Device infection is a complication of implantable cardioverter-defibrillator (ICD) therapy that significantly increases mortality. Risk factors associated with death and ICD infection are poorly understood. The purpose of this study was to identify patient characteristics associated with death after cardiovascular implantable electronic device (CIED) infection. This is a retrospective cohort study of 64,903 Medicare fee-for-service patients who received an ICD in 2007, including 1,855 with device infection. Long-term survival was significantly reduced with CIED infection (71.6% vs 85.0%, p <0.001). Regression analysis accounting for age, race, gender, and 28 co-morbidities identified only 2 patient characteristics associated with decreased long-term survival with CIED infection: female gender and human immunodeficiency virus/acquired immunodeficiency syndrome. In patients with CIED infection, women had substantially reduced long-term survival compared with men (67.3% vs 72.9%, p <0.02). The risk-adjusted hazard ratio for long-term mortality with device infection in women compared with that in men increased significantly from 0.86 (95% confidence interval [CI] 0.82 to 0.91) to 1.25 (95% CI 1.02 to 1.53), corresponding to a risk increase of >45%. Importantly, a substantial portion of this excess mortality occurred after the index admission for infection, when the hazard ratio for death in women compared with that in men increased from 0.86 (95% CI 0.82 to 0.91) to 1.20 (95% CI 0.96 to 1.51) with CIED infection, despite little gender difference in admission length of stay, disposition, and cost. In conclusion, women are significantly more likely than men to die with CIED infection. A substantial part of this excess mortality occurs after discharge. It will be important to identify and address the cause(s) of this gender difference in mortality.
Device infection is a complication of cardiovascular implantable electronic device (CIED) therapy that is associated with significant admission and long-term mortality. The demographic features and co-morbidities associated with increased mortality with CIED infection are not well understood. In this investigation, we identified patient characteristics associated with decreased survival with CIED infection in a large cohort of Medicare beneficiaries.
Methods
The study cohort was derived from the 100% Medicare Standard Analytic File (SAF) Limited Data Set version for inpatient admissions for the 2007 calendar year and consisted of all admissions that included a procedure for implantable cardioverter-defibrillator (ICD) generator implantation, replacement, or revision identified using the corresponding International Classification of Diseases Ninth Revision Clinical Modification (ICD-9-CM) procedure codes for ICDs (37.94, 37.96, and 37.98) or cardiac resynchronization therapy devices with defibrillator (00.51 and 00.54). Admissions that included only electrode implantation, replacement, or revision were not included because most CIED infections are treated with complete system explantation. Admissions including other major cardiac procedures were excluded to avoid including procedural mortality unrelated to the CIED infection ( Supplementary Table 1 ).
The study cohort was divided into admissions with and without a primary or secondary diagnosis of infection, identified using ICD-9-CM diagnosis codes for infection due to a cardiac device (996.61), infection due to a vascular device (996.62), endocarditis (421.0, 421.1, 421.9, 424.90, 424.91, and 424.99), bacteremia (790.7), septicemia (038.0, 038.2, 038.3, 038.10, 038.11, 038.19, 038.40, 038.41, 038.42, 038.43, 038.44, 038.49, 038.8, and 038.9), shock (785.50), cellulitis (682.8 and 682.9), or fever (780.6). Identification of ICD infection cases is discussed in detail in the Supplementary Data . Patients identified from these claim files using the encrypted beneficiary identifier were linked to the 2007 and 2008 Medicare Denominator Files to collect beneficiary date of death. Patient demographics were also identified from the Medicare claims file.
A total of 70,597 admissions with a qualifying ICD procedure were identified. Of these, 5,694 were excluded: 3,622 had other major cardiac procedures, 2,072 were additional admissions for patients with multiple admissions, were paid for by a Medicare Health Maintenance Organization or not associated with a Medicare payment, or had no Medicare Part A enrollment in the quarter of the index admission. If a patient had an ICD procedure admission with and another one without infection, the admission without infection was excluded.
The primary outcome of our study was long-term survival associated with ICD implantation, replacement, or revision procedures, defined as survival through the admission quarter and the subsequent 4 quarters (combination of admission and postadmission periods). Secondary outcomes were admission survival, defined as survival during the index hospitalization for the ICD procedure, and postadmission survival, defined as survival from discharge through the quarter of admission and the 4 quarters after the admission quarter. The unit of analysis for all 3 outcomes was the individual discharge.
The prevalence of specific demographic characteristics and co-morbidities in study subgroups were compared using a chi-square test. Long-term survival functions were estimated using Kaplan-Meier survival analysis and compared using a log-rank test. Kaplan-Meier mortality was derived as 1 Kaplan-Meier survival. A significance level of 0.05 was used for all tests. Statistical tests were performed using SAS, version 9.2, 2008 (SAS Institute Inc., Cary, North Carolina) and XLstat (Addinsoft, Montreal, Canada).
The rates of admission, postadmission, and long-term mortality were adjusted for age, gender, race or ethnicity, and a set of 28 co-morbidity measures associated with death in Medicare beneficiaries, originally derived by Elixhauser et al and validated for risk adjustment using administrative data. The individual patient status for this set of co-morbidities was collected from the administrative data using the Comorbidity Software (versions 3.2 and 3.3) from the Agency for Healthcare Research and Quality (Rockville, Maryland) that specifies ICD-9-CM codes corresponding to each patient co-morbidity. Peptic ulcer disease was omitted because it was rare in our study cohort. The primary condition screens, including for cardiac conditions, were not used because this study was narrowly defined to patients receiving ICD therapy.
