Bilateral Leg Edema, Obesity, Pulmonary Hypertension, and Obstructive Sleep Apnea FREE

WHEN CARDIAC conditions cause edema, it is due to the development of left or right ventricular dysfunction. When pulmonary diseases initiate edema formation, it is a result of right ventricular dysfunction. A previous study from our group identified an association between bilateral leg edema and pulmonary hypertension,¹ with many of the subjects with pulmonary hypertension having no evidence of cardiac or pulmonary disease.

Half of the subjects with elevated pulmonary artery pressures in the previous study had only mild pulmonary hypertension, their systolic pulmonary artery pressures ranging from 31 to 40 mm Hg, as estimated using Doppler echocardiography. Some cardiologists consider systolic pulmonary artery pressures of less than 35 mm Hg to be within normal limits.² Moreover, since pulmonary artery pressures increase with age in healthy individuals,^3- 6 much the same as systemic blood pressures increase with age,⁷ mild elevations of the pulmonary artery pressure are often considered to be benign. Thus, although many causes of pulmonary hypertension are recognized, often an underlying pathologic explanation is not sought when the pulmonary artery pressure is only mildly elevated. Nonetheless, it is possible that mildly elevated pulmonary artery pressures are indicative of underlying abnormalities.

In the previous study,¹ 19 (42%) of 45 subjects had pulmonary hypertension. Ten subjects with pulmonary hypertension had a simultaneous cardiac or pulmonary condition. Of the remaining 9 patients, 8 had isolated pulmonary hypertension, and 1 patient had pulmonary hypertension with right ventricular dysfunction. None of the 9 had evidence of obstructive or restrictive lung disease, recurrent pulmonary emboli, valvular heart disease, congenital heart disease, left ventricular systolic or diastolic dysfunction, collagen vascular disease, portal hypertension,^8- 9 sickle cell disease, recent pregnancy,¹⁰ choriocarcinoma or hydatidiform mole,⁷ schistosomiasis or filariasis,⁷ oral contraceptive use,^11- 12 intravenous drug abuse,¹³ methamphetamine inhalation,¹⁴ cocaine use,^15- 16 human immunodeficiency virus infection,^17- 19 or use of appetite suppressants.^20- 23

Since obstructive sleep apnea (OSA) is associated with pulmonary hypertension,²⁴ and since OSA is fairly common, being present in 2% to 4% of middle-aged adults,²⁵ we undertook this study to identify the frequency of OSA in ambulatory adult patients with isolated, unexplained pulmonary hypertension who initially presented with bilateral leg edema.

We enrolled 20 patients, 5 from an inner-city family practice in Cleveland, Ohio, from July 1, 1995, through September 30, 1997, and 15 from a suburban family practice near Cleveland from October 1, 1997, through October 31, 1998. The inner-city practice is staffed by 6 family physicians (one of whom was R.P.B.) and a nurse practitioner. Most of the patients are low-income, or blue-collar working class. Multiple ethnicities and races are represented among the patients, most commonly white and Hispanic, and less commonly African American and Asian. Two family physicians (R.P.B. being one) work in the suburban office. The suburban patients are predominantly white. They are blue- and white-collar working class, or retirees.

Ambulatory patients older than 18 years with bilateral pitting leg edema whose echocardiogram demonstrated pulmonary hypertension were eligible to participate in the study. Pulmonary artery pressures were determined via Doppler echocardiography by adding the Doppler systolic transtricuspid gradient to the estimated right atrial pressure using the inferior vena caval diameter and percentage of diameter change during respiration.²⁶ When assessments of the change in the diameter of the inferior vena cava during respiration were not available, 10 mm Hg was used as an approximation for the right atrial pressure.²⁷ For this study, pulmonary hypertension was defined as an estimated pulmonary artery systolic pressure of greater than 30 mm Hg.

