Lipid deposits or accumulation of aortic valves (AVs) has been reported to be related to aging and/or calcific changes[1-6] and tumorous fatty lesions of AVs have been rarely reported.[7-11] However, the clinicopathological features of non-tumorous AV fat cells remain poorly understood,[11, 12] although they may be scarce and represent a metaplastic condition associated with calcified or ossified valves.[12, 13] Recently, our routine examination of stenotic AVs incidentally disclosed not only fat cells, but also small bodies with eosinophilic serpiginous membranes closely resembling the histological features of membranous fat necrosis (MFN)[14-16] (unpublished data). MFN, also called membranous lipodystrophy-like changes, membranocystic changes, membranocystic fat necrosis, or lipomembranous fat necrosis,[16-19] was initially considered a morphologic hallmark of ‘membranous dystrophy,’ ‘lipomembranous polycystic osteodysplasia,’ or ‘Nasu-Hakola disease,’[12, 16-20] and is now known to be a peculiar form of fat necrosis. However, to the best of our knowledge, MFN has not been mentioned previously in AVs with or without valvular dysfunction. We questioned whether or not AV fat cells and/or MFN-like bodies are concerned with AV dysfunction. Therefore, we examined the clinicopathological characteristics of these lesions in postmortem non-dysfunctional AVs (nd-AVs) and surgically removed dysfunctional AVs (d-AVs).
We examined 152 aortic valves (AVs), which included 82 postmortem non-dysfunctional AVs (nd-AVs) and 70 surgically removed dysfunctional AVs showing aortic stenosis (AS), aortic regurgitation (AR), or combined AS and AR (AS-R). Fat cells, membranous fat necrosis (MFN), and fat-MFN-related lesions composed of fat cells and/or MFN were found in 127 (83.6%), 110 (72.4%), and 140 (92.1%) of 152 AVs, respectively, and all were associated with older age (P = 0.010, P = 0.022, and P = 0.003, respectively). MFN was associated with fibrous thickening and calcification (both, P = 0.001). Nd-AV fat cells and fat-MFN-related lesions were not correlated with body mass index. Compared with age- and sex-matched control cases, MFN in AS and AS-R cases was more frequent (P = 0.030 and P = 0.045, respectively), but MFN in AR cases showed no significant differences. Fat-MFN-related lesions, possibly representing true preceding fat cells, showed no differences in AVs with and without dysfunction or in dysfunctional types. These data suggest that AV fat cells are age-related, obesity-independent, and AV dysfunction-unrelated common phenomenon. MFN is also age-dependent and could be caused by AS and AS-R, which is probably concerned with AV thickening and calcification.
Materials and Methods
Eighty-two postmortem nd-AVs were available from 10% buffered formalin-fixed heart specimens in 82 patients without AV dysfunction (52 men and 30 women with a mean age of 70.6 years; range, 40–95 years), retrieved from the autopsy files of the Department of Pathology, Japan Self-Defense Forces Central Hospital (2008–2012), and were examined. The postmortem interval ranged from 1 to 53 h (mean, 11.5 h), and the medical charts and autopsy reports were reviewed for body mass index (BMI), heart weight, and concomitant ischemic heart diseases (IHDs), hypertension, hyperlipidemia, and diabetes mellitus. All nd-AVs were tricuspid with no tumorous features, and were longitudinally cut into 1–7 sections with free edges separately in each cusp. The total number of nd-AV specimens in each case ranged from 3 to 20 sections (mean, 15.4 sections). All sections were stained with hematoxylin and eosin (HE), and selected sections were stained with periodic acid-Schiff (PAS) and Masson's trichrome stain. In this study, we defined ‘fat-MFN-related lesions’ as when AVs showed at least one of fat cells and MFN-like bodies, and used the term ‘AV fatty lesions’ for all lesions including fat cells, MFN-like bodies, and fat-MFN-related lesions. We examined AV fatty lesions and other histological changes. We counted the number of fat cells and MFN-like bodies in each case, divided them by the total number of sections, and calculated ‘fat cell (F-C) score’ and ‘MFN-like body (MFN-B) score,’ respectively. Selected paraffin-embedded tissues containing MFN-like bodies were re-fixed in 2.5% buffered glutaraldehyde and observed ultrastructurally.
