|Home | About | Journals | Submit | Contact Us | Français|
Little is known about factors that influence survival and complications following ICD implantation in routine clinical practice. We examined patient and implanting physician factors associated with outcomes of implantable cardioverter-defibrillator (ICD) therapy in Medicare beneficiaries from 2002 through 2005.
We limited this analysis to patients aged 65 or older with Medicare fee-for-service coverage who received an ICD between January 2002 and September 2005. The main outcome measures are time to post-procedural complications within 90 days and 1-year mortality. During the study period, 8581 patients had an ICD implanted by 1959 physicians. The number of procedures increased from 1644 in 2002 to 2374 in the first three quarters of 2005. The overall complication rate declined from 18.8% in 2002 to 14.2% in 2005 (P < .001). Factors independently associated with an increased hazard of complications include chronic lung disease, dementia, renal disease, implantation by a thoracic surgeon and implantation with removal/replacement. History of congestive heart failure, outpatient implantation and more recent years of ICD implantation were associated with a lower risk of complications (P < .05 for all factors). From 2002 to 2005, we observed a decline in 1-year mortality (P < .001).
We observed an appreciable increase in the number of ICD implants which was associated with a significant decrease in the rate of complications and 1-year mortality. We identified factors associated with an increased risk of mortality and post-procedural complications that may support more nuanced treatment decisions than are currently possible.
Although the implantable cardioverter-defibrillator (ICD) has been demonstrated in randomized clinical trials to be effective at preventing sudden cardiac death, little is known about factors that influence survival and complications following ICD implantation in routine clinical practice.1–5 Identifying such factors is crucial to our efforts to reduce post-operative complications and to enhance patients’ survival and welfare. Indeed, the risk of such complications is not inconsequential; about 30% of patients receiving an ICD experience at least one complication following ICD implantation and in 10% of these patients the complication is directly related to the procedure.6
In a study of the frequency and incremental cost of major complications among Medicare beneficiaries receiving an ICD in fiscal year 2003, 10.8% experienced one or more complications that resulted in a significant increase in length of hospital stay and costs.7 However, that study did not identify factors associated with such complications nor did it examine implant-related complications that occurred after discharge from the hospital.7 Thus, more complete information is needed on the short- and long-term complications following ICD implantation as well as factors associated with their occurrence. We conducted this study to examine patient and implanting physician factors associated with outcomes of ICD therapy in Medicare beneficiaries. We also examined trends in the rates of complications of ICD implantation over the study period.
We performed this analysis using data from the Medicare inpatient, outpatient, and carrier standard analytic files (a 5% national sample) and the corresponding denominator files. Institutional claims for facility costs covered under Medicare Part A are included in the inpatient files. Claims from outpatient providers are included in the outpatient files, and non-institutional provider claims for services covered under Medicare Part B are included in the carrier files. The denominator files include demographic and enrollment information for each beneficiary enrolled in Medicare during a calendar year.
We obtained files from January, 2001 through December, 2005, from the Centers for Medicare & Medicaid Services (CMS). We limited the analysis to beneficiaries living in the United States who were 65 years or older on the date of ICD implantation. Only claims filed during periods of fee-for-service coverage were included.
From the carrier file, we selected all beneficiaries who received an ICD between January 1, 2002, and September 30, 2005. We obtained the 2001 inpatient, outpatient and carrier files to identify patient comorbid conditions up to 365 days prior to the ICD implantation date. We ended the study period on September 30, 2005, to allow us to examine the 90-day outcomes for all patients in the study. Specifically, we used Healthcare Common Procedure Coding System (HCPCS) code 33249 on a single carrier claim to identify implantation of an ICD. In the 2003–2005 data, we expanded the code list to include the temporary HCPCS codes established by Medicare for reimbursement of ICD implantations (G0297, G0298, G0299, G0300). Both new ICD implants and upgrades of existing devices to ICDs were included in these analyses. We excluded codes for epicardial ICD placement. We also excluded 531 beneficiaries whose carrier claim could not be matched to an inpatient or an outpatient Medicare claim, 68 beneficiaries whose physician’s primary specialty was missing on the AMA Physician Masterfile and 287 beneficiaries who received an ICD in conjunction with coronary artery bypass graft surgery.
