For several decades, the global prevalence of obesity has been rising. The greatest increase has been noted in the United States, where the prevalence of obesity more than doubled from 1962 to 1999.1-3 Recent data indicate that 20.9% of the adult US population is obese, reporting a body mass index (BMI) greater than 30 kg/m^sup 2^, and 2.3% is morbidly obese, with a BMI greater than 40 kg/m^sup 2^.4 In the United States, the annual estimate of deaths attributable to obesity has been reported to range between 280 000 and 325 000.5,6 In the obese population, left ventricular dysfunction, atherosclerotic heart disease, obstructive sleep apnea (OSA), asthma, and venous thromboembolism are recognized cardiopulmonary causes of morbidity and mortality.7 Although obesity is associated with an increased risk of hypertension, hypercholesterolemia, and diabetes mellitus, the impact on the respiratory system is underrecognized.8,9 Obesity is associated with sleep disordered breathing and sleep apnea; however, the relationship between OSA and pulmonary hypertension is unclear. Several studies have found a higher BMI in patients with OSA and pulmonary hypertension compared with patients with OSA but without pulmonary hypertension. 10-12 In contrast, other investigators have not found a significant difference in BMI between patients with or without pulmonary hypertension.13-15
A postmortem study of 135 overweight subjects, including 4 morbidly obese subjects, reported a linear increase in heart weight with increasing body weight up to 105 kg, above which point the heart weight increased less in relation to total body weight.16 These authors indicated that the increased heart weight was attributable to excess epicardial fat. Another autopsy study of 12 obese subjects demonstrated that the observed increased heart weight was primarily due to ventricular hypertrophy.17 In 1984, Warnes and Roberts18 reported autopsy findings in 12 massively obese subjects with weights ranging from 312 to 500 lb (ca. 141.8-227.3 kg), and they observed left ventricular hypertrophy in all subjects and right ventricular hypertrophy in 4 subjects. Furthermore, they studied 664 coronary artery segments and found that only 14% were narrowed greater than 50%, and only 3% were narrowed more than 75%.18
To our knowledge, there is only one published autopsy study comparing the cardiopulmonary pathology in morbidly obese patients.19 In this report, patients with OSA exhibited biventricular heart failure and vascular changes indicative of pulmonary hypertension. Compared with the group without OSA, individuals with morbid obesity and OSA had a higher prevalence of medial hypertrophy of the muscular pulmonary arteries, hemosiderosis, alveolar hemorrhage, and alveolar capillary proliferation. Cardiac findings included moderate to severe cardiomegaly with biventricular hypertrophy, patchy myocardial fibrosis, and absence of significant coronary atherosclerosis. Consequently, we designed a retrospective autopsy study to investigate the prevalence of cardiovascular and pulmonary disease in obese subjects.
MATERIALS AND METHODS
Study Design
A search of the autopsy database at a large university medical center during an 11-year period (1993-2003) yielded 76 individuals with BMI greater than 30 kg/m^sup 2^. Subjects were further divided into obese (>30 kg/m^sup 2^) and morbidly obese (>40 kg/m^sup 2^) groups in order to examine the clinical and pathologic features associated with morbid obesity and cardiopulmonary disease.
Medical records were examined to determine patient demographics, weight, height, comorbidities, anorexinogen use, cause of death, and clinical course. Pathologic data were collected from autopsy reports, including lung and heart weights, ventricular thickness, coronary artery atherosclerosis, presence of pulmonary edema, hemorrhage, diffuse alveolar damage, pulmonary fibrosis, and pulmonary thromboemboli. All histology slides were prepared from paraffin-embedded lung tissue, sectioned at 4.0 µm and stained with hematoxylin-eosin, as well as additional special stains for elastic tissue and Movat pentachrome stain when indicated. All lung and heart slides were examined by 2 of the authors. Lung sections were evaluated for the presence of pulmonary vascular pathology, including pulmonary arterial hypertrophy and dilatation, arteriolar thickening, pulmonary vein hypertrophy and thickening, vascular elastosis, alveolar capillary proliferation, hemosiderosis, and iron encrustation of vessels. Pulmonary vascular changes were documented using a histopathologic system of classification based on the recommendations of the Third World Symposium on Pulmonary Artery Hypertension. 20,21 All heart slides were examined for the presence of acute myocardial infarction, myocardial fibrosis, and inflammation. Coronary artery atherosclerosis was examined and graded as follows: 0 indicates no evidence of atherosclerosis; 1, less than 50% lumen stenosis; 2, 50% to 74% lumen stenosis; and 3, 75% or greater lumen stenosis.
