| Characteristic | IBD | p-value2 | |
|---|---|---|---|
| IBD, N = 1,1651 | No IBD, N = 133,6361 | ||
| Indication | <0.001 | ||
| OA | 554 (47.6%) | 66,002 (49.4%) | |
| Fracture | 93 (8.0%) | 7,063 (5.3%) | |
| AVN | 120 (10.3%) | 7,400 (5.5%) | |
| Inflammatory arthritis | 51 (4.4%) | 3,609 (2.7%) | |
| Unspecified | 347 (29.8%) | 49,562 (37.1%) | |
| Age (years) | 64.5 +/- 12.2 | 65.2 +/- 11.6 | 0.068 |
| Sex | 0.025 | ||
| Female | 702 (60.3%) | 76,161 (57.0%) | |
| Male | 463 (39.7%) | 57,475 (43.0%) | |
| Race | <0.001 | ||
| White | 1,039 (89.2%) | 109,530 (82.0%) | |
| Black | 42 (3.6%) | 10,129 (7.6%) | |
| Hispanic | 37 (3.2%) | 6,173 (4.6%) | |
| Asian | 5 (0.4%) | 821 (0.6%) | |
| Native American | 2 (0.2%) | 219 (0.2%) | |
| Other or Unknown | 40 (3.4%) | 6,764 (5.1%) | |
| Elixhauser score | 1.4 +/- 4.8 | 1.0 +/- 4.4 | 0.005 |
| Insurance | 0.013 | ||
| Private | 510 (43.8%) | 64,422 (48.2%) | |
| Medicare | 600 (51.5%) | 62,221 (46.6%) | |
| Medicaid | 28 (2.4%) | 3,714 (2.8%) | |
| Worker's Compensation | 12 (1.0%) | 1,813 (1.4%) | |
| Other | 15 (1.3%) | 1,464 (1.1%) | |
| Follow-up duration (years) | 5.3 +/- 2.3 | 5.3 +/- 2.3 | 0.495 |
| 1 n (%); Mean +/- SD | |||
| 2 Pearson’s Chi-squared test; Welch Two Sample t-test; Fisher’s Exact Test for Count Data with simulated p-value (based on 2000 replicates) | |||
THA and IBD
Tidy data
Load packages
Load data
Clean data
Methods
Data collection
The purpose of this retrospective cohort study was to evaluate the effect of IBD on outcomes following hip arthroplasty. The Statewide Planning and Research Cooperative System (SPARCS), a comprehensive database containing details related to all inpatient and outpatient encounters in New York state, was used to compare postoperative outcomes between patients with a diagnosis of IBD at the time of the arthroplasty and patients without IBD. The SPARCS database was queried for all patients who underwent primary hip arthroplasty between January 2010 and December 2020 using a comprehensive list of ICD9 procedure codes, ICD10 procedure codes, and CPT codes. The principal diagnosis associated with the procedure was considered to be the surgical indication. For all identified patients, each subsequent encounter in the SPARCS database was assessed for CPT and ICD codes associated with postoperative complications and any return to the operating room for revision arthroplasty. Patients who developed PJI were identified by the presence of a revision arthroplasty procedure code and an associated principal diagnosis related to infection. For patients who underwent both left and right primary hip arthroplasty during the study period, any subsequent complications or revision procedures were matched to the associated primary procedure based on the laterality specified by the procedure and diagnosis codes. For cases in which the laterality of a complication or revision procedure was not specified (i.e., the associated code was for “unspecified laterality”), the complication was assumed to be associated with the more recent of the two primary arthroplasty procedures. After the cohort of eligible patients was identified, diagnosis codes related to IBD were used to determine the group of patients who had a diagnosis of IBD at the time of the hip arthroplasty. Patients who were not associated with an IBD diagnosis at the time of arthroplasty but were diagnosed with IBD later in the study period were included in the control (non-IBD) group.
Summary of included patients
Initial data query yielded a total of 323,690 hip arthroplasty cases in 281,988 patients in New York state during the study period. For patients who underwent revision arthroplasty during the study period for a primary arthroplasty that was performed prior to the study period or performed outside of New York (and therefore was not found in the SPARCS database), these cases were excluded (10,775 cases in 8,466 patients). Furthermore, cases with illogical or inconsistent data, such as patients greater than 2 primary hip arthroplasty cases or more than 1 primary arthroplasty of the same laterality, were assumed to be the result of medical coding errors and were excluded from the analysis (159 cases). For this analysis, patients who underwent hemiarthroplasty were excluded (48,612 cases in 47,140 patients). Patients with less than 2 years of postoperative follow up were not included in the analysis (129,358 cases in 119,144 patients). The final cohort consisted of 134,801 THA cases in 119,094 patients (Figure X).
Analysis
The primary outcome of this study was all-cause revision hip arthroplasty. The proportion of patients with IBD who required revision arthroplasty for any reason was compared to the proportion of patients without IBD requiring revision arthroplasty. Secondary outcomes included the difference between groups in infection-related revision arthroplasty, revision-free implant survival, and short-term postoperative emergency department presentations and hospital readmissions.
Comparisons between groups were performed using Chi-squared tests, rank sum tests, or t-tests as appropriate. To further assess the effect of IBD for patients undergoing hip arthroplasty, a logistic regression model was used to evaluate the odds of all-cause revision when controlling for the effects of age, gender, overall health status as approximated by Elixhauser score, operative indication, obesity, and DM. A separate logistic regression model involving the same independent variables was used to evaluate the effect of IBD on revision arthroplasty specifically for PJI. Comparison of revision-free survival between groups was assessed using both the log-rank test and a Cox proportional hazards model to account for covariates. For all comparisons, p-values less than 0.05 were considered statistically significant. Statistical analysis was performed using R (R Foundation for Statistical Computing, Vienna, Austria).
Results
Demographics
Overall, 119,094 patients underwent 134,801 primary hip arthroplasties and were included in the analysis. Of those patients, 1,165 patients (0.9%) had a diagnosis of IBD at the time of surgery. There was no difference in age at the time of THA between the IBD group and the non-IBD group (64.5 +/- 12.2 years versus 65.2 +/- 11.6 years, p = 0.068). Patients in the IBD group had more medical comorbidities based on Elixhauser score (1.4 +/- 4.8 versus 1.0 +/- 4.4, p = 0.005). The IBD group had a greater proportion of patients undergoing hip arthroplasty for fracture (8% versus 5.3%), AVN (10.3% versus 5.5%), and inflammatory arthritis (4.4% versus 2.7%) compared to the control group (< 0.001). Mean follow-up duration for all patients was 5.3 +/- 2.3 years and there was no difference between groups (Table 1).
Outcomes
| Characteristic | IBD | p-value2 | |
|---|---|---|---|
| IBD, N = 1,1651 | No IBD, N = 133,6361 | ||
| Length of hospital stay (days) | 3.5 +/- 2.3 | 3.2 +/- 2.3 | <0.001 |
| Discharge disposition | 0.877 | ||
| Home | 690 (59.2%) | 81,085 (60.7%) | |
| Skilled Nursing Facility | 341 (29.3%) | 37,237 (27.9%) | |
| Inpatient Rehabilitation | 129 (11.1%) | 14,534 (10.9%) | |
| Transfer | 5 (0.4%) | 629 (0.5%) | |
| Against Medical Advice | 0 (0.0%) | 38 (0.0%) | |
| Discharged to Court | 0 (0.0%) | 104 (0.1%) | |
| Hospice | 0 (0.0%) | 9 (0.0%) | |
| Expired | 0 (0.0%) | 0 (0.0%) | |
| ED presentation within 3 months | 157 (13.5%) | 14,811 (11.1%) | 0.010 |
| Readmission within 3 months | 123 (10.6%) | 10,258 (7.7%) | <0.001 |
| Readmission within 12 months | 202 (17.3%) | 17,049 (12.8%) | <0.001 |
| All-cause revision | 42 (3.6%) | 3,814 (2.9%) | 0.126 |
| Revision for PJI | 9 (0.8%) | 1,240 (0.9%) | 0.582 |
| Time to first revision (years) | 3.1 (0.8, 5.6) | 2.2 (0.4, 4.8) | 0.153 |
| Reoperation within 3 months of primary | 9 (0.8%) | 833 (0.6%) | 0.520 |
| Reoperation within 12 months of primary | 12 (1.0%) | 1,300 (1.0%) | 0.843 |
| 1 Mean +/- SD; n (%); Median (IQR) | |||
| 2 Welch Two Sample t-test; Fisher’s Exact Test for Count Data with simulated p-value (based on 2000 replicates); Pearson’s Chi-squared test; Wilcoxon rank sum test | |||
GLM for risk of revision
| Characteristic | All-Cause Revision | Revision for PJI | ||||
|---|---|---|---|---|---|---|
| OR1 | 95% CI1 | p-value | OR1 | 95% CI1 | p-value | |
| IBD | 0.83 | 0.62, 1.15 | 0.242 | 1.27 | 0.70, 2.66 | 0.474 |
| Age (years) | 0.99 | 0.98, 0.99 | <0.001 | 0.98 | 0.97, 0.98 | <0.001 |
| Sex (male) | 0.95 | 0.89, 1.01 | 0.104 | 1.21 | 1.08, 1.35 | 0.001 |
| Elixhauser score | 1.01 | 1.00, 1.01 | 0.164 | 1.02 | 1.01, 1.04 | <0.001 |
| Operative indication | ||||||
| Unspecified | — | — | — | — | ||
| OA | 1.10 | 1.02, 1.18 | 0.011 | 0.97 | 0.85, 1.10 | 0.627 |
| Fracture | 1.91 | 1.67, 2.17 | <0.001 | 1.81 | 1.43, 2.27 | <0.001 |
| Inflammatory arthritis | 1.79 | 1.51, 2.11 | <0.001 | 1.91 | 1.44, 2.49 | <0.001 |
| AVN | 1.37 | 1.19, 1.56 | <0.001 | 1.49 | 1.20, 1.85 | <0.001 |
| Obesity | 1.14 | 1.05, 1.24 | 0.001 | 1.67 | 1.46, 1.89 | <0.001 |
| DM | 1.01 | 0.91, 1.11 | 0.864 | 1.22 | 1.05, 1.43 | 0.011 |
| 1 OR = Odds Ratio, CI = Confidence Interval | ||||||
:::
Table: Logistic regression models demonstrating the odds of all-cause revision following THA (left) and the odds of infection-related revision (right). A diagnosis of IBD at the time of THA was not associated with an increased risk of all-cause revision or infection-related revision when controlling for the effects of age, gender, Elixhauser score, operative indication, obesity, and DM.
Following THA, patients in the IBD group had a slightly longer length of stay (3.5 +/- 2.3 days versus 3.2 +/- 2.3 days, p < 0.001). There was no difference between groups with respect to discharge disposition. A greater proportion of patients in the IBD group presented to the ED within 3 months of THA (13.5% versus 11.1%, p = 0.01), were re-admitted for any reason within 3 months (10.6% versus 7.7%, p < 0.001), and were re-admitted within 12 months (17.3% versus 12.8%, p < 0.001).
There was no difference in the proportion of patients who underwent revision hip arthroplasty for any reason between groups (3.6% versus 2.9%, p = 0.126). The association between IBD and all-cause revision was further evaluated using a logistic regression model. Having IBD at the time of THA demonstrated no significant effect on the odds of requiring revision (OR = 0.8, 95% CI [0.6, 1.1], p = 0.242) when controlling for the effects of age, gender, Elixhauser score, operative indication, obesity, and DM (Table 3, left).
Similarly, there was no difference in the rate of revision for PJI between groups (0.8% versus 0.9%, p = 0.582). Controlling for the same covariates again demonstrated that IBD was not associated with an increased odds of requiring revision for PJI (Table 3, right). The median time to all-cause revision was 2.2 years after primary THA, and there was no difference between groups (Table X).
Survival
Figure: Kaplan-Meier plot demonstrating the probability of revision-free primary THA survival. Based on log-rank test comparing the survival curves between patients with IBD at the time of THA and those without IBD at the time of THA, patients in the IBD group did not have a significantly different survival time. Note: y-axis is limited to 90-100% survival.
| Characteristic | HR1 | 95% CI1 | p-value |
|---|---|---|---|
| IBD | 0.84 | 0.62, 1.14 | 0.256 |
| Age (years) | 0.99 | 0.99, 0.99 | <0.001 |
| Sex | 0.96 | 0.90, 1.03 | 0.224 |
| Elixhauser score | 1.01 | 1.00, 1.02 | 0.034 |
| Operative indication | |||
| Unspecified | — | — | |
| OA | 1.43 | 1.33, 1.54 | <0.001 |
| Fracture | 2.61 | 2.30, 2.97 | <0.001 |
| Inflammatory arthritis | 2.05 | 1.74, 2.41 | <0.001 |
| AVN | 1.60 | 1.40, 1.83 | <0.001 |
| Obesity | 1.32 | 1.22, 1.42 | <0.001 |
| DM | 0.97 | 0.88, 1.07 | 0.577 |
| 1 HR = Hazard Ratio, CI = Confidence Interval | |||
There was no difference between groups with respect to all-cause revision-free survival following THA (χ2 = 2.4, p = = 0.118). The effect of IBD on outcomes following THA was further evaluated using a multivariate Cox proportional hazards model. When controlling for known covariates, IBD demonstrated no effect on the risk of requiring revision hip arthroplasty (HR = 0.8, 95% CI [0.6, 1.1], p = 0.256, Table X).
Subanalysis, UC vs Crohns
| Characteristic | IBD Type | p-value2 | |
|---|---|---|---|
| Crohn’s, N = 5011 | UC, N = 6641 | ||
| Indication | 0.057 | ||
| OA | 215 (42.9%) | 339 (51.1%) | |
| Fracture | 41 (8.2%) | 52 (7.8%) | |
| AVN | 62 (12.4%) | 58 (8.7%) | |
| Inflammatory arthritis | 23 (4.6%) | 28 (4.2%) | |
| Unspecified | 160 (31.9%) | 187 (28.2%) | |
| Age (years) | 62.7 +/- 12.7 | 65.9 +/- 11.7 | <0.001 |
| Sex | 0.405 | ||
| Female | 295 (58.9%) | 407 (61.3%) | |
| Male | 206 (41.1%) | 257 (38.7%) | |
| Race | 0.576 | ||
| White | 446 (89.0%) | 593 (89.3%) | |
| Black | 22 (4.4%) | 20 (3.0%) | |
| Hispanic | 15 (3.0%) | 22 (3.3%) | |
| Asian | 3 (0.6%) | 2 (0.3%) | |
| Native American | 0 (0.0%) | 2 (0.3%) | |
| Other or Unknown | 15 (3.0%) | 25 (3.8%) | |
| Elixhauser score | 1.5 +/- 4.7 | 1.3 +/- 4.9 | 0.617 |
| Insurance | 0.027 | ||
| Private | 234 (46.7%) | 276 (41.6%) | |
| Medicare | 235 (46.9%) | 365 (55.0%) | |
| Medicaid | 17 (3.4%) | 11 (1.7%) | |
| Worker's Compensation | 6 (1.2%) | 6 (0.9%) | |
| Other | 9 (1.8%) | 6 (0.9%) | |
| Follow-up duration (years) | 5.4 +/- 2.3 | 5.2 +/- 2.3 | 0.160 |
| 1 n (%); Mean +/- SD | |||
| 2 Pearson’s Chi-squared test; Welch Two Sample t-test; Fisher’s Exact Test for Count Data with simulated p-value (based on 2000 replicates) | |||
Among the 1,165 patients in the IBD group, 507 patients (43.5%) had Crohn’s disease and 664 patients (57.0%) had ulcerative colitis. 6 patients had diagnoses of both Crohn’s disease and UC, and for the purposes of this analysis were included in the UC group. Patients with UC were slightly older at the time of THA and were more likely to have Medicare insurance, but otherwise the two subgroups based on IBD type did not differ with respect to baseline demographics.