Outcomes were evaluated between January 1, 2002, and December 31, 2005. We calculated survival as the number of days from the discharge date listed in the index procedure claim to the death date recorded in the Medicare denominator file. We used previously described algorithms to identify complications following ICD implantation.6,7 We examined pneumothorax (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] 512.1, 511.8, 34.04, or HCPCS codes 32002, 32019, or 32020), cardiac tamponade (420.x, 423.0, 423.9, 37.0, or HCPCS 33010, 33011, or 33015), hematoma (998.1x), mechanical complications (996.0, 996.04, or a combination of ICD-9-CM 996.01 and either procedure code 96.04 or HCPCS code 31500), mechanical complications with system revision (mechanical complications codes combined with 37.99 or 00.52 or HCPCS codes 33210, 33211, 33216, or 33217) and infection (996.6 and 996.61; these codes refer to infection and inflammation due to cardiac device, implant, or graft) during the index procedure stay or within 1 day following the discharge date (index complications). Mechanical complications include any malfunction on the part of the device; such as, breakdown, displacement, perforation and protrusion of the device and/or lead(s).8 We assessed hematoma, infection, mechanical complications, and mechanical complications with system revision up to 30 days after discharge (30-day complications), and infection, mechanical complications and mechanical complications with system revision up to 90 days following the discharge date (90-day complications). A complete list of all complications examined in this study is included in the Appendix.
We created two composite variables: “index complication” and “any complication”. Index complication includes pneumothorax, cardiac tamponade, other cardiac complication (defined in the ICD-9-CM as cardiac arrest, cardiac insufficiency, cardiorespiratory failure or heart failure during or resulting from a procedure), acute renal failure, pulmonary embolism, hematoma, infection, mechanical complications and mechanical complications with system revision during the index hospitalization and up to 1 day after discharge. “Any complication” includes pneumothorax, cardiac tamponade, other cardiac complication, acute renal failure, pulmonary embolism, hematoma, infection, mechanical complications and mechanical complications with system revision during the index admission and up to 1 day after discharge, pulmonary embolism, hematoma, infection, and mechanical complications with and without system revision within 30 days, and infection, mechanical complications and mechanical complications with system revisions within 90 days.
We used the American Medical Association (AMA) Physician Masterfile to obtain physician characteristics.9 The Masterfile includes current and historical data on all physicians, including AMA members and non-members. An AMA Physician Masterfile record is established when individuals enter a medical school accredited by the Liaison Committee on Medical Education or, in the case of international medical graduates, upon entry into residency programs accredited by the Accreditation Council on Graduate Medical Education or when licensed to practice medicine in the United States. For this analysis, we obtained files that included primary and secondary specialty and year of medical school graduation. Physicians’ average annual ICD implantation volumes were derived from 2002–2005 using the 5% Medicare carrier files.
Physicians were classified as electrophysiologists if electrophysiology was listed as their primary or secondary specialty in the AMA Masterfile or if they billed for an electrophysiology study (HCPCS code 93623) between 2000 and 2005. Physicians were classified as non-electrophysiology cardiologists if their primary specialty is listed as interventional cardiology, nuclear cardiology, or pediatric cardiology or their primary specialty is cardiovascular disease and their secondary specialty is internal medicine, general practice, geriatrics, family practice, interventional cardiology, nuclear cardiology, pediatric cardiology or unspecified/other specialty. Physicians were classified as thoracic surgeons if either their primary or their secondary specialty is thoracic surgery. Physicians classified as “Other” are those with any other non-missing primary specialty. Approximately two thirds of “other” physicians identified themselves primarily as internists. The remainder identified themselves across a range of specialties. We compared the resulting distribution of ICD implants by physician specialty with publicly available data from the American College of Cardiology National Cardiovascular Data Registry-ICD registry.10
We identified co-morbid conditions using coding algorithms described previously.11, 12 Specifically, we searched all inpatient, outpatient, and carrier claims for 365 days preceding the date of ICD implant for evidence of cerebrovascular disease (362.34, 430.x–438.x), chronic obstructive pulmonary disease (416.8, 416.9, 490.x–505.x, 506.4, 508.1, 508.8), congestive heart failure (428.x, 398.91, 402.x1, 404.x1, 404.x3, 425.4–425.9), coronary artery disease (ICD-9-CM codes 410.x–414.x, 429.2, V45.81), dementia (290.x, 294.1, 331.2), diabetes mellitus with complications (ICD-9-CM 250.4–250.7) or without complications (250.0–250.3, 250.8, 250.9), hypertension (401.x–405.x, 437.2), metastatic solid tumor (196.x–199.x), myocardial infarction (MI) (410.x, 412.x), peripheral vascular disease (093.0 437.3, 440.x, 441.x, 443.1–443.9, 47.1, 557.1, 557.9, V43.4), and renal disease (403.01, 403.11, 403.91, 404.02, 404.03, 404.12, 404.13, 404.92, 404.93, 582.x, 583.0–583.7, 585.x, 586.x, 588.0, V42.0, V45.1, V56.x). We used the inpatient and outpatient files to determine the setting in which the ICD implantation occurred and whether the patient was admitted emergently. We used carrier HCPCS codes 33233 and 33241 to identify ICD implant with removal/replacement.
