The impact of socio-demographic factors on health-related quality of life after coronary artery bypass surgery

Abstract

OBJECTIVES

To achieve a beneficial impact on long-term outcome after coronary artery bypass grafting (CABG), the goal of the present study was the early identification of patients at risk of impaired postoperative health-related quality of life (HRQoL), particularly evaluating the significance of socio-demographic variables.

METHODS

In this prospective, single-centre cohort study of patients having an isolated CABG (January 2004–December 2014), preoperative socio-demographic (preSOC) and preoperative medical variables as well as 6-month follow-up data including the Nottingham Health Profile were analysed in 3,237 patients.

RESULTS

All preSOC (gender, age, marriage and employment) and follow-up (chest pain, dyspnoea) variables proved to have significant influence on HRQoL (P < 0.001), male patients below 60 years being particularly impaired. The effects of marriage and employment on HRQoL are modulated by age and gender. The significance of the predictors of reduced HRQoL differs between the 6 Nottingham Health Profile domains. Multivariable regression analyses revealed explained proportions of variance amounting to 7% for preSOC and 4% for preoperative medical variables.

CONCLUSIONS

The identification of patients at risk of impaired postoperative HRQoL is decisive for providing additional support. This study reveals that the assessment of 4 preoperative socio-demographic characteristics (age, gender, marriage, employment) is more predictive of HRQoL after CABG than are multiple medical variables.

The major goals of coronary artery bypass grafting (CABG) are the prevention or treatment of myocardial ischaemia and relief from angina and dyspnoea and, thereby, the improvement of health-related quality of life (HRQoL).

INTRODUCTION

The major goals of coronary artery bypass grafting (CABG) are the prevention or treatment of myocardial ischaemia and relief from angina and dyspnoea and, thereby, the improvement of health-related quality of life (HRQoL). Previous studies have demonstrated that isolated CABG enhances postoperative HRQoL [1–3], with further improvement provided by cardiac rehabilitation programs [4]. However, some patients report remarkably reduced HRQoL independently of the medical success of CABG and the aftercare [5].

To identify those patients with long-lasting reduction of HRQoL, assessment 6 months after surgery is reasonable, because the rehabilitation program should be finished, and physical recovery should have occurred in the majority of patients [6]. For this purpose, the Nottingham Health Profile (NHP) is an established tool in cardiac surgery [1, 2, 7, 8].

It is crucial to identify patients at risk of impaired postoperative HRQoL before the impairment actually becomes evident—ideally preoperatively—because early additional interventions should be considered. For instance, Aguayo et al. reported that depression in 10–40% of patients having cardiac surgery was associated with a significantly worse outcome; thus, the early detection and management of depression are decisive in enhancing postoperative recovery and HRQoL by implementation of routine depression screening tools [9]. A study by Zywot et al. suggested that preoperative risk stratification regarding 30-day readmission is crucial to improve patient outcomes and thus HRQoL [10].

Furthermore, previous studies have reported associations between medical or socio-demographic preoperative variables and postoperative HRQoL in patients having cardiac surgery [1, 3, 7, 11, 12]. Nevertheless, the comparative significance of the different variables with regard to HRQoL affection remains unclear.

The goal of the present study was to determine preoperative risk factors for reduced postoperative HRQoL in a large cohort of patients having isolated CABG. In particular, the significance of socio-demographic variables in comparison to multiple medical characteristics was evaluated, additionally revealing the effects on each of the 6 NHP domains separately.

METHODS

Patients

A prospectively observed cohort of consecutive patients having isolated CABG (Heart Center Lahr/Baden, Germany; January 2004–December 2014) was contacted for a 6-month follow-up (FU) survey. For patient inclusion, data for each of the 6 NHP domains and for all preoperative socio-demographic variables (preSOC), preoperative medical variables (preMED) and FU variables had to be available (Supplementary Table 1), resulting in a study sample of 3,237 patients [CONSORT (CONsolidated Standards of Reporting Trials) diagram] (Supplementary Fig. 1). The independent variables were chosen according to previously published studies indicating their potential impact on HRQoL in patients having cardiac surgery [7, 8, 11–13].

Ethics statement

This study is part of a research project on HRQoL in cardiac surgery, approved by the regional ethics committee, conducted in accordance with the declaration of Helsinki and registered at the Deutsche Register Klinischer Studien (DRKS00018068). All participants signed an informed written consent form.

Health-related quality of life assessment

All patients received a questionnaire covering the symptoms chest pain and dyspnoea and additionally the first part of the NHP, a standardized international tool for the assessment of HRQoL [14]. The NHP is separated into 6 domains (emotional reaction, energy level, pain, physical abilities, sleep and social isolation) with domain scores ranging from 0 to 100, in which higher scores indicate worse HRQoL. The mean values of the NHP domain scores were defined as the primary outcome of the study.