We also examined the clinical and pathologic records of subjects between the third and eighth decades with BMI less than 25 kg/m^sup 2^ undergoing autopsy during the same study period. This control group consisted of 5 to 6 randomly selected subjects from each decade of life. Pathologic data were collected and recorded using the previously mentioned parameters. Approval for the study was obtained from the University of Texas Medical Branch Galveston Institutional Review Board.
Statistical Analysis
Values are expressed as mean ± SD. Comparison of parametric data between obese and nonobese control subjects was performed using a Student t test. Nonparametric values were compared using ?^sup 2^ and Mantel-Haenszel ?^sup 2^ tests. Obese subjects were divided into obese and morbidly obese, the latter being defined as having a BMI greater than 40 kg/m^sup 2^. We compared the clinical and pathologic differences between the groups using nonparametric tests. Significance was defined as P value of .05 or less.
RESULTS
From 1993 to 2003, 76 obese individuals were identified; mean BMI was 45 ± 13.6 kg/m^sup 2^. Demographic data, including the racial makeup of the group, are presented in Table 1. The obese group was composed of 41 males and 35 females, with mean age of 50 ± 15.3 years (range, 14- 84 years). The nonobese control group consisted of 46 subjects with a mean BMI of 23 ± 3.3 kg/m2 and a mean age of 47 ± 14.3 years (range, 20-84), with 29 men and 17 women.
Clinical Findings
The prevalence of diabetes mellitus was significantly greater in the obese group compared with the control group: 39% and 11%, respectively (P < .001). Similarly, systemic hypertension was observed more frequently in the obese group compared with the control group: 53% and 22%, respectively (P < .001). Within the obese group, 61% of the subjects were noted to be morbidly obese, with a BMI of 40 kg/m^sup 2^ or greater. The morbidly obese group had a mean age of 50 ± 14.2 years, mean BMI of 52.3 ± 13.6 kg/m^sup 2^, and 1:1 male-female ratio. A total of 7 obese individuals were diagnosed antemortem with OSA by polysomnography, 9 with chronic obstructive pulmonary disease by spirometry, and 3 with interstitial lung disease. Anorexinogen use was not reported or identified. The causes of mortality in the obese group were related to pulmonary disease in 27 subjects (35%) and cardiac disease in 24 (31%) (Table 2). The specific pulmonary cause of death was observed to be pulmonary thromboembolus (12), bronchopneumonia (6), diffuse alveolar damage (5), and cardiogenic pulmonary edema (4). The cardiac cause of death included acute myocardial infarction (10), sudden death due to cardiac arrhythmia/ventricular tachycardia (9), and congestive heart failure (5). Other notable causes of death were sepsis (8), acute pancreatitis (4), and heat stroke (2). The principal causes of death in the nonobese subjects were pulmonary (9), cardiac (8), trauma (6), complications of cancer (9), sepsis (6), and other (8).
Pulmonary Gross Findings
No significant difference was observed in the mean combined lung weights of the obese and the control groups: 1613 ± 136 g (range, 590-4010 g) and 1461 ± 82 g (range, 760-2650 g), respectively (P = .15). The obese group, however, had a significantly higher occurrence of pulmonary edema compared with the control group: 76% and 37%, respectively (P < .001). Similarly, pulmonary hemorrhage occurred more commonly in the obese compared with the control group: 62% and 19.5%, respectively (P < .001). Pulmonary pathologic findings are shown in Table 3.
Pulmonary Histologic Findings
Lung sections were evaluated for morphologic changes in arteries, arterioles, venules, and veins. Histologic features of pulmonary arterial and venous hypertension were present in 72% of obese subjects. A striking finding observed in 72% of the obese group was thickening of septal and acinar pulmonary veins; 6% of the control group demonstrated these pulmonary venous changes (P < .001). The pulmonary veins were focally dilated, tortuous, and showed medial thickening and fibrosis, as well as intimal fibrosis (Figures 1 and 2). The interstitial lobular septa showed varying degrees of edema, thickening, fibrosis, and lymphatic dilatation (Figures 3 and 4).
Pulmonary arteries also showed hypertensive changes, although these were less severe compared with the venous hypertensive changes.22 Pulmonary arterial medial hypertrophy, defined as an increase in the cross-sectional area of the arterial media due to smooth muscle hypertrophy, was seen in 72% of obese subjects compared with 26% in controls (P < .001). Figures 5 through 7 show representative sections of an arteriole and a small pulmonary artery, with intimal thickening and fibrosis, as well as medial hypertrophy. Rare small collections of vessels resembled plexiform and dilatation lesions; however, these vascular lesions lacked the typical diagnostic histologic features of plexiform lesions (Figure 8). There was also focal extension of smooth muscle into arterioles and intraacinar arteries. Thickening and muscularization of the pulmonary arterioles (<100 µm in diameter) were signifi- cantly greater in the obese group compared with the control group: 45% and 17%, respectively (P = .007). Medial hypertrophy and intimal atherosclerosis of the large elastic pulmonary arteries were mild and were seen in a few obese subjects; no significant difference was noted between obese and control groups (17% and 13%, respectively). Pulmonary alveolar capillary proliferation, histologically similar to pulmonary capillary hemangiomatosis (PCH), was almost exclusively observed in the obese group compared with the controls: 50% and 2%, respectively (P < .001). Figures 9 through 11 demonstrate the characteristic histologic features of PCH. The alveolar capillaries were dilated and congested, forming double layers and lining both sides of the alveolar septa, with focal tufting. In contrast, pulmonary congestion in nonobese subjects showed a single layer of capillaries without tufting (Figure 12). Elastosis of intra-acinar arterioles and venules, seen as homogenous pale, eosinophilic expansion of the vessel wall, was noted in 17% of obese subjects and 6% of controls. The areas of vascular elastosis did not stain with special stains for either mucin (mucicarmine, Alcian blue, periodic acid-Schiff) or amyloid (Congo red). In the obese group, pulmonary hemosiderosis, defined as large collections of hemosiderin-filled alveolar macrophages, was more frequently observed than in the control group: 22% and 9%, respectively (P = .05). Iron encrustation of the pulmonary venous elastic tissue was not detected in either group. Importantly, plexiform lesions or fibrinoid necrosis of vessels were not observed in the obese or morbidly obese subjects. Eight obese subjects had thrombosed and recanalized vessels (Arnold Rich lesions), vascular lesions often associated with left ventricular diseases (Figure 13); none of the controls had these lesions.23
Cardiac Pathologic Findings
Cardiomegaly (defined as heart weight >360 g in males and >280 g in females) was present in all 76 obese subjects (Table 4). The obese group had a significantly higher mean heart weight compared with the control group: 541 ± 41 g (range, 300-1040 g) and 444 ± 32 g (range, 220- 1000 g), respectively (P < .002). In the obese group, right ventricular hypertrophy (>0.5 cm) was present in 30%, and left ventricular hypertrophy (>1.5 cm) was present in 64%. There was no significant difference in the mean right ventricular thickness between the obese group (0.42 cm) and the control group (0.32 cm). Mean left ventricular thickness was also not statistically different between the obese group (1.65 cm) and the control group (1.55 cm). Acute myocardial ischemic changes and/or acute myocardial infarction were found in 14% of obese and 10% of control subjects. Coronary artery atherosclerosis was present in 71% of the obese group and 100% of the control group. However, more subjects in the obese group had severe atherosclerosis (>75% luminal stenosis) compared with the controls: 30% and 22%, respectively. Almost one third of the obese subjects (29%) were free of coronary artery atherosclerosis, whereas all controls had some degree of coronary atherosclerosis (P < .001). The mean age of the obese subjects without atherosclerosis was lower (37 years) compared with the mean age for the entire obese group (50 years).