We examined the baseline characteristics of patients and physicians. Categorical variables are presented as percentages, and continuous variables are presented as means with SDs. To test for any differences among physician specialties, we calculated P values for all categorical variables using the Cochran-Mantel-Haenszel test for general association, and for age we used the Kruskal-Wallis test.
We determined the frequency of each outcome by year, across all years, and by physician specialty. To test for a temporal trend in complication rates, we calculated P values using the Cochran-Mantel-Haenszel test for non-zero correlation. We calculated P values using the Cochran-Mantel-Haenszel test for general association to test for differences in complication rates by physician specialty. We used the Kaplan-Meier survival analysis method to calculate mortality rates. To test for a temporal trend in 1-year mortality rates, we calculated the P value using a log-rank test for trend. To test for differences among physician specialties, we also calculated the P value using a log-rank test. We examined univariate and multivariable relationships between patient and provider characteristics and outcomes. We used Cox proportional hazards models to model mortality and any ICD complication. Variables included patient characteristics (gender, age, race/ethnicity, geographic location, comorbid conditions) and provider characteristics (specialty, years since graduation, Medicare procedure volume). Robust standard errors were used to account for clustering of similar patients within hospitals.13 We used SAS version 9.1.5 for all analyses (SAS Institute Inc, Cary, North Carolina).
This study was approved by the institutional review board of the Duke University Health System. The authors had full access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript as written.
During the study period, 8581 patients had an ICD implanted by 1959 physicians. Of these physicians, 1840 (93.9%) were male. Table 1 shows the number of patients in each specialty category and the baseline characteristics of patients by physician specialty. The distribution of ICD implants by physician specialty is consistent with the distribution of physician specialty from the American College of Cardiology National Cardiovascular Data Registry-ICD registry (electrophysiologists, 79%; non-electrophysiology cardiologists, 17%; cardiac surgeons, 3%; other, 0.25%).10 Notably, patients who had their device implanted by a non-EP cardiologist were significantly more likely to have CHF than patients who had their ICD implanted by a thoracic surgeon. Patients who had their ICD implanted by a thoracic surgeon were more likely to have cerebrovascular disease than patients who had their ICD implanted by physicians of other specialties. The regional distribution of ICD patients varied by physician specialty. For example, 46% of all ICD implants performed by thoracic surgeons occurred in the South region as compared with 39% of ICD implants by electrophysiologists. In addition, the volume of ICDs with cardiac resynchronization therapy was significantly lower for thoracic surgeons than for electrophysiologists or non-EP cardiologists.
Table 2 shows unadjusted ICD complications and mortality by year of ICD implantation. The number of ICD implants increased from 1644 in 2002 to 2374 in the first three quarters of 2005. The rates of any complication following ICD implantation declined from 18.8% in 2002 to 14.2% in 2005 (P < .001). A decline in mechanical complications was observed (P < .001). Two other important observations stand out. First, the majority of complications following ICD implantation occur during the index hospital stay or up to 1 day following discharge. Second, mechanical complications are the most common index complication with 7.2% in 2002 and 3.8% in 2005.
The unadjusted rate of index ICD complication differed by gender (10.3% in men and 12.1% in women, P = .03). No significant differences in the rates of any ICD complication were observed by georgraphic region (15.4% in the West, 15.5% in the South, 17.3% in the Midwest, 16.2% in the Northeast), or by race (15.5% in blacks, 14.2% in other, and 16.3% in whites). The rate of any ICD complication was significantly higher if the ICD was implanted following an admission from the emergency room (17.2%) versus other inpatient (16.5%) or outpatient (12.3%) setting (P = .002).
The distribution of ICD complications and mortality by physician specialty is shown in Table 3. The rate of any complication following ICD implantation was significantly higher for thoracic surgeons than for physicians of any other specialty.
