Quality of Life Research

, Volume 17, Issue 10, pp 1229–1238

Health-related quality of life in unselected outpatients with heart failure across Spain in two different health care levels. Magnitude and determinants of impairment: The INCA study

Authors

    • Medical DepartmentAstraZeneca Farmacéutica Spain, S.A.
  • Gaietà Permanyer-Miralda
    • Epidemiology Unit, Cardiology Department. Hospital Vall d’HebronCIBER de Epidemiología y Salud Pública (CIBERESP)
  • Carlos Brotons
    • Research UnitSardenya Primary Care Center
  • Joaquín Aznar
    • Hospital Nuestra Señora de Gracia
  • Eduardo Sobreviela
    • MediClin
Article

DOI: 10.1007/s11136-008-9397-3

Cite this article as:
de Rivas, B., Permanyer-Miralda, G., Brotons, C. et al. Qual Life Res (2008) 17: 1229. doi:10.1007/s11136-008-9397-3

Abstract

Aims

To assess health-related quality of life (HRQL) in Spanish outpatients with chronic heart failure (CHF).

Methods

Cross-sectional study carried out in a sample of CHF patients (echocardiography was performed in all of them) followed either in Primary Care (PC) centres or Cardiology outpatient clinics throughout Spain. HRQL was evaluated using the EuroQol 5D (EQ-5D) and Minnesota Living with Heart Failure (MLWHF) Questionnaire.

Results

The study subjects were 2161 CHF patients (1412 PC; 749 Cardiology). Patients were older and had more severe disease in PC than in Cardiology settings. Their HRQL scores were likewise worse. After adjusting for clinical variables, the differences in global and physical MLWHF disappeared, but persisted to a smaller degree in EQ-5D and mental MLWHF. HRQL was worse than in a representative sample of the Spanish population and in other chronic conditions such as rheumatoid arthritis or type 2 diabetes, being only comparable to severe chronic obstructive pulmonary disease (COPD).

Conclusion

All domains of HRQL were significantly impaired in CHF patients. Differences found in HRQL between PC and Cardiology should possibly be attributed to a large extent to the different clinical characteristics of the patients attended. In spite of the differences between EQ-5D and MLWHF, our results suggest that both questionnaires adequately reflect the severity of the disease.

Keywords

Health-related quality of lifeHeart failure

Introduction

Heart failure (HF) is an growing health problem around the world [1, 2]. Patients with HF face significant impairment in functional status, multiple hospital admissions, high mortality, multiple physical and psychological symptoms and a diminished quality of life [3, 4]. Although recent advances in therapy for HF have improved functional capacity and survival, it is becoming increasingly clear that, for many HF patients, improving quality of life is at least as important as the survival benefit that a pharmacological treatment may provide [5]. For this reason, increasingly, more studies are taking quality of life into account, and clinical trials have included its measurement to evaluate the effectiveness of different treatment strategies and the course of the disease.

Previous studies indicate that heart failure leads to significant impairment in all health-related quality of life (HRQL) dimensions and that patients with HF have worse HRQL than the general population and than patients with other chronic diseases [69]. Worse HRQL in patients with HF has also been associated with hospital readmission and death, suggesting that HRQL questionnaires could be a helpful tool to identify patients who are at increased risk of hospital readmission [1013].

However, most published data on HRQL have been obtained from selected, hospital-based patients participating in clinical trials. It is not known how representative they are of patients in the community. Few studies have reported the impact of HF on HRQL in the community, and there is even less information about the comparison of HRQL in HF patients attended in different health care settings.

The objectives of the INCA (Insuficiencia Cardíaca) study were to estimate the global impact of heart failure in a large number of unselected outpatients in Spain, to compare HRQL between patients followed in Primary Care (PC) and in Cardiology outpatient clinics, and to investigate the clinical determinants of HRQL. The study also evaluated the burden of HF on HRQL through comparisons of the EQ-5D assessments from the INCA with results from a sample of the Spanish general population and from patients with other chronic conditions.

Patients and methods

Participation in the study was offered to general practitioners and cardiologists in outpatients clinics all over Spain. Specific heart failure clinics were not included.

HF diagnosis was considered when Framingham criteria (2 major or 1 major and 2 minor) were present or there was a hospital discharge report with the diagnosis of HF. In both situations, an echocardiogram consistent with HF diagnosis was required. All patients were outpatients (≥18 years old) who had been stable for at least 3 months, had routine basic laboratory testing performed in the last year and were capable of answering the HRQL questionnaires.

