 13 | 12 | 1 | .49b |
| Major or minor complication | 85 | 71 | 14 | .41b |
| Readmission | 85 | 68 | 17 | >.99b |
| Hospital stay (days) | 1.77 ± 1.91a | 1.72 ± 1.82a | 1.98 ± 2.20a | .03c |
| Hospital stay >7 days | 38 (1.9 %) | 24 (1.5 %) | 14 (3.5 %) | .02b |
3 Results
The median age of included patients was 69 years in both the diabetic and non-diabetic groups (Table 1). The average BMI was 34 in the diabetic group and 30 in the non-diabetic group (P < .001 by one-way analysis of variance). The diabetic and non-diabetic groups were 49 % and 56 % women, respectively (P = .03 by Fisher's exact test). The distribution of ASA classification did not differ significantly between groups, with ASA classification III (severe systemic disease) comprising the majority of cases in both groups. Two-component replacements constituted 57 % of surgeries among patients with diabetes and 64 % of surgeries among those without diabetes (P = .007 by Fisher's exact test). Among the comorbidities studied, hypertension, bleeding disorders, and hematocrit <35 % were significantly more prevalent in the diabetic cohort. Total operation time did not differ significantly between the groups.
Using multivariate analyses, hypertension was the only significant risk factor identified for major complications (odds ratio [OR] 4.0, 95 % CI 1.3, 17.1) (Table 3). Risk factors for prolonged hospital stay included ASA >2 (OR 3.6, 95 % CI 1.4, 12.6) and preoperative hematocrit <35 % (OR 4.3, 95 % CI 2.1, 8.6). Notably, diabetes was not identified as a significant factor in either model, even though univariate analyses indicated a higher rate of prolonged hospital stays in the diabetes cohort.
| Variable | Odds Ratio (95 % Confidence Interval) | |
| Major Complication | Prolonged Stay (>7days) | |
| Age >75 years | 1.1 (0.5, 2.3) | 2.0 (0.9, 4.1) |
| BMI >30 | 1.7 (0.8, 3.8) | 1.3 (0.6, 2.7) |
| Female sex | 1.4 (0.7, 3.0) | 1.3 (0.6, 2.6) |
| ASA > II | 1.4 (0.6, 3.6) | 3.6 (1.4, 12.6) |
| 2 components replaced | 0.9 (0.4, 1.9) | 1.1 (0.5, 2.1) |
| Comorbidities | ||
| Diabetes | 0.7 (0.3, 1.7) | 1.8 (0.9, 3.7) |
| Hypertension | 4.0 (1.3, 17.1) | 0.7 (0.3, 1.6) |
| Current smoker | 0.0 (0.0, 10073709.9) | 1.4 (0.5, 3.7) |
| History of severe COPD | 2.6 (0.9, 6.4) | 1.3 (0.4, 3.3) |
| Congestive heart failure | 1.5 (0.1, 8.7) | 1.0 (0.1, 6.3) |
| Bleeding disorder | 1.7 (0.3, 6.1) | 0.6 (0.0, 2.9) |
| Hematocrit <35 % | 1.0 (0.3, 2.4) | 4.3 (2.1, 8.6) |
4 Discussion
In our study, patients with diabetes who underwent aseptic revision TSA had similar rates of 30-day major complications as patients without diabetes. They also had similar rates of prolonged hospital stay. Hypertension was associated with increased likelihood of 30-day major complications among all patients, while ASA > II and hematocrit <35 % were associated with increased likelihood of prolonged hospital stay. In our study, we found that patients with and without DM did not exhibit significantly different 30-day complication rates after aseptic revision TSA surgery. Similarly, patients with DM did not experience a difference in prolonged hospital stays compared to patients without DM. These findings challenge previous assumptions regarding the impact of diabetes on postoperative outcomes in revision TSA patients.
Previous studies have also characterized the risk of postoperative complications in patients with DM compared to those without DM. One such study, a 2017 retrospective cohort study of 5918 patients, used the NSQIP database to analyze postoperative complications in patients with noninsulin-dependent DM, insulin-dependent DM, and no DM in the first 30 days after TSA.9 The study revealed that the odds of experiencing one or more complications were greater among patients with insulin-dependent DM than among non-diabetic patients, but patients with DM had no greater risk of major complications. Moreover, the authors reported that non-insulin-dependent DM patients did not differ from the non-diabetic group in any study end point metrics. Although our paper did not delineate between insulin- and non-insulin-dependent patients, our findings align with the 2017 study in that patients with diabetes, regardless of insulin dependence, who underwent aseptic revision TSA demonstrated rates of 30-day major complications that were comparable to those without diabetes.