From 2002 to 2005, we observed a decline in 1-year mortality (P < .001). This decline is illustrated in Figure 1 with Kaplan-Meier mortality curves stratified by year of ICD implantation.
Table 4 displays the results of a multivariable model of 1-year mortality following ICD implantation. After controlling for other confounders, factors associated with an increased risk of 1-year mortality are older age, history of MI, CHF, chronic lung disease, dementia, diabetes with and without complications, metastatic cancer, peripheral vascular disease, renal disease, and admission from the ER. Hypertension, an outpatient implantation setting, and more recent years of ICD implantation were associated with a lower risk of 1-year mortality. None of the physician specialties was associated with increased mortality.
Table 5 displays the results of a multivariable model of time to any ICD complication. Independent factors associated with an increased hazard of any complication include chronic lung disease, dementia, renal disease, implantation of the device by a thoracic surgeon and implantation with removal/replacement. History of CHF, an outpatient implantation setting and more recent years of ICD implantation were associated with a lower risk of complications. In multivariable models, men were at a significantly lower risk of index pneumothorax (HR, 0.50; 95% CI, 0.34–0.75; P < .001) than women. As measured in the 5% sample, physician volume was not significantly associated with the risk of complications even after testing for thresholds of < 20 implants per year, < 15 implants per year, and < 10 implants per year in sensitivity analyses.
Our study has four key findings. First, the appreciable increase in the number of ICD implants between 2002 and 2005 was associated with a significant decrease in the rate of any complication following ICD implantation. Second, the majority of complications following ICD implantation occur either during the index hospital stay or up to 1 day following discharge with mechanical complications being the most common index complication. Third, the rate of any complication following ICD implantation was significantly higher for thoracic surgeons than for physicians of any other specialty. Fourth, from 2002 to 2005, there was a decline in the risk of 1-year mortality.
With the expanding indications for ICD therapy, it is not surprising that the number of ICD implantations has increased substantially since 2002.1–5,14 In our study, this increase was accompanied by a decrease in the risk of any post-procedural complication. The risk of infection remained unchanged over the study period. Although one study showed a significant increase in the risk of cardiac device infections among Medicare beneficiaries from 1990 to 1999, that study was done in a different time period than our study and it included mechanical valves and left ventricular assist devices in addition to pacemakers and defibrillators.15 An analysis of the National Hospital Discharge Survey from 1996 to 2003 showed a 6-fold increase in the number of hospitalizations for ICD infection paralleling the significant rise in the number of ICD implantations, but this observation is not supported by our findings.16
Our study provides new insights into post-ICD implantation complications. Not only were complications highest during the index hospitalization and within 1 day after discharge, but the most common complication was mechanical complications with a high annual incidence rate of 7.2% in 2002 and 3.8% in 2005. In addition, the risk of complications was significantly higher for thoracic surgeons than for electrophysiologists. This increased risk of complications was still significant even after adjusting for other confounders. Although we could not determine the reasons for this finding, it is possible that cardiothoracic surgeons implant a lower volume of ICDs, use a different technique for ICD implantation or operate on patients that are different from patients implanted by non-cardiothoracic surgeons in ways that we can not measure. That mortality was not significantly different for thoracic surgeons and for electrophysiologists likely indicates that patients operated on by thoracic surgeons were not significantly sicker than patients operated on by electrophysiologists.
From 2002 to 2005, we observed a decline in 1-year mortality. Although reasons for this improvement are not readily analyzable, it may have resulted from better patient care due to implementation of evidence-based therapies, better patient selection, better ICD and lead technology, and/or improved ICD implantation techniques and follow-up.
Another important contribtution of our study is the identification of factors associated with an increased risk of mortality including older age, the presence of cardiac and non-cardiac comorbidities, admission from the ER, and more remote years of ICD implantation. We also identified factors associated with an increased risk of any post-procedural complication. Among several factors, implantation of the device by a thoracic surgeon stands out and raises a question about whether this finding is related to a lower volume of ICD implants performed by thoracic surgeons. Although we found no significant association between physician volume and the risk of complications, this finding likely reflects imprecision in the volume variable as it is based on the 5% sample. A prior study based on the 20% sample showed an inverse relationship between post-procedural complications and physician volume.17 In addition, ICD removal/replacement was associated with an increased risk of any complication primarily driven by an increased risk of infection and mechanical complications. This finding is in concordance with other studies that have reported a higher risk of infection and other complications from device removal/replacement.18,19 It is surprising that the risk of complications was not higher for ICDs with cardiac resynchronization therapy as implanting these devices is technically more challenging than implanting regular ICDs; however, our study may not have had enough power to address this question.
