Patients with Cirrhosis Have Significantly Increased Risk of PJI and All-Cause Revision Following THA.
Methods
Summary of included patients
Initial data query yielded a total of 332,657 cases in 281,979 patients. Revision cases where the corresponding primary case was not found in the database (e.g., because the primary occurred prior to the study period or outside of New York state) were excluded (15,441 cases in 12,763 patients). Additional cases with illogical or inconsistent data, likely to be the result of medical coding errors, were removed as necessary (e.g., patients greater than 2 primary cases or more than 1 primary case of the same laterality). Patients who underwent THA for fracture were excluded. A final cohort of 242,962 primary cases in 209,439 patients were included in the analysis. Among patients in the final cohort, a total of 126,417 primary cases in 111,553 patients had a minimum of 2 years postoperative follow-up.
Results
Demographics
| Characteristic | Cirrhosis | p-value2 | |
|---|---|---|---|
| Cirrhosis, N = 6651 | No Cirrhosis, N = 242,2971 | ||
| Indication | <0.001 | ||
| OA | 374 (56.2%) | 162,194 (66.9%) | |
| AVN | 132 (19.8%) | 13,628 (5.6%) | |
| Inflammatory arthritis | 24 (3.6%) | 6,541 (2.7%) | |
| Unspecified | 135 (20.3%) | 59,934 (24.7%) | |
| Age (years) | 62.6 +/- 10.5 | 64.4 +/- 11.5 | <0.001 |
| Sex | <0.001 | ||
| Female | 311 (46.8%) | 131,262 (54.2%) | |
| Male | 354 (53.2%) | 111,032 (45.8%) | |
| Race | 0.002 | ||
| White | 512 (77.0%) | 196,911 (81.3%) | |
| Black | 66 (9.9%) | 18,974 (7.8%) | |
| Hispanic | 41 (6.2%) | 11,323 (4.7%) | |
| Asian | 8 (1.2%) | 2,048 (0.8%) | |
| Native American | 5 (0.8%) | 438 (0.2%) | |
| Other or Unknown | 33 (5.0%) | 12,603 (5.2%) | |
| Elixhauser score | 13.0 +/- 5.4 | 0.6 +/- 4.2 | <0.001 |
| Insurance | <0.001 | ||
| Private | 289 (43.5%) | 123,225 (50.9%) | |
| Medicare | 308 (46.3%) | 105,296 (43.5%) | |
| Medicaid | 57 (8.6%) | 7,645 (3.2%) | |
| Worker's Compensation | 3 (0.5%) | 3,284 (1.4%) | |
| Other | 8 (1.2%) | 2,845 (1.2%) | |
| Follow-up duration (years) | 3.0 +/- 2.8 | 3.0 +/- 2.9 | 0.824 |
| 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) | |||
| Characteristic | Cirrhosis | p-value2 | |
|---|---|---|---|
| Cirrhosis, N = 3721 | No Cirrhosis, N = 126,0451 | ||
| Indication | <0.001 | ||
| OA | 187 (50.3%) | 66,222 (52.5%) | |
| AVN | 68 (18.3%) | 7,331 (5.8%) | |
| Inflammatory arthritis | 12 (3.2%) | 3,596 (2.9%) | |
| Unspecified | 105 (28.2%) | 48,896 (38.8%) | |
| Age (years) | 61.9 +/- 10.5 | 64.9 +/- 11.5 | <0.001 |
| Sex | <0.001 | ||
| Female | 177 (47.6%) | 70,675 (56.1%) | |
| Male | 195 (52.4%) | 55,370 (43.9%) | |
| Race | 0.002 | ||
| White | 289 (77.7%) | 103,357 (82.0%) | |
| Black | 41 (11.0%) | 9,711 (7.7%) | |
| Hispanic | 21 (5.6%) | 5,759 (4.6%) | |
| Asian | 6 (1.6%) | 694 (0.6%) | |
| Native American | 3 (0.8%) | 199 (0.2%) | |