A total of 70 surgically removed d-AVs, excluding infectious endocarditis, were retrieved from the surgical pathology files of the Department of Pathology, Japan Self-Defense Forces Central Hospital (1982–2012), and all HE-stained slides were examined. No tumorous lesions were detected preoperatively in any cases. The patients were 41 men and 29 women, ranging in age from 41 to 88 years (mean, 69.2 years). Patient charts and request forms for pathological examination were reviewed for AV dysfunction and IHDs. AV dysfunction had been determined using echocardiography and/or cardiac angiography. Mildly regurgitant AVs with moderate to severe aortic stenosis (AS) were included in the AS group, mildly stenotic AVs with moderate to severe aortic regurgitation (AR) in the AR group, and those with both AS and AR of moderate or severe degree in the combined AS and AR (AS-R) group. Accordingly, the patients were diagnosed as follows: 27 with AS, 24 with AR, and 19 with AS-R. The 70 d-AVs consisted of 59 tricuspid and 11 bicuspid valves, and the latter included 3 congenital bicuspid AVs and 8 secondarily fused bicuspid AVs, which had been assessed by intraoperative findings. On the basis of AV dysfunction and morphological features, we proposed underlying diseases,[4-6, 21] such as ‘rheumatic AVs’ including bicuspid and tricuspid AVs, ‘nodular calcific tricuspid AVs,’ ‘nodular calcific bicuspid AVs,’ ‘regurgitant floppy AVs,’ and ‘regurgitant AVs with other specified etiology,’ and classified the surgical cases into these categories. All d-AVs were fixed with 10–20% buffered formalin, longitudinally cut into 1 to 10 sections, and paraffin-embedded. The total number in each case ranged from 2 to 26 sections (mean, 15.2 sections). Selected sections were additionally cut and stained with HE, PAS, Masson's trichrome stain, and elastica van Gieson (EVG) stain, and immunostained with S-100 protein (S-100; Nichirei Corp., Tokyo, Japan).
The association of AV fatty lesions with clinicopathological findings was analyzed using the chi-square test and Mann-Whitney U-test. The correlation of patient age with F-C score and MFN-B score was examined using Spearman's coefficient test. The differences in AV fatty lesions were evaluated between AS and AR cases, and between rheumatic AVs and non-rheumatic AVs. In addition, we designed age- and sex-matched control AVs selected from 82 postmortem nd-AVs in each AV dysfunctional type, and compared d-AV fatty lesions with those in the respective control AVs. Statistical significance was set at P < 0.05.
In 75 of 82 postmortem nd-AVs (91.5%), fat cells were focally or multifocally scattered, chiefly in the spongiosa of the middle to proximal cuspal portion (proximal, 42.3%; middle, 56.6%; and distal portion, 1.1%), with no tumorous features (Fig. 1a,b). The nd-AV F-C score ranged from 0 to 48.3 (mean, 12.3) and F-C scores in 17 cases (20.7%) were >20. In 52 nd-AVs (63.4%), MFN-like bodies were identified singly or in some clusters always intermingled with fat cells (Fig. 1c), and the incidence of fat-MFN-related lesions was equal to that of fat cells. MFN-like bodies were composed of fat cell-sized rounded or contracted bodies lined by eosinophilic serpiginous or crenulated membranes. The nd-AV MFN-B score ranged from 0 to 11.9 (mean, 0.76). MFN-like bodies were generally scarce and microscopic, and their recognition seemed to be difficult on HE-stained sections alone. However, PAS and Masson's trichrome staining clearly revealed them (Fig. 1d,e). Ultrastructurally, MFN-like bodies showed markedly laminated thinned and/or thickened membranes, and the latter were composed of moderately electron-dense material containing numerous minute tubular structures (Fig. 2). Moderate fibrous thickening, moderate calcification, and myxoid material deposition of moderate degree were found in 10 (12.2%), 9 (11.0%), and 21 nd-AVs (25.6%), respectively, but no neovascularization, ossification, or lymphoplasmacytic infiltration was found in any nd-AVs.
Seventy d-AVs were variously accompanied by morphological changes and/or AV deformity, and could be divided into 35 rheumatic AVs, 19 nodular calcific tricuspid AVs, 1 nodular calcific bicuspid AV, 12 regurgitant floppy AVs, and 3 regurgitant AVs with other specified etiology (2 aortic root dilatation and 1 aortitis). In 52 d-AVs (74.3%), fat cells were focally or multifocally found in the less prominent deformed areas (Fig. 3), and the F-C score ranged from 0 to 29.1 (mean, 3.02). In 58 d-AVs (82.9%), MFN-like bodies were observed singly or in some clusters with or without fat cells, and were highlighted by Masson's trichrome stain and PAS (Fig. 4), but were negative for EVG (Fig. 4f). Fat cells were positive for S-100 protein, but MFN-like bodies were negative (Fig. 4h). The d-AV MFN-B score ranged from 0 to 15.5 (mean, 1.35). Only MFN-like bodies without fat cells were observed in 13 d-AVs (18.6%) and fat-MFN-related lesions were found in 65 d-AVs (92.9%), which was not statistically different from that (91.5%) in 82 nd-AVs (P = 0.987). Fatty bone marrow surrounded by metaplastic bones was found in 6 d-AVs, but was distinctly distant from scattered fat cells.
D-AV histological findings in dysfunctional types and underlying diseases are shown in Tables 1 and 2, respectively. Between the AS and AR group, there were no significant differences in AV fat cells, MFN-like bodies, and fat-MFN-related lesions (P = 0.253, P = 0.571, and P = 0.524, respectively). However, histologically, MFN-like bodies in AS and AS-R cases appeared to be found more frequently in the thickened fibrous stroma compared with AR cases. MFN-like bodies in AR cases seemed to be identified more commonly in the non-thickened cuspal middle to proximal portion. Between rheumatic and non-rheumatic AVs, there were no significant differences in AV fat cells, MFN-like bodies, or fat-related lesions (P = 0.056, P = 0.751, or P = 0.353, respectively).
|AS, No. (%) (n = 27)||AR, No. (%) (n = 24)||AS-R, No. (%) (n = 19)|
|Fat cells||19 (70.4)||21 (87.5)||12 (63.2)|
|MFN-like bodies||24 (88.9)||19 (79.2)||15 (78.9)|
|Fat-related lesions†||26 (96.3)||23 (95.8)||16 (84.2)|
|Nodular fibrous thickening‡||26 (96.3)||4 (16.7)||18 (94.7)|
|Neovascularization||12 (44.4)||4 (16.7)||12 (63.2)|
|Calcification‡||25 (92.6)||0 (0)||16 (84.2)|
|Ossification||7 (25.9)||0 (0)||3 (15.8)|
|Lymphoplasmacytic infiltration‡||19 (70.4)||2 (8.3)||12 (63.2)|
|Myxoid material deposition‡||20 (74.1)||22 (91.7)||12 (63.2)|
|Rheumatic AVs,§ No. (%) (n = 35)||Non-rheumatic AVs (n = 35)|
|Total, No. (%) (n = 35)||Nodular calcific tricuspid AVs, No. (%) (n = 19)||Nodular calcific bicuspid AVs, No. (%) (n = 1)||Regurgitant floppy AVs, No. (%) (n = 12)||Regurgitant AVs with other specified etiology, No. (%) (n = 3)|
|Fat cells||22 (62.9)||30 (85.7)||15 (78.9)||0 (0)||12 (100)||3 (100)|
|MFN-like bodies||29 (82.9)||29 (82.9)||16 (84.2)||1 (100)||10 (83.3)||2 (66.7)|
|Fat-related lesions†||31 (88.6)||34 (97.1)||18 (94.2)||1 (100)||12 (100)||3 (100)|
|Fibrous thickening‡||27 (77.1)||21 (60.0)||19 (100)||1 (100)||1 (8.3)||0 (0)|
|Neovascularization||28 (80.0)||0 (0)||0 (0)||0 (0)||0 (0)||0 (0)|
|Calcification‡||21 (60.0)||20 (57.1)||19 (100)||1 (100)||0 (0)||0 (0)|
|Ossification||9 (25.7)||1 (2.9)||1 (5.3)||0 (0)||0 (0)||0 (0)|
|Lymphoplasmacytic infiltration‡||20 (57.1)||13 (37.1)||12 (63.2)||0 (0)||0 (0)||0 (0)|
|Myxoid material deposition‡||24 (68.6)||30 (85.7)||15 (78.9)||0 (0)||12 (100)||3 (100)|
