EURO PEAN SO CIETY O F CARDIOLOGY ®
Original scientific paper
Positive affect moderates the effect of negative affect on cardiovascular disease-related hospitalizations and all-cause mortality after cardiac rehabilitation
European Journal of Preventive Cardiology 2015, Vol. 22(10) 1247–1253 ! The European Society of Cardiology 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/2047487314549745 ejpc.sagepub.com
Fiorenza Angela Meyer1,2, Roland von Ka¨nel2,3,4, Hugo Saner5, Jean-Paul Schmid6 and Stefanie Stauber7
Abstract Background: Little is known as to whether negative emotions adversely impact the prognosis of patients who undergo cardiac rehabilitation. We prospectively investigated the predictive value of state negative affect (NA) assessed at discharge from cardiac rehabilitation for prognosis and the moderating role of positive affect (PA) on the effect of NA on outcomes. Methods: A total of 564 cardiac patients (62.49 11.51) completed a comprehensive three-month outpatient cardiac rehabilitation program, filling in the Global Mood Scale (GMS) at discharge. The combined endpoint was cardiovascular disease (CVD)-related hospitalizations plus all-cause mortality at follow-up. Cox regression models estimated the predictive value of NA, as well as the moderating influence of PA on outcomes. Survival models were adjusted for sociodemographic factors, traditional cardiovascular risk factors, and severity of disease. Results: During a mean follow-up period of 3.4 years, 71 patients were hospitalized for a CVD-related event and 15 patients died. NA score (range 0–20) was a significant and independent predictor (hazard ratio (HR) 1.091, 95% confidence interval (CI) 1.012–1.175; p ¼ 0.023) with a three-point higher level in NA increasing the relative risk by 9.1%. Furthermore, PA interacted significantly with NA (p < 0.001). The relative risk of poor prognosis with NA was increased in patients with low PA (p ¼ 0.012) but remained unchanged in combination with high PA (p ¼ 0.12). Conclusion: The combination of NA with low PA was particularly predictive of poor prognosis. Whether reduction of NA and increase of PA, particularly in those with high NA, improves outcome needs to be tested.
Keywords Cardiovascular disease, cardiac rehabilitation, negative affect, positive affect, mortality, prospective longitudinal study Received 11 April 2014; accepted 12 August 2014
1 Division of Psychosomatic Medicine, Department of General Internal Medicine, Inselspital, Bern University Hospital and University of Bern, Switzerland 2 Department of Clinical Research, University of Bern, Switzerland 3 Department of Neurology, Inselspital, Bern University Hospital, and University of Bern, Switzerland 4 Department of Psychosomatic Medicine, Clinic Barmelweid, Barmelweid, Switzerland 5 Cardiovascular Prevention, Rehabilitation and Sports Medicine, Department of Cardiology, Inselspital, Bern University Hospital and University of Bern, Switzerland
6 Cardiology Clinic, Tiefenauspital, Bern University Hospital, Bern, Switzerland 7 Heart Failure and Transplantation, Department of Cardiology, Inselspital, Bern University Hospital and University of Bern, Switzerland
Corresponding author: Fiorenza Angela Meyer, Department of General Internal Medicine, Inselspital, Bern University Hospital, Freiburgstrasse, CH-3010 Bern, Switzerland. Email: [email protected]
Downloaded from cpr.sagepub.com at Karolinska Institutets Universitetsbibliotek on November 16, 2015
1248
European Journal of Preventive Cardiology 22(10)
Introduction Cardiovascular disease (CVD) is a major cause of mortality.1 Although there has been a decline in the number of CVD-related deaths during the last decade, the consequences of the disease and the hospital admission rate due to recurrent CVD-related events are still high.2 There are well-established modifiable and non-modifiable cardiovascular risk factors (CVRFs), including age, sex, family history, diabetes, hypertension, tobacco use, hypercholesterolemia, and obesity.3 In addition, there is compelling evidence for the role of psychosocial factors in increasing the risk of CVD.4,5 In particular, a broad spectrum of negative emotions such as depressive symptoms,6 anxiety,7 exhaustion,8 anger,9 Type D personality,10 and negative affect (NA)11 have been shown to independently increase the risk of a CVD-related event and/or mortality in different cardiac populations. A recently published position paper by Pogosova and colleagues provides a succinct overview regarding the role of different psychosocial risk factors in cardiac rehabilitation.12 Comparatively little is known about psychosocial factors with a cardioprotective