Dyspnea is frequently reported by the elderly, with an estimated prevalence of 20-60%.1,2 It is the symptom most commonly reported by patients with heart failure (HF),3,4 although with low specificity, since it is also frequent in prevalent conditions such as chronic obstructive pulmonary disease (COPD), depression, obesity, anemia and coronary artery disease.2

The pathophysiology of dyspnea is complex, with various etiologies and mechanisms.5–8 Among the different types of dyspnea, the most common are exertional dyspnea, paroxysmal nocturnal dyspnea (PND) and orthopnea.9–11 A new type has recently been described in patients with HF: ‘bendopnea’ (shortness of breath when bending forward),12 which is associated with an increase in echocardiographic indices of left-sided filling pressures.13

The aim of the present study was to determine the association of exertional dyspnea, orthopnea, PND and bendopnea with chronic disease, especially HF, and their phenotypes in primary care.

This cross-sectional study is part of the DIGITALIS study and included 633 individuals aged between 45 and 99 years enrolled in a primary care program in Niteroi, Rio de Janeiro state, Brazil. The individuals invited to participate were selected according to previously established methods to be representative of the population under study. Data were collected between July 2011 and December 2012. Selected individuals were seen at their nearest health center, where they completed a questionnaire and underwent a clinical, anthropometric and nutritional assessment, blood pressure (BP) measurement, electrocardiogram and echocardiogram, and blood and urine samples were collected for laboratory tests, all on the same day. The study design and the results of the main study have been previously published.14

BP was measured using an Omron HEM-711 HC14 monitor, three measurements being taken at 1-min intervals with the subject seated and arm supported at the level of the heart. If the difference between any measurements was greater than 5 mmHg a fourth reading was taken. The first measurement was discarded and the mean of the others was used in the analysis.15

Individuals with hypertension were defined as those who answered affirmatively to question 14.4 of the DIGITALIS questionnaire, were under antihypertensive medication, or had mean systolic blood pressure (SBP) ≥140 mmHg or mean diastolic blood pressure (DBP) ≥90 mmHg. Patients previously diagnosed as hypertensive were considered to be controlled if they had BP <140/90 mmHg.15

Tissue Doppler imaging (TDI) was performed on an Acuson Cypress 20 (Siemens) or an Esaote AU3 Partner by two experienced echocardiographers who were blinded to the results of other exams, following the American Society of Echocardiography/European Association of Electrocardiography guidelines for chamber quantification.16

Systolic function was assessed by measuring left ventricular (LV) ejection fraction (LVEF) by Simpson's method and longitudinal strain (S’). Left atrial volume was determined by the biplane disc method (modified Simpson's rule) in apical 4- and 2-chamber views at end-systole and indexed to body surface area. Diastolic function parameters were assessed as the mean of five consecutive cardiac cycles. Early (E) and late (A) transmitral flow and E-wave deceleration time were calculated and myocardial relaxation velocity during early diastole (E’) was measured by TDI in the septal segment of the mitral annulus.16

The study endpoints were the four types of dyspnea. They were considered to be present if the answers to the following questions were affirmative: (1) “Do you feel breathless on exertion?” (exertional dyspnea); (2) “Do you feel breathless when lying down?” (orthopnea); (3) “Do you wake up at night feeling breathless after being asleep for several hours?” (PND); and (4) “Do you find it difficult to bend over or kneel?” with the answers “Slightly difficult” or “Very difficult” (bendopnea).

A diagnosis of HF with reduced ejection fraction (HFrEF) was made in patients with a history of HF, or signs and symptoms of HF, and LVEF <50%.17 A diagnosis of HF with preserved ejection fraction (HFpEF) was made in patients with a history of HF and LVEF ≥50% and diastolic dysfunction.18

HFpEF was defined in accordance with the criteria of the European Society of Cardiology as the presence of signs or symptoms of HF, LVEF ≥50%, LV end-diastolic volume index <97 ml/m2, and evidence of diastolic dysfunction.17 Diastolic dysfunction was defined as E/E’ ratio >15 on TDI; if E/E’ is suggestive of diastolic dysfunction (8-15), other echocardiographic parameters should be used to confirm the diagnosis such as LV mass index (≥122 and ≥149 g/m2 for women and men, respectively), left atrial volume index (≥40 ml/m2), or E/A ratio <0.5 with E deceleration time >280 ms. An electrocardiogram showing atrial fibrillation together with E/E’ ratio 8-15 also confirms a diagnosis of HFpEF.17,18

As well as those diagnosed with HFrEF on the basis of a history of HF and LVEF <50%, individuals without a history of HF, but with signs and symptoms of HF at enrollment in the study, New York Heart Association class >II, B-type natriuretic peptide >35 pg/ml and LVEF <50%, were also considered to have HFrEF.