Fat cells, MFN-like bodies, and fat-MFN-related lesions were found in 127 (83.6%), 110 (72.4%), and 140 (92.1%) of 152 AVs, respectively, and all of them appeared to increase with older age (Table 3). The association of overall AV fatty lesions with clinicopathological features is summarized in Table 4. AV fat cells, MFN-like bodies, and fat-MFN-related lesions were statistically associated with older age (P = 0.010, P = 0.022, and P = 0.003, respectively). MFN-like bodies were associated with female predominance (P = 0.031), although fat cells and fat-MFN-related lesions were not. No AV fatty lesions were associated with IHDs. AV fat cells were inversely associated with fibrous thickening (P = 0.002), neovascularization (P < 0.001), and lymphoplasmacytic infiltration (P = 0.003), but were not associated with other histological changes. MFN-like bodies were positively associated with fibrous thickening (P = 0.001), calcification (P = 0.001), lymphoplasmacytic infiltration (P = 0.013), and myxoid material deposition (P = 0.018), but not with other histological changes. Fat-MFN-related lesions were not associated with any histological changes. In Spearman's coefficient analysis, patient age was correlated with F-C score (r = 0.20, P = 0.013) and MFN-B score (r = 0.16, P = 0.043). Table 5 is a summary of the association of nd-AV fatty lesions with other clinicopathological features. Nd-AV fatty lesions were not associated with other concomitant diseases, postmortem interval, BMI, or heart weight. There was no favored cusp of any AV fatty lesions.
|Age (years)||All cases, No.||Fat cells, No. (%)||MFN-like bodies, No. (%)||Fat-MFN-related lesions,† No. (%)|
|40–49||8||46 (75.0)||5 (62.5)||7 (87.5)|
|50–59||27||20 (74.1)||15 (55.6)||21 (77.8)|
|60–69||25||19 (76.0)||17 (68.0)||22 (88.0)|
|70–79||60||53 (88.3)||44 (73.3)||58 (96.7)|
|80–89||30||27 (90.0)||28 (93.3)||30 (100)|
|90–99||2||2 (100)||1 (50.0)||2 (100)|
|Total||152||127 (83.6)||110 (72.4)||140 (92.1)|
|Fat cells + (n = 127)||Fat cell − (n = 25)||P||MFN-like bodies + (n = 110)||MFN-like bodies − (n = 42)||P||Fat-MFN-related lesions† + (n = 140)||Fat-MFN-related lesions† − (n = 12)||P|
|Mean age, years (range)||71.4 (40–95)||65.7 (41–88)||0.010*||71.6 (43–90)||67.2 (40–95)||0.022*||70.9 (40–95)||59.8 (41–79)||0.003*|
|Concomitant IHDs, +/−||33/94||3/22||0.213||30/80||6/36||0.141||34/106||2/10||0.809|
|Fibrous thickening, +/−||41/86||17/8||0.002**||51/59||7/35||0.001*||54/86||4/8||0.961|
|Lymphoplasmacytic infiltration, +/−||22/105||11/14||0.003**||30/80||3/39||0.013*||32/108||1/11||0.420|
|Myxoid material deposition, +/−||66/61||9/16||0.215||62/48||14/28||0.018*||72/68||4/8||0.367|
|Fat cells (+) (n = 75)†||Fat cells (−) (n = 7)†||P†||MFN-like bodies (+) (n = 52)||MFN-like bodies (−) (n = 30)||P|
|Diabetes mellitus, +/−||22/53||0/7||0.219||17/35||5/25||0.187|
|Mean postmortem interval, hours (range)||11.5 (1–53)||11.3 (2–40)||0.759||11.0 (1–53)||12.4 (2–40)||0.604|
|Mean body mass index, kg/m2 (range)||21.4 (12.9–38.1)||20.6 (15.4–27.7)||0.624||21.8 (14.3–38.1)||20.4 (12.9–28.6)||0.342|
|Mean heart weight, g (range)||366.7 (205–613)||300.4 (249–404)||0.058||370.2 (205–613)||345.1 (238–582)||0.206|
|The distribution in 3 cusps, L/R/N‡||63/71/68||18/11/13||0.428||28/32/36||53/50/45||0.529|
The results of a control study regarding AV dysfunction are shown in Table 6. Fat cells in AS and AS-R cases were less frequent compared with those in the respective control cases (P = 0.028 and P = 0.042, respectively) and MFN-like bodies in AS and AS-R cases were more frequent (P = 0.030 and P = 0.045, respectively). However, there were no significant differences in fat-MFN-related lesions between them. AV fatty lesions showed no significant differences between AR cases and the control cases. AV fibrous thickening, neovascularization, calcification, lymphoplasmacytic infiltration, myxoid material deposition in AS and AS-R cases were significantly more frequent than those in the control cases, and ossification in AS was more common than that in AS-R cases. Regarding AR cases, myxoid material deposition was significantly more frequent than that in the control cases.