| Other or Unknown | 12 (3.2%) | 6,325 (5.0%) | |
| Elixhauser score | 12.7 +/- 5.1 | 0.8 +/- 4.2 | <0.001 |
| Insurance | <0.001 | ||
| Private | 167 (44.9%) | 62,161 (49.3%) | |
| Medicare | 164 (44.1%) | 57,465 (45.6%) | |
| Medicaid | 34 (9.1%) | 3,409 (2.7%) | |
| Worker's Compensation | 2 (0.5%) | 1,650 (1.3%) | |
| Other | 5 (1.3%) | 1,358 (1.1%) | |
| Follow-up duration (years) | 4.9 +/- 2.3 | 5.4 +/- 2.3 | <0.001 |
| 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) | |||
Outcomes
Patients with minimum 2 year follow-up
| Characteristic | Cirrhosis | p-value2 | |
|---|---|---|---|
| Cirrhosis, N = 3721 | No Cirrhosis, N = 126,0451 | ||
| Length of hospital stay (days) | 3.6 +/- 2.4 | 3.1 +/- 1.8 | <0.001 |
| Discharge disposition | 0.704 | ||
| Home | 237 (63.7%) | 78,381 (62.2%) | |
| Skilled Nursing Facility | 94 (25.3%) | 33,883 (26.9%) | |
| Inpatient Rehabilitation | 38 (10.2%) | 13,088 (10.4%) | |
| Transfer | 3 (0.8%) | 557 (0.4%) | |
| Against Medical Advice | 0 (0.0%) | 33 (0.0%) | |
| Discharged to Court | 0 (0.0%) | 95 (0.1%) | |
| Hospice | 0 (0.0%) | 8 (0.0%) | |
| Expired | 0 (0.0%) | 0 (0.0%) | |
| ED presentation within 3 months | 56 (15.1%) | 13,279 (10.5%) | 0.005 |
| Readmission within 3 months | 57 (15.3%) | 8,923 (7.1%) | <0.001 |
| Readmission within 12 months | 93 (25.0%) | 15,051 (11.9%) | <0.001 |
| All-cause revision | 20 (5.4%) | 3,552 (2.8%) | 0.003 |
| Revision for PJI | 10 (2.7%) | 1,114 (0.9%) | 0.002 |
| Time to first revision (years) | 2.8 +/- 3.0 | 3.0 +/- 2.7 | 0.704 |
| Reoperation within 3 months of primary | 5 (1.3%) | 737 (0.6%) | 0.070 |
| Reoperation within 12 months of primary | 9 (2.4%) | 1,179 (0.9%) | 0.009 |
| 1 Mean +/- SD; n (%) | |||
| 2 Wilcoxon rank sum test; Fisher's Exact Test for Count Data with simulated p-value (based on 2000 replicates); Pearson's Chi-squared test; Fisher's exact test | |||
Perioperative outcomes
| Characteristic | Cirrhosis | p-value2 | |
|---|---|---|---|
| Cirrhosis, N = 6651 | No Cirrhosis, N = 242,2971 | ||
| ED presentation within 3 months | 105 (15.8%) | 21,389 (8.8%) | <0.001 |
| Readmission within 3 months | 102 (15.3%) | 14,023 (5.8%) | <0.001 |
| Readmission within 12 months | 164 (24.7%) | 23,318 (9.6%) | <0.001 |
| Mortality within 3 months | 12 (1.8%) | 376 (0.2%) | <0.001 |
| Mortality within 12 months | 25 (3.8%) | 942 (0.4%) | <0.001 |
| 1 n (%) | |||
| 2 Pearson's Chi-squared test; Fisher's exact test | |||
Survival analysis
Kaplan-Meier plots
Figure: Kaplan-Meier plots plots demonstrating probability of PJI-free survival (A) and overall revision-free survival (B) over time between groups. Patients with cirrhosis demonstrated a significantly lower PJI-free survival (< 0.001) and significantly lower all-cause revision-free survival (< 0.001) following THA. Note: y-axis is limited to 90-100% survival for PJI-free survival and 75-100% survival for revision-free survival to visualize detail of curves.