Obese Versus Morbidly Obese Group
The data were further analyzed to examine differences in the clinical and pathologic findings between the obese (BMI, 30-40 kg/m^sup 2^) and morbidly obese (BMI, >40 kg/ m^sup 2^). A total of 46 of the 76 subjects were morbidly obese. The mean age for the morbidly obese group was similar to the obese group: 49 and 51 years, respectively. Furthermore, no differences in sex, race, prevalence of diabetes mellitus, systemic hypertension, or lung weights were noted among these groups. The morbidly obese group had significantly greater incidence of PCH compared with the obese: 61% and 39%, respectively (P = .03; Table 5). There was no difference in other gross and histologic pulmonary findings between the obese and morbidly obese subjects. With respect to cardiac findings, the mean heart weight and the left ventricular thickness of the obese and morbidly obese were similar. Comparison of mean right ventricular thickness revealed a lower value in the morbidly obese compared with the obese group: 0.40 cm and 0.45 cm, respectively (P < .001). Coronary artery atherosclerosis was equally absent in both groups (29%), and fewer morbidly obese subjects had severe atherosclerosis compared with obese subjects: 22% versus 40%, respectively.
COMMENT
The main pathologic findings of this retrospective autopsy study were the presence of pulmonary vascular changes consistent with pulmonary hypertension in 72% and cardiomegaly in 100% of obese subjects. Pulmonary hypertension includes a variety of cardiopulmonary conditions with similar clinical presentation but different etiologies. Previous classifications of pulmonary hypertension were based on the assumption that all subsets of pulmonary hypertension had a similar spectrum of pathologic lesions; that is, medial hypertrophy, intimal thickening, and fibrosis of muscular arteries, and atherosclerosis and dilation of large elastic arteries, often associated with right ventricular hypertrophy. However, morphometric studies have revealed differences in the distribution and prevalence of pulmonary vascular changes in pulmonary hypertension of different etiologies, as well as coexistence of hypertensive pulmonary venous changes in certain cases of pulmonary hypertension.23-28
Earlier reports have described the presence of pulmonary hypertension in morbidly obese individuals and suggested that morbid obesity was a cause for the pulmonary hypertension.16,29 Subsequent reports confirmed these earlier observations and described morbidly obese individuals with hypercapneic respiratory failure and right ventricular dysfunction.30,31 Eventually, Burwell and colleagues32 coined the term Pickwickian syndrome to describe the syndrome of severe obesity, alveolar hypoventilation, and cor pulmonale. A few recent clinical studies have reported pulmonary hypertension in 15% to 41% of obese patients.10,33-35 While these earlier reports described the clinical course of cardiopulmonary failure associated with severe obesity, they did not describe the morphologic changes in the pulmonary vasculature associated with severe obesity. In 1997, cardiomegaly and pulmonary vascular changes were described in a small group of obese subjects with OSA undergoing autopsy.19 A novel finding of our study is the high prevalence of pulmonary vascular pathology, detected in 72% of the obese subjects. Medial hypertrophy of pulmonary arteries found in 44% of our subjects correlates with the incidence of pulmonary hypertension reported in previous clinical studies.10,33,34 Since we found a higher prevalence of pulmonary hypertensive changes compared with the prevalence reported in clinical studies, we speculate that pulmonary vascular histologic changes may precede clinical recognition and diagnosis of pulmonary hypertension.
We also found hypertensive changes in pulmonary venous vasculature, in addition to the pulmonary arterial vasculature. Concomitant changes in the pulmonary venous vasculature have been reported in case series of pulmonary hypertension.23,26 Obstruction of pulmonary venous outflow has been described to lead to medial hypertrophy and muscularization of pulmonary veins, arteriolar muscularization, hemosiderosis, and interstitial edema.36 Conditions such as systolic and/or diastolic heart failure as well as mitral and aortic valve disease can contribute to the development of pulmonary venous hypertension.
The epidemic of obesity has brought into focus the cardiac structural changes and hemodynamic alterations resulting in impaired pulmonary function, leading to a clinical syndrome in severe and longstanding obesity, the obesity cardiomyopathy.37 Obesity is associated with an increase in circulating blood volume and cardiac output because of the high metabolic activity of excessive fat that leads to left ventricular dilatation, increased left ventricular wall stress, compensatory (eccentric) left ventricular hypertrophy, and left ventricular systolic and diastolic dysfunction with ultimate left ventricular failure.37 Thus, in the setting of left ventricular dysfunction, one would expect to observe pulmonary venous hypertension with subsequent development of pulmonary arterial hypertension. 37 Obesity cardiomyopathy has been proposed as a pathologic entity in morbidly obese subjects, and this condition may have contributed to the pulmonary venous changes seen in our study.37 Importantly, pulmonary hypertension in many obese individuals is most likely multifactorial, stemming from left ventricular dysfunction secondary to longstanding hypertension, OSA, and obstructive lung disease.