|AS cases No. (%) (n = 27)||Control cases† No. (%) (n = 27)||P||AR cases No. (%) (n = 24)||Control cases† No. (%) (n = 24)||P||AS-R cases No. (%) (n = 19)||Control cases† No. (%) (n = 19)||P|
|Mean age, years (range)||73.3 (52–88)||73.3 (52–87)||66.3 (46–88)||66.3 (46–88)||67.2 (41–83)||67.2 (40–84)|
|Fat cells||19 (70.4)||26 (96.3)||0.028**||21 (87.5)||19 (79.2)||0.699||12 (63.2)||18 (94.2)||0.042**|
|MFN-like bodies||24 (88.9)||16 (59.3)||0.030*||19 (79.2)||17 (70.8)||0.739||15 (78.9)||8 (42.1)||0.045*|
|Fat-MFN-related lesions‡||26 (96.3)||26 (96.3)||0.471||23 (95.8)||19 (79.2)||0.188||16 (84.2)||18 (94.2)||0.604|
|Fibrous thickening§||26 (96.3)||6 (22.2)||<0.001*||4 (16.7)||1 (4.2)||0.345||18 (94.2)||1 (5.3)||<0.001*|
|Neovascularization||12 (44.4)||0 (0)||<0.001*||4 (16.7)||0 (0)||0.118||12 (63.2)||0 (0)||<0.001*|
|Calcification§||25 (92.6)||0 (0)||<0.001*||0 (0)||0 (0)||1||16 (884.2)||1 (5.3)||<0.001*|
|Ossification||7 (25.9)||0 (0)||0.010*||0 (0)||0 (0)||1||3 (15.8)||0 (0)||0.230|
|Lymphoplasmacytic infiltration§||19 (70.4)||0 (0)||<0.001*||2 (8.3)||0 (0)||0.470||12 (63.2)||0 (0)||<0.001*|
|Myxoid material deposition§||20 (74.1)||7 (25.9)||0.001*||22 (91.7)||4 (16.7)||<0.001*||12 (63.2)||5 (26.3)||0.049*|
AV MFN-like bodies were microscopic and not cystic, but were lined by characteristic eosinophilic crenulated membranes. The presence of nd-AV MFN-like bodies was not associated with a lengthy postmortem interval, which rules out possible time-dependent postmortem degeneration. MFN-like bodies were PAS-positive, stained red with Masson's trichrome stain, and negative with EVG stain. Ultrastructurally, the tortuous thickened membranes of MFN-like bodies contained numerous minute tubular structures. These histopathological features were identical to those of MFN.[14-18] Hence, MFN-like bodies are true MFN of non-cystic type, implying a necrotic form of preceding fat cells. Only MFN-like bodies correspond to extensive necrosis of AV fat cells. Therefore, fat-MFN-related lesions, composed of viable and/or necrotic fat cells, represent true pre-existing fat cells. In the current study, the incidence of fat-MFN-related lesions showed no significant differences between nd-AVs and d-AVs or between d-AVs and respective age- and sex-matched control cases, regardless of AV dysfunctional type or underlying disease. On that basis, no AV dysfunction or underlying diseases plays a role in the development of AV fat cells themselves. The true incidence of AV fat cells in individuals over 40 years old is probably equal to the incidence of AV fat-MFN-related lesions, which is estimated to be 92–93%. This is strikingly higher than originally expected.