potential13,14 such as positive affect (PA), commonly referring to mood states like cheerfulness, activity and joy.15 Importantly, PA is not the mere opposite of NA, because both affect states can be experienced simultaneously.16 High PA has been independently associated with a low incidence of CVD,17 whereas low PA predicted, for instance, major adverse cardiovascular events following percutaneous coronary intervention (PCI).18 Most of the above-mentioned studies explored risk factors for CVD-related events in cardiac patients who received standard care. We were particularly interested in investigating prognostic risk factors in patients who had undergone cardiac rehabilitation. Patients with CVD may benefit from rehabilitation in numerous ways.12,19 However, comparatively few studies have examined the prognostic value of psychosocial risk factors within a setting of cardiac rehabilitation.12,20 To our knowledge, this is the first study to assess the impact of NA plus the moderating role of PA on future CVD-related hospitalizations plus all-cause mortality within the setting of cardiac rehabilitation. We decided to measure NA at discharge from rehabilitation because previous studies had shown a stronger predictive value for scores measured at discharge compared with scores measured at admission.20 Instead of investigating individual aversive mood states such as, for instance, depression and anxiety, we decided to measure the broad spectrum of NA taken as a whole. For this purpose, we chose the Global Mood Scale (GMS), which allows assessment of different dimensions of NA, including depression, anxiety, exhaustion, and anger.21
The primary aim of our study was to examine the predictive value of NA assessed at discharge from outpatient cardiac rehabilitation on the combined endpoint of future CVD-related hospitalizations plus all-cause mortality. A secondary aim was to determine whether PA has any effect on the prognostic impact of NA. We hypothesized that NA is an independent predictor of poor prognosis in patients after cardiac rehabilitation. Furthermore, we assumed that PA can act as a ‘‘buffering’’ variable by mitigating the negative impact of NA on prognosis.
Methods Participants and study design Between January 2004 and December 2010, patients with coronary heart disease (CHD) and chronic heart failure (CHF) who had completed a comprehensive 8–12-week outpatient cardiac rehabilitation program at the Bern University Hospital, Switzerland were enrolled in the study. A more detailed description of the rehabilitation program can be found in the study by Blum and colleagues.22 The ethics committee of the state of Bern approved the study protocol and all participants gave informed consent. Patients completed questionnaires at discharge from rehabilitation. Demographic and medical data were collected through hospital charts and a physical examination. We conducted a follow-up investigation of 618 patients with a structured telephone interview in order to gather data on hospitalizations due to cardiovascular events that had occurred since discharge from cardiac rehabilitation. At least three attempts were made to contact patients who could not be reached initially. We lost 54 patients (8.7%) to follow-up. These exclusions led to a final sample of 564 patients with a complete data set for the analysis. Patients who had dropped out were similar to the final sample in parameters of sex, education, CVRFs, medication, disease severity, and NA and PA scores.
Assessment of the main outcome measure and covariates In order to measure state affect, participants completed the German version of the GMS.21 The GMS comprises 20 items yielding a NA and a PA subscale. Typical items are ‘‘insecure’’, ‘‘helpless’’, ‘‘feeble’’, or ‘‘weakened’’ under the NA subscale and ‘‘cheerful’’, ‘‘lively’’, ‘‘active’’, or ‘‘dynamic’’ under the PA subscale. Patients rate all items on a 5-point Likert scale according to their perceived mood state in the preceding week, ranging from 0 (not at all) to 4 (extremely).
Downloaded from cpr.sagepub.com at Karolinska Institutets Universitetsbibliotek on November 16, 2015
Meyer et al.
1249
We gathered information about sex, age, educational background, and the severity of CHD (i.e. number of diseased coronary vessels) during rehabilitation. At discharge from rehabilitation a medical history was taken, traditional CVRFs were assessed, and the body mass index (BMI) was calculated. We defined obesity (BMI >30 kg/m2) as a CVRF. We specified diabetes, hypercholesterolemia, smoking (defined as currently smoking at least one cigarette per day), positive family history of CHD, and hypertension as further CVRFs.