Statistics

After calculating the patients’ NHP domain scores, we subtracted the age- and gender-specific norm values [15] as recently published [8], receiving the individual deviations (normalized NHP domain scores). Finally, we calculated the mean deviation of all 6 domains (normalized NHP score), receiving an HRQoL score normalized to age- and gender-matched healthy persons, in which negative values represent better HRQoL.

Descriptive statistics present continuous variables as mean ± standard error of the mean and categorical variables as number of patients and percentage. The univariable statistics comprise independent sample t-tests with Welch’s correction if indicated or one-way analysis of variance for categorical variables and linear regression analyses for continuous variables, each with normalized NHP scores as the dependent variable. The subgroup analyses regarding interactions between socio-demographic risk factors were conducted using Mann-Whitney U tests. In order to address different patterns of HRQoL impairment with regard to the 6 domains in patients with adverse preSOC and FU characteristics, matched one-way analysis of variance of the normalized NHP domain scores with the Greenhouse-Geisser correction and the Dunnett post hoc tests using the normalized NHP scores as a reference were performed. With regard to prediction of reduced HRQoL 6 months after CABG, a multivariable evaluation of preSOC, preMED and FU variables was conducted using hierarchical regression analyses with the normalized NHP score and the normalized NHP domain scores, respectively, as the dependent variable. For the normalized NHP score, 2 models were computed: model I (hierarchy steps preSOC+preMED+FU) and model II (hierarchy steps preSOC+FU). For all hierarchical regression analyses, the adjusted R² values for the preSOC, preMED and FU steps, as well as the unstandardized regression coefficients with 95% confidence intervals for all variables, are reported. Significance was assumed if P < 0.05. Analyses were exploratory in nature. Data analysis was conducted with SPSS Statistics for Windows (IBM, Armonk, NY, USA).

RESULTS

Patient characteristics

Data from 3,237 patients undergoing isolated CABG (2,082 off-pump cases) were analysed. The preoperative characteristics of the patients are presented in Table 1.

Table 1:

Characteristics of the coronary artery bypass graft cohort and results of univariable analyses

PreSOC variables
Gender, male	2,684 (82.9%)	P < 0.001
Age	<60 years: 746 (23.0%)	P < 0.001
	60-74 years: 1,823 (56.3%)
	≥75 years: 668 (20.6%)
Not married	680 (21.0%)	P = 0.103
Not employed	2,331 (72.0%)	P < 0.001
PreMED variables
BMI	28.2 ± 0.08(kg*m-2)	P < 0.001
ASA	I: 4 (0.1%)	P = 0.006
	II: 135 (4.2%)
	III: 2,436 (75.3%)
	IV: 648 (20.0%)
	V: 14 (0.4%)
NYHA	I: 202 (6.2%)	P = 0.150
	II: 917 (28.3%)
	III: 1,936 (59.8%)
	IV: 182 (5.6%)
CCS classification	0: 1,549 (47.9%)	P = 0.165
	I-III: 1,289 (39.8%)
	IV: 399 (12.3%)
LVEF < 30%	120 (3.7%)	P = 0.167
AF	121 (3.7%)	P = 0.037
PVD	264 (8.2%)	P = 0.003
COPD	368 (11.4%)	P < 0.001
Creatinine level	1.10 ± 0.01(mg*dl-1)	P = 0.012
DM II w/o insulin	665 (20.5%)	P = 0.017
DM II+insulin	376 (11.6%)	P < 0.001
AMI < 48 h	412 (12.7%)	P = 0.013
CPR < 48 h	8 (0.2%)	P = 0.757
Previous stroke	234 (7.2%)	P = 0.024
Redo surgery	178 (5.5%)	P = 0.342
Emergency CABG	306 (9.5%)	P = 0.319
FU symptoms
Chest pain	421 (13.0%)	P < 0.001
Dyspnoea	720 (22.2%)	P < 0.001
NYHA	I: 2,032 (62.8%)	P < 0.001
	II: 676 (20.9%)
	III: 529 (16.3%)

Descriptive and univariable statistics for preoperative and follow-up variables. P-values result from unpaired t-tests or one-way analysis of variance for categorical variables and linear regression analyses for continuous variables each with health-related quality of life as the dependent variable. AF: atrial fibrillation; AMI: acute myocardial infarction; ASA: American Society of Anesthesiologists; BMI: body mass index; CABG: coronary artery bypass graft; CCS: Canadian Cardiovascular Society; COPD: chronic obstructive pulmonary disease; CPR: cardiopulmonary resuscitation; DM: diabetes mellitus; FU: follow-up; LVEF: left ventricular ejection fraction; NYHA: New York Heart Association; preMED: preoperative medical; preSOC: preoperative socio-demographic; PVD: peripheral vascular disease.

