Article Text

Download PDFPDF

Original article
Ethnic differences in 1-year mortality among patients hospitalised with heart failure
  1. Padma Kaul1,2,3,
  2. Finlay A McAlister2,3,4,
  3. Justin A Ezekowitz1,2,3,
  4. Vaneeta Kaur Grover3,
  5. Hude Quan5
  1. 1Division of Cardiology, Department of Medicine, University of Alberta, Edmonton, Alberta, Canada
  2. 2Mazankowski Alberta Heart Institute, University of Alberta, Edmonton, Alberta, Canada
  3. 3Canadian VIGOUR Center, University of Alberta, Edmonton, Alberta, Canada
  4. 4Division of General Internal Medicine, Department of Medicine, University of Alberta, Edmonton, Alberta, Canada
  5. 5Department of Community Health Sciences, University of Calgary, Calgary, Alberta
  1. Correspondence to Dr Padma Kaul, Department of Medicine, University of Alberta, 331 Environmental Engineering Building, Edmonton, Alberta T6G 2M8, Canada; pkaul{at}ualberta.ca

Abstract

Objective The incidence of cardiovascular disease and the prevalence of risk factors have been shown to differ significantly across ethnic groups. The objective of this study was to examine the impact of ethnicity on 1-year mortality among patients with heart failure in a single payer healthcare system with universal access.

Design, setting and patients Alberta residents aged 20 years or older hospitalised with heart failure between 1 April 1999 and 31 December 2005 are included. Previously validated algorithms were used to assign ethnicity based on patient surname. Patients were categorised as white, Chinese or East Indian.

Main outcome measure One-year mortality after adjusting for baseline differences.

Results 52 980 white, 851 Chinese, and 377 East Indian individuals were hospitalised with heart failure. Chinese patients were the oldest and had the highest rates of renal disease. East Indian patients were the youngest and had the highest rates of ischaemic heart disease and diabetes. One-year mortality rates were 31.0% among white patients, 38.7% among Chinese and 26.5% among East Indian patients (p<0.01). Adjusted HR (and 95% CI) for 1-year mortality among Chinese compared with white patients was 1.34 (1.20 to 1.49) and among East Indian compared with white patients it was 1.04 (0.85 to 1.27). These findings were consistent across various subgroups, including patients with incident heart failure.

Conclusions Ethnicity appears to modulate patient outcomes in heart failure. Chinese patients have significantly higher 1-year mortality rates compared with white patients; there appear to be no differences in mortality among East Indian and white patients.

  • Ethnic groups
  • heart failure
  • mortality

Statistics from Altmetric.com

Heart failure is a debilitating and chronic condition associated with significant morbidity and mortality. The prevalence of heart failure is expected to increase globally in the coming years as a result of several factors, including ageing populations, improved survival rates after myocardial infarction, and the development of new therapies for more effective treatments.1–6 There is evidence to suggest that ethnicity is a modulator of heart failure incidence and outcomes.7–15 The Multi-ethnic Study of Atherosclerosis, which enrolled patients without known cardiovascular disease and followed them for a median of 4 years found that African-American individuals had the highest incidence of heart failure, followed by Hispanic, white and Chinese-American patients.7 A majority of the studies examining the importance of ethnicity in heart failure have been restricted to comparisons between white and African-American patients, and have shown the latter to have a worse prognosis compared with their white counterparts.10–12 Studies from the UK have reported similar or slightly better outcomes among South Asian individuals (people of Indian, Pakistani, Sri Lankan, or Bangladeshi origin) compared with white patients with heart failure.13–15 However, no previous study has examined mortality differences, at a population level, among white, Chinese and East Indian patients with heart failure.

The province of Alberta, Canada, consists of approximately 3.7 million people, of which 86% are white, 3.7% are Chinese and 3.2% are South Asian (predominantly East Indian).16 All residents have universal health care, which includes free access to inpatient, outpatient and physician services and to prescription medications for patients aged 65 years or older. Using population-based administrative healthcare databases maintained by the provincial health department (Alberta Health and Wellness), we examined the extent to which ethnicity was associated with 1-year mortality outcomes among patients hospitalised with heart failure.