Follow-up assessment The follow-up period was defined as the time interval between discharge from rehabilitation and the semistructured telephone interview. The combined endpoint was defined as hospitalizations due to a CVD-related event plus all-cause mortality. A patient’s claim to have been hospitalized post-discharge from rehabilitation was confirmed by obtaining medical records from the attending general practitioner or from the hospital to which he/she was admitted.
Statistical analysis Data were analyzed using the SPSS 21.0 statistical software package (SPSS Inc., Chicago, IL). Significance level was set at p < 0.05 (2-tailed). Missing items (5%) were replaced by the expectation maximum algorithm.23 Patient characteristics were examined applying a median split on scores of NA and PA measured at discharge after rehabilitation. We calculated the group differences between high or low NA and high or low PA using Chi-square test or univariate analysis of variance (ANOVA) for categorical and continuous variables, respectively. We ran Cox proportional hazard models to estimate the predictive value of NA for the relative risk of the combined endpoint hospitalizations due to a CVD-related event plus all-cause mortality that had occurred during follow-up. The GMS measures affect continuously, so the relative risk would refer to a one-point higher score on the NA scale. As a one-point interval is unlikely to be of clinical relevance, we expressed the change in the hazard ratio for a three-point higher level on the NA scale and also for age intervals of five years. Furthermore, we added PA and the NA-by-PA interaction to the model to test for a moderating effect of PA on NA. In the case of a significant interaction, we applied the approach described by Holmbeck24 to identify the effect of high PA (1.5 SD above the sample mean) versus low PA (1.5 SD below the sample mean) on the predictive value of NA. We measured the correlation between PA and NA with Kendall’s tau-b. Control variables were defined a priori due to their known impact on prognosis.
To prevent model overfitting, given the number of outcomes,25 we performed data collapsing for CVRFs by assigning one point each for obesity, diabetes, hypercholesterolemia, smoking, positive family history, and hypertension. For analysis, these points were added up to form a total CVRF score ranging from 0 to 6. Missing CVRFs at discharge were replaced by CVRFs obtained at the start of rehabilitation. The final model was adjusted for sex, age, educational background, the CVRF score, and disease severity.
Results Patient characteristics Table 1 presents the sociodemographic and clinical variables for the total sample as well as for the subgroups with (a) low NA (NA) plus low PA (PA), (b) NA plus high PA (PAþ), (c) high NA (NAþ) plus PA, and (d) NAþ plus PAþ. Cut-off value for the median split was 8 for the subscale NA and 27 for PA. In terms of sociodemographic variables and clinical parameters, patients with NA plus PA were significantly older than those in the NA plus PAþ subgroup. There were no other group differences except for the expected ones in NA and PA.
Mortality and CVD-related events On average, the follow-up telephone interview was conducted 41.0 13.6 months (range 11.9–77.5) after discharge from rehabilitation. Of the 564 patients, 71 (12.6%) had been hospitalized for a CVD-related event and 15 had died (2.6%) during the follow-up period. Five deaths occurred due to a CVD-related event, while causes of mortality for the remainder could not be verified. The CVD-related event occurred on average 22.6 16.6 months (range 0.3–64.9) after discharge. We registered 16 different CVD-related events resulting in hospitalizations: these were elective PCI (40.7%), non-elective PCI (9.3%), cardiac arrhythmia (4.6%), cerebrovascular insult (5.8%), sudden cardiac arrest (4.6%), pacemaker implantation (4.6%), non-elective coronary angiography without intervention (2.3%), CHF (2.3%), pulmonary embolism (3.5%), endocarditis (2.3%), coronary artery bypass surgery (1.1%), hypertensive crisis (2.3%), aneurysm (1.1%), Takotsubo cardiomyopathy (1.1%), and conduit change (1.1%).