The normalized NHP score of the entire patient cohort amounted to -1.6 ± 0.26. Whereas patients with HRQoL superior to the reference group presented with a score of -9.5 ± 0.11, a total of 1,031 patients (31.9%) exhibited positive normalized NHP scores, which amounted to 15.4 ± 0.43 among those patients with impaired HRQoL.

Preoperative predictors and follow-up risk factors of impaired health-related quality of life

Univariable statistics showed a significant influence of most of the preSOC and FU variables on HRQoL, while several preMED characteristics also had a relevant impact [Table 1].

Multivariable regression analyses revealed that all preSOC and FU variables were independent predictors of HRQoL (each P < 0.001) (Fig. 1, Supplementary Table 2), whereas male patients and patients < 60 years were particularly impaired. On the one hand, preMED chronic comorbidities, such as diabetes mellitus type II (P = 0.018 without insulin, P = 0.003 with insulin), chronic obstructive pulmonary disease (P = 0.007) or previous stroke (P = 0.001), showed a significant effect in terms of reduction of HRQoL. Other preMED variables, such as symptoms categorized by the New York Heart Association (NYHA) and the Canadian Cardiovascular Society (CCS) classification systems or reduced left ventricular ejection fraction, did not predict HRQoL.

Figure 1:

Predictors of health-related quality of life. Multivariable statistics of preSOC, preMED and FU analysed by a hierarchical regression analysis predicting reduced HRQoL 6 months after CABG. Unstandardized regression coefficients with 95% confidence intervals are plotted. For exact values, see Supplementary Table 2. AF: atrial fibrillation; AMI: acute myocardial infarction; ASA: American Society of Anesthesiologists; BMI: body mass index; CABG: coronary artery bypass graft; CCS: Canadian Cardiovascular Society; COPD: chronic obstructive pulmonary disease; CPR: cardiopulmonary resuscitation; DM: diabetes mellitus; FU: follow-up; HRQoL: health-related quality of life; LVEF: left ventricular ejection fraction; NYHA: New York Heart Association; preMED: preoperative medical; preSOC: preoperative socio-demographic; PVD: peripheral vascular disease; w/o: without; y: year.

Hierarchical regression analyses were conducted with 2 different models to examine in particular the relevance of the preMED variables: model I showed a low R² value for preMED (0.04), whereas preSOC (0.07) and FU (0.21) had higher values, resulting in a total value of 0.32 (Fig. 2). In model II, the preSOC R² value was 0.07, and the FU value was 0.24, so the total R² value was 0.31.

Figure 2:

Significance of health-related quality of life predictors. Adjusted R² values for preoperative socio-demographic, preoperative medical and follow-up variables analysed by hierarchical regression analyses (black arrow, hierarchical direction) predicting reduced health-related quality of life 6 months after coronary artery bypass graft. Model I hierarchy steps: preoperative socio-demographic+preoperative medical+follow-up. Model II hierarchy steps: preoperative socio-demographic+follow-up. Significant F-change values <0.001 for all steps.

Nottingham Health Profile domain analysis of risk factors of impaired health-related quality of life

Multivariable hierarchical regression analyses revealed a significant association of male gender and younger age with worse values in all NHP domains, whereas the status “unmarried” impaired only the social isolation domain, and the status “not employed” did not impair the energy level or the physical abilities (Table 2). None of the preMED variables significantly influenced any of the NHP domains, and only 3 variables (ASA classification, creatinine and previous stroke) impaired at least 3 domains. Physical abilities were most frequently influenced by preMED variables (by 10 out of 16). In the FU, the symptoms chest pain and dyspnoea, particularly NYHA III, showed a major impact on HRQoL reduction in all NHP domains.

Table 2:

Predictors of health-related quality of life

	Emotional reaction	Energy level	Pain	Physical abilities	Sleep	Social isolation
PreSOC variables
Gender, male	4.01*** [2.49-5.53]	7.54*** [4.93-10.16]	6.35*** [4.87-7.84]	4.42*** [3.00-5.84]	5.47*** [3.10-7.84]	2.83*** [1.64-4.02]
Age < 60 years	12.21*** [10.19-14.23]	18.82*** [15.35-22-29]	12.49*** [10.51-14.46]	13.65*** [11.76-15.53]	16.65*** [13.51-19.80]	9.25*** [7.68-10.83]
Age, 60-74 years	5.74*** [4.30-7.18]	8.63*** [6.15-11.11]	6.40*** [4.99-7.81]	8.27*** [6.92-9.61]	5.40*** [3.16-7.65]	5.14*** [4.01-6.26]
Not married	0.48 [-0.89-1.86]	0.84 [-1.53-3.21]	−0.13 [-1.47-1.22]	−0.14 [-1.42-1.15]	1.23 [-0.91-3.38]	5.27*** [4.20-6.35]
Not employed	1.73* [0.19-3.27]	1.13 [-1.52-3.77]	2.23** [0.72-3.73]	1.29 [-0.14-2.73]	3.29** [0.90-5.69]	1.72** [0.52-2.93]
PreMed variables
BMI	−0.11 [-0.24-0.01]	0.21 [0.00-0.43]	0.20** [0.08-0.33]	0.33*** [0.21-0.45]	−0.05 [-0.24-0.15]	0.00 [-0.09-0.10]
ASA	1.14 [-0.02-2.31]	0.87 [-1.13-2.88]	1.89** [0.75-3.03]	1.49** [0.41-2.58]	0.46 [-1.35-2.27]	1.01* [0.10-1.92]
NYHA II	−1.32 [-3.72-1.08]	−0.12 [-4.25-4.02]	0.01 [-2.34-2.36]	−0.78 [-3.02-1.46]	−3.22 [-6.96-0.52]	−1.66 [-3.54-0.22]
NYHA III	−0.22 [-2.52-2.08]	−0.31 [-4.27-3.64]	−0.54 [-2.79-1.70]	−0.99 [-3.13-1.16]	−3.24 [-6.82-0.34]	−1.08 [-2.88-0.72]
NYHA IV	−0.27 [-3.62-3.07]	1.62 [-4.13-7.37]	0.34 [-2.93-3.60]	−1.10 [-4.22-2.02]	−1.88 [-7.08-3.33]	−1.28 [-3.89-1.34]
CCS I-III	0.21 [-0.97-1.39]	0.57 [-1.46-2.59]	0.94 [-0.21-2.09]	−0.34 [-1.44-0.76]	0.91 [-0.92-2.75]	0.24 [-0.68-1.16]
CCS IV	1.36 [-0.48-3.20]	2.68 [-0.48-5.84]	1.58 [-0.22-3.37]	1.46 [-0.25-3.18]	0.84 [-2.02-3.70]	−0.44 [-1.87-1.00]
LVEF < 30%	−1.24 [-4.20-1.71]	2.04 [-3.04-7.12]	−0.19 [-3.07-2.70]	4.81*** [2.05-7.56]	−2.03 [-6.63-2.56]	1.13 [-1.18-3.44]
AF	1.54 [-1.37-4.45]	7.71** [2.71-12.72]	0.78 [-2.06-3.63]	2.95* [0.23-5.67]	2.10 [-2.44-6.63]	0.53 [-1.74-2.81]
PVD	0.66 [-1.42-2.74]	1.53 [-2.05-5.10]	4.16*** [2.12-6.19]	2.79** [0.85-4.73]	−1.12 [-4.36-2.11]	0.72 [-0.90-2.35]
COPD	1.24 [-0.51-2.99]	3.93* [0.92-6.94]	1.00 [-0.71-2.71]	2.44** [0.80-4.07]	2.65 [-0.07-5.38]	−0.19 [-1.56-1.18]
Creatinine level	−0.29 [-1.08-0.50]	1.44* [0.09-2.80]	0.64 [-0.13-1.41]	1.11** [0.37-1.84]	2.55*** [1.32-3.78]	0.18 [-0.43-0.80]
DM II w/o insulin	0.78 [-0.61-2.16]	3.5** [1.12-5.88]	0.99 [-0.36-2.34]	1.39** [0.10-2.68]	0.56 [-1.60-2.71]	0.53 [-0.55-1.61]
DM II+insulin	0.37 [-1.41-2.16]	5.05** [1.98-8.13]	1.73 [-0.02-3.47]	4.82*** [3.15-6.49]	0.38 [-2.40-3.16]	0.34 [-1.05-1.74]
AMI < 48h	−1.30 [-3.03-0.43]	0.57 [-2.40-3.54]	1.07 [-0.62-2.75]	0.42 [-1.19-2.03]	0.96 [-1.73-3.65]	0.08 [-1.27-1.43]
CPR < 48h	2.40 [-8.66-13.46]	−5.15 [-24.17-13.88]	−4.85 [-15.66-5.95]	−2.92 [-13.25-7.41]	−8.21 [-25.44-9.02]	−1.74 [-10.39-6.90]
Previous stroke	2.29* [0.05-4.53]	4.16* [0.30-8.01]	1.47 [-0.72-3.66]	5.95*** [3.86-8.04]	0.88 [-2.60-4.37]	2.22* [0.47-3.97]
Redo surgery	−0.61 [-3.01-1.79]	−0.75 [-4.88-3.39]	−0.50 [-2.85-1.84]	−0.68 [-2.93-1.56]	1.85 [-1.89-5.59]	0.31 [-1.57-2.19]
Emergency CABG	−0.61 [-2.70-1.49]	−0.39 [-3.99-3.22]	−1.42 [-3.47-0.62]	0.26 [-1.70-2.21]	0.62 [-2.64-3.88]	0.10 [-1.53-1.74]
FU
Chest pain	7.42*** [5.75-9.09]	13.28*** [10.41-16.15]	8.84*** [7.21-10.47]	4.19*** [2.64-5.75]	10.88*** [8.28-13.48]	2.83*** [1.52-4.13]
Dyspnoea	0.87 [-0.76-2.51]	11.48*** [8.67-14.28]	2.1** [0.51-3.7]	4.77*** [3.24-6.29]	2.65* [0.11-5.19]	−0.87 [-2.15-0.40]
NYHA II	2.71*** [1.22-4.21]	7.22*** [4.65-9.78]	0.95 [-0.5-2.41]	1.69* [0.30-3.08]	5.46*** [3.14-7.78]	1.64** [0.48-2.81]
NYHA III	11.01*** [9.12-12.89]	27.22*** [23.97-30.47]	8.29*** [6.44-10.13]	14.82*** [13.06-16.58]	15.04*** [12.10-17.98]	5.63*** [4.15-7.11]

