Beyond BMI

New Perspectives on
Obesity
Virend K. Somers MD, PhD

New Perspectives on Obesity

BMI and the obesity paradox
How good is BMI in diagnosing obesity?
Implications of central obesity
Does sleeping less make us fat?
Management of the high risk obese
patient

Obesity
Defined by the WHO/NIH as an excess of
body fat percentage of :
> 25% men
> 35% women

Body fat percentage varies importantly

according to sex, age and race.

BMI
Weight
height2

How do we define Obesity?

For most clinicians and

epidemiologists
obesity
=

increased BMI
(> 30 kg/m2)

Background
Obesity Definition

Clinicians and epidemiologists often

rely on body mass index (BMI) to
diagnose obesity

BMI

Low = < 18.5 kg/m2

Normal = 18.5 - 24.9 kg/m2
Overweight = 25 - 29.9 kg/m2
Obese = 30 - 39.9 kg/m2
Extremely obese = > 40 kg/m2

BMI Advantages

Simple and reproducible

Associated with cardiovascular (CV)
risk factors

Low and high BMI are associated

with increased mortality and CV
events

Do you know your BMI?

BMI Disadvantages
Obesity Paradox
General
Population

Flegal, K. M. et al. JAMA 2007.

CAD
Population

Romero-Corral, A. et al. Lancet 2006.

Results

Underweight: BMI (< 20) strongly associated

with an increased long-term mortality and other
CV events.

Overweight: BMI (>25 29.9) associated with

better survival and fewer CV events.

Obese: BMI 30, not significantly associated

with long-term mortality and other CV events in
patients with CAD.

Interpretation.

Obese is better ???

Possible Explanations
Low BMI:
Older patients
Cancer and other co-morbidities
Low lean mass and high body fat

Elevated BMI:
Younger
Better treatment?
Preserved or increased lean mass rather than
excess in body fat

Conclusions
Mechanisms underlying this apparently

paradoxical relationship are unclear, but

they may relate to the definition of obesity
itself (BMI)

BMI unable to differentiate between BF %

and lean muscle mass

New Perspectives on Obesity

BMI and the obesity paradox
How good is BMI in diagnosing obesity?
Implications of central obesity
Does sleeping less make us fat?
Management of the high risk obese
patient

Objective
Assess the diagnostic performance of
BMI in detecting excess in BF% in
patients with CAD *.
> 25 % in men.
> 35 % in women.
Diagnostic Performance: Sensitivity,
Specificity, correlation coefficients between
BMI and both, BF % and LM.
* BF % estimated using air displacement plethysmography

Methods
A cross-sectional design of 13,601 subjects

(age 20-79.9 years; 48% men) from the Third

National Health and Nutrition Examination
Survey (NHANES III).

Bioelectrical impedance analysis was used

to estimate BF % and LM. Height and weight

measurements were used to calculate BMI
(kg/m2).

Body Fat vs. Lean Mass

Romero-Corral, A. et al. Int J Obes (Lond) 2008.

2016 Superbowl - Denver Broncos

2011 MFMER | slide18

Conclusions
Obesity defined as a BMI > 30 kg/m2 misses

more than half of the patients with CAD with

real BF %-obesity.

Although BMI had a good correlation with

BF % it also had a good correlation with
lean mass.

In men BMI correlated better with lean mass

than with BF % while in women BMI
correlated better with BF % than lean mass.
This might explain the better diagnostic
performance of BMI in women.

Conclusions

BMI fails to discriminate between fat

mass and lean mass in the US
population, has a poor diagnostic
performance and poor association
with adverse events.

Other measures that discriminate

between fat mass and lean mass,

especially those accounting for central
obesity (eg WHR) are needed.

Variability of BF%
in People with
BMI 25 kg/m2
(n = 54)

(n = 54)

~55%
~20%

~62%
~18%

Range: 13.8 to 35.3 %

Range: 26.2 to 42.8 %

Romero-Corral, A. et al. Int J Obes (Lond) 2008.