Obesity is also associated with vasodilatation of the heart, kidney, gastrointestinal tract, and skeletal muscle vascular bed, presumably due to increased metabolic rate and local accumulation of vasodilator metabolites, as well as growth of the organs in response to increased metabolic demand.38 Thus, one may postulate that a similar mechanism may result in vasodilatation of the pulmonary veins and capillaries.
Muscularization of the arterioles, one of the changes seen with pulmonary hypertension and commonly associated with chronic hypoxia, was present in 45% of our cases and has been reported in high-altitude inhabitants, patients with emphysema, kyphoscoliosis, and obese hypoxic patients.39 It is postulated that the site of hypoxic pulmonary vasoconstriction is in the pulmonary arterioles, leading to muscularization of the pulmonary arterioles.40,41 The pulmonary vascular changes associated with chronic hypoxia thus may be different from the changes observed with idiopathic pulmonary hypertension or congenital heart disease with an intracardiac shunt, which primarily affect the pulmonary muscular arteries.
Of note was the presence of PCH in 50% of morbidly obese subjects. Pulmonary capillary hemangiomatosis is characterized by diffuse or localized proliferation of alveolar capillaries on both sides of the alveolar walls, with formation of glomeruloid tufts or nodules.42,43 The capillary endothelial cells are cytologically bland and do not exhibit excessive mitosis. Interestingly, this histologic feature was also observed in another series of autopsied obese patients, and it was ascribed to chronic hypoxiaassociated pulmonary hypertension.19 The pathogenesis of PCH remains controversial, as some investigators consider it to be neoplastic, whereas others surmise that it is a response to an unknown angiogenic stimulus, such as hypoxia. 42,43
Our study also demonstrated the presence of cardiomegaly in all obese subjects, almost entirely from biventricular hypertrophy. Although some of the earlier reports suggested the cardiomegaly to be secondary to increased adiposity,16 other studies have shown that the increased heart weight was primarily due to ventricular hypertrophy and not adiposity.17-19 The absence of coronary artery atherosclerosis in almost one third (29%) of our obese subjects is similar to another study that described a 14% incidence of coronary atherosclerosis.18 A recent forensic autopsy study of 166 individuals also reported absence of coronary artery atherosclerosis in 38% to 44% of severely obese individuals.44
The role of obesity in the development of pulmonary hypertension has been debated. Obesity is considered an unlikely risk factor by some investigators,21 whereas other investigators have found mild pulmonary hypertension in a small percentage (17%) of 220 patients with obesity and OSA.12 Hypoxia has been postulated as a contributing factor in obesity-related pulmonary hypertension. In a clinical study of morbidly obese subjects, lower PaO2 value was observed in the supine position compared with the upright position; this was ascribed to abdominal content compression of the pulmonary lower lobes, resulting in atelactasis and hypoxemia.45 Furthermore, another study of regional distribution of ventilation and perfusion in obese subjects described diminished ventilation to wellperfused lower lobe segments, noticed to be even more pronounced in morbidly obese subjects.46 These investigators demonstrated that in morbidly obese subjects, ventilation- perfusion abnormalities resulting in hypoxemia may be attributable to lower lobe ventilation perfusion mismatch. The histologic findings of pulmonary hypertension with arteriolar muscularization and PCH in our study may be related to hypoxia.36,47
In summary, our autopsy study found pulmonary hypertension, particularly venous hypertension and pulmonary capillary hemangiomatosis, in 72% of obese individuals. Cardiomegaly with biventricular hypertrophy was present in all obese subjects. Morbidly obese subjects had a higher prevalence of pulmonary capillary hemangiomatosis compared with nonmorbid obese subjects. We postulate that the pulmonary venous hypertensive changes and pulmonary capillary proliferation in the obese are most likely related to chronic hypoxia and left ventricular dysfunction.
© 2008 College of American Pathologists Provided by ProQuest LLC. All Rights Reserved.
Copyright 2008 Archives of Pathology & Laboratory Medicine