^* P < 0.05;

^** P < 0.01;

^*** P < 0.001.

Multivariable statistics of preoperative socio-demographic, preoperative medical and follow-up variables were analysed by a hierarchical regression analysis predicting reduced health-related quality of life separately analysed in all NHP domains 6 months after CABG. Unstandardized regression coefficients with 95% confidence intervals re displayed. AF: atrial fibrillation; AMI: acute myocardial infarction; ASA: American Society of Anesthesiologists; BMI: body mass index; CABG: coronary artery bypass graft; CCS: Canadian Cardiovascular Society; COPD: chronic obstructive pulmonary disease; CPR: cardiopulmonary resuscitation; DM: diabetes mellitus; FU: follow-up; LVEF: left ventricular ejection fraction; NYHA: New York Heart Association; preMED: preoperative medical; preSOC: preoperative socio-demographic; PVD: peripheral vascular disease; w/o: without.

R² values were low for the preMED variables in 5 of the 6 domains, whereas only the physical abilities presented a higher value (0.07) (Fig. 3). The preSOC variables influenced all normalized NHP domain scores, whereas the effect was largest for social isolation (0.06), pain (0.06) and physical abilities (0.05). The FU variables had a major impact on the energy level (0.21) and the physical abilities (0.16); social isolation (0.02) was marginally influenced.

Figure 3:

Significance of health-related quality of life predictors. Adjusted R² values for preoperative socio-demographic, preoperative medical and follow-up variables analysed by hierarchical regression analyses (model I; black arrow, hierarchical direction) predicting reduced health-related quality of life analysed separately in all NHP domains 6 months after coronary artery bypass graft. Significant F-change values <0.001 for all steps.

The impact of each domain score on HRQoL was evaluated in the subgroups. All subgroups except age ≥75 years presented a heterogeneous extent of the impact of HRQoL among the 6 domains (Table 3). Patients with male gender or age <60 years showed particularly impaired energy levels, whereas the domains social isolation and pain were less affected. At the ages of 60 to 74, physical abilities have the largest negative impact on HRQoL. Unmarried patients suffered most from social isolation, and unemployed patients, from low energy levels. Patients who had CABG and who had chest pain or dyspnoea were most severely impaired in the energy level domain, and patients who were NYHA III were additionally impaired in the physical abilities domain.

Table 3:

Influence of the different Nottingham Health Profile domains on health-related quality of life