| Characteristic | Group | p-value2 | ||
|---|---|---|---|---|
| Crohn’s, N = 5011 | UC, N = 6641 | No IBD, N = 133,6361 | ||
| Length of hospital stay (days) | <0.001 | |||
| Mean +/- SD | 3.6 +/- 2.5 | 3.4 +/- 2.1 | 3.2 +/- 2.3 | |
| Median (IQR) | 3.0 (2.0, 4.0) | 3.0 (2.0, 4.0) | 3.0 (2.0, 3.0) | |
| Discharge disposition | 0.753 | |||
| Home | 293 (58.5%) | 397 (59.8%) | 81,085 (60.7%) | |
| Skilled Nursing Facility | 154 (30.7%) | 187 (28.2%) | 37,237 (27.9%) | |
| Inpatient Rehabilitation | 50 (10.0%) | 79 (11.9%) | 14,534 (10.9%) | |
| Transfer | 4 (0.8%) | 1 (0.2%) | 629 (0.5%) | |
| Against Medical Advice | 0 (0.0%) | 0 (0.0%) | 38 (0.0%) | |
| Discharged to Court | 0 (0.0%) | 0 (0.0%) | 104 (0.1%) | |
| Hospice | 0 (0.0%) | 0 (0.0%) | 9 (0.0%) | |
| Expired | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | |
| ED presentation within 3 months | 77 (15.4%) | 80 (12.0%) | 14,811 (11.1%) | 0.007 |
| Readmission within 3 months | 68 (13.6%) | 55 (8.3%) | 10,258 (7.7%) | <0.001 |
| Readmission within 12 months | 102 (20.4%) | 100 (15.1%) | 17,049 (12.8%) | <0.001 |
| All-cause revision | 21 (4.2%) | 21 (3.2%) | 3,814 (2.9%) | 0.180 |
| Revision for PJI | 5 (1.0%) | 4 (0.6%) | 1,240 (0.9%) | 0.721 |
| Time to first revision (years) | 3.3 (0.2, 6.0) | 2.9 (1.0, 5.3) | 2.2 (0.4, 4.8) | 0.356 |
| Reoperation within 3 months of primary | 6 (1.2%) | 3 (0.5%) | 833 (0.6%) | 0.223 |
| Reoperation within 12 months of primary | 6 (1.2%) | 6 (0.9%) | 1,300 (1.0%) | 0.792 |
| 1 n (%); Median (IQR) | ||||
| 2 One-way ANOVA; Fisher’s Exact Test for Count Data with simulated p-value (based on 2000 replicates); Pearson’s Chi-squared test; Kruskal-Wallis rank sum test | ||||
Pairwise comparison demonstrated that patients with Crohn’s had a longer LOS than patients without IBD (p = 0.008). Similarly, patients with UC had a longer LOS than patients without IBD (p = 0.007). There was no difference in LOS between patients with Crohn’s and patients with UC (p = 0.785).
Patients with Crohn’s were more likely to present to the ED within 3 months of THA (p = 0.003), more likely to be readmitted within 3 months (p < 0.001), and more likely to be readmitted within 12 months (p < 0.001) compared to patients without IBD. Patients with UC did not demonstrate a difference in 3 month ED presentation (p = 0.467), 3 month readmission (p = 0.608), or 12 month readmission (p = 0.086) compared to non-IBD patients.
When stratifying based on subtype of IBD, there was no difference in the risk of all-cause revision (p = 0.18) or PJI (p = 0.721) between groups.
| Characteristic | All-Cause Revision | Revision for PJI | ||||
|---|---|---|---|---|---|---|
| OR1 | 95% CI1 | p-value | OR1 | 95% CI1 | p-value | |
| Group | ||||||
| No IBD | — | — | — | — | ||
| Crohn's | 1.37 | 0.86, 2.07 | 0.161 | 0.97 | 0.34, 2.10 | 0.940 |
| UC | 1.07 | 0.67, 1.62 | 0.748 | 0.64 | 0.20, 1.49 | 0.370 |
| Age (years) | 0.99 | 0.98, 0.99 | <0.001 | 0.98 | 0.97, 0.98 | <0.001 |
| Sex (male) | 0.95 | 0.89, 1.01 | 0.104 | 1.21 | 1.08, 1.35 | 0.001 |
| Elixhauser score | 1.01 | 1.00, 1.01 | 0.165 | 1.02 | 1.01, 1.04 | <0.001 |
| Operative indication | ||||||
| Unspecified | — | — | — | — | ||
| OA | 1.10 | 1.02, 1.18 | 0.011 | 0.97 | 0.85, 1.10 | 0.629 |
| Fracture | 1.91 | 1.67, 2.17 | <0.001 | 1.81 | 1.43, 2.27 | <0.001 |
| Inflammatory arthritis | 1.79 | 1.51, 2.11 | <0.001 | 1.91 | 1.44, 2.49 | <0.001 |
| AVN | 1.37 | 1.19, 1.56 | <0.001 | 1.49 | 1.20, 1.85 | <0.001 |
| Obesity | 1.14 | 1.05, 1.24 | 0.001 | 1.67 | 1.46, 1.89 | <0.001 |
| DM | 1.01 | 0.91, 1.11 | 0.867 | 1.22 | 1.05, 1.43 | 0.011 |
| 1 OR = Odds Ratio, CI = Confidence Interval | ||||||
When stratifying based on type of IBD and comparing patients with no IBD to patients with Crohn’s and patients with UC, neither type of IBD was associated with an increased risk of all-cause revision or PJI.