Cox proportional hazards
| Characteristic | Time to PJI | Time to All-Cause Revision | ||||
|---|---|---|---|---|---|---|
| HR1 | 95% CI1 | p-value | HR1 | 95% CI1 | p-value | |
| Cirrhosis | ||||||
| No Cirrhosis | — | — | — | — | ||
| Cirrhosis | 2.24 | 1.30, 3.87 | 0.004 | 1.73 | 1.18, 2.54 | 0.005 |
| Age (years) | 0.99 | 0.98, 0.99 | <0.001 | 1.0 | 0.99, 1.00 | <0.001 |
| Sex | ||||||
| Female | — | — | — | — | ||
| Male | 1.25 | 1.13, 1.37 | <0.001 | 0.95 | 0.90, 1.00 | 0.062 |
| Elixhauser score | 1.01 | 1.00, 1.02 | 0.087 | 1.00 | 1.00, 1.01 | 0.234 |
| Operative indication | ||||||
| Unspecified | — | — | — | — | ||
| OA | 1.47 | 1.31, 1.65 | <0.001 | 1.69 | 1.58, 1.81 | <0.001 |
| AVN | 1.79 | 1.47, 2.17 | <0.001 | 1.94 | 1.73, 2.17 | <0.001 |
| Inflammatory arthritis | 2.42 | 1.91, 3.06 | <0.001 | 2.40 | 2.09, 2.76 | <0.001 |
| Obesity | ||||||
| No | — | — | — | — | ||
| Yes | 1.77 | 1.59, 1.98 | <0.001 | 1.32 | 1.23, 1.41 | <0.001 |
| DM | ||||||
| No | — | — | — | — | ||
| Yes | 1.15 | 1.01, 1.32 | 0.040 | 0.96 | 0.89, 1.05 | 0.404 |
| History of smoking | ||||||
| No | — | — | — | — | ||
| Yes | 1.27 | 1.10, 1.48 | 0.001 | 1.25 | 1.14, 1.36 | <0.001 |
| 1 HR = Hazard Ratio, CI = Confidence Interval | ||||||
Table: Cox proportional hazards models evaluating the effect of cirrhosis on time to PJI (left) and time to all-cause revision (right). When controlling for the covariates listed, the main effect of cirrhosis significantly increases the risk of PJI and all-cause revision.
GLM for risk of revision
| Characteristic | Odds of PJI | Odds of All-Cause Revision | ||||
|---|---|---|---|---|---|---|
| OR1 | 95% CI1 | p-value | OR1 | 95% CI1 | p-value | |
| Cirrhosis | ||||||
| No Cirrhosis | — | — | — | — | ||
| Cirrhosis | 2.37 | 1.15, 4.35 | 0.010 | 1.93 | 1.18, 2.99 | 0.006 |
| Age (years) | 0.98 | 0.98, 0.99 | <0.001 | 0.99 | 0.99, 0.99 | <0.001 |
| Sex | ||||||
| Female | — | — | — | — | ||
| Male | 1.24 | 1.10, 1.40 | <0.001 | 0.93 | 0.87, 0.99 | 0.034 |
| Elixhauser score | 1.01 | 0.99, 1.02 | 0.282 | 0.99 | 0.98, 1.00 | 0.155 |
| Operative indication | ||||||
| Unspecified | — | — | — | — | ||
| OA | 1.03 | 0.90, 1.18 | 0.657 | 1.13 | 1.05, 1.21 | 0.002 |
| AVN | 1.59 | 1.28, 1.97 | <0.001 | 1.41 | 1.23, 1.61 | <0.001 |
| Inflammatory arthritis | 2.02 | 1.51, 2.64 | <0.001 | 1.88 | 1.59, 2.21 | <0.001 |
| Obesity | ||||||
| No | — | — | — | — | ||
| Yes | 1.64 | 1.43, 1.88 | <0.001 | 1.09 | 1.01, 1.19 | 0.034 |
| DM | ||||||
| No | — | — | — | — | ||
| Yes | 1.19 | 1.01, 1.40 | 0.038 | 1.0 | 0.90, 1.10 | 0.919 |
| Length of hospital stay (days) | 1.04 | 1.02, 1.06 | <0.001 | 1.04 | 1.03, 1.05 | <0.001 |
| 1 OR = Odds Ratio, CI = Confidence Interval | ||||||
Table. Logistic regression models demonstrating the odds of revision for PJI (A) and all-cause revision (B) for patients undergoing THA with at least 2 years of follow-up. Patients with cirrhosis have a significantly increased risk of all-cause revision or infection-related revision compared to patients without cirrhosis when controlling for underlying covariates.