Subjects were excluded from participation in the study if their echocardiogram identified valvular heart disease, congenital heart disease, left ventricular systolic dysfunction, or left ventricular diastolic dysfunction. Subjects were excluded if they had a known diagnosis of interstitial lung disease, emphysema, or chronic bronchitis before enrolling in the study, or if pulmonary function evaluation before enrolling in the study indicated the presence of chronic obstructive pulmonary disease or restrictive lung disease. Subjects were also excluded from the study if they were pregnant, had unilateral edema or idiopathic cyclic edema, or were hospital-based patients. The protocol was approved by the institutional review board at the MetroHealth Medical Center, Cleveland. Informed consent was obtained from each subject.

The percentage of predicted forced vital capacity, percentage predicted forced expiratory volume in 1 second, and forced expiratory volume in 1 second in relation to the forced vital capacity were determined by means of spirometry (Spiroscan 2000; Brentwood Medical Technology, Torrance, Calif). Oxygen saturations on room air were determined via oximetry (N-20; Nellcor, Inc, Hayward, Calif). Polysomnography was performed on all subjects, and the average number of episodes of apneas and hypopneas per hour of sleep (apnea-hypopnea index) was calculated. Sleep apnea was defined as an apnea-hypopnea index of no less than 20 events per hour. Levels of serum albumin, antinuclear antibody, and rheumatoid factor; results of liver function tests; and sedimentation rate were obtained. African American and Hispanic subjects underwent a sickle cell blood test. Men were considered obese if they had a body mass index (calculated as weight in kilograms divided by the square of height in meters) of greater than 27.8, and women were considered obese if they had a body mass index greater than 27.3.²⁸

The medical history of each patient was reviewed for emphysema, chronic bronchitis, asthma, interstitial lung disease, scoliosis, sickle cell disease, cirrhosis, rheumatological disorder, human immunodeficiency virus infection, hypertension, diabetes, kidney disease, pulmonary emboli, deep venous thrombosis, cancer, parasitic infection, hepatitis, alcoholism, and use of intravenous drugs, inhaled methamphetamine, cocaine, and appetite suppressants. The medical history of each woman was reviewed for oral contraceptive use, recent pregnancy, choriocarcinoma, and hydatidiform mole. The medications of each subject were reviewed for digoxin, diuretics, nitrate medications, theophylline products, oxygen, inhaled bronchodilators, nonsteroidal anti-inflammatory medications, corticosteroids, estrogens, progesterone, testosterone, calcium channel blockers, and other antihypertensive agents.

Patients were asked to report the duration of the edema; cigarette smoking status; whether they had orthopnea, paroxysmal nocturnal dyspnea, or dyspnea on exertion; whether they snored; and whether there were pauses in their snoring. Patients answered the questions comprising the Epworth Sleepiness Scale.²⁹

Descriptive statistics include percentages for categorical variables and means and SDs for continuous measures. To test the association of the categorical demographic and historical information with OSA, χ² tests were used. Because of the small sample size involved, Fisher exact test was used when appropriate. Mean values were compared with t test.

Twenty patients enrolled in the study, 5 of whom declined to complete a polysomnogram. For the remaining 15 patients, ages ranged from 41 to 81 years (Table 1). Ten of the subjects were women and 13 were white. Most patients had mild pitting edema, usually of 1+ or 2+ severity. Nine subjects reported that the edema had been present for more than 2 years. Fourteen of the subjects were obese.

The comparison of the subjects who completed a polysomnogram with those who did not revealed no significant differences in age, sex, race, marital status, education, duration of edema, body mass index, pulmonary artery pressure, spirometric measurements, oxygen saturation, and sedimentation rate. There were no significant differences between groups for most of the symptoms evaluated, except the subjects who completed a polysomnogram were more likely to experience dyspnea on exertion (for those patients who provided this information, 10/14 vs 0/4; P = .02), were more likely to snore (for those patients who provided this information, 12/14 vs 1/5; P = .02), and were more likely to report daytime sleepiness (Epworth Sleepiness Scale score, 8.9 ± 4.9 vs 3.0 ± 1.9; P = .03).