Conversely, our finding of similar rates of prolonged hospital stay in patients with DM differs from findings in previous studies. A 2017 retrospective cohort study that examined 44,050 patients who underwent elective TSA demonstrated that a diagnosis of DM was associated with significantly increased length of hospital stay.10 The authors also noted that patients with DM were significantly more likely than those without DM to have 2 or more in-hospital complications. Another retrospective cohort study also demonstrated that uncontrolled diabetes was a risk factor for prolonged length of stay.11
Additionally, our finding that hypertension was associated with increased odds of 30-day complications after revision TSA mirrors previous literature demonstrating the risks of these comorbidities. Studies characterizing the association between hypertension and complication status after TSA are limited in number. However, a study by Fassler et al. demonstrated metabolic syndrome to be a significant risk factor for numerous postoperative complications, including postoperative pneumonia, renal insufficiency, acute renal failure, myocardial infarction, and longer length of stay.12
Accurately accounting for the impact of comorbidities on complications, readmissions, and increased length of stay is critical as the movement toward bundled care continues. This is particularly important given that physician reimbursement rates for performing TSA are decreasing,13 and profit margins for TSA are lower than those of hip or knee arthroplasty.14 As policymakers and payers construct bundles, it behooves surgeons to identify which subsets of their patients carry elevated risk. Many research studies have started to investigate potential risk factors in the primary shoulder arthroplasty realm; thus, we sought to study a critical comorbidity in the revision setting. Readmission and revision costs are exorbitantly high, challenging the notion of “cost containment”.15,16 Previous studies have identified several risk factors for readmission.17 Farronato et al. even used a bundled-payment TSA population to identify patients at risk of 90-day readmission in the hope of directing appropriate risk-adjusted reimbursement.18 Another study detailed an accurate comorbidity risk score to predict short-term complications.19 Based on our results, we submit that patients with anemia or hypertension who are undergoing revision TSA might warrant additional reimbursement in a bundled care model because of the increased likelihood of prolonged length of stay (hematocrit <35 %) or medical complications (hypertension). Notably, while DM may be thought to contribute risk, our findings suggest that these patient populations may not require additional reimbursement for aseptic revision TSA simply because of their DM status alone.
We acknowledge several limitations of the current study. The NSQIP database tracks patient data up to 30 days postoperatively, but it does not record complications that occurred beyond 30 days. Typical limitations specific to database studies also apply to this study, including coding issues that may skew data. One CPT code encompasses both anatomic and reverse TSA, so we were unable to determine complication rates for these 2 surgeries separately. Moreover, NSQIP does not provide hemoglobin A1C lab values, so there was no way to differentiate between those with controlled and uncontrolled diabetes, which some previous studies have done. Lastly, our study became underpowered when we attempted to split the DM cohort into type 1 and type 2 diabetes cohorts. One strength of our study was the use of a large database, which allowed for longitudinal data from 2006 to 2021. The NSQIP database also allowed us to include only aseptic revision TSA; previous studies have been unable to isolate this variable.
While our study contributes valuable insights into the outcomes of patients undergoing aseptic revision TSA, there are several avenues for future research. Understanding the heterogeneity within the diabetic population is essential, and further studies should delineate any differences in the likelihood of postoperative complications among patients with type 1 versus type 2 diabetes. Additionally, investigating the costs associated with caring for patients with DM and examining changes in physician behavior when managing DM patients undergoing revision TSA would provide valuable insights into optimizing patient care and healthcare resource allocation.
By addressing these research gaps, we can enhance our understanding of the factors influencing outcomes in revision TSA patients, ultimately improving patient management strategies and healthcare delivery in this population.
Guardian/patient's consent
No guardian/patient's consent was required or sought, as this was a retrospective database study using only deidentified patient information.
Institutional review board/ethical committee approval
Approval from an ethical review committee was not required for this study, as it used data from a national deidentified database.
Author contribution
All authors are involved in Conceptualization, Investigation,Methodology, Writing – original draft, Supervision, Project administration,Writing – review & editing.
Funding statement
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
References
- Prevalence of shoulder arthroplasty in the United States and the increasing burden of revision shoulder arthroplasty. JB JS Open Access. 2021;6
- [Google Scholar]
- Postoperative adverse events in patients with diabetes undergoing orthopedic and general surgery. Medicine (Baltim). 2019;98
- [Google Scholar]
- Diabetes and its negative impact on outcomes in orthopaedic surgery. World J Orthop. 2015;6:331-339.
- [Google Scholar]
- Perioperative hyperglycemia management: an update. Anesthesiology. 2017;126:547-560.
- [Google Scholar]
- Effect of glycaemic control on complications following cardiac surgery: literature review. J Cardiothorac Surg. 2018;13:10.
- [Google Scholar]
- The impact of insulin dependence on short-term postoperative complications in diabetic patients undergoing total shoulder arthroplasty. J Shoulder Elbow Surg. 2017;26:2091-2096.
- [Google Scholar]
- Impact of diabetes on perioperative complications in patients undergoing elective total shoulder arthroplasty. Bull Hosp Jt Dis. 2013;75:173-179.
- [Google Scholar]
- Modifiable risk factors increase length of stay and 90-day cost of care after shoulder arthroplasty. J Shoulder Elbow Surg. 2022;31:2-7.
- [Google Scholar]
- Components of metabolic syndrome as significant risk factors for postoperative complications following total shoulder arthroplasty: hypertension, diabetes, and obesity. JSES Int. 2024;8:141-146.
- [Google Scholar]
- Recent trends in Medicare utilization and surgeon reimbursement for shoulder arthroplasty. J Shoulder Elbow Surg. 2021;30:120-126.
- [Google Scholar]
- Variation in the profit margin for different types of total joint arthroplasty. J Bone Joint Surg Am. 2022;104:459-464.
- [Google Scholar]
- Prediction of total healthcare cost following total shoulder arthroplasty utilizing machine learning. J Shoulder Elbow Surg. 2022;31:2449-2456.
- [Google Scholar]
- Cost analysis and complication profile of primary shoulder arthroplasty at a high-volume institution. J Shoulder Elbow Surg. 2020;29:1337-1345.
- [Google Scholar]
- Characteristics and risk factors for 90-day readmission following shoulder arthroplasty. J Shoulder Elbow Surg. 2022;31:324-332.
- [Google Scholar]
- Effects of patient comorbidities and demographics on episode-of-care costs following total shoulder arthroplasty. J Am Acad Orthop Surg. 2023;31:451-457.
- [Google Scholar]
- A novel comorbidity risk score for predicting postoperative 30-day complications in total shoulder arthroplasty and elucidation of potential racial disparities. JSES Int. 2022;6:867-873.
- [Google Scholar]