Author + information
- Received July 13, 2015
- Revision received September 14, 2015
- Accepted September 19, 2015
- Published online March 1, 2016.
- Karthik Murugiah, MDa,
- Yun Wang, PhDb,c,
- Nihar R. Desai, MD, MPHa,b,
- Erica S. Spatz, MD, MHSa,b,
- Sudhakar V. Nuti, BAd,
- Rachel P. Dreyer, PhDa,b and
- Harlan M. Krumholz, MD, SMa,b,e,∗ ()
- aSection of Cardiovascular Medicine, Department of Internal Medicine, Yale School of Medicine, New Haven, Connecticut
- bCenter for Outcomes Research and Evaluation, Yale-New Haven Hospital, New Haven, Connecticut
- cDepartment of Biostatistics, Harvard T. H. Chan School of Public Health, Boston, Massachusetts
- dYale School of Medicine, New Haven, Connecticut
- eRobert Wood Johnson Foundation Clinical Scholars Program, Department of Internal Medicine, Yale School of Medicine and the Department of Health Policy and Management, Yale School of Public Health, New Haven, Connecticut
- ↵∗Reprint requests and correspondence:
Dr. Harlan M. Krumholz, Yale School of Medicine, 1 Church Street, Suite 200, New Haven, Connecticut 06510.
Objectives The aim of this study was to assess trends in hospitalizations and outcomes for Takotsubo cardiomyopathy (TTC).
Background There is a paucity of nationally representative data on trends in short- and long-term outcomes for patients with TTC.
Methods The authors examined hospitalization rates; in-hospital, 30-day, and 1-year mortality; and all-cause 30-day readmission for Medicare fee-for-service beneficiaries with principal and secondary diagnoses of TTC from 2007 to 2012.
Results Hospitalizations for principal or secondary diagnosis of TTC increased from 5.7 per 100,000 person-years in 2007 to 17.4 in 2012 (p for trend < 0.001). Patients were predominantly women and of white race. For principal TTC, in-hospital, 30-day, and 1-year mortality was 1.3% (95% confidence interval [CI]: 1.1% to 1.6%), 2.5% (95% CI: 2.2% to 2.8%), and 6.9% (95% CI: 6.4% to 7.5%), and the 30-day readmission rate was 11.6% (95% CI: 10.9% to 12.3%). For secondary TTC, in-hospital, 30-day, and 1-year mortality was 3% (95% CI: 2.7% to 3.3%), 4.7% (95% CI: 4.4% to 5.1%), and 11.4% (95% CI: 10.8% to 11.9%), and the 30-day readmission rate was 15.8% (95% CI: 15.1% to 16.4%). Over time, there was no change in mortality or readmission rate for both cohorts. Patients ≥85 years of age had higher in-hospital, 30-day, and 1-year mortality and 30-day readmission rates. Among patients with principal TTC, male and nonwhite patients had higher 1-year mortality than their counterparts, whereas in those with secondary TTC, mortality was worse at all 3 time points. Nonwhite patients had higher 30-day readmission rates for both cohorts.
Conclusions Hospitalization rates for TTC are increasing, but short- and long-term outcomes have not changed. At 1 year, 14 in 15 patients with principal TTC and 8 in 9 with secondary TTC are alive. Older, male, and nonwhite patients have worse outcomes.
Takotsubo cardiomyopathy (TTC), also known as stress cardiomyopathy or apical ballooning syndrome, is a rare but increasingly recognized cardiac syndrome that mimics acute myocardial infarction (AMI) (1). It is frequently triggered by an acute medical illness or intense emotional or physical stress, predominantly affecting older women. TTC is usually characterized by transient systolic dysfunction of segments of the left ventricle in the absence of angiographic evidence of obstructive coronary artery disease (CAD) or acute plaque rupture (2).
Our knowledge regarding outcomes of patients with TTC is largely limited to case series because of the rarity of the disease (3–5). These studies suggest that the short-term prognosis for this enigmatic condition is favorable, with most patients surviving the acute episode and experiencing recovery of ventricular function. However, wide variation (0% to 8%) in in-hospital mortality rates has been reported (6), and there is a paucity of nationally representative data. Furthermore, little is known about longer-term clinical outcomes, including mortality and readmission, after TTC, or trends in hospitalizations and outcomes.