The study was approved by the Clinical Research Ethics Committee of the Hospital Clinic in Barcelona. Data collection was undertaken after written informed consent was obtained from all patients.

About 100 cardiologists (1–2 in each Spanish province, according to population density), and to 480 PC physicians (4–5 for every cardiologist participant) were invited to take part in the study. The invitation was accepted by 93 cardiologists (93%) and 415 PC physicians (86.4%). Participating physicians were asked to include consecutive patients, 5 by PC physicians and 12 by cardiologists, who attended the clinic and fulfilled the inclusion criteria during June 2006. For the sake of simplicity, these two patient groups are mentioned as PC and CA, respectively.

All data needed for the study were obtained during a single visit. Patients completed the HRQL questionnaires during their office visit, and sociodemographic and clinical characteristics were obtained via patient interview and medical record review. Sociodemographic data collected were age, gender, marital status, educational level and living situation. Clinical data included NYHA classification, heart failure aetiology, time since diagnosis of HF, ejection fraction (preserved HF was considered with EF > 40%), hospitalisations during the last year, co-morbidities, data from the last laboratory test, blood pressure measurement and current medications. Doctors were asked to give their opinion about their patients’ compliance with treatment.

The patients’ HRQL was measured with a generic instrument, the EuroQol 5D (EQ-5D) [14], and an HF-specific instrument, the Minnesota Living with Heart Failure (MLWHF) questionnaire [15].

The EQ-5D is a self-administered, generic, health-related quality-of-life (HRQL) instrument with a 5-question multi-attribute questionnaire (EQ-5D descriptive) and a visual analogue self-rating scale (EQ-VAS). Respondents were asked to rate severity of their current problems (level 1 = no problems, level 2 = moderate problems and level 3 = severe problems) for five dimensions of health: mobility, self-care, usual activities, pain/discomfort and anxiety/depression. With this information, patients were classified into 243 (35) health statuses. EQ-5D health statuses may be converted into a EQ-5D index score ranging from −0.594 to 1.0 (where 1 is full health and 0 is dead) using a set of weighted preferences produced from the Spanish population [16]. The EQ-VAS consists of a 20-cm, vertical 0–100 scale, where 0 represents the worst imaginable health status and 100 represents the best imaginable health status. The respondents marked a point on the scale to reflect their overall health on the day of the interview.

The MLWHF questionnaire [15] is a 21-item disease-specific instrument whose aim is to determine how HF affects the physical, psychological and socio-economic condition of patients. The answer for each question is chosen from a scale of 0 (none) to 5 (very much). The total MLWHF score is obtained by adding the scores for all 21 items (range, 0–105); the higher the score, the worse the HRQL. In addition, it is possible to calculate a summary of the impact of HF on physical dimensions, constructed on the basis of 8 items (questions 2–7, 12–13; maximum score = 40), and another summary of its impact on emotional dimensions, constructed on the basis of 5 items (questions 17–21; maximum score = 25).

Both questionnaires have been previously translated into Spanish, used in Spain, and information regarding their validity has been published [11, 1719].

Statistical analysis

Data management was performed in the MediClin Biometrics Department, where data were entered into an Access 2002 database, with edit checks to ensure accurate entry of the data.

Statistical results were obtained using the statistical software SAS v.8.02 for Windows.

Categorical variables are given as frequencies and percentages, and continuous variables as mean ± standard deviation, median, percentiles 25–75, and maximum and minimum values. Two-sided confidence intervals at 95% level were also calculated.

Difference between the quality of life score in PC and CA

  • Significance related to EQ-5D and MLWHF dimensions was assessed with the chi-squared test or Fisher’s exact test.

  • Significance related to mean EQ-5D index, EQ-VAS, and MLWHF score was assessed with the Student’s t test for normally distributed continuous variables and the Wilcoxon test (Rank Sums) for non-normally distributed variables. In order to avoid the bias for confounding variables, an analysis of covariance adjusting for clinical variables (functional class, gender, age and non-cardiovascular co-morbidity) was performed.

Multivariate analysis

In order to identify the factors most strongly associated with Minnesota score and EQ-VAS score,
  1. (1)

    First, we analysed the mean of the scores for each of the factors. To evaluate significant differences among groups we used Student’s t test in case of two groups and a one-way analysis of variance in case of more than two groups (normal distribution was assumed, since population size was large enough).

     
  2. (2)

    Second, two multivariate linear regression models were fitted, one with Minnesota score as dependent variable and the other with the EQ-VAS score as dependent variable. Factors included in the model were those having a significant association with the dependent variable at greater than the 80% level (P < 0.2) in the bivariate analysis. Finally, the factors kept in the model were those with P-values less than 0.1. For each of the factors, least squared means with the two-sided 95% confidence intervals were calculated.

     

Difference between the quality of life score in INCA study and other studies

It was assessed with the Student’s t test for the mean of EQ-VAS and the chi-squared test or Fisher’s exact test for the EQ-5D dimensions.

The differences were considered statistically significant only if P < 0.05.

Results

Of the total population included in the INCA study (2,709 patients: 1,839 PC and 870 CA), 548 were excluded from the analysis, the majority of them due to absence of echocardiographic data. Some patients were also excluded when at least one of the questionnaires was not completed adequately. The remaining 2,161 (79.8%) patients (1,412 PC and 749 CA) were suitable for the objectives of the present study.

The baseline characteristics of patients enrolled in the study are shown in Table 1. To summarise, patients followed in CA were younger, there were more men, they had better functional class, lower ejection fraction, and fewer co-morbidities than those followed in PC. Although treatment specifically directed to HF seems to be better adjusted to guidelines recommendations in CA, blood pressure and diabetes mellitus were better controlled in PC.
Table 1

Baseline characteristics of the total population, patients followed in PC and by cardiologist

 

Total

PC

Cardiologist

P

Number of patients

2,161

1,412

749

 

Age, mean, (CI 95%)

70.9 (70.5, 71.4)

72.2 (71.7, 72.8)

68.5 (67.7, 69.3)

<0.0001

Sex (% male), (CI 95%)

55.62 (53.5, 57.7)

51.58 (49.0, 54.2)

63.32 (59.8, 66.8)

<0.0001

Living situation (%)

   

0.0017

  Alone

13.05 (11.6, 14.5)

14.46 (12.6, 16.3)

10.33 (8.08, 12.6)

 

  With family/friends

82.79 (81.2, 84.4)

80.69 (78.6, 82.8)

86.85 (84.4, 89.3)

 

  Hospice care

4.16 (3.30, 5.02)

4.85 (3.71, 5.99)

2.83 (1.61, 4.05)

 

Marital status (%)

   

<0.0001

  Single

6.29 (5.23, 7.34)

6.35 (5.05, 7.66)

6.16 (4.38, 7.94)

 

  Married

61.54 (59.4, 63.7)

57.1 (54.4, 59.8)

70.07 (66.7, 73.5)

 

  Widowed

30.06 (28.1, 32.1)

34.38 (31.8, 36.9)

21.78 (18.7, 24.8)

 

  Divorced

2.11 (1.49, 2.74)

2.17 (1.39, 2.95)

2.01 (0.97, 3.05)

 

NYHA functional class (%), (CI 95%)

   

<0.0001

  I

16.00 (14.4, 17.6)

16.05 (14.1, 18.0)

15.92 (13.3, 18.6)

 

  II

53.38 (51.2, 55.5)

49.48 (46.8, 52.1)

60.54 (57.0, 64.1)

 

  III

27.93 (26.0, 29.9)

30.92 (28.5, 33.4)

22.45 (19.4, 25.5)

 

  IV

2.68 (1.99, 3.38)

3.55 (2.56, 4.54)

1.09 (0.34, 1.84)

 

Left ventricular ejection fraction: mean, (CI 95%)

47.5 (46.8, 48.1)

49.5 (48.7, 50.3)

44.2 (43.2, 45.2)

<0.0001

% preserved EF (> 40%), (CI 95%)

61.7 (59.5, 63.9)

68.28 (65.6, 70.9)

51.09 (47.5, 54.7)

<0.0001

Hospitalisations for HF in the previous year: median, (IR)

0.0 (1.0)

0.0 (1.0)

0.0 (1.0)

0.0871

Co-morbidities (%), (CI 95%)

93.98 (93.0, 95.0)

96.74 (95.8, 97.7)

88.79 (86.5, 91.0)

<0.0001

Non CV co-morbidities (%), (CI 95%)

47.84 (45.8, 50.0)

51.7 (49.1, 54.3)

40.72 (37.2, 44.2)

<0.0001

Hypertension (%), (CI 95%)

76.35 (74.6, 78.1)

80.17 (78.1, 82.2)

69.16 (65.9, 72.5)

<0.0001

Blood pressure < 130/80 mmHg in hypertensive patients (%)

19.48 (16.9, 22.1)

22.4 (19.2, 25.6)

12,05 (8.00, 16.1)

0.0005

Diabetes (%), (CI 95%)

35.45 (33.4, 37.5)