	NHP	NHP domains						P-value
	Mean	Emotional reaction	Energy level	Pain	Physical abilities	Sleep	Social isolation
All	−1.56 ± 0.26	−1.14 ± 0.30	-0.01*** ± 0.55	-2.72*** ± 0.29	−1.16 ± 0.30	−2.24 ± 0.46	−2.09 ± 0.23	<0.001
PreSOC variables
Gender, male	−1.07 ± 0.27	−0.73 ± 0.32	0.42** ± 0.59	-1.80* ± 0.30	−0.71 ± 0.31	−1.71 ± 0.48	-1.88** ± 0.23	<0.001
Age < 60 years	4.29 ± 0.57	4.07 ± 0.78	8.47*** ± 1.16	2.35*** ± 0.62	3.63 ± 0.49	5.95 ± 0.98	1.28*** ± 0.56	<0.001
Age, 60-74 years	−2.05 ± 0.31	−1.58 ± 0.34	−1.19 ± 0.69	−2.75 ± 0.36	-0.82*** ± 0.36	-3.91*** ± 0.58	−2.07 ± 0.26	<0.001
Age ≥ 75 years	−6.76 ± 0.59	−5.76 ± 0.60	−6.27 ± 1.33	-8.29* ± 0.70	−7.46 ± 0.88	−6.83 ± 1.03	−5.90 ± 0.50	0.135
Not married	−0.68 ± 0.62	−0.95 ± 0.69	1.52 ± 1.30	-3.62*** ± 0.70	−1.28 ± 0.74	−1.26 ± 1.10	1.51** ± 0.70	<0.001
Not employed	−2.33 ± 0.31	−1.88 ± 0.34	-1.16* ± 0.67	-3.41*** ± 0.37	−1.80 ± 0.39	−3.23 ± 0.55	−2.50 ± 0.26	<0.001
Follow-up
Chest pain	9.31 ± 0.93	8.01 ± 1.12	20.66*** ± 1.87	7.28 ± 1.19	6.78** ± 1.02	11.66 ± 1.53	1.48*** ± 0.88	<0.001
Dys-pnoea	7.13 ± 0.62	4.41*** ± 0.76	20.59*** ± 1.40	2.87*** ± 0.77	8.19 ± 077	6.95 ± 1.11	-0.18*** ± 0.58	<0.001
NYHA I	−5.20 ± 0.25	-3.68*** ± 0.30	-8.29*** ± 0.52	−4.79 ± 0.31	−4.76 ± 0.31	-6.49** ± 0.49	-3.19*** ± 0.22	<0.001
NYHA II	−0.75 ± 0.53	−0.73 ± 0.64	3.05*** ± 1.23	-3.01*** ± 0.62	−1.84 ± 0.59	0.01 ± 1.03	-1.98* ± 0.51	<0.001
NYHA III	11.3 ± 0.82	8.09*** ± 1.02	27.88*** ± 1.72	5.62*** ± 1.04	13.49* ± 0.98	11.19 ± 1.44	1.98*** ± 0.84	<0.001

^* P < 0.05;

^** P < 0.01;

^*** P < 0.001).

Normalized scores (mean±SEM) for all Nottingham Health Profile domains. P-values from matched one-way analysis of variance with the Greenhouse-Geisser correction, addressing interdomain differences. Bold domain scores significantly differ from the respective mean scores, as calculated by Dunnett post hoc tests.

FU: follow-up; NYHA: New York Heart Association; preSOC: preoperative socio-demographic.

Interactions among socio-demographic risk factors

When stratified by age, the status “unmarried” resulted in significantly reduced HRQoL only in patients younger than 60 years (Fig. 4A). Gender-specific analyses revealed that below 60 years, unmarried female patients had decreased HRQoL, whereas among older unmarried patients, men showed decreased HRQoL compared to married men (Fig. 4B) and to unmarried women (P < 0.001).

Figure 4:

Interactions between socio-demographic predictors. Age- and gender-grouped analyses of the effects of marital status (A, B) and employment (C, D) on health-related quality of life. **P<0.01;***P<0.001.

The status of being unemployed decreased the HRQoL only in male patients below 60 years (Fig. 4C, D). Unemployed men had significantly reduced HRQoL compared to unemployed women <60 years (P = 0.002) and ≥60 years (P < 0.001). The normalized NHP scores of unemployed men below 60 years did not differ in dependency on unemployment versus retirement [9.3 (interquartile range 1.4–26.9) versus 7.1 (3.3–27.0); P = 0.776].

DISCUSSION

The present study focused on the significance of preSOC characteristics in comparison to medical variables with regard to postoperative HRQoL in patients undergoing isolated CABG. By means of uni- and multivariable analyses in a large cohort, we found that 4 easily available socio-demographic variables (gender, age, marriage and employment) significantly influence the postoperative HRQoL and are more predictive than a multitude of preoperative medical parameters that are typically analysed as potential risk factors of morbidity, mortality and HRQoL in patients having cardiac surgery [8, 12, 16].

Socio-demographic predictors

Although age [2] and gender [17] have been previously identified as influencing the HRQoL after CABG, the impact of marital and employment status and the potential interactions between these socio-demographic variables have remained unclear. Our study revealed that younger age, male gender, not married and unemployed status are independent predictors of impaired postoperative HRQoL.