Accordingly, we assessed trends in hospitalization rates and outcomes for TTC in the United States among Medicare fee-for-service beneficiaries between 2007 and 2012. We evaluated in-hospital, 30-day, and 1-year mortality; 30-day readmission; and risk-adjusted trends in these outcomes. We also evaluated trends in length of stay (LOS) and discharge disposition. Finally, we assessed differences in outcomes by subgroups of age, sex, and race.
Because TTC can present as an acute coronary syndrome or occur in conjunction with an acute medical illness, we reported outcomes separately for principal and secondary diagnoses of TTC. We reasoned that hospitalizations with principal diagnoses of TTC more likely represent primary coronary presentations of TTC, whereas hospitalizations with secondary diagnoses of TTC more likely represent cases in which a significant acute medical illness was the primary reason for admission, and TTC was precipitated by the acute medical illness. This distinction is important because outcomes for secondary diagnosis of TTC are likely to be driven by the primary reason for hospitalization.
We used Medicare inpatient claims data and Medicare enrollment data from the Centers for Medicare & Medicaid Services to identify all Medicare fee-for-service beneficiaries ages 65 years and older hospitalized with principal or secondary diagnoses of TTC in acute care hospitals in the United States from January 1, 2007, to December 31, 2012, using the International Classification of Diseases-Ninth Revision-Clinical Modification (ICD-9-CM) code 429.83. Because the specific ICD-9-CM code for TTC was introduced in October 2006, we used only data from 2007 onward. Although there is no worldwide consensus on the diagnostic criteria for TTC, most criteria for diagnosing TTC require the absence of obstructive CAD or angiographic evidence of acute plaque rupture (2,7). Thus, we included only those patients in our study sample who underwent coronary angiography (ICD-9-CM codes 37.22, 37.23, 88.55, 88.56, and 88.57) and did not receive revascularization therapy (i.e., percutaneous coronary intervention [ICD-9-CM codes 36.01, 36.02, 36.05, 36.07, 36.09, and 00.66] or coronary artery bypass grafting [ICD-9-CM codes 36.1x]). This method has been used previously (8). For the hospitalization analysis, the unit of analysis was at the Medicare fee-for-service beneficiary level; for reporting patient characteristics and the outcome analysis, the unit of analysis was at the patient level. Thus, if a patient had >1 admission for TTC during the year, we randomly selected 1 hospitalization. Institutional review board approval was obtained from the Yale University Human Investigation Committee.
We examined patients’ age, sex, and race (white, black, or other) as well as common comorbidities and trends in the demographic and clinical profiles separately for patients with principal and secondary diagnoses of TTC. Comorbidities included those used in risk adjustment models of 30-day mortality measures for AMI and heart failure (9,10). They were identified from secondary discharge diagnosis codes in the index hospitalization for TTC as well as principal or secondary diagnosis codes of all inpatient hospitalizations up to 1 year prior. For patients hospitalized in 2007, comorbidities were assessed using inpatient claims data from 2006.
Hospitalization rates and outcomes
We reported hospitalization rates separately for principal diagnosis, secondary diagnosis, and any diagnosis (principal or secondary) of TTC. We calculated the annual rate of hospitalization for TTC by dividing the number of hospitalizations with discharge diagnoses of TTC by the corresponding person-years for all fee-for-service Medicare beneficiaries in that year, to account for new enrollment, disenrollment, or death during an index year.
We reported outcomes separately for principal and secondary diagnoses of TTC. We evaluated the following outcomes associated with these hospitalizations: all-cause in-hospital and 30-day mortality, LOS, proportion discharged to home, all-cause 30-day readmission rates, and 1-year mortality and reported trends in these outcomes. To permit complete follow-up, we restricted the 1-year mortality analysis to 2011 discharges and the 30-day all-cause readmission analysis to November 30, 2012, discharges. Among strata of age, sex, and race (white vs. nonwhite), we only reported outcomes for the entire time period combined. For patients admitted with secondary diagnoses of TTC, we listed the top 10 principal diagnoses.