Methods

Data sources

Data for the study consist of databases described previously and maintained by the Alberta Ministry of Health and Wellness.1 17–20 For the current study, the following data sources were linked using a unique anonymous patient identifier: (1) the discharge abstract database, which contains diagnostic and treatment information, length of stay and discharge status for patients admitted to an acute care hospital in Alberta; (2) the ambulatory care classification system database, which records all outpatient clinic visits (including emergency department visits); (3) the physician claims database, which records all physician services; (4) the Alberta healthcare insurance registry database, which records the status of all residents of Alberta including year of birth, gender and postal code; (5) the vital status database, which includes the date of death; and (6) the Alberta Blue Cross medication database, which includes pharmaceutical claims for patients over 65 years of age.

Study population

Our cohort consists of all residents of Alberta aged 20 years or older hospitalised with a discharge diagnosis of heart failure between 1 April 1999 and 31 December 2005 (subsequently referred to as the ‘index hospitalisation’). Vital status as of 31 December 2006 was available for all patients. The diagnosis of heart failure is based on International Classification of Diseases, 9th revision, clinical modification (ICD-9-CM) code 428 or ICD-10 code I50 (after 1 April 2002). The accuracy of administrative coding for heart failure has previously been validated.21–23 In this study's database, the specificity for heart failure coding was 98.7%.21 22

Ethnicity, comorbidities, and cardiovascular procedures

The ethnicity of patients hospitalised with heart failure was assessed using a combination of previously validated algorithms based on patients' surnames.24 25 The name lists by Quan et al24 and Lauderdale and Kestenbaum25 have been shown to have positive predicted values of 81% and 77% for Chinese and East Indian ethnicity, respectively. After excluding patients of other less common ethnicities, the remaining population was categorised into three groups: white, Chinese and East Indian.

Comorbid conditions of interest, such as diabetes and hypertension, were considered to be present at baseline if they were coded in the index hospitalisation record and in any inpatient or outpatient record in the preceding 365 days. Previously established ICD-9-CM and ICD-10 code-based definitions were used to identify comorbid conditions.22 26 A complete list of ICD-9-CM and ICD-10 codes used in the analyses is provided in online appendix 1.

The heart failure incidence date was calculated as the first occurrence of the diagnosis in an inpatient, outpatient, or physician office visit. Data on all hospitalisations starting from 1 April 1992; on all outpatient clinic visits starting from 1 April 1997; and on all physician office visits starting from 1 April 1993 were available for the study population. The duration of heart failure was calculated as the difference between the admission date for the index hospitalisation and the incidence date and was categorised as less than 1 year and 1 year and over.

The index hospitalisation and subsequent hospitalisations, outpatient records and physician claims within 1 year of the index hospitalisation were examined to identify patients who underwent specific cardiac procedures, including cardiac catheterisation, percutaneous coronary intervention (PCI), coronary artery bypass graft surgery, implantable cardioverter defibrillators, pacemakers and dialysis (see online appendix 1 for codes).

Outcomes of interest

The primary outcome of interest was 1-year mortality. Secondary outcomes related to resource utilisation included the number of re-hospitalisations, emergency department visits, outpatient clinic visits and physician office visits within 1 year following discharge. We also examined the rates of medications (such as ACE inhibitors, β-blockers and statins) among patients aged 65 years or older. A patient was considered to be on a drug if he/she had a pharmaceutical claim for the drug in the year before the index heart failure hospitalisation.

Statistical analysis

Continuous variables are presented as means with SD or medians with IQR and are compared across groups using analysis of variance and Kruskal–Wallis tests, respectively. Categorical variables are presented as percentages and are compared across groups using χ2 tests.