The reports in the literature regarding the influence of age are contradictory. Whereas Herlitz et al. observed a tendency toward lower HRQoL in older versus younger patients who had CABG [18], Peric et al. reported superior postoperative improvement in older patients [2], whereas age was not an independent predictor of HRQoL deterioration, which may be attributed to only 226 analysed patients. Our study in 3,237 patients who had CABG supports the hypothesis that younger patients are at higher risk of reduced age- and gender-matched HRQoL.

The finding of a higher risk for young patients has different facets. From a mathematical point of view, it has to be mentioned that in the healthy reference group, people <60 years have low NHP scores (<10), whereas the scores for people >75 years range up to 29 [15]. Considering that the HRQoL in our study was assessed as the difference of the individual NHP scores from the reference group scores (normalized NHP score) and that the NHP score ranges from 0 to 100, subjectively unimpaired patients (with an individual score of 0), if they are older, reach much higher NHP score differences and thus have much more improved normalized NHP scores. In turn, younger patients with severely impaired HRQoL can reach mathematically substantially worse normalized NHP scores than older patients. When focusing on psychological aspects, young patients may judge postoperative restrictions more drastic than elderly patients who might be more indolent, particularly in the face of physical impairment. For instance, there is evidence that the somatosensory pain threshold increases in elderly people compared to young people [19, 20], indicating that young people are more sensitive to pain.

In our study, male patients had remarkably reduced HRQoL compared to women. At first glance, this finding seems to contradict findings from previous studies: Martin et al. reported that the baseline of female patients having CABG and the 6-month FU HRQoL were lower than in men [17]; Peric et al. showed improved performance in the NHP sleep domain for men versus women 2 years after CABG [7]; and Falcoz et al. found lower HRQoL 2 years after CABG in women compared to men, whereas the postoperative improvement was significant for both genders [11]. In contrast to previous studies, our analyses are based on normalization, considering age- and gender-matched healthy persons as reference groups [15]; additionally, women of the reference population have substantially higher NHP score values, indicating worse perceived HRQoL. Thus, in order to reveal gender effects, normalization based on a reference population, or at least on the course of the score over time, is required.

A beneficial effect of marriage on postoperative HRQoL in patients who have had CABG has not been described before. Interestingly, only younger women (<60 years) and elderly men (≥60 years) showed a significant benefit when married. It may be speculated that emotional patterns in a relationship and age-dependent variations of the need for mutual psychological and physical care (particularly in older people) might explain this age-dependent shift of the protective effect of marriage for coronary patients. Besides the age and gender of patient having CABG, the personalities of the patient and of the spouse have been shown to play an important role for the postoperative HRQoL of the married patient as well as his/her spouse [21]. Predominantly neuroticism and depressive symptoms have been identified as negatively influencing factors.

In the context of patients undergoing CABG, employment thus far has mainly been addressed only as a dependent variable, coding as postoperative return to work and not as a potential predictor of HRQoL. The present study revealed a significant impairment from missing employment only for male patients below 60 years. This finding may be explained by higher perceived significance of employment among men compared to women, whereas "no employment" seems to be an issue even if retirement is the underlying reason. The gender difference might be due to the fact that until today, men on average earn more money and work in more prestigious positions compared to women [22].

Age and gender seem to have a higher impact on postoperative HRQoL than employment and marital status. This finding may be due to the fact that, besides psychosocial and psychosomatic implications, age and gender also have physical effects on cardiac pathology. For instance, younger patients who require CABG frequently exhibit faster development and progression of atherosclerosis than patients who reach the stage of surgical indication at an advanced age. This situation might also impair HRQoL during the postoperative course.

Beyond employment status, socio-economic deprivation has been reported to influence the outcome after CABG. In a recent analysis performed by Thompson et al., the Michigan Value Collaborative and the Michigan Society of Thoracic and Cardiovascular Surgeons Quality Collaborative, patients with high socio-economic deprivation were shown to suffer from increased morbidity and even in-hospital death [23]. The interplay between employment status, actual socio-economic deprivation and postoperative HRQoL should be subject to further investigation.

Medical risk variables

Among multiple preoperative medical characteristics, especially angina, dyspnoea and severely impaired cardiac function, peripheral vascular disease and myocardial infarction or cardiopulmonary resuscitation preceding CABG were not independent predictors of HRQoL, although these parameters have been previously reported to be relevant risk factors for higher morbidity after CABG [24, 25]. These findings could be explained by a higher acceptance of cardiovascular disease and associated symptoms by these patients, which may result in lower perceived impairment. Non-cardiac chronic diseases, such as diabetes, chronic obstructive pulmonary disease or stroke, exhibited larger negative impacts on HRQoL. Naturally, a severe chronic disease in addition to the cardiac pathology impacts the HRQoL to a greater extent than the coronary disease alone.

