Obesity Paradox Revisited: Lower BMI Linked to Higher Death, CV Events After PCI

Download this article's Factoid (PDF & PPT for Gold Subscribers)

Yet another study appears to confirm the ‘obesity paradox’ in a large group of South Korean patients undergoing percutaneous coronary intervention (PCI). In the study, published online March 26, 2013, ahead of print in Circulation: Cardiovascular Interventions, low body mass index (BMI) was significantly associated with an increased risk of major cardiovascular events and all-cause mortality.

Seung-Jung Park, MD, of Asan Medical Center (Seoul, South Korea), and colleagues pooled patient-level data from 11 clinical studies (8 randomized trials, 3 registries). All studies were conducted in South Korea and involved a total of 23,181 Asian patients undergoing PCI. Patients were categorized by baseline BMI, ranging from low (< 18.5; n = 339) to obese (≥ 30; n = 1,113). The mean BMI for the entire cohort was 24.9 ± 3.0.

During a mean follow up of 2.1 years, 2,381 patients had a major cardiovascular event (10.3%), and 1,004 patients died (4.3%).

Major cardiovascular events and mortality increased with decreasing BMI (log-rank P < 0.001 for both). The cumulative incidence of major cardiovascular events over time showed an initial steep rise, followed by a continuous separation of the curves, with a significantly higher rate of events in the lowest BMI group. The same pattern was seen with mortality.

After adjustment for multiple variables, the inverse relationship between BMI and major cardiovascular events and all-cause mortality was maintained (table 1).

Table 1. Risk of Major CV Events, Mortality by BMI


Major CV Events HR (95% CI)

All-Cause Mortality HR (95% CI)

< 18.5

1.52 (1.16-1.99)

2.93 (2.63-3.27)


1.05 (0.83-1.33)

2.44 (1.95-3.05)


1.03 (0.92-1.17)

1.39 (1.24-1.56)





0.97 (0.87-1.07)

0.79 (0.72-0.87)


0.97 (0.85-1.11)

0.76 (0.67-0.85)


0.78 (0.62-0.98)

0.79 (0.61-1.04)

Sensitivity analyses excluding events within 7 days also consistently showed an inverse relationship between BMI and major cardiovascular events and all-cause mortality. However, patients with the highest BMI did not demonstrate higher rates of mortality as previous studies have shown.

Explaining the Paradox

According to the authors, previous studies supporting a protective effect of higher BMI, termed the ‘obesity paradox,’ have been hampered by small patient populations, short follow-up periods, or retrospective observational study designs.

“Our study does not fully clarify the exact mechanism of an absence of association or an inverse association of BMI and clinical outcomes after PCI. However, there could be some possible explanations for this phenomenon,” they write.

These include:

  • High BMI may reflect a preserved or increased lean body mass as opposed to body fat
  • The high prevalence of coexisting cardiovascular conditions in those with higher BMI may indicate use of more aggressive secondary preventive drug therapies
  • Higher BMI may be related to larger vessels treated with larger stent diameter or to the cardioprotective effect of adipokines

Consistent with Previous Studies

Carl J. Lavie, MD, of Ochsner Medical Center (New Orleans, LA), noted in an e-mail communication with TCTMD that the results are consistent with most other similar studies.

“The worst prognosis is at the lowest, underweight BMIs, and the overweight and mildly obese have a better prognosis than do the ‘normal’ BMI patients. We have shown this with percent body fat as well,” he added. “Those with higher body fat and coronary heart disease or heart failure do better than [those with] low body fat, whereas lean mass or non-fat mass is as one would expect: the higher do better.”

Dr. Lavie agreed with the study authors that “it is just as much a ‘lean paradox,’ meaning that the lean patients with heart disease—especially those with low BMI and low body fat, and low body fat and low lean mass—are particularly doing poorly.”

But fitness can substantially alter the relationship, he added. High fitness seems to improve prognosis regardless of body composition in chronic heart disease and heart failure, Dr. Lavie said. But again, a paradox is present in the case of low fitness where thinner patients have poorer outcomes than overweight or obese patients.

Dr. Lavie cited a recent meta-analysis (Flegal KM, et al. JAMA. 2013; 309:71-82) showing that overweight patients have a lower mortality rate than those with so-called normal BMI. In the study, even class 1 obese patients (BMI of 30 to 34.9) had a 5% lower mortality compared with those with normal BMI, although the difference was not statistically significant, he reported.

No Excuse to Gain Weight

In a telephone interview with TCTMD, Robert M. Minutello, MD, of Weill Cornell Medical College (New York, NY), said the study confirms prior data by presenting it in a 100% Asian population. An explanation for the lack of increased mortality in the very obese in this study, he added, is that there were few of them to begin with, which is typical of an Asian population.

“At this point, all the data that we have is thought provoking, but it’s been thought provoking for about the last 12 years and doesn’t influence how we treat patients,” he said. “This study doesn’t provide any new explanations for why the higher BMI patients have better outcomes. If there is a physiologic explanation, we don’t have it yet.”

While it may seem logical to conclude that being overweight gives patients an advantage, Dr. Minutello said he believes the data simply show associations that are likely influenced by any number of confounders. He compared it to bare-metal stent data suggesting smokers have lower rates of restenosis.

“Should we be telling patients who get stents to start smoking? Absolutely not. This is a similar issue. We know that patients who are obese are more unhealthy overall and more likely to develop coronary disease,” Dr. Minutello said. “Whether lower BMI is a marker of chronic illness or whether there truly is a protective effect of obesity is unclear at this point.”

Study Details

Patients had a mean age of 62 years, 70% were men, 30% had diabetes, and 58% presented with ACS. PCI consisted primarily of DES (82%).

Patients categorized as obese were younger overall than those with low BMI (59 vs. 68 years). Patients with higher BMIs were more likely to have diabetes, hypertension, hyperlipidemia, and histories of MI and previous PCI, while those with lower BMIs had higher rates of previous congestive heart failure, renal dysfunction, and left main disease.


Park D-W, Kim Y-H, Yun S-C, et al. Association of body mass index with major cardiovascular events and with mortality after percutaneous coronary intervention. Circ Cardiovasc Interv. 2013;Epub ahead of print.



Related Stories:

  • Drs. Park, Lavie, and Minutello report no relevant conflicts of interest.

We Recommend