Cox regression model NA at discharge from rehabilitation was a significant predictor for CVD-related hospitalizations plus allcause mortality during follow-up (hazard ratio (HR)
Downloaded from cpr.sagepub.com at Karolinska Institutets Universitetsbibliotek on November 16, 2015
1250
European Journal of Preventive Cardiology 22(10)
Table 1. Characteristics of patients at discharge from rehabilitation. Subgroups NA Variable Age Male gender (%) Highest level of education (%) Primary school Vocational training College or university Diabetes (%) Hypercholesterolemia (%) Smoking (%) Hypertension (%) Positive family history (%) Obesity (BMI>30) (%) 0 -, 1 -, 2 -, 3 -vessel disease (%) Aspirin (%) Beta blockers (%) Statins (%) Diuretics (%) Angiotensin-converting enzyme inhibitors (%) Clopidogrel (%) CVD-related hospitalizations or death (%) Negative affect, M SD Positive affect M SD
Total sample (N ¼ 564)
NAþ
PA (N ¼ 46) b
PAþ (N ¼ 230)
PA (N ¼ 223)
PAþ (N ¼ 65)
p-value
61.29 10.67 83.9
62.98 12.26 76.4
62.08 11.73 76.9
0.024 0.226 0.783
62.50 11.49 79.7
66.78 10.39 78.3
5.5 70.3 24.2 12.7 66.1 42.2 61.8 29.9 18.6 16.8, 33.7, 21.6, 27.9 86.6 83.2 85.2 19.1 64.3
4.3 73.9 21.7 15.2 73.9 41.3 71.7 28.3 19.6 13.0, 28.3, 23.9, 34.8 91.3 89.1 91.3 26.1 69.6
3.0 73.0 23.9 8.3 67.8 39.6 62.6 28.7 17.0 18.3, 33.5, 20.9, 27.4 86.1 82.2 86.1 15.2 61.3
8.9 66.2 24.9 16.4 64.4 44.0 61.3 28.4 18.2 15.1, 34.7, 21.8, 28.4 87.6 83.6 84.4 19.1 66.7
3.1 72.3 24.6 13.8 60.0 46.2 53.8 40.0 24.6 20.0, 35.4, 21.5, 23.1 81.5 81.5 80.0 27.7 63.1
58.5 15.2
58.7 10.9
57.4 14.8
58.7 16.4
61.5 15.4
9.33 7.85 25.97 7.04
4.30 2.07 b,c 21.74 5.27 a,c
2.77 2.24 d,e 31.41 3.25 d
16.55 6.51f 19.98 5.54f
11.14 3.93 30.42 2.70
0.064 0.407 0.711 0.291 0.308 0.571 0.397 0.470 0.688 0.395 0.078 0.565 0.947 0.809
Original scientific paper
Positive affect moderates the effect of negative affect on cardiovascular disease-related hospitalizations and all-cause mortality after cardiac rehabilitation
European Journal of Preventive Cardiology 2015, Vol. 22(10) 1247–1253 ! The European Society of Cardiology 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/2047487314549745 ejpc.sagepub.com
Fiorenza Angela Meyer1,2, Roland von Ka¨nel2,3,4, Hugo Saner5, Jean-Paul Schmid6 and Stefanie Stauber7
Abstract Background: Little is known as to whether negative emotions adversely impact the prognosis of patients who undergo cardiac rehabilitation. We prospectively investigated the predictive value of state negative affect (NA) assessed at discharge from cardiac rehabilitation for prognosis and the moderating role of positive affect (PA) on the effect of NA on outcomes. Methods: A total of 564 cardiac patients (62.49 11.51) completed a comprehensive three-month outpatient cardiac rehabilitation program, filling in the Global Mood Scale (GMS) at discharge. The combined endpoint was cardiovascular disease (CVD)-related hospitalizations plus all-cause mortality at follow-up. Cox regression models estimated the predictive value of NA, as well as the moderating influence of PA on outcomes. Survival models were adjusted for sociodemographic factors, traditional cardiovascular risk factors, and severity of disease. Results: During a mean follow-up period of 3.4 years, 71 patients were hospitalized for a CVD-related event and 15 patients died. NA score (range 0–20) was a significant and independent predictor (hazard ratio (HR) 1.091, 95% confidence interval (CI) 1.012–1.175; p ¼ 0.023) with a three-point higher level in NA increasing the relative risk by 9.1%. Furthermore, PA interacted significantly with NA (p < 0.001). The relative risk of poor prognosis with NA was increased in patients with low PA (p ¼ 0.012) but remained unchanged in combination with high PA (p ¼ 0.12). Conclusion: The combination of NA with low PA was particularly predictive of poor prognosis. Whether reduction of NA and increase of PA, particularly in those with high NA, improves outcome needs to be tested.