36.54 (34.0, 39.1)

33.38 (30.0, 36.8)

0.1431

Diabetes control (HbA1c < 6.5%)

32.33 (28.9, 35.8)

36.33 (32.1, 40.6)

22.82 (17.1, 28.5)

0.0005

Anemia (Hb < 12 g/L) (%), (CI 95%)

15.27 (13.7, 16.8)

14.87 (13.0, 16.8)

16.02 (13.4, 18.7)

0.4851

Possible cause of HF (%), (CI 95%)

  Hypertensive heart disease

56.26 (54.2, 58.4)

60.34 (57.8, 62.9)

48.58 (45.0, 52.2)

<0.0001

  Coronary arterial disease

31.5 (29.5, 33.5)

27.44 (25.1, 29.8)

39.14 (35.6, 42.7)

<0.0001

  Heart valve disease

19.41 (17.7, 21.1)

21.26 (19.1, 23.4)

15.92 (13.3, 18.6)

0.003

  Dilated cardiomyopathy

17.25 (15.6, 18.9)

13.86 (12.0, 15.7)

23.62 (20.6, 26.7)

<0.0001

  Alcoholic cardiomyopathy

2.44 (1.78, 3.09)

1.94 (1.22, 2.66)

3.37 (2.07, 4.67)

<0.0001

  Other

5.77 (4.78, 6.76)

6.11 (4.85, 7.37)

5.13 (3.54, 6.72)

0.3562

HF treatment (%), (CI 95%)

  Diuretics

84.91 (83.4, 86.4)

83.78 (81.9, 85.7)

87.05 (84.6, 89.5)

0.0434

    Thiazide

26.28 (24.4, 28.1)

28.54 (26.2, 30.9)

22.03 (19.1, 25.0)

0.0011

    Loop

57.75 (55.7, 59.8)

53.4 (50.8, 56.0)

65.95 (62.6, 69.3)

0.0001

  Spironolactone

21.47 (19.7, 23.2)

19.12 (17.1, 21.2)

25.9 (22.8, 29.0)

0.0003

  Eplerenone

3.98 (3.16, 4.81)

0.99 (0.48, 1.51)

9.61 (7.50, 11.7)

<0.0001

  Digoxin

33.5 (31.5, 35.5)

34.63 (32.1, 37.1)

31.38 (28.1, 34.7)

>0.05

  Beta-blockers

43.96 (41.9, 46.1)

35.13 (32.6, 37.6)

60.61 (57.1, 64.1)

<0.0001

  ACEi

49.63 (47.5, 51.8)

46.88 (44.3, 49.5)

54.87 (51.3, 58.4)

0.0004

  ARB

45.83 (43.7, 47.9)

46.28 (43.7, 48.9)

44.99 (41.4, 48.6)

>0.05

  ACEi or ARB

89.4 (88.1, 90.7)

88.3 (86.7, 90.1)

91.32 (89.3, 93.3)

0.0348

Treatment compliance (%), (CI 95%)

92.97 (91.9, 94.1)

92.67 (91.3, 94.1)

93.54 (91.7, 95.3)

0.4631

CI Confidence intervals, NYHA New York Heart Association, EF ejection fraction, HF heart failure, CV cardiovascular, Hb haemoglobin, ACEi angiotensin converting enzyme inhibitor, ARB angiotensin receptor blocker

Quality of life

Response rates for quality of life assessments

More than 95% of patients completed all items of the EQ-5D descriptive, the visual analogue scale, and the MLWHF questionnaire.

Results from HRQL questionnaires

Table 2 shows the percentages of patients reporting any problems on the EQ-5D dimensions and the mean EQ-5D index, EQ-VAS, and MLWHF score. The percentage of patients with any problem on the five dimensions of the descriptive part of EQ-5D was higher among patients followed in PC than in patients followed in CA. HRQL mean scores for both EQ (index and VAS) and MLWHF (summary, physical and mental) were also worse in PC than in CA.
Table 2

Results of the HRQL questionnaires, comparing PC and Cardiologist

 

Total patients

PC

Cardiologist

P

EQ-5D descriptive [% patients with any problem, (CI 95%)]

Mobility

64.8 (62.8%, 66.9%)

66 (63.4%, 68.4%)

62.7 (59.3%, 66.2%)

0.18

Self-care

34 (32.0%, 36.1%)

37 (34.4%, 39.5%)

28.6 (25.4%, 31.9%)

<0.05

Usual activities

61.2 (59.2%, 63.4%)

65.2 (62.7%, 67.7%)

53.8 (50.3%, 57.4%)

