Original Research

Bilateral leg edema, pulmonary hypertension, and obstructive sleep apnea

Author and Disclosure Information

A cross-sectional study



This study was undertaken to clarify whether pulmonary hypertension is a useful marker for underlying obstructive sleep apnea in patients with edema. Twenty-eight ambulatory adults with bilateral leg edema and a normal echocardiogram were enrolled. Sixteen subjects had pulmonary hypertension, and 12 subjects had normal pulmonary artery pressures. Spirometry, pulse oximetry on room air, and polysomnography were obtained for each subject. Ten of 16 (63%) pulmonary hypertension subjects and 9 of 12 (75%) nonpulmonary hypertension subjects had obstructive sleep apnea (P = .48). Eleven of 16 (69%) pulmonary hypertension subjects and 11 of 12 (92%) nonpulmonary hypertension subjects were obese (P = .20). If these results are generalizable, obstructive sleep apnea is frequently associated with bilateral leg edema and obesity, regardless of the presence of pulmonary hypertension. Thus, especially in obese patients, bilateral leg edema may be a useful clinical marker for underlying obstructive sleep apnea.

We previously found an association between bilateral leg edema and pulmonary hypertension in primary care patients.1 After consideration of the differential diagnosis of pulmonary hypertension, obstructive sleep apnea was deemed the most likely explanation for the high frequency of pulmonary hypertension.2 Subsequently, we identified an association among leg edema, obesity, pulmonary hypertension, and obstructive sleep apnea in ambulatory patients with normal left ventricular function.3

Our earlier data failed to clarify whether leg edema, obesity, pulmonary hypertension, or a combination thereof is the most useful marker for obstructive sleep apnea. This cross-sectional study was undertaken to determine whether subjects with bilateral leg edema and pulmonary hypertension have a higher frequency of obstructive sleep apnea than edematous subjects with normal pulmonary artery pressures.


A single physician (R.P.B.) enrolled a convenience sample of subjects from an inner city group family practice in Cleveland OH, from July 1995 to September 1997, and from a 2-physician suburban family practice near Cleveland, OH, from October 1997 to July 2000. Ambulatory patients older than 18 years with bilateral pitting leg edema, no clinically overt lung disease, no echocardiographic evidence of a cardiac abnormality, and an echocardiogram that permitted an estimation of the pulmonary artery pressure were eligible to participate in the study. The methodology for estimating the pulmonary artery pressures has been described previously.3-5 For this study, pulmonary hypertension was defined as an estimated pulmonary artery systolic pressure > 30 mm Hg, whereas an estimated pulmonary artery systolic pressure 30 mm Hg was considered normal.

Subjects were excluded if their echocardiogram revealed valvular heart disease, congenital heart disease, or left ventricular systolic or diastolic dysfunction; if they used dihydropyridine calcium antagonists; if they had a known pulmonary condition; or if pulmonary function evaluation indicated the presence of obstructive or restrictive lung disease. Individuals with asthma were included as long as the asthma was well controlled. The protocol was approved by the Institutional Review Board at the MetroHealth Medical Center (Cleveland, OH).

The medical history of each subject was reviewed for risk factors recognized as being associated with pulmonary hypertension,3 and subjects answered the Epworth sleepiness scale questions.6 The percent predicted forced vital capacity (FVC), the percent predicted forced expiratory volume in 1 second (FEV1), and the FEV1 in relation to the FVC were determined by spirometry (Brentwood Spiroscan 2000, Hoks Electronics, Inc, Japan). Oxygen saturations on room air were determined by oximetry (N-20, Nellcor, Inc, Hayward, CA). Polysomnography was performed on all subjects in a sleep laboratory, and the average number of episodes of apneas and hypopneas per hour of sleep (apnea-hypopnea index) was calculated.

No universally accepted criteria exist for diagnosing obstructive sleep apnea.7 For this study, obstructive sleep apnea was defined as an apneahypopnea index of ≥ 20 events per hour,8 or a rapid eye movement-specific apnea-hypopnea index of ≥ 20 events per hour. Levels of serum albumin, antinuclear antibody, rheumatoid factor, and thyroid stimulating hormone were obtained on all subjects, as were sedimentation rate and results of liver function tests. Subjects were considered obese if they had a body mass index (weight in kg/height in m2) of more than 30 kg/m2.9

Mean values between study groups were compared with Student’s t-test, and 2 statistics were used to compare differences between proportions. A final regression analysis was conducted to test whether controlling for potential confounding variables altered the univariate association observed. A hierarchical logistic regression analysis was performed by first regressing obstructive sleep apnea status on potential confounding variables as the first level, and then allowing pulmonary hypertension status to enter the equation as the second level. These analyses compared the extent to which pulmonary hypertension is associated with obstructive sleep apnea status before and after adjusting for confounding variables.

Next Article: