Disparities in prevalence of heart failure between the genders in relation to age, multimorbidity and socioeconomic status in southern Sweden: a cross-sectional study

Abstract

Objective

Prior studies have reported that heart failure typically affects elderly, multimorbid and socioeconomically deprived men. Women with heart failure are generally older, have a higher EF (ejection fraction) and have more heart failure-related symptoms than men. This study explored the disparities in the prevalence of heart failure between men and women in relation to age, multimorbidity level and socioeconomic status of the population in southern Sweden.

Design

A register-based, cross-sectional cohort study.

Setting and subjects: The inhabitants from 20 years of age onwards (N = 981,383) living in southern Sweden in 2015.

Main outcome measure: Prevalence and mean probability of having heart failure in both genders. CNI (Care Need Index) percentiles depend on the socioeconomic status of their listed primary healthcare centres.

Results

Men had a higher OR for HF − 1.70 (95% CI 1.65–1.75) - than women. The probability of men having heart failure increased significantly compared to women with advancing age and multimorbidity levels. At all CNI levels, the multimorbid patients had a higher prevalence of heart failure in men than in women. The disparity in the mean probability of heart failure between the most affluent and deprived CNI percentile was more apparent in women compared to men, especially from 80 years.

Conclusions

The prevalence of heart failure differs significantly between the genders. Men had an increasing mean probability of heart failure with advancing age and multimorbidity level compared to women. Socioeconomic deprivation was more strongly associated with heart failure in women than in men. The probability of having heart failure differs between the genders in several aspects.

Key Points

• Independently of socioeconomic status, men had a higher prevalence of heart failure than women among the multimorbid patients.

• The mean probability of men having heart failure increased significantly compared to women with advancing age and multimorbidity level.

• Socioeconomic status was more strongly associated with heart failure in women than in men.

Keywords:

• Heart failure (HF)
• multimorbidity (MM)
• prevalence
• probability
• primary health care

Introduction

HF represents a global health problem that has a high impact on healthcare resources and affects approximately 26 million adults worldwide [1]. HF is one of the leading causes of morbidity and mortality among elderly individuals, and the mortality rate remains high despite the implementation of new treatment strategies. With increased life expectancy, combined with innovative treatments for cardiovascular comorbidities, HF is a major burden for healthcare services in many high-income countries.

HF presents with similar symptoms in both men and women. However, females are generally older and have a higher EF (ejection fraction) have and more HF-related symptoms than men [1]. Three criteria are required for the diagnosis of HF: typical clinical symptoms, such as dyspnoea, fatigue, exertional intolerance and oedema of the lower body; objective findings of elevated BNP value; and impaired cardiac function on echocardiography, myocardial scintigraphy, magnet resonance tomography or other imaging [2].

HF is classified as HFrEF (heart failure with reduced ejection fraction), HFmrEF (heart failure with mildly reduced ejection fraction), and HFpEF (heart failure with preserved ejection fraction). All subtypes of HF have the same clinical symptoms, but different pathophysiology and prognosis. HFrEF is established when the left ventricle loses its ability to contract normally, whereas HFpEF is established when the left ventricle loses its ability to relax normally. HFmrHF has a mixture of characteristics from both HFrEF and HFpEF regarding aetiology, pathophysiology and comorbidities [3]. The symptoms of HF are usually not specific like wheezing, coughing, and shortness of breath, which can be misinterpreted as bronchial asthma and can delay the diagnosis of cardiac asthma caused by congestive HF [4]. Women have double the risk for incident HFpEF and are more likely to have a background of hypertension and valve dysfunctions as HF aetiologies compared to men [1,5]. Men are predisposed to HFrEF with ischaemic aetiology and have an earlier onset of HFrEF and a higher mortality rate compared to women [6]. Although cardiovascular risk factors predispose both genders to HFrEF, diabetes and obesity significantly increase the risk of HFrEF in women compared to men. Generally, it is observed that female HF patients tend to have more comorbidities such as atrial fibrillation, diabetes, hypertension, anaemia, iron deficiency, renal disease, arthritis, depression, and thyroid abnormalities [1].

Multimorbidity is common in all subtypes of HF, but slightly more severe in HFpEF. A majority of deaths in patients with HFpEF are cardiovascular-related, but the proportion of non-cardiovascular related deaths is higher in patients with HFpEF than HFrEF [7].

Low socioeconomic status is one of the most common predictors of morbidity and premature mortality in the world, even when taking traditional risk factors into consideration [8], not least for the morbidities that contribute most to the mortality rate, i.e. cancer and cerebrovascular diseases [9, 10]. Prior studies reported inequalities in HF risk according to neighbourhood, socioeconomic deprivation, education and occupational social class. Notably, the increased HF risk in the socioeconomically deprived population was only partly influenced by established cardiovascular risk factors, including hypertension, hyperlipidaemia, diabetes, smoking, and physical inactivity [11–16]. Marital status is also one of the important socioeconomic factors; living alone is often associated with a lower income. Unmarried men have double the age-adjusted relative mortality risk and higher incidence of HF compared to their married counterparts. There is a continuous gradient in Western countries, including Sweden, between cardiovascular morbidity and mortality and socioeconomic status (SES), with higher SES being more favourable [17–20]. The aim of this study was to explore the disparities in the prevalence of HF between men and women in relation to age, multimorbidity level and socioeconomic status of the population listed at primary healthcare centres in southern Sweden.

Material and methods

Setting and study population

Most residents in Sweden are listed at a primary healthcare centre. Scania is the southernmost county of Sweden and had approximately 1.3 million inhabitants in 2015. The whole study population was recruited from Scania. The biggest city in Scania is Malmö, which is Sweden’s third largest city, and had approximately 320,000 inhabitants during the study period. In 2015, about a third of the residents in Malmö were born abroad with most countries in the world being represented, whilst approximately 25% of the whole study population were born abroad [21]. Almost half of the inhabitants (48.40%) in Malmö were under 35 years of age in 2015 [22].

The study population was divided into age groups from 20 to 80. The age group 20 included all individuals between 20 to 29 years, the age group 40 included all individuals between 40 to 49 years, and so on. The age group 80 included all individuals aged 80 years and older. Gender was categorised as female and male.

Data source and measurements

The data we used in this study were retrieved from the Scania County Council healthcare register which contains anonymised registry information from the population, including age, gender, socioeconomic status and diagnostic data. Data were collected concerning diagnoses at each consultation in both primary and secondary health care. Diagnoses were recorded according to the International Statistical Classification of Diseases and Related Health Problems version 10 (ICD 10). HF was identified if the diagnosis code I50 was recorded, which comprised all subtypes of HF. A total of 152 primary healthcare centres were in operation in Scania during 2015, with an average of 8587 listed patients (95% CI 7971 − 9292) including 133 patients with HF (95% CI 122 – 143) at each primary healthcare centre.

Multimorbidity

To measure multimorbidity we used a method developed by A Calderòn-Larrañaga et al. at The Aging Research Centre in Stockholm [23]. They analysed the full list of ICD-10 codes on a four-digit level to define if a diagnosis is chronic or not in an elderly population. A disease or condition was considered to be chronic if it had a prolonged duration and either (a) left residual disability or worsening quality of life or (b) required a long period of care, treatment, or rehabilitation. This measure of multimorbidity was designed for persons older than 60 years and was based on a clinical assessment of medical diagnoses by specialists in geriatrics and family medicine.

To determine if a condition is chronic or not the following key features were identified and discussed concerning their pertinence and sustainability: duration, course, reversibility, treatment, and consequences. The diagnoses were then grouped into 60 chronic disease categories, which is the most comprehensive list of chronic conditions for measuring multimorbidity thus far. The broad scope of diagnoses included conditions such as blood diseases, lens diseases, chromosomal abnormalities, chronic infectious diseases, chronic ulcers of the skin, allergies, autoimmune and connective tissue diseases, glaucoma, multiple sclerosis, peripheral neuropathy, venous and lymphatic disorders etc. We applied their definition and list of chronic conditions to estimate multimorbidity in the study population. Multimorbidity was then estimated by counting the number of chronic conditions in each patient. All diagnoses from the last week of 2015 were obtained from the electronic medical record database in Scania County Council. Those individuals who had at least two of these chronic conditions were considered multimorbid. To study the degree of multimorbidity in relation to the prevalence of HF, the patients were then further divided into groups MM0 (less than two chronic conditions), MM1 (two to four chronic conditions), MM2 (five to nine chronic conditions), and MM3 (ten chronic conditions or more).