Annual hospitalization rates are expressed per 100,000 person-years. We calculated hospitalization rates for TTC overall and for the first (2007) and last (2012) years of the cohort. To simplify the presentation, we assessed patient demographics, clinical characteristics, and outcomes of unique patients hospitalized for TTC in 2-year intervals. Rates of mortality, readmission, and proportion discharged to home are expressed as percentages. LOS is represented in days. We used the Mantel-Haenszel chi-square test to determine the significance of temporal changes in hospitalization rates, patient characteristics, and outcomes. In addition, we fitted a linear mixed-effects model with a logit link function and hospital-specific random intercepts, adjusting for patient demographics and comorbidities, as in the Centers for Medicare & Medicaid Services mortality measures (9–11), to estimate temporal changes in rates of 30-day and 1-year mortality. For 30-day readmission, we conducted a survival analysis to calculate the proportion of patients who were readmitted for any cause to acute care hospitals, censoring those who died before readmission. We constructed a Cox proportional hazards model to assess the change in readmission rate over time. We fitted a separate model for each outcome. All models included an ordinal time variable, ranging from 0 to 5, corresponding to each year of the study period, after the visual inspection of crude rates revealed a linear pattern. All statistical tests were 2-sided at a significance level of p < 0.05. All analyses were performed with SAS version 9.3 64-bit (SAS Institute, Cary, North Carolina).
Overall, there were 9,442 hospitalizations with principal diagnoses of TTC from 2007 to 2012. Of those, there were 8,090 (85.7%) hospitalizations contributed by 8,068 unique Medicare beneficiaries in which the patients underwent angiography but did not receive revascularization therapy (from here on called the principal TTC cohort). The annual number of principal TTC hospitalizations increased from 634 in 2007 to 2,008 in 2012, and the corresponding incidence of TTC increased from 2.3 hospitalizations per 100,000 person-years in 2007 to 7.1 in 2012 (p for trend <0.001).
From 2007 to 2012, 26,048 hospitalizations were associated with secondary diagnoses of TTC. Of those, there were 11,971 (46%) hospitalizations, contributed by 11,898 unique Medicare beneficiaries, in which patients with secondary diagnoses of TTC underwent angiography but did not receive revascularization therapy (from here on called the secondary TTC cohort). The annual number of secondary TTC hospitalizations increased from 944 in 2007 to 2,921 in 2012, and the corresponding incidence of TTC increased from 3.4 hospitalizations per 100,000 person-years in 2007 to 10.3 in 2012 (p for trend <0.001).
Combined, the incidence of TTC hospitalizations with principal or secondary diagnoses of TTC in which patients underwent angiography but did not receive revascularization therapy increased from 5.7 hospitalizations per 100,000 person-years in 2007 to 17.4 in 2012 (p for trend <0.001).
Patients were predominantly women (94.3% for principal TTC and 90.3% for secondary TTC) and of white race (92.6% for principal TTC and 90.3% for secondary TTC) in both principal and secondary TTC, but patients with secondary TTC were more likely to be male and nonwhite (p < 0.001 for both). Cardiovascular risk factors were common and increased over time. A large proportion had prior diagnoses of atherosclerotic disease (46.1% among those with principal TTC and 47.5% among those with secondary TTC) (Table 1).
Overall, the in-hospital mortality rate for the principal TTC cohort was 1.3% (95% confidence interval [CI]: 1.1% to 1.6%), and the 30-day mortality rate was 2.5% (95% CI: 2.2% to 2.8%). The average LOS was 4.0 ± 3.3 days. The majority of patients were discharged home (73.6%; 95% CI: 72.7% to 74.6%). Within 30 days of discharge, 11.6% (95% CI: 10.9% to 12.3%) of patients were rehospitalized. The 1-year mortality rate was 6.9% (95% CI: 6.4% to 7.5%). Outcomes by study year are shown in Table 2. There were significant decreases over time in LOS (p < 0.001) and the proportion of patients discharged directly home (p = 0.045). In the risk-adjusted model, there was no significant difference in 30-day and 1-year mortality or in 30-day readmission between 2007 to 2008 and 2011 to 2012 (Figure 1).
For the secondary TTC cohort, in-hospital mortality, 30-day mortality, and 1-year mortality was 3% (95% CI: 2.7% to 3.3%), 4.7% (95% CI: 4.4% to 5.1%), and 11.4% (95% CI: 10.8% to 11.9%). Within 30 days of discharge, 15.8% (95% CI: 15.1% to 16.4%) of patients were rehospitalized. There was no significant difference in 30-day and 1-year mortality or 30-day readmission between 2007 to 2008 and 2011 to 2012. There was a significant decrease over time in the proportion of patients discharged directly home (p = 0.001) (Table 2).
Outcomes by strata of age, sex, and race are shown in Table 3. In both the principal and secondary TTC cohorts, older patients (≥85 years of age) had higher rates of mortality (in-hospital, 30-day, and 1-year) and readmission, had longer LOS, and were more likely to require post–acute care services compared with their counterparts. In the principal TTC cohort, both male and nonwhite patients had no appreciable difference in short-term mortality compared with their counterparts but had higher 1-year mortality. In the secondary TTC cohort, male and nonwhite patients had higher mortality at all 3 time points compared with their counterparts.
Among the secondary TTC cohort, the top 10 principal diagnoses by 3-digit ICD-9-CM codes were AMI (58.2%), heart failure (4.3%), other disease of the lung (including respiratory failure) (4.2%), other forms of chronic ischemic heart disease (3.1%), septicemia (3.0%), cardiac dysrhythmias (2.5%), chronic bronchitis (2.0%), symptoms involving respiratory system and other chest symptoms (1.5%), pneumonia (1.0%), and conduction disorders (0.9%).
In the secondary TTC cohort, a substantial proportion of patients were coded with principal diagnoses of AMI, and it is likely that these patients were similar to those in the principal TTC cohort. Thus, as a subanalysis of the secondary TTC cohort, we reported outcomes separately for those with principal diagnoses of AMI and all other principal diagnoses. Patients with principal diagnoses other than AMI had much higher mortality at all 3 time points and higher readmission rates, whereas those coded AMI for the principal diagnosis had mortality rates comparable with those in the principal TTC cohort (Table 4).
During the cohort construction, a notable observation was that a substantial proportion of patients coded for secondary diagnoses of TTC did not have obstructive CAD excluded by angiography (54%), and over time, there was an increasing trend of such coding (from 42.5% in 2007 to 57.7% in 2012, p for trend <0.001). The outcomes for these patients (in-hospital mortality 7.0% [95% CI: 6.5% to 7.5%], 30-day mortality 11.9% [95% CI: 11.2% to 12.5%], and 1-year mortality 24.5% [95% CI: 23.6% to 25.3%]) were much worse than for those in the secondary TTC cohort.
Among hospitalizations coded with principal diagnoses of TTC, 14.3% did not have obstructive CAD excluded by angiography. The outcomes for these patients were in-hospital mortality of 3.0% (95% CI: 2.7% to 3.3%), 30-day mortality of 4.7% (95% CI: 4.4% to 5.2%), and 1-year mortality of 11.4% (95% CI: 10.8% to 11.9%).
To our knowledge, this is the largest nationally representative study of post-discharge outcomes up to 1 year among patients hospitalized with TTC as well as the first to provide information on trends in hospitalizations and outcomes nationwide. We found that among Medicare fee-for-service beneficiaries, hospitalizations for both principal and secondary diagnoses of TTC increased by 3-fold from 2007 to 2012. For the principal TTC cohort, the outcomes were excellent. In-hospital mortality was low, with almost all patients surviving the acute episode. These patients continued to do well after discharge: 1-year mortality rates were low, and only 1 in 10 patients was rehospitalized within 30 days of discharge. There was no change in outcomes over time. The secondary TTC cohort had worse outcomes compared with the principal cohort. Excluding those who were coded with principal diagnoses of AMI, the mortality of the remainder of the secondary cohort was nearly 3× worse than that of the principal cohort. In both the principal and secondary cohorts, older, male, and nonwhite patients had worse outcomes. We also found that a large proportion of patients were being coded with secondary diagnoses of TTC who did not undergo angiography.
The prevalence of TTC reported in prior studies is approximately 1% to 2% of all patients presenting with initial diagnoses of either acute coronary syndromes or AMI (2,12). In our study, in 2012, the incidence of principal TTC among Medicare fee-for-service beneficiaries was 7.1 per 100,000 person-years, which, when compared with the hospitalization rate for AMI in the same population in 2010 to 2011 of 816 per 100,000 person-years (13), is about 0.9%. Adding the patients in the secondary TTC cohort who were coded with AMI for the principal diagnosis field, this proportion rises to 1.6%, which is similar to that observed in the prior case series of TTC (2,14,15). We found that over the study period, the incidence of TTC steadily increased. This is likely due to both increased recognition of the syndrome and increased recognition of the specific ICD-9-CM code for TTC, which was introduced in October 2006 (16).