COPD was defined as a previous history of the disease reported by the patient during the assessment in response to the question “Has any doctor ever told you that you have…?” Anemia was defined as hemoglobin <13 g/dl for men and <12 g/dl for women.19 The Patient Health Questionnaire-9,20 validated for use in Brazil,21 was used to identify patients with depression, defined as those with a score of 10, 15 or 20 (moderate, moderately severe, or severe depression).

Subjects’ weight and height were measured to calculate their body mass index (BMI) (weight in kg divided by height in m squared), with obesity defined as BMI ≥30 kg/m2.22 Subjects who answered affirmatively to the question “Has any doctor ever told you that you have hypertension (high blood pressure)?”, were under antihypertensive medication, or had mean SBP ≥140 mmHg or mean DBP ≥90 mmHg, were classified as hypertensive.15 Diabetes was defined as a previous history as shown by a positive reply to the question “Has any doctor ever told you that you have diabetes (high blood sugar)?”, fasting blood glucose of ≥126 mg/dl, or reported use of oral antidiabetic medication and/or insulin.23

The study included 633 individuals enrolled in a primary care program. Most were women (61.8%), aged between 45 and 69 years (80.7%), with monthly income below €278.77 (75.6%), and hypertensive (72.7%) (Table 1).

Dyspnea was present in 43.8% of the subjects. The most common types were exertional dyspnea (26.5%) and bendopnea (26.2%); the prevalence of orthopnea was 8.8% and of PND 7.1%.

Tables 2 and 3 show the unadjusted ORs for exertional dyspnea, orthopnea, PND and bendopnea according to comorbidities. COPD was associated with exertional dyspnea, orthopnea and PND, but not with bendopnea. Depression was the only comorbidity associated with all types of dyspnea, with ORs ranging between 2.68 and 4.19. HF was significantly associated with PND and bendopnea: HFpEF with all types except exertional dyspnea but HFrEF with none (although it presented an OR >2 with PND, probably due to the small number of subjects with this condition).

Table 4 shows the results of multiple logistic regression analysis of all variables that reached p<0.1 in bivariate analysis. Separate models were constructed for HF and HFpEF. After adjustment, HF only maintained an association with PND (OR 3.68, 95% CI 1.43-9.46), while the association with bendopnea (OR 1.84) was without statistical significance (p<0.1). HFpEF lost its significant association with the three types of dyspnea included in the logistic regression analysis following adjustment, with ORs of 3.15 for PND and 2.04 for bendopnea (p<0.1). Depression was strongly associated (p<0.05) in the five logistic regression models, while COPD showed a strong association with orthopnea (OR 6.86) when HFpEF was included but was associated with PND (OR >2) for both models (HF and HFpEF). Age was a protective factor against PND and was not associated with bendopnea.

In the present study, mapping the different types of dyspnea to common morbidities in adults and the elderly revealed a pattern. COPD showed a strong unadjusted association with exertional dyspnea, orthopnea and PND, and a non-significant association with bendopnea. The opposite was true for HF and HFpEF, which showed strong unadjusted associations with bendopnea. Another interesting finding was the weak association of HFrEF with all types of dyspnea except PND. There is plentiful evidence in the literature that COPD as well as HF is a common cause of dyspnea.24

PND is one of the major Framingham criteria for a diagnosis of HF and is also one of the Boston criteria.25,26 Dyspnea is caused by the decreased ability of the heart to fill and empty, producing elevated pressures in the pulmonary blood vessels.27 Albert et al.28 assessed 276 patients with HF and reported a prevalence of self-reported PND of 23.6%, compared to 14.1% in the present study. The greater prevalence of PND observed by Albert et al. may be explained by the fact that their study included both outpatients and hospitalized patients, unlike our study population, which consisted exclusively of non-hospitalized volunteers who were not necessarily ill.

The study by Thibodeau et al.12 of 102 individuals with HFrEF found a prevalence of bendopnea of 29%, compared to 37.9% in our study. They showed that patients with bendopnea had a hemodynamic profile characterized by greater LV filling pressures and reduced cardiac index. The difference between their results and ours may be explained by the fact that we only used answers to the question “Do you find it difficult to bend over or kneel?” to diagnose bendopnea, a less specific method than that used by Thibodeau et al., who asked whether subjects reported shortness of breath within 30 s of bending forward such as when putting on their shoes.

Thibodeau et al. did not investigate patients with HFpEF. In our study, the prevalence of bendopnea in these patients was greater than in HFrEF (51.5% vs. 37.9%), and this association persisted in multivariate analysis, although with a p value between 0.05 and 0.1.

Depression was the only variable in the univariate analysis that showed a strong association with all four types of dyspnea and with three types in the multivariate analysis. There is ample evidence of a link between anxiety and depression and respiratory symptoms,29,30 but the cause/effect relationship between psychological factors and dyspnea is not fully understood.31 Some studies indicate that dyspnea can trigger psychiatric disorders, while others suggest that psychiatric disorders, particularly depression, intensify the subjective experience of dyspnea.31–34

Although they differ in certain aspects, the pathophysiological processes and the clinical signs and symptoms of COPD and HF often overlap,35 as do the terms used by COPD and HF patients to describe their shortness of breath, and the frequency with which they experience dyspnea is similar.7,36,37 This can make it difficult to distinguish between the two diseases. In the present study, the self-reported prevalences of HF and COPD were low (8.2% and 3.5%, respectively), and only three (5.8%) of the 52 participants who stated they had been previously diagnosed with HF had also been diagnosed with COPD. This is in agreement with community studies that report prevalences of COPD in HF patients ranging between 7 and 13%.35

In the present study, subjects with BMI ≥30 kg/m2 were more likely to present exertional dyspnea or bendopnea in univariate analysis, an association that was only maintained with bendopnea in multivariate analysis. Furthermore, waist circumference of ≥94 cm for men and ≥80 cm for women was significantly associated with bendopnea in univariate analysis. An association between obesity and dyspnea is reported in the literature.38 The mechanisms involved in dyspnea in the obese include reduced end-expiratory lung volume39 and increased respiratory work for all levels of exertion.40 The patients with bendopnea in the study by Thibodeau et al.12 also had higher BMI than those without, which may have increased their discomfort when bending over. Similarly, in the present study there was a greater prevalence of BMI ≥30 kg/m2 in subjects with bendopnea than with other types of dyspnea. Invasive hemodynamic monitoring by Thibodeau et al. showed that bending over increased venous return and filling pressures, causing shortness of breath. The authors pointed out that obese subjects presented higher filling pressures before leaning over, and thus were more likely to reach the pressure threshold that would cause symptoms.12

The present study used self-reporting to determine certain important variables such as the presence of COPD. Self-reporting of chronic diseases is generally reasonably accurate,41–43 and using previous diagnoses reported by patients themselves is a method that has been validated in clinical trials in other countries, in both men and women.44,45 In Brazil, a household survey on risk behavior and morbidity from non-communicable diseases used self-reporting to establish the prevalence of ischemic heart disease in the population.46 An important advantage of this method is that it facilitates comparisons with other countries, since the considerable differences in health care systems between countries mean that data from medical records are often not easily comparable.43 Studies have found self-reported prevalences of COPD to be fairly reliable but probably underestimated.47 In general, it can be stated that the findings of the present study are reasonably reproducible.

The study has certain limitations. Its cross-sectional nature makes it impossible to establish whether dyspnea was the result of HF or of other comorbidities. It should also be borne in mind that the definition of bendopnea used differs from that of Thibodeau et al., which is more specific; in the present study individuals with musculoskeletal problems may have been incorrectly classified as having bendopnea.

Furthermore, the question “Do you feel breathless on exertion?” did not take the time factor into account and thus may have excluded individuals with this symptom who at the time of the survey were compensated and thus less symptomatic.

Diagnosing COPD without performing objective respiratory tests is another limitation, although these individuals were being followed by their family physician and would generally have been aware of their condition.

The criteria used to classify subjects as having diabetes did not include a second measurement of blood glucose, which may have led to overestimation of its prevalence.

Since only 22 participants presented COPD, 35 HFpEF and 29 HFrEF, the lack of statistical significance for some findings may reflect inadequate statistical power.

Bendopnea was the only type of dyspnea not linked to respiratory disease or coronary heart disease. Even after adjusting for depression and BMI, the association remained with HF and HFpEF, and bendopnea thus appears to be a promising symptom to differentiate heart failure from the other two disease groups. These associations could also help improve triage of patients with suspected HF and referral for further investigation.

The authors have no conflicts of interest to declare.