Keywords Cardiovascular disease, cardiac rehabilitation, negative affect, positive affect, mortality, prospective longitudinal study Received 11 April 2014; accepted 12 August 2014
1 Division of Psychosomatic Medicine, Department of General Internal Medicine, Inselspital, Bern University Hospital and University of Bern, Switzerland 2 Department of Clinical Research, University of Bern, Switzerland 3 Department of Neurology, Inselspital, Bern University Hospital, and University of Bern, Switzerland 4 Department of Psychosomatic Medicine, Clinic Barmelweid, Barmelweid, Switzerland 5 Cardiovascular Prevention, Rehabilitation and Sports Medicine, Department of Cardiology, Inselspital, Bern University Hospital and University of Bern, Switzerland
6 Cardiology Clinic, Tiefenauspital, Bern University Hospital, Bern, Switzerland 7 Heart Failure and Transplantation, Department of Cardiology, Inselspital, Bern University Hospital and University of Bern, Switzerland
Corresponding author: Fiorenza Angela Meyer, Department of General Internal Medicine, Inselspital, Bern University Hospital, Freiburgstrasse, CH-3010 Bern, Switzerland. Email: [email protected]
Downloaded from cpr.sagepub.com at Karolinska Institutets Universitetsbibliotek on November 16, 2015
1248
European Journal of Preventive Cardiology 22(10)
Introduction Cardiovascular disease (CVD) is a major cause of mortality.1 Although there has been a decline in the number of CVD-related deaths during the last decade, the consequences of the disease and the hospital admission rate due to recurrent CVD-related events are still high.2 There are well-established modifiable and non-modifiable cardiovascular risk factors (CVRFs), including age, sex, family history, diabetes, hypertension, tobacco use, hypercholesterolemia, and obesity.3 In addition, there is compelling evidence for the role of psychosocial factors in increasing the risk of CVD.4,5 In particular, a broad spectrum of negative emotions such as depressive symptoms,6 anxiety,7 exhaustion,8 anger,9 Type D personality,10 and negative affect (NA)11 have been shown to independently increase the risk of a CVD-related event and/or mortality in different cardiac populations. A recently published position paper by Pogosova and colleagues provides a succinct overview regarding the role of different psychosocial risk factors in cardiac rehabilitation.12 Comparatively little is known about psychosocial factors with a cardioprotective potential13,14 such as positive affect (PA), commonly referring to mood states like cheerfulness, activity and joy.15 Importantly, PA is not the mere opposite of NA, because both affect states can be experienced simultaneously.16 High PA has been independently associated with a low incidence of CVD,17 whereas low PA predicted, for instance, major adverse cardiovascular events following percutaneous coronary intervention (PCI).18 Most of the above-mentioned studies explored risk factors for CVD-related events in cardiac patients who received standard care. We were particularly interested in investigating prognostic risk factors in patients who had undergone cardiac rehabilitation. Patients with CVD may benefit from rehabilitation in numerous ways.12,19 However, comparatively few studies have examined the prognostic value of psychosocial risk factors within a setting of cardiac rehabilitation.12,20 To our knowledge, this is the first study to assess the impact of NA plus the moderating role of PA on future CVD-related hospitalizations plus all-cause mortality within the setting of cardiac rehabilitation. We decided to measure NA at discharge from rehabilitation because previous studies had shown a stronger predictive value for scores measured at discharge compared with scores measured at admission.20 Instead of investigating individual aversive mood states such as, for instance, depression and anxiety, we decided to measure the broad spectrum of NA taken as a whole. For this purpose, we chose the Global Mood Scale (GMS), which allows assessment of different dimensions of NA, including depression, anxiety, exhaustion, and anger.21
The primary aim of our study was to examine the predictive value of NA assessed at discharge from outpatient cardiac rehabilitation on the combined endpoint of future CVD-related hospitalizations plus all-cause mortality. A secondary aim was to determine whether PA has any effect on the prognostic impact of NA. We hypothesized that NA is an independent predictor of poor prognosis in patients after cardiac rehabilitation. Furthermore, we assumed that PA can act as a ‘‘buffering’’ variable by mitigating the negative impact of NA on prognosis.