<0.05

Pain/discomfort

59.6 (57.5%, 61.7%)

64.6 (62.1%, 67.1%)

50.1 (46.5%, 53.7%)

<0.05

Anxiety/depression

51.6 (49.5%, 53.8%)

55.2 (52.6%, 57.9%)

44.9 (41.3%, 48.5%)

<0.05

Mean EQ-5D index, (CI 95%) (min HRQL −0.594–max 1)

0.63 (0.62, 0.64)

0.60 (0.59, 0.61)

0.68 (0.66, 0.69)

<0.05

Mean EQ-VAS score, (CI 95%) (min HRQL 0–max 100)

57.6 (56.9, 58.3)

55.8 (54.9, 56.7)

61.0 (59.8, 62.2)

<0.05

Mean summary MLWHF score (min HRQL 105–max 0), (CI 95%)

39.9 (39.0, 40.8)

40.8 (39.8, 41.9)

38.1 (36.7, 39.6)

<0.05

Physical score (min 40–max 0)

17.8 (17.4, 18.2)

18.4 (17.9, 18.8)

16.8 (16.2, 17.5)

<0.05

Mental score (min 25–max 0)

9.3 (9.1, 9.6)

9.8 (9.5, 10.1)

8.4 (8.0, 8.8)

<0.05

HRQL Health-related quality of life, EQ-5D EuroQol 5D, CI confidence intervals, EQ-VAS EuroQol visual analogue scale, MLWHF Minnesota Living with Heart Failure questionnaire

After adjusting for clinical variables (functional class, gender, age and non-cardiovascular co-morbidity) the differences between PC and CA disappeared in the summary and physical MLWHF score, but persisted to a smaller degree in EQ-5D index, EQ-VAS and mental MLWHF score (Table 3).
Table 3

Adjusted results of the HRQL questionnaires, comparing PC and Cardiologist

 

PC

Cardiologist

P

Adjusteda mean EQ-5D index, (CI 95%)

0.54 (0.53, 0.56)

0.57 (0.55, 0.59)

<0.05

Adjusteda mean EQ-VAS score, (CI 95%)

52.5 (51.3, 53.7)

55.1 (53.7, 56.6)

<0.05

Adjusteda mean summary MLWHF score, (CI 95%)

   

Physical score

44.9 (43.5, 46.4)

45.1 (43.3, 46.8)

0.87

Mental score

20.1 (19.5, 20.7)

20.0 (19.3, 20.8)

0.80

 

10.8 (10.3, 11.2)

10.2 (9.7, 10.8)

0.03

aAdjusted for functional class, gender, age and non cardiovascular co-morbidity

HRQL Health-related quality of life, EQ-5D EuroQol 5D, CI confidence intervals, EQ-VAS EuroQol visual analogue scale, MLWHF Minnesota Living with Heart Failure questionnaire

Relationship between EQ-5D and MLWHF scores

The relationship between the MLWHF summary score and the EQ-5D index revealed an association with a Spearman correlation coefficient of −0.6664 (P < 0.0001). As expected, increasing MLWHF score was associated with a decrease in EQ-5D index score. The Spearman correlation coefficient between MLWHF summary score and EQ-VAS was −0.5801 (P < 0.0001) and between EQ-5D index and EQ-VAS was 0.6568 (P < 0.0001).

Clinical determinants of HRQL

The bivariate analysis showed significant associations between score (for both instruments) and functional class, gender (women had worse scores), age, type 2 diabetes, anaemia, number of hospital admissions in the previous year, left ventricular ejection fraction (worse in patients with FE ≤ 40%). Worse scores were also observed in widowed patients and those on hospice care.

In the multivariate analysis, there were 10 independent clinical determinants of HRQL in the MLWHF and 9 independent clinical determinants of HRQL in the EQ-VAS after adjusting for clinical factors (Tables 4 and 5). The main continuous determinants of HRQL in both instrument (higher MLWHF summary score and lower EQ-VAS score) included advanced age and lower ejection fraction. Categorical variables associated with HRQL included female gender, worse NYHA class, medical history of diabetes, renal dysfunction, and chronic obstructive pulmonary disease (COPD). Clinical determinants (only in the MLWHF) were medical history of hypertension, hyperkalaemia and heart valve disease as the cause of HF, and those only found in the EQ-VAS were: anemia and follow up in PC.
Table 4

Independent determinantsa of quality of life in heart failure patients (MLWHF)

 

Adjusted mean

Mean differences (CI 95%)

P

NYHA

I: 33.5

 