In contrast to preoperative dyspnoea and angina, the symptoms of dyspnoea and chest pain at 6 months FU were relevant risk factors for impaired HRQoL, which is not surprising, because both can severely impact daily life [26, 27]. The rate of patients with chest pain at 6 months (13%) was slightly higher than that reported in the literature [28]. The symptom “chest pain” can not only reflect the diagnosis “angina” but also muscle- or bone-related or psychosomatic pain. As we recently reported for cardiosurgical patients [8], dyspnoea and particularly chest pain at 6 months should be evaluated in terms of preSOC variables, because their relevance for HRQoL seems to interact with the presence of socio-demographic risk factors.

Nottingham Health Profile domain analyses

To the best of our knowledge, this is the first study to investigate the influence of socio-demographic and medical factors on the separate NHP domains. Interesting results showed that the socio-demographic variables had the largest impact in the social isolation domain. This finding can be partially explained by the facts that not being able to work and the disruption to the routines of one's marriage contribute to the preSOC parameters. Whereas sleep, emotional reactions and even pain and physical abilities were remarkably influenced by the preSOC variables, the preMED characteristics exhibited a relevant effect only in the domain of physical abilities. It is of particular interest that the preSOC variables significantly impacted not only psychological aspects of HRQoL, but also predominantly somatic outcomes, such as pain and physical abilities, which reflect a complex interplay between socio-demographic, psychosomatic and physical aspects. This finding is all the more noteworthy because improvement of the somatic outcome is a primary goal of cardiac surgery. Thus, attention should be paid to the preSOC parameters in patients who have had CABG.

Patients at risk and early intervention

Our data confirm that CABG is an operation that allows for good postoperative HRQoL despite the fact that the patients are diagnosed with severe cardiac disease. Actually, 6 months after CABG, 2 out of 3 patients in our study presented with perceived HRQoL not inferior to the status of the healthy reference population. Nevertheless, at 6 months, 32% of the patients who had CABG still exhibited reduced HRQoL. Independently of the medical success aspects of the operation, such as revascularization, myocardial preservation or avoidance of surgery-associated complications, a perceived reduction of the patient’s HRQoL is a major issue that also impairs the net benefit of the therapy. Therefore, it is crucial to identify patients at risk of impaired postoperative HRQoL as early as possible.

Psychological interventions provide additional opportunities to optimize HRQoL and thereby the whole postoperative outcome. In a recent study, it was demonstrated that a preoperative psychological intervention beneficially influences the outcome expectations after CABG and thereby improves the physical and mental postoperative outcomes [29]. However, it was also shown that such interventions are more effective when tailored to the individual risk profile [30, 31], underlining the importance of identifying HRQoL predictors.

Limitations

This study is an analysis of prospectively collected data in a single centre in Germany. Other health systems or different cultures with patients potentially exhibiting different risk factors, particularly regarding socio-demographic characteristics or perceived disability, were not included. Out of 8,202 patients who underwent CABG, only 39.5% provided FU data that could be included in the analyses, whereas a significant quantity of HRQoL data still exist. However, a potential selection bias has to be considered, because the actual HRQoL may have influenced the patients’ decisions to respond to the FU survey. Furthermore, the NHP survey was performed only at 1 timepoint postoperatively; in contrast, other groups have reported surveys taken prior to the operation [32, 33]. However, the goal of our approach was to normalize age- and gender-matched reference groups so that postoperative HRQoL deviations of patients who had CABG versus healthy people could be demonstrated.

Conclusions

The identification of patients at risk of impaired postoperative HRQoL is decisive in order to provide additional and, in particular, patient-adjusted support. This study reveals that the assessment of only 4 preSOC characteristics (age, gender, marriage and employment) is more predictive of HRQoL after CABG than multiple medical variables, whereas psychological as well as somatic aspects of HRQoL are influenced. Furthermore, age and gender seem to modulate the influence of marriage and employment on postoperative HRQoL.

Glossary

Abbreviations

CABG

Coronary artery bypass grafting

FU

Follow up

HRQoL

Health-related quality of life

NHP

Nottingham Health Profile

NYHA

New York Heart Association

preMed

Preoperative medical variables

preSOC

Preoperative socio-demographic variables

SEM

Standard error of the mean

Data availability

The data underlying this article will be shared in response to a reasonable request to the corresponding author.

Author contributions

Anna Kathrin Assmann: data analysis, drafting the manuscript; Alexander Assmann: conception of the work, data analysis, drafting the manuscript; Sebastian Waßenberg: statistical analysis; Besnik Kojcici: data acquisition, review and editing; Nora K. Schaal: review and editing; Artur Lichtenberg: supervision; Jürgen Ennker: supervision; Alexander Albert: conception of the work, review and editing.