Methods Participants and study design Between January 2004 and December 2010, patients with coronary heart disease (CHD) and chronic heart failure (CHF) who had completed a comprehensive 8–12-week outpatient cardiac rehabilitation program at the Bern University Hospital, Switzerland were enrolled in the study. A more detailed description of the rehabilitation program can be found in the study by Blum and colleagues.22 The ethics committee of the state of Bern approved the study protocol and all participants gave informed consent. Patients completed questionnaires at discharge from rehabilitation. Demographic and medical data were collected through hospital charts and a physical examination. We conducted a follow-up investigation of 618 patients with a structured telephone interview in order to gather data on hospitalizations due to cardiovascular events that had occurred since discharge from cardiac rehabilitation. At least three attempts were made to contact patients who could not be reached initially. We lost 54 patients (8.7%) to follow-up. These exclusions led to a final sample of 564 patients with a complete data set for the analysis. Patients who had dropped out were similar to the final sample in parameters of sex, education, CVRFs, medication, disease severity, and NA and PA scores.
Assessment of the main outcome measure and covariates In order to measure state affect, participants completed the German version of the GMS.21 The GMS comprises 20 items yielding a NA and a PA subscale. Typical items are ‘‘insecure’’, ‘‘helpless’’, ‘‘feeble’’, or ‘‘weakened’’ under the NA subscale and ‘‘cheerful’’, ‘‘lively’’, ‘‘active’’, or ‘‘dynamic’’ under the PA subscale. Patients rate all items on a 5-point Likert scale according to their perceived mood state in the preceding week, ranging from 0 (not at all) to 4 (extremely).
Downloaded from cpr.sagepub.com at Karolinska Institutets Universitetsbibliotek on November 16, 2015
Meyer et al.
1249
We gathered information about sex, age, educational background, and the severity of CHD (i.e. number of diseased coronary vessels) during rehabilitation. At discharge from rehabilitation a medical history was taken, traditional CVRFs were assessed, and the body mass index (BMI) was calculated. We defined obesity (BMI >30 kg/m2) as a CVRF. We specified diabetes, hypercholesterolemia, smoking (defined as currently smoking at least one cigarette per day), positive family history of CHD, and hypertension as further CVRFs.
Follow-up assessment The follow-up period was defined as the time interval between discharge from rehabilitation and the semistructured telephone interview. The combined endpoint was defined as hospitalizations due to a CVD-related event plus all-cause mortality. A patient’s claim to have been hospitalized post-discharge from rehabilitation was confirmed by obtaining medical records from the attending general practitioner or from the hospital to which he/she was admitted.
Statistical analysis Data were analyzed using the SPSS 21.0 statistical software package (SPSS Inc., Chicago, IL). Significance level was set at p < 0.05 (2-tailed). Missing items (5%) were replaced by the expectation maximum algorithm.23 Patient characteristics were examined applying a median split on scores of NA and PA measured at discharge after rehabilitation. We calculated the group differences between high or low NA and high or low PA using Chi-square test or univariate analysis of variance (ANOVA) for categorical and continuous variables, respectively. We ran Cox proportional hazard models to estimate the predictive value of NA for the relative risk of the combined endpoint hospitalizations due to a CVD-related event plus all-cause mortality that had occurred during follow-up. The GMS measures affect continuously, so the relative risk would refer to a one-point higher score on the NA scale. As a one-point interval is unlikely to be of clinical relevance, we expressed the change in the hazard ratio for a three-point higher level on the NA scale and also for age intervals of five years. Furthermore, we added PA and the NA-by-PA interaction to the model to test for a moderating effect of PA on NA. In the case of a significant interaction, we applied the approach described by Holmbeck24 to identify the effect of high PA (1.5 SD above the sample mean) versus low PA (1.5 SD below the sample mean) on the predictive value of NA. We measured the correlation between PA and NA with Kendall’s tau-b. Control variables were defined a priori due to their known impact on prognosis.
To prevent model overfitting, given the number of outcomes,25 we performed data collapsing for CVRFs by assigning one point each for obesity, diabetes, hypercholesterolemia, smoking, positive family history, and hypertension. For analysis, these points were added up to form a total CVRF score ranging from 0 to 6. Missing CVRFs at discharge were replaced by CVRFs obtained at the start of rehabilitation. The final model was adjusted for sex, age, educational background, the CVRF score, and disease severity.
Results Patient characteristics Table 1 presents the sociodemographic and clinical variables for the total sample as well as for the subgroups with (a) low NA (NA) plus low PA (PA), (b) NA plus high PA (PAþ), (c) high NA (NAþ) plus PA, and (d) NAþ plus PAþ. Cut-off value for the median split was 8 for the subscale NA and 27 for PA. In terms of sociodemographic variables and clinical parameters, patients with NA plus PA were significantly older than those in the NA plus PAþ subgroup. There were no other group differences except for the expected ones in NA and PA.
Mortality and CVD-related events On average, the follow-up telephone interview was conducted 41.0 13.6 months (range 11.9–77.5) after discharge from rehabilitation. Of the 564 patients, 71 (12.6%) had been hospitalized for a CVD-related event and 15 had died (2.6%) during the follow-up period. Five deaths occurred due to a CVD-related event, while causes of mortality for the remainder could not be verified. The CVD-related event occurred on average 22.6 16.6 months (range 0.3–64.9) after discharge. We registered 16 different CVD-related events resulting in hospitalizations: these were elective PCI (40.7%), non-elective PCI (9.3%), cardiac arrhythmia (4.6%), cerebrovascular insult (5.8%), sudden cardiac arrest (4.6%), pacemaker implantation (4.6%), non-elective coronary angiography without intervention (2.3%), CHF (2.3%), pulmonary embolism (3.5%), endocarditis (2.3%), coronary artery bypass surgery (1.1%), hypertensive crisis (2.3%), aneurysm (1.1%), Takotsubo cardiomyopathy (1.1%), and conduit change (1.1%).
Cox regression model NA at discharge from rehabilitation was a significant predictor for CVD-related hospitalizations plus allcause mortality during follow-up (hazard ratio (HR)
Downloaded from cpr.sagepub.com at Karolinska Institutets Universitetsbibliotek on November 16, 2015
1250
European Journal of Preventive Cardiology 22(10)
Table 1. Characteristics of patients at discharge from rehabilitation. Subgroups NA Variable Age Male gender (%) Highest level of education (%) Primary school Vocational training College or university Diabetes (%) Hypercholesterolemia (%) Smoking (%) Hypertension (%) Positive family history (%) Obesity (BMI>30) (%) 0 -, 1 -, 2 -, 3 -vessel disease (%) Aspirin (%) Beta blockers (%) Statins (%) Diuretics (%) Angiotensin-converting enzyme inhibitors (%) Clopidogrel (%) CVD-related hospitalizations or death (%) Negative affect, M SD Positive affect M SD
Total sample (N ¼ 564)
NAþ
PA (N ¼ 46) b
PAþ (N ¼ 230)
PA (N ¼ 223)
PAþ (N ¼ 65)
p-value
61.29 10.67 83.9
62.98 12.26 76.4
62.08 11.73 76.9
0.024 0.226 0.783
62.50 11.49 79.7
66.78 10.39 78.3
5.5 70.3 24.2 12.7 66.1 42.2 61.8 29.9 18.6 16.8, 33.7, 21.6, 27.9 86.6 83.2 85.2 19.1 64.3
4.3 73.9 21.7 15.2 73.9 41.3 71.7 28.3 19.6 13.0, 28.3, 23.9, 34.8 91.3 89.1 91.3 26.1 69.6
3.0 73.0 23.9 8.3 67.8 39.6 62.6 28.7 17.0 18.3, 33.5, 20.9, 27.4 86.1 82.2 86.1 15.2 61.3
8.9 66.2 24.9 16.4 64.4 44.0 61.3 28.4 18.2 15.1, 34.7, 21.8, 28.4 87.6 83.6 84.4 19.1 66.7
3.1 72.3 24.6 13.8 60.0 46.2 53.8 40.0 24.6 20.0, 35.4, 21.5, 23.1 81.5 81.5 80.0 27.7 63.1
58.5 15.2
58.7 10.9
57.4 14.8
58.7 16.4
61.5 15.4
9.33 7.85 25.97 7.04
4.30 2.07 b,c 21.74 5.27 a,c
2.77 2.24 d,e 31.41 3.25 d
16.55 6.51f 19.98 5.54f
11.14 3.93 30.42 2.70
0.064 0.407 0.711 0.291 0.308 0.571 0.397 0.470 0.688 0.395 0.078 0.565 0.947 0.809