<0.0001

II, III, IV: 49.6

16.1 (12.6, 19.6)

COPD

No: 37.7

 

<0.0001

Yes: 45.1

7.1 (4.2, 10.1)

Gender

Male: 38.7

 

<0.0001

Female: 44.3

5.7 (3.1, 8.311)

Hypertension

No: 39.1

 

0.001

Yes: 44.0

4.9 (1.9, 7.9)

Serum potassium

<5:39.3

 

0.009

≥5:43.8

4.5 (1.1, 8.0)

Ejection fraction

≤40:43.6

 

0.001

>40:39.5

−4.1 (−6.7, −1.6)

Renal dysfunction

No: 39.5

 

0.06

Yes: 43.5

4.0 (−0.3, 8.2)

Valve disease

No: 39.6

 

0.01

Yes: 43.5

3.9 (0.7, 7.2)

Age

<75:39.7

 

0.004

 

≥75:43.3

3.6 (1.1, 6.0)

Diabetes mellitus

No: 40

 

0.04

 

Yes: 43

3.0 (0.1, 6.0)

R2 = 21.63

aFactors included in the model [association with the dependent variable >80% level (P < 0.2)]: gender, age, living situation, hypertension, COPD, diabetes mellitus, atrial fibrillation, health care setting, functional class (NYHA), heart failure aetiology (coronary arterial disease, heart valve disease), diabetes mellitus control, haemoglobin levels, potassium levels, renal dysfunction and ejection fraction.

Factors kept in the model (P < 0.1): NYHA, COPD, gender, hypertension, serum potassium, ejection fraction, renal dysfunction, valve disease, age, diabetes mellitus

MLWHF Minnesota Living with Heart Failure questionnaire, CI confidence intervals, NYHA New York Heart Association, COPD chronic obstructive pulmonary disease

Table 5

Independent determinantsa of quality of life in heart failure patients (EQ-VAS)

 

Adjusted mean

Mean differences (CI 95%)

P

NYHA

I: 57.9

 

<0.0001

II, III, IV: 49.9

−8.0 (−10.9, −5.2)

Gender

Male: 56.6

 

<0.0001

Female: 51.2

−5.4 (−7.6, −3.3)

Renal dysfunction

No: 56.5

 

0.0037

Yes: 51.2

−5.3 (−8.8, −1.7)

Clinical followed up

PC: 51.6

 

<0.0001

Cardiology: 56.26

4.6 (2.5, 6.8)

COPD

No: 56.2

 

0.0005

Yes: 51.6

−4.6 (−7.2, −2.0)

Ejection fraction

≤40:51.8

 

<0.0001

>40:56.1

4.4 (2.3, 6.6)

Diabetes mellitus

No: 56

 

0.0001

Yes: 51.9

−4.1 (−6.2, −2.0)

Age

<75:55.7

 

0.0006

≥75:52.1

−3.6 (−5.7, −1.5)

Haemoglobin

<12:52.5

 

0.04

≥12:55.3

2.9 (0.1, 5.6)

R2 = 17.88

aFactors included in the model [association with the dependent variable >80% level (P < 0.2)]: gender, age, living situation, hypertension, COPD, diabetes mellitus, atrial fibrillation, health care setting, functional class (NYHA), heart failure aetiology (dilated cardiomyopathy, heart valve disease), diabetes mellitus control, haemoglobin levels, potassium levels, renal dysfunction and ejection fraction

Factors kept in the model (P < 0.1): NYHA, gender, renal dysfunction, clinical followed up, COPD, ejection fraction, diabetes mellitus, age, haemoglobin

EQ-VAS EuroQol visual analogue scale, CI confidence intervals, NYHA New York Heart Association, COPD chronic obstructive pulmonary disease

Comparison of the EQ-5D scores of patients enrolled in the INCA study with Spanish populations norms and with patients with other chronic conditions

In order to assess the impact of chronic heart failure (CHF) on HRQL, we compared the mean EQ VAS score and the proportion of patients reporting any problem in EQ-5D descriptive dimensions in the INCA study with a representative sample of Spanish population [17] and with the results from available studies with other chronic conditions. This comparison indicates that HRQL of the HF patients enrolled in the INCA was worse than in a representative sample of the Spanish population [17], and was also apparently worse than in patients with other chronic conditions such as rheumatoid arthritis [20] or type 2 diabetes [21], being only comparable to the HRQL of patients with very severe COPD [22] (Table 6).
Table 6

Comparison of EQ-VAS INCA results with normal population and patients with chronic diseases other than HF

 

Mean age

Sex (male %)