Admission, postadmission, and long-term mortality were modeled using the binomial complementary log-log regression specification. Terms for an interaction between device infection and age, gender, race/ethnicity, and the 28 co-morbidities were included in the regression models for mortality. A discussion of the rationale for these model choices can be found in the Supplementary Data . For some analyses, certain co-morbidities were not present in either men or women and were dropped for that device. Risk-adjusted mortality hazard ratios (HRs), defined as mortality rate in women with device infection or mortality rate in men with device infection, for admission, postadmission, and long-term mortality were calculated by exponentiating the ratio of the sum of the parameter estimates for device infection, gender, and the gender-device infection interaction term in the corresponding mortality regression models for women with device infection and men with device infection. Risk-adjusted mortality HR without device infection, defined as mortality rate in women without device infection or mortality rate in men without device infection, were calculated by exponentiating the ratio of the parameter estimates for gender in corresponding mortality regression models for women without device infection and men without device infection. The regressions were estimated using Stata, version 11 (Stata Corp LP, College Station, Texas). Results were robust to heteroskedasticity.
The Goodwyn Institutional Review Board (Cincinnati, Ohio) determined that the study was exempt from full human subjects research review.
Results
The study cohort consisted of 64,903 Medicare fee-for-service patients who received an ICD in 2007 ( Table 1 ). In the study cohort, 84.1% of patients were aged ≥65 years, 26.8% were women, 16.1% were non-white, 18.0% had renal failure, 27.6% had diabetes mellitus, and 42% received a cardiac resynchronization therapy device with defibrillator.
| Characteristic | All, n = 64,903 (%) | Infection | p | |
|---|---|---|---|---|
| No, n = 63,048 (%) | Yes, n = 1,855 (%) | |||
| Age (yrs) | ||||
| <65 | 10,335 (16) | 9,977 (16) | 358 (19) | <0.001 |
| 65–69 | 12,120 (19) | 11,807 (19) | 313 (17) | 0.04 |
| 70–74 | 13,128 (20) | 12,772 (20) | 356 (19) | 0.26 |
| 75–79 | 14,170 (22) | 13,803 (22) | 367 (20) | 0.03 |
| 80–84 | 10,525 (16) | 10,223 (16) | 302 (16) | 0.94 |
| >84 | 4,625 (7) | 4,466 (7) | 159 (9) | 0.01 |
| Women | 17,400 (27) | 16,958 (27) | 442 (24) | 0.003 |
| Men | 47,503 (73) | 46,090 (73) | 1,413 (76) | 0.003 |
| White | 54,429 (84) | 52,925 (84) | 1,504 (81) | <0.001 |
| Non-white | 10,474 (16) | 10,123 (16) | 351 (19) | <0.001 |
| Congestive heart failure | 28,948 (45) | 27,938 (44) | 1,010 (54) | <0.001 |
| Valvular heart disease | 13,805 (21) | 13,425 (21) | 380 (21) | 0.40 |
| Hypertension | 33,080 (51) | 32,489 (52) | 591 (32) | <0.001 |
| Peripheral vascular disease | 4,690 (7) | 4,269 (7) | 61 (3) | <0.001 |
| Pulmonary circulation disorders | 2,866 (4) | 2,827 (5) | 39 (2) | <0.001 |
| Chronic pulmonary disease | 13,433 (21) | 13,004 (21) | 429 (23) | 0.01 |
| Renal failure | 11,684 (18) | 11,166 (18) | 518 (28) | <0.001 |
| Fluid and electrolyte disorders | 5,767 (9) | 5,411 (9) | 356 (19) | <0.001 |
| Diabetes mellitus, without CC | 15,915 (25) | 15,640 (25) | 275 (15) | <0.001 |
| Diabetes mellitus, with CC | 2,009 (3) | 1,956 (3) | 53 (3) | 0.55 |
| Hypothyroidism | 3,965 (6) | 3,816 (6) | 59 (3) | <0.001 |
| Blood loss anemia | 295 (0.5) | 269 (0.4) | 20 (1) | <0.001 |
| Deficiency anemia | 2,523 (4) | 2,465 (4) | 58 (3) | 0.09 |
| Coagulation deficiency | 1,601 (3) | 1,483 (2) | 118 (6) | <0.001 |
| Liver disease | 405 (0.6) | 386 (0.6) | 19 (1) | 0.03 |
| HIV infection or AIDS | NA | 43 (0.1) | ∗ | |
| Lymphoma | 375 (0.6) | 363 (0.6) | 12 (0.6) | 0.69 |
| Metastatic cancer | NA | 119 (0.2) | ∗ | |
| Solid tumor without metastasis | 519 (0.8) | 502 (0.8) | 17 (1) | 0.57 |
| Rheumatoid arthritis/CVD | 800 (1) | 780 (1) | 20 (1) | 0.54 |
| Obesity | 2,114 (3) | 2,096 (3) | 18 (1) | <0.001 |
| Weight loss | 513 (0.8) | 445 (0.7) | 68 (4) | <0.001 |
| Paralysis | 411 (0.6) | 399 (0.6) | 12 (0.6) | 0.53 |
| Other neurologic disorders | 1,507 (2) | 1,459 (2) | 48 (3) | 0.44 |
| Alcohol abuse | 727 (1) | 709 (1) | 18 (1) | 0.53 |
| Drug use | 297 (0.5) | 283 (0.4) | 14 (0.8) | 0.05 |
| Psychoses | NA | 465 (0.7) | ∗ | |
| Depression | 1,296 (2) | 1,280 (2) | 16 (1) | <0.001 |
| CRT-D | 27,261 (42) | 26,532 (42) | 729 (39) | 0.02 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