TTC affects post-menopausal women disproportionately (8,17), for reasons that remain unclear, and 94% of our principal TTC cohort was female. A lower proportion of patients with TTC (46.1%) had histories of atherosclerotic disease compared with that reported for patients hospitalized with AMI (73.4%) (13). In fact, studies of angiographic findings in patients with TTC have reported a 40% to 80% prevalence of CAD (18,19). Although a lower incidence than for AMI, this is still a substantial burden of CAD and one that would be expected, given that TTC occurs commonly in older patients. This has important implications for the diagnosis of TTC, emphasizing the frequent coexistence of CAD with TTC and the need for careful assessment of the characteristics and distribution of echocardiographic and angiographic findings for a diagnosis of TTC.
Our method of isolating a cohort of patients with principal diagnoses of TTC who underwent angiography but did not undergo revascularization therapy provides a high degree of specificity to identify true TTC hospitalizations. Furthermore, this cohort likely represents the classic coronary presentations of TTC in which the outcome is driven largely by TTC instead of a primary acute illness. The other important advantage of this approach is that the secondary diagnosis field in administrative data often represents a history of the condition rather than the condition occurring in the same hospitalization. Consequently, it is possible that some hospitalizations with secondary diagnoses of TTC may in fact represent histories of the illness. However, our requirement of angiography without revascularization therapy should help minimize this bias in the secondary TTC cohort.
One prior national-level study, using data from the Healthcare Cost and Utilization Project (20), examined in-hospital outcomes for TTC. However, that study did not use the additional criteria, such as receipt of cardiac catheterization and the absence of revascularization, used in our study to reliably identify TTC, and thus the outcomes may not be robust. Exclusion of obstructive CAD or acute plaque rupture by angiography is a prerequisite for diagnosing TTC (2,7). Furthermore, that study did not provide long-term outcomes.
We found outcomes for our principal TTC cohort to be excellent, with almost all patients surviving to discharge. Our observed in-hospital mortality rate of 1.3% in the principal TTC cohort is at the lower end of the entire spectrum of reported rates (0% to 8%) in case series on TTC (6,21). In contrast to outcomes for principal TTC in 2011 to 2012, the in-hospital mortality for AMI in 2010 to 2011 for the white, female subgroup, which is the demographic for TTC and thus more comparable, was 6× higher (Figure 2) (13). Furthermore, 3 in 4 patients were discharged directly home without the need for post–acute care services. These patients continued to do well after discharge, with only 1 in 9 needing to be rehospitalized within 30 days, compared with 1 in 5 who were post-AMI. The long-term outcomes of TTC were excellent; in 2012, only 1 in 15 patients with TTC did not survive to 1 year, compared with 1 in 4 patients with AMI (13). In summary, although TTC closely mimics AMI, the outcomes are vastly better. Studies show that patients who survive acute episodes of TTC typically recover ventricular function within weeks to months (5).
The outcomes for the secondary TTC cohort were, expectedly (6), worse compared with those of the principal TTC cohort. An important and surprising finding was that nearly 3 in 5 patients with secondary diagnoses of TTC were coded with AMI for the principal diagnosis. These patients were either being coded with AMI in the principal diagnosis field incorrectly, or possibly intentionally for billing purposes. These patients were likely clinically similar to those in the principal TTC cohort, and we found their outcomes to be similar to those of the principal TTC cohort. In future studies of TTC using administrative data, investigators may want to account for this observation when defining principal versus secondary TTC. Excluding patients with principal diagnoses of AMI, 1 in 15 patients of the remainder of the secondary TTC cohort did not survive the hospitalization, and 1 in 6 did not survive to 1 year. However, this is lower than the in-hospital mortality reported in prior studies for TTC secondary to acute medical conditions (10%) (6). Respiratory conditions, such as respiratory failure, bronchitis, and pneumonia, and sepsis were the most common acute conditions among patients with secondary TTC.
Older patients were at higher risk for adverse outcomes in both principal and secondary TTC cohorts, as were nonwhite and male patients, which warrants further study and increased attention to these subgroups of patients with TTC. There is some evidence that men presenting with TTC tend to have higher disease acuity (22,23), and a previous meta-analysis showed that they may be at higher risk for mortality. Our study, using nationally representative data, corroborates this observation. Furthermore, we observed that over the study period, there was no change in outcomes for TTC. The optimal medical regimen for TTC is not defined, and there are no controlled data available. Management is largely supportive, and patients are usually treated with standard medications for left ventricular systolic dysfunction.