Kaplan–Meier survival curves of white, Chinese and East Indian patients were compared using the log-rank statistic. We used Cox proportional hazard regression model to examine the independent association between patient ethnicity and 1-year mortality after adjusting for differences in baseline characteristics. The model included age, sex, ethnicity, ischaemic heart disease, diabetes, hypertension, cerebrovascular disease, cancer, renal disease, peripheral vascular disease, dementia, chronic obstructive pulmonary disease, atrial fibrillation, duration of heart failure and median household income. Household income was based on census estimates of neighbourhood-level income and patients were categorised into quartiles of income. The cut-off values for the quartiles were derived from the entire population of Alberta.

Subgroup analyses

We undertook several subgroup analyses to examine whether the association of ethnicity and mortality was consistent across specific patient groups. The first subgroup consisted of patients with incident heart failure. For this analysis, the patient population was restricted to patients whose incident heart failure hospitalisation date was the same as the index hospitalisation date. In order to ensure that the association between ethnicity and mortality was not confounded by terminal illness, our second subgroup consisted of patients who did not have cancer. We also examined ethnic differences in mortality after excluding patients with renal disease and those who had either cancer or renal disease.

Finally, we re-examined the independent association between ethnicity and 1-year mortality after adjusting for medication use. For this analysis, we restricted the patient population to those aged 65 years and older (as mentioned previously, data on medication use is available only for the elderly). In addition to baseline characteristics, this model adjusted for differences in the rates of ACE inhibitors/angiotensin receptor blockers (ARB), β-blockers, statins, sprinolactone, warfarin, digoxin and diuretics across the three ethnic groups.

All analyses were conducted using SAS version 9.2. The health ethics board of the University of Alberta approved this study.

Results

Between 1 April 1999 and 31 December 2005 55 938 patients were hospitalised with heart failure, of whom 97% (54 208) were identified as either white (n=52 980), Chinese (n=851) or East Indian (n=377). The baseline characteristics of patients by ethnic group are presented in table 1. Chinese patients were the oldest and had the highest rates of renal disease. East Indian patients were the youngest, had the highest rates of ischaemic heart disease and diabetes and the lowest rates of dementia, peripheral vascular disease, chronic obstructive pulmonary disease, cancer and atrial fibrillation. White patients had the highest rates of dementia, chronic obstructive lung disease and the lowest rates of hypertension, renal disease, diabetes and lower household income.

Table 1

Baseline characteristics of patients hospitalised with heart failure by ethnicity

The mortality rate during the index heart failure hospitalisation was 13.9% among white patients, 21.4% among Chinese patients and 12.5% among East Indian patients (p<0.01). Figure 1 shows the Kaplan–Meier 1-year survival curves for each ethnic group. Cumulative 1-year mortality rates were 31.0% among white, 38.7% among Chines, and 26.5% among East Indian patients. The unadjusted and adjusted association between baseline characteristics and 1-year mortality are presented in figure 2. After baseline adjustment, Chinese patients had a significantly higher hazard of 1-year mortality compared with white patients (adjusted HR 1.34; 95% CI 1.20 to 1.49); however, the mortality hazard among East Indian patients was similar to those of white patients (adjusted HR 1.04; 95% CI 0.85 to 1.27).

Figure 1

Kaplan–Meier estimates of 1-year mortality among white, Chinese and East Indian patients hospitalised with heart failure.

Figure 2

Baseline characteristics associated with 1-year mortality among patients hospitalised with heart failure.

East Indian patients had the highest rates of cardiac catheterisation and PCI within 1 year of their index heart failure admission (table 2). Dialysis rates among Chinese and East Indian patients were more than twice those among white patients. White patients had higher rates of re-hospitalisations and emergency department visits; whereas Chinese and East Indian patients had higher numbers of visits to outpatient clinics and physicians offices. Medication rates among patients aged 65 years and older by ethnic group are presented in table 3. Rates of ACE inhibitors/ARB, β-blockers and statins were highest and digoxin lowest among East Indian patients. White patients had the highest rates of digoxin, warfarin and diuretics.

Table 2

Resource utilisation within 1 year of hospitalisation for heart failure by ethnicity

Table 3

Medication use among patients aged 65 years or older

The association between ethnicity and mortality among specific subgroups of patients is presented in table 4. The percentage of patients with new-onset (incident) heart failure was similar across the three ethnic groups (78% among white and 81% among Chinese and East Indian patients). Mortality differences among incident patients were similar to those observed in the overall cohort. The adjusted hazard of 1-year mortality remained significantly higher among Chinese patients compared with white patients among subgroups of patients with no cancer and/or renal disease. Differences in medication use did not appear to account for the differences in mortality rates across the three groups: the 1-year mortality hazard among elderly (age ≥65 years) Chinese patients compared with elderly white patients, after adjusting for age, sex, comorbidities and medication use, was 1.32 (95% CI 1.18 to 1.48). There continued to be no difference in mortality risk among elderly East Indian patients compared with elderly white patients (adjusted HR 1.03; 95% CI 0.83 to 1.27). The use of ACE inhibitors/ARB, β-blockers, statins and warfarin was associated with a lower hazard of 1-year mortality.

Table 4

Association between ethnicity and mortality among specific subgroup of patients

Discussion

To the best of our knowledge, this study is the first to examine 1-year mortality among white, Chinese and East Indian patients hospitalised with heart failure at a population-level. We found that Chinese patients had a significantly higher mortality rate at 1 year (39%) compared with white (31%) and East Indian patients (27%, p<0.01). The higher mortality risk among Chinese patients persisted after adjustment for baseline covariates (HR for Chinese compared with white patients 1.34; 95% CI 1.20 to 1.49), while the mortality risk among East Indian and white patients was nearly identical (HR for East Indians compared with whites 1.04; 95% CI 0.85 to 1.27). Even when our analysis was restricted to patients with a new diagnosis of heart failure, Chinese patients had the highest 1-year mortality rates (38% vs 30% among white and 25% among East Indian patients), which persisted after multivariate adjustment (adjusted HR for Chinese compared with white patients 1.35; 95% CI 1.19 to 1.52). The excess risk among Chinese patients was also maintained after adjusting for medication use in a subgroup of elderly patients aged 65 years or older (adjusted HR for Chinese compared with white patients 1.32; 95% CI 1.18 to 1.48). These mortality differences cannot be attributed to the differential access to health care as Chinese and East Indian patients had significantly more outpatient clinic and physician visits than their white counterparts.

Our study is particularly relevant given the increasing presence of ethnic minorities in North America and Europe. In Canada, according to the 2006 census, ethnic minorities accounted for 16.2% of the population.27 South Asian individuals (consisting of people from India, Pakistan, Sri Lanka and Bangladesh) and Chinese individuals are among the largest and fastest growing minority groups. It is therefore important to examine whether the outcomes of chronic diseases such as heart failure differ across ethnic groups so that treatment and management strategies can take into consideration patients' cultural and social environment.

Our finding of increased mortality among Chinese patients with heart failure is surprising given the results of previous studies examining ethnic differences in other patient populations. Among patients with end-stage renal disease, Hammelgarn et al28 found Chinese patients to have better survival after initiating dialysis treatment compared with white patients. Similarly, among patients who underwent coronary angiography, Quan et al29 found that Chinese patients were slightly less likely to undergo PCI or coronary artery bypass graft but had similar long-term mortality as whites. However, both those studies were focused on select populations undergoing specific procedures, which may have been subject to referral bias. In a recent study examining outcomes after acute myocardial infarction in Chinese, South Asian and white patients, Khan et al30 found Chinese patients had a higher 30-day mortality compared with white patients; however, there was no difference in 1-year mortality between the two groups.