The disease states and conditions associated with bilateral leg edema and pulmonary hypertension are shown in Table 2. None of the 9 subjects with OSA demonstrated an obstructive pattern on spirometric evaluation, but 9 of 15 subjects had a restrictive spirometry pattern, consistent with their obesity. One subject had asthma, which was well controlled with medication. One subject with OSA also had right ventricular dysfunction. Neither subject who had used appetite suppressants could remember the name of the medication. None of the other risk factors associated with pulmonary hypertension were identified in any of the subjects.

Table 3 compares demographic, clinical, and symptom characteristics between the 9 subjects who had OSA and the 6 subjects who did not. In addition to having significantly higher apnea-hypopnea indexes, the OSA group had significantly higher pulmonary artery pressures. Six of the subjects with OSA had mild pulmonary hypertension: 3 subjects had a pulmonary artery systolic pressure ranging from 31 to 35 mm Hg, and 3 subjects had a pulmonary artery systolic pressure ranging from 36 to 40 mm Hg. Three subjects with OSA had pulmonary artery systolic pressures of greater than 40 mm Hg. Subjects with OSA were significantly more likely to have systemic hypertension than subjects without OSA, and there was a trend toward the subjects in the OSA group being more obese.

There were no significant differences between subjects with and without OSA in age, race, sex, marital status, level of education, spirometry measurements, percentage of oxygen saturation, cigarette smoking, duration of edema, dyspnea on exertion, orthopnea, paroxysmal nocturnal dyspnea, snoring patterns, feeling tired on awakening, having difficulty staying awake during the day, scores on the Epworth Sleepiness Scale, and use of appetite suppressants. There were no significant differences between groups in any prescription medications.

We found that OSA is associated with pulmonary hypertension in patients with bilateral leg edema, most of whom are obese. Since a previous study demonstrated that bilateral leg edema is associated with isolated pulmonary hypertension,¹ the implication of both studies is that bilateral leg edema is associated with OSA in ambulatory adults. Furthermore, if these findings are generalizable, then bilateral leg edema may be a useful clinical marker for underlying OSA.

Owing to the small sample size and the possibility of selection bias, our results should be interpreted with caution. These findings need to be replicated with a larger sample to confirm the association. Moreover, lacking a control group, our data do not allow an estimation of the magnitude of the risk for sleep apnea conferred by pulmonary hypertension with bilateral leg edema compared with bilateral leg edema in the absence of pulmonary hypertension.

Nonetheless, our data raise the question of a causal relationship among sleep apnea, pulmonary hypertension, and bilateral leg edema. The absence of other known risk factors for pulmonary hypertension—in particular, the absence of daytime hypoxemia, significant spirometric abnormalities, and left ventricular dysfunction—suggests that the relationship may be a valid one. We speculate that the hypoxemia caused by apnea during sleep induces the pulmonary hypertension.

Most of the participants in our study have not been using nasal continuous positive airway pressure for long, but 2 subjects who have been using nightly nasal continuous positive airway pressure for more than 2 years have experienced reduced leg edema, without using any diuretic medication. One of these individuals has undergone an echocardiogram since starting the nasal continuous positive airway pressure therapy, and the pulmonary artery pressure has normalized in this patient. Although these 2 cases do not prove a causal relationship between OSA, pulmonary hypertension, and bilateral leg edema, they are suggestive.

Although 15% to 25% of patients with OSA have pulmonary hypertension,^30- 34 the bulk of the research literature argues against there being a causal relationship. Only 3 studies suggest that sleep apnea itself may cause pulmonary hypertension.^33,35- 36 Most research suggests that daytime hypoxemia attributable to abnormal pulmonary functioning in patients with sleep apnea is the cause of the pulmonary hypertension.^31,37- 41 These latter studies found that the pulmonary hypertension correlates better with daytime hypoxemia than with the severity of the OSA. However, a recent study³⁶ indicates that the severity of the sleep apnea may not be the critical variable in the development of pulmonary hypertension, but that some individuals with OSA and pulmonary hypertension have heightened pulmonary artery pressor responses to hypoxia.