First, we studied Medicare fee-for-service beneficiaries, which limits the generalizability of the findings to younger populations. Nevertheless, TTC usually affects older patients, a population captured in our cohort.
Second, we relied on the ICD-9-CM code to identify TTC. To our knowledge, there are no validation studies for this code (429.83). However, the code has good face validity, as the code description reads “Takotsubo syndrome” and is unlikely to be applied incorrectly. To improve the specificity of TTC, we limited the study cohort to those who underwent cardiac catheterization only. This method has been used previously (8).
Third, it is possible that cases of TTC may have been coded incorrectly as only AMI, and we may have underestimated the incidence of TTC. In contrast, cases with true acute coronary syndromes but minimal angiographic findings, which were unrecognized by the physicians, may lead to overcoding of TTC.
Fourth, there may have been patients in the secondary TTC cohort in whom the secondary diagnosis of TTC represented histories of the condition rather than the condition occurring in the same hospitalization.
Among Medicare fee-for-service beneficiaries, hospitalization rates for TTC are increasing over time. Although the syndrome mimics AMI at presentation, we found short- and long-term outcomes of principal TTC to be excellent compared with those for AMI, with nearly all patients with TTC surviving to discharge and 14 in 15 surviving to 1 year. Patients with secondary TTC had worse outcomes, likely driven by their primary acute medical condition. Older, male, and nonwhite subgroups were more vulnerable to adverse outcomes after TTC.
COMPETENCY IN MEDICAL KNOWLEDGE: Hospitalization rates for TTC are increasing in the United States. This condition predominantly affects women of white race. Short- and long-term outcomes of primary TTC are excellent, with 14 in 15 patients surviving to 1 year. Outcomes for TTC secondary to other medical conditions are worse, and 8 in 9 survive to 1 year. Older, male, and nonwhite patients have worse outcomes.
COMPETENCY IN PATIENT CARE: Although TTC is frequently precipitated by an episode of stress, short- and long-term outcomes for principal TTC are excellent. Making patients aware of their good prognosis is important to avoid further stress from the illness. Outcomes for secondary TTC are worse and driven largely by the primary acute medical condition.
TRANSLATIONAL OUTLOOK: Further study is required to elucidate why male and nonwhite patients have comparatively poorer outcomes than their counterparts.
During the time the work was conducted, Dr. Murugiah and Mr. Nuti were affiliated with the Center for Outcomes Research and Evaluation, Yale-New Haven Hospital, New Haven, Connecticut, and Mr. Nuti was affiliated with the Section of Cardiovascular Medicine, Department of Internal Medicine, Yale School of Medicine, New Haven, Connecticut.
This project was supported by grant 1U01 HL105270-05 (Center for Cardiovascular Outcomes Research at Yale University) from the National Heart, Lung, and Blood Institute. Dr. Krumholz is the recipient of research agreements from Medtronic and Johnson & Johnson (Janssen), through Yale University, to develop methods of clinical trial data sharing and is chair of a cardiac scientific advisory board for UnitedHealth. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- acute myocardial infarction
- coronary artery disease
- confidence interval
- International Classification of Diseases-Ninth Revision-Clinical Modification
- length of stay
- Takotsubo cardiomyopathy
- Received July 13, 2015.
- Revision received September 14, 2015.
- Accepted September 19, 2015.
- American College of Cardiology Foundation
- Sharkey S.W.,
- Windenburg D.C.,
- Lesser J.R.,
- et al.
- Krumholz H.M.,
- Wang Y.,
- Mattera J.A.,
- et al.
- Krumholz H.M.,
- Wang Y.,
- Mattera J.A.,
- et al.
- Gianni M.,
- Dentali F.,
- Grandi A.M.,
- Sumner G.,
- Hiralal R.,
- Lonn E.
- Krumholz H.M.,
- Normand S.L.,
- Wang Y.
- Akashi Y.J.,
- Goldstein D.S.,
- Barbaro G.,
- Ueyama T.
- ↵Centers for Disease Control and Prevention. ICD-9-CM Coordination and Maintenance Committee Meeting. September 30, 2005. Available at: http://www.cdc.gov/nchs/data/icd/agendasept05.pdf. Accessed September 11, 2015.
- Prasad A.,
- Dangas G.,
- Srinivasan M.,
- et al.