EQ-VAS (mean ± SD)

P vs INCA

INCA

70.9 ± 10.6

55.6

57.6 ± 16.7

 

Age group

15–44 (63.7 ± 26.1)

45–64 (61.5 ± 15.5)

≥65 (56.4 ± 16.7)

Spanish population [17]

Age group

46.85

Total 71.1 ± 18

<0.05

15–44 (50.6%)

15–44 (76.8)

<0.05

45–64 (29.8%)

45–64 (68)

<0.05

≥65 (19.6%)

≥65 (60.6)

<0.05

Rheumatoid arthritis [20]

61.5 ± 25.9

21.7

65.0 ± 19.3

<0.05

COPD [22]

64.5 ± 8.4

73.0

  

Moderate

64 ± 8.4

71.4

67.7 ± 15.7

<0.05

Severe

65.6 ± 8.2

74.7

62.4 ± 17.0

<0.05

Very severe

61.6 ± 8.4

75.8

57.8 ± 16.2

0.91

Diabetes mellitus [21]

67.3 ± 10

44.5

61.7 ± 18.6

<0.05

EQ-VAS EuroQol visual analogue scale, HF heart failure, COPD chronic obstructive pulmonary disease

Discussion

The recent demonstration that HRQL scores in heart failure patients are related with the severity of the disease [7, 8], have prognostic value [1012] and are also modified by treatment with several drugs (ACEIs, BBs, ARBs) [2325] has aroused interest to evaluate if, in addition to the usual clinical variables, they can be used to identify patients who are at special risk for suffering an event (death/hospitalisation) and thus are candidates for closer follow-up and more intensive treatment.

However, most of the available evidence comes from clinical trials or studies carried out in the hospital setting in patients with acute or severe disease.

One of the original features of the INCA study is that it not only evaluates the impact of CHF on quality of life in stable patients, but also does so in the two health care settings where these patients are generally followed, PC centres and CA outpatient clinics. Additionally, data of the INCA study suggest that HF with preserved EF (EF > 40%) may be the most prevalent form of HF in outpatients in Spain (61.7%). Although the diagnosis of HF with preserved EF (strictly normal EF) has some limitations in our view this is a shortcoming of the heart failure definition itself, which is still unsatisfactory. There is agreement that some diagnostic inaccuracy in occasional individual patients is inescapable in the present state of our knowledge of HF even if, on the whole, the diagnosis of HF were correct for most patients when INCA criteria were followed. Besides, this finding agrees with recent data published in Spain were HF with preserved EF was present in 61% of patients [26].

HRQL and HF

The results of the INCA study show that CHF significantly affects all dimensions of quality of life in patients with stable disease who are attended in both PC and CA.

The significant impact of heart failure on quality of life has been previously shown in other studies, where it was also found that quality of life of patients with HF was worse than in the general population [7] and equally or more affected than in other chronic diseases [79]. A particular finding of the INCA study was that the impact of the disease on HRQL was only comparable to that reported by patients with very severe COPD [22].

In spite of the difference in age and sex between INCA patients and other chronic diseases, we believe that the differences in HRQL found are clinically valid because these differences may reflect real differences of the characteristics of patients with different diseases.

Determinants of HRQL

The scores obtained on the HRQL questionnaires were independently associated with several clinical variables: gender, age, severity of HF and presence of concomitant diseases.

Similarly to most previous studies, quality of life of the women included in the INCA study was worse than in men [7, 19, 27].

In the INCA study, as in the Spanish general population using the EQ-5D to measure HRQL [17], worse quality of life was observed in older patients. This result is not in agreement with most previous studies that have used EQ-5D or MLWHF which found no association [7] or even observed worse quality of life in younger patients [27]; only Parajón et al. [19] observed a weak trend to worse quality of life with increasing patient age. The association with age found in the INCA study may be related to the different characteristics of the population included (older age, greater co-morbidity and more patients in PC) compared to previous studies.

As in other studies, quality of life decreased as patients’ functional class worsened [79, 19, 27]. Functional class was also the major determinant of a worse quality of life. Juenger et al. [8] reported similar findings, and called attention to the fact that the alterations in functional capacity can only explain 51% of the impairment in quality of life, thus indicating that functional class and quality of life do not measure exactly the same dimension. Our study points to the same direction.

With regard to ejection fraction and its relationship to HRQL, the results of the studies published so far are contradictory. While in the INCA study and in another observational study [28] patients with depressed EF showed higher scores on the MLWHF, in clinical trials such as CHARM [27] no significant difference was seen between the scores obtained in patients with depressed or preserved EF. In view of the limitations (variability) of echocardiographic measurement of EF and the retrospective nature of this examination in the INCA study, the validity of the results in this subgroup may be questionable.