Our results suggested that AV fat cells are not so rare because 21–48 fat cells per nd-AV cusp section were detectable in 20% of the cases. However, AV fat cells did not exhibit tumorous configuration or capsular structures, and were considered a metaplastic condition rather than of a neoplastic or hamartomatous nature. Because of a lack of distinct tumorous features, we believe that AV fat cells do not cause valvular dysfunction. Steiner et al. noted metaplastic fatty bone marrow in AVs. However, the currently detected AV fat cells were not accompanied by metaplastic bones. Fat cells probably arise from perivascular pluripotent cells. In addition, AV is avascular,[2, 6, 22] and AV fat cells were statistically associated with absence of neovascularization. These findings suggest that metaplastic AV fat cells possibly arise from normally present fibroblasts or mesenchymal cells rather than from perivascular cells.[12, 23] Pantanowits et al. described a case of nd-AV fatty infiltration or aggregates, incidentally detected as an echogenic mass. The published microphotographs in his case resembled those of the currently detected AV fat cells without tumorous features. Therefore, AV fat cells may be rarely detectable by imaging, although no such cases were included in our series.
Usually detectable epicardial fat cells are known to be increased in association with obesity, steroid therapy, and older age,[3, 6, 23] but such associations are unknown in AV fat cells. In this study, the older age of the patient was significantly associated with overall AV fat cells and fat-MFN-related lesions, and was positively correlated with F-C score. These findings suggest that AV fat cells are an age-dependent phenomenon. The current analysis indicates that nd-AV fat cells are not associated with BMI or obesity, and are non-specific for concomitant diseases.
Other than generalized Nasu-Hakola disease,[16, 20] MFN most commonly involves the subcutaneous tissue[14, 16, 17] and breast, and could be found in bone marrow and intra-articular loose bodies. In the current study, MFN was identified in 58 d-AVs (83%) and in 63% of nd-AVs, which are higher than or almost equal to the reported incidences (38–75%) of MFN in subcutaneous tissues affected by some underlying diseases. Therefore, AV is one of the previously unrecognized common target organs of MFN. Moreover, older patient age was associated with AV MFN, and was correlated with MFN-B score, implying that MFN is also a senescent condition. AV MFN was statistically associated with females, which may be similar to the female predominance of MFN-related panniculitis.[16, 19]
In avascular AVs, the presence of MFN indicates a morbid condition that disturbs nutritious permeation directly from the blood flow. The current study demonstrated a close relationship of MFN with AV fibrous thickening and calcification, which could explain the disturbance in such nutritious permeation. Moreover, these histological changes, together with MFN, in AS and AS-R cases were more frequent than those in age- and sex-matched control cases. MFN in AS and AS-R cases histologically appeared to be present commonly in the thickened areas. These findings suggest that AS and AS-R could induce FMN through AV fibrous thickening and calcification. Previous articles[22, 26, 27] have also described the relationship among AS-related blood flow-dynamics and AV morphological changes, such as AV sclerotic/fibrotic changes initiating in sites suffering from lower shear stress and disturbed oscillatory flow, and the correlation of AV calcification with high bending stress. On the other hand, AR-related backward flow injury can cause focal fibrous thickening of distal closing edges, where fat cells were rare in the current study. MFN in AR cases seemed to be frequently found in less deformed areas, resembling those in nd-AVs, and was not statistically different from that in the age- and sex-control cases. On that basis, MFN in AR cases may be chiefly contributed to a senescent change. On the other hand, AV lymphoplasmacytic infiltration and myxoid material deposition were closely associated with MFN, but the pathogenetic association between them was unclear in the current study. Age-related environmental alteration influencing AV fat cells and MFN were also undetermined. To elucidate these points, further investigation is required.
In conclusion, AV fat cells are an age-dependent and BMI-unrelated common condition and are not be rare. Their development is not associated with the presence or type of AV dysfunction. Moreover, this article is the first to describe another age-related MFN in AVs, which could be induced by AS and AS-R, through AV fibrous thickening and calcification.
The authors thank Akio Kinoshita, Kenji Okada, Takanori Kakimoto, and Hiroaki Yamaguchi for excellent technical assistance, and Daniel Mrozek for editing the manuscript.