Obesity-hypoventilation syndrome (OHS) is a combination of daytime hypercapnia and obesity in the absence of other causes of hypoventilation. Pathogenesis is multifactorial, and often associated with sleep apnea–hypopnea syndrome (SAHS). It is a risk factor for heart failure (HF)1,2 and various abnormalities in cardiac morphology and function.3,4

Although HF is diagnosed in clinical practice on the basis of clinical, radiological, analytical and echocardiographic criteria, it in fact reflects a situation that involves inadequate cardiac output (CO) and increased pressures in the heart cavities1; these 2 variables constitute the hemodynamic expression of HF and are probably the most objective way of measuring it. Individuals with OHS often receive treatment with positive pressure devices that can affect their hemodynamic status,5 but experience is limited and sometimes contradictory6,7; although the co-existence of HF in patients with OHS2,8,9 is well known, there is no established consensus for the evaluation and management of this situation in clinical practice guidelines.10,11

Cardiothoracic impedance (CTI) provides a non-invasive estimation of hemodynamic data. It consists of the application of a low frequency alternating current to detect changes in volume and blood flow in the chest. The analysis of these flows has helped validate this technique and provides reliable parameters such as: CO, cardiac index (CI), systolic volume (SV), etc.12–16

Our hypothesis is that CTI can be used to explore the hemodynamics of patients with OHS in a non-invasive manner. Our primary objective was to describe the baseline hemodynamic situation in patients with OHS (as a marker for HF) and secondarily to assess the effect that non-invasive ventilation might have on these parameters.

Patients were consecutively included using a cross-sectional and observational repeated measures design. Hemodynamic monitoring was performed during the inclusion visit using NICCOMO bioimpedance equipment (Medis Medizinische Messtechnik GmbH, Ilmenau, Germany), providing a continuous (beat-to-beat) non-invasive measurement of cardiac function. Two pairs of specific electrodes were placed on the laterocervical and flank regions above the midaxillary line. The procedure took place with the patient reclining at 45° for a period of 60min, the first 30min at baseline and then another 30min receiving NIV with ventilatory parameters prescribed for the patient for home treatment. Hemodynamic data collected were: heart rate (HR), mean arterial pressure (MAP), CO, CI, SV, and mean SpO2; these variables were then analyzed with specific software (Cardiovascular Lab®, Ilmenau, Germany) provided by the manufacturer.

CO was taken as a reference parameter for the diagnosis of heart failure and the value obtained was expressed as a percentage of the lower limit of normal (LLN) calculated for each patient using the Cardiovascular Lab® software, based on age, sex, and anthropometric data. HF was defined as CO below the LLN. Two groups were established for the analysis of results: those with normal CO (CO ≥100% LLN) and those with a decreased CO (CO <100% LLN). To determine the effect of NIV on the patient's hemodynamics, the same parameters were measured after the application of NIV, and compared with baseline measurements.

The estimated sample size was 40 patients with OHS, with a confidence level of 95% and an accuracy of 15%. It was assumed that the prevalence of OHS in the general population is 0.6%6 and that the heart failure rate is unknown (P=0.5).

Values are expressed as mean and ranges for quantitative variables, and as percentages for qualitative variables. Statistical analysis of quantitative variables was performed using the Mann–Whitney U test for comparison of independent variables and the Wilcoxon test for paired variables. The Chi-squared test was used for qualitative variables. A P-value of <0.05 was considered significant.

A total of 42 patients were studied, 6 of whom were excluded due to poor quality studies. The final sample was 36 cases, 66 years of age (±8), 19 (52%) males. Twenty-four (66%) had high blood pressure, 18 (50%) were diabetic, and 18 (50%) had had at least one previously diagnosed episode of heart failure. Twenty cases (56%) had severe SAHS with an average AHI of 62/h (38–110) and 16 had mild–moderate SAHS with an AHI of 15/h (8–23).

Table 1 shows the general, functional, and hemodynamic characteristics and ventilatory parameters of all patients included in the study. Heart failure was detected in 17 cases (46%), with a mean CO of 3.7l/min (66% LLN).

In the HF group (N=17), 11 subjects were women and 6 were men (P=0.024), 9 patients had hypertension (P=0.15), 9 were diabetics (P=0.31), 8 had previous episodes of HF (P=0.62), and 7 had severe SAHS (P=0.1). Mean age was higher and BMI significantly lower than in patients with low CO. There were no differences in lung function, ventilatory parameters, or LVEF obtained by echocardiography. Table 1 shows the comparative analysis between the two groups.

Table 2 shows the hemodynamic changes that occurred in each patient group after the application of NIV. The group of patients with low CO showed a significant improvement in CO, CI, and SV, while those with normal CO did not experience changes.

The echocardiographic study was carried out in 26 cases, while the remaining cases were unevaluable due to poor echocardiographic views and failure to complete all measurements. Mean pulmonary artery pressure was 36mmHg; in 7 (26%) patients, figures were consistent with pulmonary hypertension (>40mmHg), and LVEF was normal in all cases (greater than 50%).

Almost half (46%) of our patients with stable OHS receiving NIV had hemodynamic dysfunction, with low CO consistent with heart failure. The use of NIV improved CO in patients in HF, but it did not significantly affect patients with normal CO.

Obesity is known to be associated with the presence of cardiological comorbidity and increased risk of episodes of heart failure,1,17,18 and cardiac changes have also been described in OHS patients, especially right heart overload and signs of PHT.3,4 Castro-Añon et al.,19 using echocardiography in patients with OHS, detected right ventricular overload in 43% of cases with preserved LVEF, and Corral et al.20 found echocardiographic alterations in more than half of the patients in their series; progress in both studies was favorable with ventilatory treatment. In our series, pulmonary artery pressure could only be determined in 26 cases, of which 27% had figures indicative of PHT.

The most important finding of our study is that we detected HF in a substantial percentage of patients with OHS based on the measurement of CO. Our series consisted of stable patients who attended scheduled appointments; subjects with HF were predominantly women, older, with a lower BMI than the group without HF. We found no significant differences in the existence of previous episodes of HF, lung function, AHI, LVEF measured by echocardiography, pCO2 values, or ventilator parameters. From a hemodynamic point of view, patients with HF had low SV and CI values, while MAP and heart rate were similar between the 2 groups. One hypothesis for this might be that hemodynamic changes are associated with abnormalities in cardiac function18,20 and would not be related to the underlying ventilatory disease.

Two types of HF are distinguished according to echocardiographic findings: HF with normal LVEF (diastolic HF) and HF with diminished LVEF (low contractility).1 In our series, all cases had LVEF >60%, as observed in the study by Corral et al.20 Otair et al.,9 using echocardiography, also detected diastolic dysfunction in 67% of patients with OHS. These data suggest that HF associated with OHS is predominantly diastolic, with preserved left ventricular contractility.

The application of positive airway pressure may alter hemodynamics in patients with lung disease,6,21,22 although there are conflicting results in this regard. Lukacsovits et al.,6 using non-invasive measurements of CO (Finometer-PRO) in COPD patients, showed that in cases where high intensity ventilation with mean IPAP of 27cmH2O was used, greater reductions were observed in CO than with the use of more moderate pressures (17cmH2O). The authors proposed using lower pressures in cases of underlying heart disease. Leontine et al.,7 using echocardiography in patients with COPD, did not detect changes in CO using slightly lower pressures (23cmH2O). In our study of patients with OHS using maximum pressures of around 20cmH2O in patients with normal LVEF, we found that NIV had different effects depending on the patient's underlying hemodynamic situation: all parameters improved in patients with hemodynamic dysfunction, while those with normal CO showed no changes. Although we believe that our study is the first to describe this situation in a series of patients with OHS, Yoshida et al.23 described the case of a patient with OHS and severe SAHS with hypercapnic respiratory failure monitored with cardiothoracic impedance in whom low CO figures improved rapidly and significantly with the application of positive pressure ventilation. Although there is no definitive explanation for this phenomenon, it could be related to the improvement in oxygenation, decrease in preload and postload, a reduction in cardiac transmural pressure, and in pulmonary circulation resistances.23,24

In our study, patients with normal CO had a higher BMI than those with decreased CO. The association between obesity and HF shows that patients with higher BMI have a higher risk of presenting HF17; this result could be explained by the reduced sample size and should be confirmed in larger series.

Our study has its limitations: we could not perform echocardiography in all patients included and the echocardiography and the CTI procedures were not performed at the same time, so there may be differences in the clinical status of the patients between both examinations; moreover, although the number of patients estimated in the sample calculation were included (40), the final figure assessed was slightly lower (36).

We can conclude that including cardiac and hemodynamic examinations in the management of patients with OHS provides useful information that can help better define their clinical status and their therapeutic needs, and assess their progress and indicate complementary treatments. We found that NIV impacts positively on patients with more altered hemodynamics and normal LVEF, but given our limited data, these findings will have to be confirmed in future studies.

The authors state that they have no conflict of interests.