If pulmonary hypertension associated with obesity and OSA is a cause of bilateral leg edema, then the results of our study may contradict conventional notions of cardiopulmonary physiology. Although right ventricular dysfunction, or cor pulmonale, is recognized as one of the manifestations of congestive heart failure, our findings indicate that pulmonary hypertension without echocardiographically demonstrable right or left ventricular failure is associated with bilateral leg edema. The major textbooks of cardiology and internal medicine do not include pulmonary hypertension or OSA in the differential diagnosis of bilateral leg edema.^42- 47

One of several mechanisms of edema formation seems possible. Since the hypoxemia that occurs during OSA increases sympathetic nervous system activity, thereby contributing to systemic hypertension,^48- 50 a possible mechanism to explain edema formation is that nocturnal hypoxemia leads to neuroendocrine activation, which in turn leads to salt and water retention. If this mechanism is correct, then the differential diagnosis of edema should be expanded to include OSA.

A second conceivable mechanism is that obesity, without any intervening cardiopulmonary factors, is the cause of edema formation. In this scenario, the edema might result from increased hydrostatic pressure in the venous circulation, perhaps due to increased venous pressure in the organs of the abdomen or pelvis, or due to obesity-induced effects on the veins of the lower extremities. If this explanation is correct, lymphedema due to increased hydrostatic pressure in the lymphatic system may contribute to the swelling.

Another plausible hydrostatic mechanism is that edema results from the pulmonary hypertension as the increased pressure is transmitted from the pulmonary artery to the right ventricle, the right atrium, and ultimately through the vena cava and into the venous circulation of the legs, perhaps contributing to lymphedema in the process. If this mechanism is correct, then the differential diagnosis of edema should be expanded to include pulmonary hypertension.

Previous research indicates that pulmonary hypertension in and of itself is inadequate to cause edema, at least in some instances. Some healthy individuals living at high altitudes have pulmonary hypertension without evidence of heart failure.⁵¹ In these individuals, chronic pulmonary hypertension allows right ventricular hypertrophy to maintain blood flow.⁵² Among patients with primary pulmonary hypertension, virtually none have edema at the time of initial diagnosis,^13,53 but, with time, edema develops in increasing numbers of these individuals.⁵³ These observations suggest that a factor apart from pulmonary hypertension is necessary for edema formation.

Last, a fourth hypothetical mechanism is that edema results from intermittent right ventricular failure. During episodes of sleep-associated hypoxemia, acute elevations in pulmonary artery pressure occur.^54- 55 The thin-walled right ventricle does not tolerate pressure overload well, and the response of the right ventricle to pulmonary hypertension depends on the acuity and the severity of the pressure load.⁴⁵ It is possible that in some patients with OSA and baseline pulmonary hypertension, right ventricular failure occurs nocturnally during periods of hypoxemia-induced elevations of the pulmonary artery pressure, and then normalizes during waking hours when the hypoxemia improves. If this explanation is correct, then echocardiograms obtained during wakefulness would identify the baseline pulmonary hypertension but fail to detect the right ventricular dysfunction that occurs only during periods of sleep.

Theoretical issues aside, we believe that our study has immediate clinical relevance. If bilateral leg edema occurred only rarely, then this research would be applicable to only a few patients and the details of the cardiovascular pathophysiology would be of interest to only a few specialists and researchers. However, it has been our experience that bilateral leg edema associated with symptoms of sleep-disordered breathing is common in ambulatory medical practice; with a single exception, all of the patients in our study were identified by a single primary care physician (R.P.B.).

Our results support the notion that echocardiography has a role in the evaluation of bilateral leg edema. Recently published guidelines do not recommend obtaining an echocardiogram to evaluate leg edema unless there is clinical evidence of elevated jugular venous pressure.⁵⁶ Since most of the subjects in our study lacked jugular venous distension or were too obese to allow identification of jugular venous distension, using an echocardiogram to evaluate bilateral leg edema in the absence of any other evidence of cardiac disease would expand the indications for echocardiography.

