In our community‐based sample, more than a third of individuals had high TAT. In the absence of high VAT, excess TAT identified a subset of individuals with a higher prevalence of adverse cardiometabolic characteristics compared with individuals without high TAT even after adjustment for total volume of VAT. This included a higher prevalence of CVD in both women and men, lower HDL levels in women, and a higher prevalence of current smoking in men. These findings provide a better understanding of the risk factor correlates of TAT, a specific subtype of perivascular fat. Further experimental studies are required to elucidate whether TAT is pathogenic.
Multiple basic science and small clinical studies have suggested a local effect of perivascular fat that changes with the development of obesity.5–6,8
Previous work in the Framingham Heart Study has demonstrated a novel and reliable method of quantifying thoracic periaortic fat by MDCT.9
We now present a comprehensive age‐ and sex‐specific description of TAT distribution in a community‐based sample and report differences in cardiometabolic features among subcategories of TAT and VAT.
Our findings that more than one third of individuals in the community had high TAT is similar to our previous reports of fat depots, including VAT, pericardial, and intrathoracic fat.18–19
Despite this similar prevalence among various fat depots, our findings suggest that a high volume of TAT is associated with adverse cardiometabolic features among the subset of individuals with normal VAT. Prior work has highlighted the existence of “metabolically obese normal weight individuals” who exhibit glucose intolerance and hyperinsulinemia despite a normal BMI.20
Differences in visceral adiposity have been postulated to contribute to this phenotype.21
Our findings suggest that high thoracic periaortic fat may also identify a “metabolically obese” group among individuals who do not meet criteria for excess VAT. Some of these differences, such as the higher prevalence of CVD, are present in both sexes. However, others are sex specific, with a prevalence of low HDL higher in women than men. This finding is consistent with prior literature suggesting stronger associations of ectopic fat and metabolic risk factors in women compared with men. However, the prevalence of current smoking was higher in men than in women with high TAT/normal VAT compared with normal TAT/normal VAT.
TAT is directly wrapped around the aorta, and this distinct anatomic location may explain the specific association between high TAT and CVD among individuals with normal VAT. TAT may serve as a marker of perivascular fat throughout the body including smaller blood vessels, and perivascular fat has been postulated to have adverse effects on the vasculature.4–5
Supporting this, prior work in the Framingham Heart Study has demonstrated an association between TAT and both abdominal aortic calcium and coronary artery calcium among individuals without known cardiovascular disease. These associations persisted after adjustment for VAT and standard cardiovascular risk factors.10
Alternatively, our findings that high TAT was associated with CVD among individuals with normal VAT may reflect that these individuals were more likely to have a higher prevalence of certain CVD risk factors, including low HDL in women and smoking in men.
Our findings of a difference in the prevalence of smoking among men with normal VAT but discordant for TAT deserves specific comment. Smokers tend to have a lower body weight but are known to have a higher risk of cardiovascular disease. Ectopic fat distribution, specifically higher volumes of VAT, is already known to differ between smokers and nonsmokers.22
We have extended these findings among individuals with normal VAT by demonstrating a higher prevalence of male smokers among individuals with high versus normal TAT. These findings are noteworthy given the potential modulating role of nicotine on perivascular fat previously demonstrated in animal models.23
Specifically, exposure of rats to nicotine prenatally and during lactation led to an increase in total adiposity as well as perivascular fat in the offspring compared with controls. Furthermore, the normal anticontractile effect of perivascular fat on blood vessels was no longer present in nicotine‐exposed offspring, but was restored with the transfer of the medium surrounding normal fat. Thus, nicotine was associated with higher volumes as well as dysfunction of perivascular fat. These findings are consistent with prior work suggesting that excess perivascular adipose tissue disrupts the normal contribution of perivascular fat to vascular tone24
and further suggest that nicotine may contribute to this process. Other studies also support a potential role of nicotine on adipose tissue.25–26
For example, differences in adipose tissue lipoprotein lipase activity have been found between smokers and nonsmokers.25
Thus, elucidation of the full expression profile of perivascular fat in response to nicotine may help to extend our observational findings.
Strengths and Limitations
The major strength of our study is the relatively large sample size and community‐based nature of the cohort. This allowed exploration of differences within subgroups of TAT and VAT and limited referral bias. We assessed TAT and VAT using a highly reproducible CT volumetric assessment. Certain limitations warrant discussion. The cross‐sectional and observational design of the analysis prevents inferences of causality or temporality. The Framingham Heart Study is predominantly white, and results cannot be generalized to other ethnic groups. We did not have information on the severity of CVD. Our analyses were exploratory, and we did not account for multiple testing. Replication of our findings in independent cohorts is warranted. There were temporal differences between the MDCT scans and the risk factor assessments among the Offspring cohort. However, our sensitivity analysis limited to the Third‐Generation cohort (in which MDCT scans and risk factor assessment occurred during the same period) demonstrated overall similar findings. Finally, our findings do not suggest that TAT quantification should be used as a clinical tool.