That the rate of penumothorax was significantly higher in women than in men is not surprising. A higher risk of complications in women has been reported in relation to other cardiac procedures.20–22 This observation may help physicians provide gender-specific procedural information to their patients.
Our study has some limitations. First, because we included only Medicare patients, our results may not be generalizable to non-Medicare patients, especially younger patients who are likely to have fewer post-procedural complications. Second, our analyses are dependent upon accurate coding in the Medicare Claims Database and identification of those codes in the database. To the extent that complications of interest are not consistently coded, we may have underestimated the prevalence of complications. Third, because we used claims data that lack important clinical parameters such as ejection fraction, we may not have accurately characterized the severity of illness. Fourth, results regarding the physician volume variable should be interpreted cautiously as they are based on the 5% sample. In addition, our analyses did not take into account the possible variation in the annual volume of ICD implants for each physician over the years of the study.
Between 2002 and 2005, we observed an appreciable increase in the number of ICD implants which was associated with a significant decrease in the rate of any complication following ICD implantation. We identified factors associated with an increased risk of mortality and post-procedural complications. These factors may allow the characterization of patients with large versus small expected net clinical benefits from ICD therapy, thereby supporting more nuanced treatment decisions than are currently possible.
We thank Damon M. Seils, MA, of Duke University for assistance with manuscript preparation. Mr Seils did not receive compensation for his assistance apart from his employment at the institution where the study was conducted.
Funding Sources: Supported by grant 1R01AG026038-01A1 from the National Institute on Aging and grant 5U01HL66461-05 from the National Heart, Lung, and Blood Institute. The funding bodies had no role in the design and conduct of the study; in the collection, analysis, and interpretation of the data; or in the preparation, review, or approval of the manuscript.
Financial Disclosures: Dr Al-Khatib reported receiving research support from Medtronic; and honoraria for presentations from Medtronic. Dr Hernandez reported receiving research grants from Scios, Medtronic, GlaxoSmithKline, and Roche Diagnostics; and serving on the speaker’s bureau or receiving honoraria in the past 5 years from Novartis. Dr Schulman reported receiving research and salary support from Actelion Pharmaceuticals, Allergan Pharmaceuticals, Amgen, Bristol-Myers Squibb, Ernst & Young, Genentech, GlaxoSmithKline, IBM Center for Healthcare Management, Inspire Pharmaceuticals, Johnson & Johnson, Kureha Corporation, Lilly Foundation, Medtronic, NABI Biopharmaceuticals, Novartis, OSI Eyetech, Pfizer, Pharmacia, Purdue Pharma, Sanofi-Aventis, Scios, Theravance, Wyeth, and Yamanouchi USA Foundation; receiving personal income for consulting from Genentech, The Health Strategies Consultancy, and the National Pharmaceutical Council; having equity in and serving on the board of directors of Cancer Consultants; having equity in and serving on the executive board of Faculty Connection LLC; and having equity in Alnylam Pharmaceuticals. Dr Schulman has made available online a detailed listing of financial disclosures (http://www.dcri.duke.edu/research/coi.jsp). Dr Curtis reported receiving research and salary support from Allergan Pharmaceuticals, GlaxoSmithKline, Lilly, Medtronic, Novartis, Ortho Biotech, OSI Eyetech, Pfizer, and Sanofi-Aventis. Dr Curtis has made available online a detailed listing of financial disclosures (http://www.dcri.duke.edu/research/coi.jsp). Ms Greiner and Dr Peterson did not report any financial disclosures.
Publisher's Disclaimer: This is an un-copyedited author manuscript that was accepted for publication in Circulation: Arrhythmia and Electrophysiology, copyright The American Heart Association. This may not be duplicated or reproduced, other than for personal use or within the “Fair Use of Copyrighted Materials” (section 107, title 17, U.S. Code) without prior permission of the copyright owner, The American Heart Association. The final copyedited article, which is the version of record, can be found at ://circep.ahajournals.org/cgi/content/full/1/4/240. The American Heart Association disclaims any responsibility or liability for errors or omissions in this version of the manuscript or in any version derived from it by the National Institutes of Health or other parties.