Socioeconomics

We used the term Care Need Index (CNI) [24] to divide the primary healthcare centres into 10 groups depending on their socioeconomic status. CNI is based on different measures of a group, which in our study were the patients listed at different primary healthcare centres in Scania. CNI 1 percentile was assigned to those listed patients at primary healthcare centres that belonged to the most socioeconomically affluent population; CNI 10 percentile was assigned to those patients listed at primary healthcare centres that belonged to the most deprived population [24].

Statistical analyses

We used frequencies, percentages and cross-tabulations for descriptive analysis. Multivariable logistic regression was used to analyse the multivariate models. Only the linear predictions of the fully adjusted models are shown in the figures. A p-value of < 0.05 was considered statistically significant. The predicted mean probability of HF was calculated as average marginal effects and contrasts using the Delta-method. The Delta-method is a statistical method, which we used to calculate the mean probability of HF in different categories [25]. We used STATA version 16.0 and 17.0 (Stata Corporation, Texas, USA) for statistical analyses and to create the artwork.

Results

We analysed data from 981,383 inhabitants (about a tenth of the Swedish population) aged 20 years and older living in Scania in the last week of 2015. The study population consisted of 50.85% women and 49.15% men. The total prevalence of HF was 1.93% (9,630 patients) in women and 2.19% (10,563 patients) in men. Men had a higher odds ratio for HF − 1.70 (95% CI 1.65–1.75) - than women. Men had a higher prevalence and mean probability of HF than women across all age groups, and the difference in prevalence and mean probability of HF between the genders increased continuously with advancing age. In the age group 80, the difference in prevalence reached 3.01% and the mean probability was 0.03 between the genders (Figure 1).

Figure 1. The mean probability of heart failure in women and men adjusted for age with 95% confidence intervals, using Delta method.

The total prevalence of multimorbidity was 42.82% in women (213,685 patients) and 33.89% in men (163,470 patients) in 2015. The women had a higher prevalence of multimorbidity in all age groups and CNI percentiles, but included a lower prevalence of HF, when compared to men. Men had also a higher prevalence of HF among the multimorbid patients than women in all CNI percentiles (Table 1).

Table 1. Prevalence of heart failure and multimorbidity in all age groups and CNI (Care Need Index) percentiles, and prevalence of heart failure in the multimorbid patients in women and men.

	Women					Men
CNI percentiles	HF		MM		HF/MM %	HF		MM		HF/MM %
	N	%	N	%		N	%	N	%
CNI 1
age
20	2	0.03	1243	20.29	0.16	2	0.03	780	11.57	0.26
30	4	0.04	2403	25.01	0.17	6	0.07	1138	13.74	0.53
40	9	0.07	3670	29.55	0.25	25	0.20	2408	19.53	1.04
50	20	0.23	3669	41.58	0.55	68	0.76	3070	34.18	2.21
60	102	1.01	6059	60.10	1.68	209	2.25	5437	58.61	3.84
70	240	3.53	5317	78.20	4.51	442	6.75	5129	78.35	8.62
80	507	15.13	2987	89.14	16.97	442	19.53	2016	89.09	21.92
CNI 2
age
20	3	0.04	1538	19.00	0.20	4	0.05	879	10.88	0.46
30	1	0.01	2091	25.06	0.05	3	0.04	1143	14.74	0.26
40	9	0.09	3134	29.68	0.29	24	0.24	2119	20.79	1.13
50	29	0.30	4057	42.23	0.71	65	0.70	3231	34.80	2.01
60	127	1.26	5767	57.43	2.20	234	2.42	5578	57.58	4.20
70	257	3.89	5017	75.97	5.12	381	6.20	4640	75.47	8.21
80	548	14.66	3221	86.17	17.01	429	1.89	2222	87.48	19.31
CNI 3
age
20	1	0.01	1609	19.48	0.06	5	0.06	936	10.75	0.53
30	6	0.08	1977	25.55	0.30	4	0.05	1063	14.15	0.38
40	7	0.08	2763	33.23	0.25	24	0.29	1787	21.58	1.34
50	35	0.47	3216	43.37	1.09	67	0.93	2577	35.68	2.60
60	95	1.22	4660	59.70	2.04	218	2.88	4409	58.19	4.94
70	253	4.71	4176	77.72	6.06	371	7.92	3607	77.02	10.29
80	586	16.81	3059	88.76	18.93	415	20.09	1801	87.17	23.04
CNI 4
age
20	1	0.01	1486	21.10	0.07	5	0.07	788	11.15	0.63
30	7	0.10	1754	25.99	0.40	4	0.06	1009	15.10	0.40
40	9	0.12	2577	33.12	0.35	34	0.43	1774	22.21	1.92
50	36	0.49	3261	44.59	1.10	85	1.16	2772	37.74	3.07
60	105	1.45	4297	59.38	2.44	245	3.23	4505	59.36	5.44
70	252	5.34	3601	76.28	7.00	378	8.06	3587	76.51	10.54
80	560	18.06	2708	87.33	20.68	423	20.93	1762	87.18	24.01
CNI 5
age
20	0	0.00	1301	20.53	0.00	2	0.03	700	10.84	0.29
30	3	0.05	1788	27.37	0.17	4	0.06	925	13.94	0.43
40	11	0.16	2244	33.36	0.49	24	0.34	1605	22.44	1.50
50	18	0.30	2680	45.06	0.67	56	0.90	2289	36.96	2.45
60	76	1.31	3443	59.26	2.21	180	3.04	3407	57.62	5.28
70	183	4.86	2866	76.06	6.39	284	7.97	2763	77.50	10.28
80	451	17.31	2306	88.49	19.56	354	22.52	1380	87.79	25.65
CNI 6
age
20	1	0.01	1698	19.47	0.06	2	0.02	1071	11.38	0.19
30	2	0.03	1989	25.51	0.10	6	0.08	1116	14.04	0.54
40	13	0.14	2942	32.42	0.44	28	0.30	2025	21.99	1.38
50	41	0.50	3620	43.99	1.13	68	0.82	3075	37.05	2.21
60	127	1.54	4781	57.80	2.66	205	2.51	4709	57.67	4.35
70	272	4.43	4709	76.71	5.78	347	6.53	4077	76.66	8.51
80	642	14.73	3768	86.46	17.04	467	18.41	2180	85.96	21.42
CNI 7
age
20	3	0.04	1484	17.44	0.20	7	0.07	936	9.82	0.75
30	4	0.06	1659	23.86	0.24	6	0.08	1019	13.23	0.59
40	12	0.17	2226	31.61	0.54	21	0.29	1582	21.50	1.33
50	27	0.43	2705	42.79	1.00	66	1.05	2291	36.51	2.88
60	91	1.37	3809	57.17	2.39	185	2.87	3609	55.90	5.13
70	229	4.74	3651	75.57	6.27	314	7.66	3084	75.24	10.18
80	605	17.12	3025	85.62	20.00	388	19.06	1751	86.00	22.16
CNI 8
age
20	2	0.02	2390	20.33	0.08	3	0.03	1211	11.37	0.25
30	3	0.03	2778	26.73	0.11	16	0.15	1582	14.89	1.01
40	14	0.15	3224	33.45	0.43	31	0.32	2214	22.99	1.40
50	50	0.54	4139	44.86	1.21	95	1.11	3174	37.21	2.99
60	114	1.28	5235	58.71	2.18	232	2.86	4543	56.10	5.11
70	269	4.58	4450	75.77	6.04	378	7.91	3581	74.95	10.56
80	625	15.58	3464	86.36	18.04	387	19.08	1730	85.31	22.37
CNI 9
age
20	0	0.00	2194	18.32	0.00	3	0.03	1136	10.20	0.26
30	4	0.04	2411	23.78	0.17	7	0.06	1565	13.74	0.45
40	10	0.13	2468	31.44	0.41	29	0.34	1847	21.64	1.57
50	29	0.41	3185	44.51	0.91	74	0.97	2744	35.84	2.70
60	117	1.82	3661	56.87	3.20	217	3.49	3287	52.94	6.60
70	253	5.16	3687	75.18	6.86	307	7.65	2882	71.85	10.65
80	626	15.47	3344	82.63	18.72	353	17.50	1670	82.80	21.14
CNI 10
age
20	6	0.04	2311	17.06	0.26	4	0.03	1239	9.75	0.32
30	6	0.06	2334	24.46	0.26	10	0.09	1612	13.72	0.62
40	18	0.26	2601	37.30	0.70	30	0.37	1907	23.46	1.57
50	29	0.49	2879	48.68	1.01	90	1.35	2578	38.55	3.49
60	130	2.80	2712	58.50	4.79	175	3.75	2528	54.16	6.92
70	260	8.20	2319	73.13	11.21	268	10.39	1787	69.26	15.00
80	443	19.24	1891	82.15	23.43	228	19.86	895	77.96	25.47