East Indian patients were younger, had higher rates of diabetes and ischaemic heart disease and lower rates of atrial fibrillation at presentation. The clinical profile of East Indian patients in our study is similar to that reported by earlier studies.31 32 Studies have documented similar or improved survival among South Asian patients compared with white patients.13–15 There is also some evidence to suggest that East Indian patients may have improved survival compared with Chinese patient. In a study of heart failure mortality among Chinese, Indian and Malay individuals in Singapore, Indian patients had a lower mortality rate compared with Chinese patients.33

So what could explain the higher mortality among Chinese patients with heart failure? One potential explanation is the prevalence of comorbidities, which differed significantly across the three ethnic groups. The most striking of these was the high rate of renal disease among the Chinese (29%) compared with white (15%) and East Indian (19%) patients. The dialysis rate was similar among Chinese and East Indian individuals, and was twice the rate among white individuals. However, in a subgroup analysis excluding patients with renal disease, 1-year mortality among Chinese patients remained significantly higher compared with white patients. Similarly, ethnic differences in mortality remained after patients with cancer were excluded from the analysis. Therefore, differences in identified comorbidities do not appear to explain entirely the higher mortality observed among the Chinese: the 1-year mortality hazard among Chinese patients remained approximately 35% higher than white patients after adjusting for baseline differences.

It has been suggested that Chinese patients may not understand English well and may have issues with cultural assimilation.8 34 To the extent that these issues serve as barriers to appropriate and timely medical care, they could potentially explain the higher mortality. However, this does not explain why mortality among the Chinese was significantly higher than among East Indian individuals, who may experience similar issues. Also, our examination of resource utilisation in the year following the index hospitalisation reveals that rates of cardiovascular procedures among Chinese patients were similar or higher than among white patients. Although Chinese patients had lower rates of re-hospitalisation and emergency department visits, they had significantly higher rates of outpatient clinic and physician office visits. A similar pattern was observed among East Indian patients: they had higher rates of cardiac catheterisation, PCI, outpatient clinic and physician visits compared with white individuals. These data suggest that in a universal healthcare system, ethnic minorities do have similar access to healthcare resources.

A third potential explanation for the higher mortality rate is differences in medication use and disease management. In our examination of medication use among elderly patients (aged ≥65 years), rates of ACE inhibitors/ARB, β-blockers and statins were similar or slightly higher among Chinese patients compared with white patients. Chinese ethnicity continued to be associated with higher mortality compared with white patients even after adjusting for medication use. However, we have no information on the use of alternative therapies in our patient population. Chinese patients are more likely to rely on herbal and traditional Chinese medicines.34–36 The extent to which these interact with western medicine and their impact on outcomes is not fully understood.36

In our opinion, none of the potential explanations discussed above adequately explain our intriguing finding of higher mortality in Chinese patients with heart failure. One of the major limitations of our study is the unavailability of detailed laboratory (creatinine, sodium, haemoglobin, glucose) or clinical data (severity/type of heart failure, extent of left ventricular dysfunction, data on biomarkers or smoking status). Our study is therefore hypothesis generating, and future studies incorporating these important measures may be better able to identify the underlying causes of the observed ethnic differences. Another study limitation is the use of surname analysis to categorise patients into ethnic groups. Although this is an established methodology, it is subject to a certain degree of misclassification.32 37 38 We also assumed that patients who were not identified as belonging to an ethnic minority were white. A small percentage of these patients may belong to other ethnic groups. However, misclassification is likely to bias the results towards the null, and the mortality differences we observed may thus be underestimates of the true differences. A major strength of our study is that it is population based and includes a comprehensive assessment of ethnic differences in demographic, clinical, healthcare resource utilisation and socioeconomic factors. However, data on other measures of social position, such as employment and education, which may account for some of the observed differences in outcome, were unavailable.

Conclusion

Ethnicity is a modulator of mortality outcomes among patients with heart failure. Chinese patients have significantly higher mortality compared with both white and East Indian patients. Mortality among white and East Indian patients was similar. While there are differences in the prevalence of comorbidities across the three ethnic groups, they do not entirely explain the higher risk among Chinese patients. Further research is required to identify the reasons for the higher mortality among Chinese patients.

References

Footnotes

  • Funding This study was funded by a Canadian Institutes of Health Research (CIHR) operating grant. PK and JAE are supported by population health investigator awards and FAM and HQ by senior health scholar awards from Alberta Innovates – Health Solutions. JAE is supported by a CIHR new investigator award and FAM by a patient health management chair at the University of Alberta.

  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the health ethics board of the University of Alberta.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.