The clear association of HRQL with clinical determinants observed in the INCA suggests a close relationship between them. It should not be forgotten, however, that HRQL represents a set of dimensions relatively independent of these clinical determinants since they did not completely explain its variance in the analysis performed.

Difference in quality of life between the two health care settings

As with the clinical characteristics, there are differences in HRQL between patients with CHF attended in the different health care settings. This is one of the most important findings of the present study. The scores obtained indicate that patients attended in PC centres had worse HRQL than those attended in CA outpatient clinics. However, after adjusting for clinical variables (age, gender, functional class and non cardiovascular co-morbidity) the differences disappeared in the summary and physical MLWHF score and were attenuated in the EQ-5D and mental MLWHF score. The differences in the clinical characteristics of the patients attended in one setting or the other probably accounted for, to a large extent, the differences in quality of life observed. However, the small differences seen after adjustment, especially in the EQ-5D, also raise other possibilities of interpretation. First, it should be asked if a difference of 3 points on the scale has clinical significance. This may be so if we consider that this difference was observed in the multivariate analysis between groups with a clearly different clinical relevance. On the other hand, the fact that the differences disappeared on the specific questionnaire rather than on the generic one may suggest that the generic questionnaire captures general health perceptions better than the specific questionnaire, which focuses on the functional limitations specific to the disease. The fact that a small difference persists after adjustment in precisely the emotional dimension of MLWHF questionnaire is also consistent with this interpretation. Accordingly, patients attended in PC could have an impairment of global perception of their health, especially in the emotional sphere, slightly greater than that of patients attended in CA and independently of the clinical determinants analyzed.

Generic and specific HRQL measurement instruments

The use of a specific instrument complements the information provided by a generic instrument, providing information on how HF specifically affects patients’ quality of life.

A statistically significant correlation was found in the INCA study between the score obtained on the EQ-5D index and the MLWHF. Both questionnaires showed that they adequately reflected the severity of the disease. This good correlation between the two questionnaires has already been previously described [7]. An additional reported advantage of specific questionnaires (greater sensitivity to clinical change) falls beyond the scope of this cross-sectional study performed at a single moment in time.

In the INCA study, response rates on both the EQ-5D and the MLWHF were very high, possibly because investigators supervised that patients completed all questions on the questionnaires. This practice is to be recommended for avoiding the serious problem caused by missing data when evaluating the score obtained on the questionnaires.

The high response rate obtained on the EQ-5D in this and other studies, along with the simplicity and rapidity of its administration and the demonstration of its relationship with the score on the MLWHF questionnaire and the severity of heart failure, suggest that the EQ-5D may be a valid and acceptable way to measure quality of life of patients with heart failure particularly when the attending physicians, especially in PC, have little time available for performing them. However, further studies are needed to evaluate the prognostic value of the EQ-5D and its applicability to individual patients to determine its true clinical usefulness in the evaluation and follow-up of these patients. Their respective sensitivity to clinical change should also be evaluated.

Limitations

Given the cross-sectional nature of the design, the interpretation of the results of the burden of HF on the HRQL is restricted. Future research with a longitudinal approach would be valuable. However, the fact that the INCA study aimed to evaluate HRQL under usual clinical practice conditions, and that this was done in a wide sample of unselected and consecutively recruited patients collected by a large number of physicians (508) throughout Spain, suggests that the results presumably reflect HRQL of the population with stable CHF attended on an outpatient basis in Spain at a specific time. Some degree of unavoidable bias may have occurred in the inclusion of the patients in this study, as it is well known that the patients capable of completing a questionnaire are usually healthier than those who are not.

Conclusions

Despite improvements in the treatment for heart failure, all domains of HRQL are significantly impaired in CHF patients. Their HRQL is poorer than in the Spanish general population and, when the mean scores are compared with those found in other chronic conditions, they are only similar to those in very severe COPD. Differences found in HRQL between PC and CA should possibly be attributed to a large extent to the different clinical characteristics of patients attended in each health care setting. In spite of the differences between EQ-5D and MLWHF, our results suggest that both questionnaires adequately reflect the severity of the disease. Due to its simplicity of administration, the EQ-5D could be useful in evaluating quality of life of patients with CHF attended on an outpatient basis, although further studies are needed to evaluate the value of this questionnaire along time and its true clinical usefulness.

Copyright information

© Springer Science+Business Media B.V. 2008