If the results of our study are generalizable, and if echocardiograms are to aid primary care clinicians in the diagnosis of pulmonary hypertension, the cause of which may be OSA, then cardiologists can aid the task of primary care physicians by identifying and reporting the presence of elevated pulmonary artery pressures, even if the pulmonary artery pressure is only mildly elevated. Some cardiologists do not consider an echocardiographically estimated systolic pulmonary artery pressure of less than 35 mm Hg to be abnormal,² but one third of the patients with OSA in our study had a systolic pulmonary artery pressure of less than 35 mm Hg. In our opinion, if an echocardiogram is to be a sensitive guide for primary care physicians caring for patients with bilateral leg edema, then pulmonary hypertension should be identified when the pulmonary artery systolic pressure exceeds 30 mm Hg.

As primary care physicians receive echocardiogram reports indicating the presence of pulmonary hypertension, it will be important to keep in mind that 6 subjects in our study (40%) had unexplained pulmonary hypertension. Three of these individuals had intermediate apnea-hypopnea indexes of 18, 15, and 13. Sleep apnea is sometimes diagnosed when the apnea-hypopnea index is less than 20 mm Hg,^57- 61 especially if patients have symptoms of hypersomnolence. The 3 subjects in our study with intermediate apnea-hypopnea indexes had Epworth Sleepiness Scale scores of 1, 7, and 8. Since a score from 6 through 15 indicates some daytime sleepiness and is consistent with mild OSA,²⁹ 1 or 2 of these individuals may have OSA that failed to meet the diagnostic criteria of this study.

The other 3 individuals had apnea-hypopnea indexes of 2, 4, and 5. Clearly, none of them have OSA. Curiously, their Epworth Sleepiness Scale scores were 19, 9, and 12, respectively, indicating that all 3 had some symptoms of daytime sleepiness. One of these 3 individuals used an appetite suppressant in the past. For the 2 individuals with low apnea-hypopnea indexes who did not use appetite suppressants, and perhaps even for the one who did, the pulmonary hypertension may be age related.^3- 6

Finally, it is worth emphasizing a point that has practical relevance for clinicians evaluating leg edema: sleep apnea may be present even when the echocardiogram does not detect pulmonary hypertension. Often, the echocardiogram is unable to provide an estimation of the pulmonary artery pressure, especially if there is no tricuspid valve regurgitation present. In our study, 3 patients with bilateral leg edema who had symptoms consistent with sleep apnea had initial echocardiographic findings that failed to report an estimation of the pulmonary artery pressure. When the echocardiography technician repeated the portion of the examination estimating pulmonary artery pressure, all 3 had pulmonary artery pressures of greater than 30 mm Hg. Two of the 3 underwent polysomnography, and both received a diagnosis of sleep apnea.

In addition to the 3 patients who required a second echocardiogram to document the presence of pulmonary hypertension, 7 patients with bilateral leg edema were not included in the study because their echocardiograms failed to allow any estimations of the pulmonary artery pressures (n = 3) or because no echocardiogram was ordered (n = 4). Because each of these 7 individuals had symptoms consistent with OSA, sleep studies were ordered. All 7 had apnea-hypopnea indexes of greater than 20, confirming the diagnosis of OSA. Thus, for patients with bilateral leg edema who lack symptoms or signs of congestive heart failure, but who have symptoms consistent with sleep-disordered breathing, the most prudent and cost-effective diagnostic strategy may be to bypass the echocardiogram and refer these individuals directly for a sleep evaluation.

Accepted for publication February 1, 2000.

The authors would like to acknowledge the assistance of Louise Wiatrak, MA, Robert Bahler, MD, and Robert Finkelhor, MD.

Reprints: Robert P. Blankfield, MD, MS, University Hospitals Primary Care Physician Practice, Williamsport Plaza, 398 W Bagley Rd, Suite 1, Berea, OH 44017.