HF = heart failure.

MM = Multimorbidity.

N = number of patients.

Men had a higher mean probability of HF in all multimorbidity levels and the difference increased with each multimorbidity level, except the MM0 group, when compared to women. The predicted mean probability of HF as a comorbidity in men belonging to the age group 70 was 0.06 (95% CI 0.05 − 0.06) in the MM2 (5-9 chronic conditions) group, and 0.10 (95% CI 0.07 − 0.13) in the MM3 (>10 chronic conditions) group, if compared to women as reference (Figure 2).

Figure 2. The difference in mean probability of heart failure between the genders with increasing level of multimorbidity used women as reference with 95% confidence intervals, using Delta method. MM0 = less than two conditions (not multimorbid); MM1 = 2-4 chronic conditions; MM2 = 5-9 chronic conditions; MM3 = more than 10 chronic conditions.

The women listed at the primary healthcare centres with the most affluent CNI percentile had the lowest prevalence of HF in the whole study population. The men listed at the primary healthcare centres with the most deprived CNI percentile had the highest prevalence of HF between 50 to 80 years of age, while the women belonging to this CNI percentile had the highest prevalence of HF from 60 years of age, if compared to the more affluent CNI percentiles (Table 1). The disparity in the mean probability of HF between the most affluent and deprived CNI percentile was more apparent in women compared to men (Figure 3). The most obvious disparity was observed in the HF patients from 80 years: Women had about a 14 times higher disparity of mean probability of HF than men between the most affluent and deprived percentile (Figure 3).

Figure 3. Disparities in mean probability of heart failure between the genders in the most affluent and deprived CNI (Care Need Index) percentile with 95% confidence intervals, using Delta method.

Men had an increasing mean probability of HF and confidence interval with advancing age compared to women in most CNI percentiles. Most values were statistically significant from 50 years of age with the exception of the age group 80 in CNI percentiles 7 and 10, indicating no difference in mean probability of having HF between the genders in these groups (Figure 4).

Figure 4. The difference in mean probability of heart failure in all age groups and CNI (Care Need Index) percentiles between the genders used women as reference with 95% confidence intervals, using Delta method.

Discussion

The total prevalence of HF was 1.93% (9,630 patients) in women and 2.19% (10,563 patients) in men. Men had a higher OR for HF − 1.70 (95% CI 1.65–1.75) - than women. The mean probability of men, compared to women, having HF increased significantly with advancing age and multimorbidity level. The women had a higher prevalence of multimorbidity in all age groups and CNI percentiles, but a lower prevalence of HF, when compared to men. Men had also a higher prevalence of HF among the multimorbid patients than women in all CNI percentiles. The disparity in the mean probability of having HF between the most affluent and deprived CNI percentile was more apparent in women, which suggests that socioeconomic deprivation was more strongly associated with HF in women than in men. Growing evidence supports the cardioprotective effect of oestrogen [26], which could explain the lower prevalence of HF in women within each CNI percentile, although our results indicate that socioeconomic deprivation was independently associated with HF as well as gender and multimorbidity.

Strengths and weaknesses of the study

The main strength of this study was the large study population comprising almost one million inhabitants, corresponding to about 10% of the Swedish population. HF is known to be strongly associated with advancing age, but our results even analysed a crude difference between the genders having HF in relation to multimorbidity level and socioeconomic status.

This study has several limitations. All patients with the diagnosis code I50 were included, which did not distinguish the subtypes of HF characterised by different aetiologies and multimorbidity levels. Since HF might appear with non-specific symptoms, we suspect that this condition is underdiagnosed. In particular, HFpEF, is more difficult to diagnose than HFrEF, due to a lack of objective findings of reduced ejection fraction during echocardiography.

We have no data on the severity of HF, which is an important factor regarding the outcome and mortality. Furthermore, we had no access to many risk factors of HF, such as heredity, congenital heart disease, drug abuse or obesity. The data on comorbidities with a high impact on mortality were also lacking. Many patients were most likely underdiagnosed concerning the conditions, which had no impact on their quality of life or prognosis, thus resulting in a lower prevalence of multimorbidity and multimorbidity levels. The low availability of cardiologic consultations in primary health care could influence the treatment outcome and the prognosis of HF. We had no data on the quality of health care in the neighbourhood.

Findings in relation to other studies

Our findings have similarities with studies conducted in other countries [27,28], which increases the credibility and validity of our results. A prior study reported that men were approximately five years younger than women at the time of HF diagnosis [27]. Some improvements for men have been observed, probably due to improved primary coronary interventions, but this sex difference persisted over time and is still consistent with our findings [29]. An American epidemiologic cohort study also reported that HF is far more prevalent in men than women across all age groups from the age of 20, with more apparent differences among elderly people [28]. Various chronic disorders tend to coexist and predominate as possible etiologies for HF including coronary disease, hypertension, and diabetes mellitus [30]. All these disorders are associated with socioeconomic deprivation [31], which contributes to increased multimorbidity levels as well.

With the prevention of cardiovascular diseases, obesity and smoking [32], HF occurs later in their life but still affects mostly male, elderly, multimorbid and socioeconomically deprived individuals [33]. It is most likely a complex pathway between socioeconomic deprivation and disparities in the prevalence of HF between the genders. Further studies are warranted to investigate the specific risk factors for HF and multimorbidity associated with age, gender and socioeconomic deprivation to improve general health. The more we learn about the risk factors and diagnoses constituting multimorbidity, the more we can improve public health. Multimorbidity is most likely a continuous scale of interactive chronic disorders rather than single diagnoses. Analyzing the composition of multimorbidity in HF patients should be the next step towards increased knowledge about their risk factors. Our data revealed a high prevalence of multimorbidity in young women, who could have diagnoses such as thyroid disease, asthma and psychiatric disorders, which are common chronic disorders in young people [34]. Socioeconomic inequality is a political issue and is known to have health consequences. Healthcare staff could offer more preventive strategies for the socioeconomically deprived population, in order to diminish this inequality indirectly.

Meaning of the study and conclusions

The prevalence of HF differs significantly between the genders. Men had an increasing mean probability of HF with advancing age and multimorbidity level compared to women. HF constitutes one of the SES-related conditions and most likely worsens the prognosis of multimorbid patients. Standardized pathways should be implemented nationally to diagnose and provide care for HF patients in primary health care. Prevention of socioeconomic inequality could be an important approach to reducing the prevalence of HF, especially in women.

Consent for publication

Not applicable.

Statement of ethics

Data in the present study are based on anonymised information provided by Scania County Council. They provided anonymised information for research purposes once the study had been approved by the Ethics Committee at Lund University (application no. 2018/778). All analyses were performed in accordance with relevant guidelines and regulations.

Due to the requirement for anonymised data, each individual could not be asked for consent to participate; the active refusal of participation was instead applied. This was done by publishing information about the planned study in the Swedish local newspaper “Sydsvenskan”. The advertisement outlined the study and contained information on how to contact the research manager (first author) to opt out of the study.

Acknowledgements

We thank Scania County Council for providing the patient data to enable this study to take place. We are indebted to Patrick O’Reilly for his expertise and invaluable advice in proofreading the manuscript.

Disclosure statement

The authors declare that they have no competing interests.

Additional information

Funding

Lund University, Sweden, funded the supervisors as a resource for this study.