Hypotheses

Subjects considered to have normal

body weight based on BMI, could still

have a high body fat (BF) content

Normal Weight Obesity, NWO

NWO subjects have a higher prevalence
of metabolic syndrome and
cardiometabolic dysregulation

Definitions

Normal Weight Obesity

Normal BMI = 18.5-24.9 kg/m2
BF % > 20% men
> 30% women

Metabolic Syndrome
ATP III criteria

Traditional CV risk factors

Dyslipidemia (NCEP ATP-III), HTN (JNC 7), DM (ADA),

smoking, CVD (MI & stroke)

Prevalence of NWO
57.5% in men and 53.5% in women
70.4%
61.4%

61.4%

53.2%

Weighted prevalences

Dose Response

* Adjusted for age and race

Sex-Specific lowest quartile used as the reference

NWO and Metabolic Syndrome

*OR=1.38 (1.02 - 1.87)
*OR=1.66 (1.16 - 2.4)
*OR=1.68 (1.02 - 2.84)

* Odds ratio adjusted for age and race

NWO and Metabolic Syndrome

*OR=1.61 (1.07 - 2.46)
*OR=1.77 (1.18 - 2.7)
*OR=11.26 (5.46 - 27.3)
*OR=2.8 (1.62 - 5.09)

* Odds ratio adjusted for age and race

NWO and CV Mortality

MEN

WOMEN
Body fat < 30%
Body fat < 20%

Body fat > 20%

(NWO)

Body fat > 30%

(NWO)

Coutinho TLopezJimenez F, JACC

2012

Discussion
Main Findings

1 out of 2 subjects with a normal BMI has an

excess in BF (NWO)

~ 45 million Americans

NWO associated with cardiometabolic

dysregulation (MetSx)

Unaware of increased risk due to their

normal BMI

NWO might be related to increased CV

mortality, especially women

New Perspectives on Obesity

BMI and the obesity paradox
How good is BMI in diagnosing obesity?
Implications of central obesity
Does sleeping less make us fat?
Management of the high risk obese
patient

New Perspectives on Obesity

BMI and the obesity paradox
How good is BMI in diagnosing obesity?
Implications of central obesity
Does sleeping less make us fat?
Management of the high risk obese
patient

Central
vs.
Non-central obesity

Visceral Adipose Tissue

5-15% of total body fat

More metabolically active
than subcutaneous fat

Associated with higher

prevalence of
cardiovascular and
metabolic risk factors
and disorders

Vega et al., J Clin Endocrinol Metab 2006

Mahabadi et al., Eur Heart J 2009

BMI vs. WHR

Yusuf, S. Lancet 2005;366:1640-9.

Study Design and Population

Double blinded Randomized controlled
trial

lean healthy humans

Randomization - 1 Control:4 Gainers
Visit 1
Sleep study
Blood draw
Body composition

8-weeks

Visit 2
Sleep study
Blood draw
Body composition

8-weeks

Visit 3
Sleep study
Blood draw
Body composition

Flow Mediated Vasodilatation

Corretti, MC et al. JACC 2002

Image at baseline

5 min

15-30 sec
(image acquisition)

Cuff inflation

Cuff deflation

Change in brachial artery diameter

FMD (%)

Endothelial Function in Fat

Gainers
p = 0.75

p = 0.003

p = 0.07

10
8
6
4
2
0

Baseline

Fat gain

Recovery

Romero-Corral A et al, J Am Coll Cardiol. 2010 17;56(8):662-6

Case Example
27 yr male, BMI 24.5 kg/m2
Weight gainer

Baseline

Follow-up

Visceral
Adipose Tissue

Subcutaneous
Adipose Tissue

Body weight: + 4.5 %

Visceral fat: + 44 %
Subcutaneous fat: + 35 %
24-h SBP/DBP/MAP: + 5/1/3 mmHg

Endothelial Dysfunction by Tertiles of Visceral

Fat Gain in the Fat Gain Group

Reduction in brachial artery

diameter reactivity FMD (%)

p < 0.0001
p = 0.006

4.5

3.5

p = 0.05

2.5
1.5
0.5
-0.5
-1.5

< 8 cm2

8 16.5 cm2

> 16.5 cm2

Visceral Fat Gain

Romero-Corral A et al, J Am Coll Cardiol. 2010 17;56(8):662-6

Potential Mechanisms
Visceral fat accumulation may cause

increase in BP through greater secretion of:

FREE-FATTY
ACIDS

PRO-INFLAMMATORY
CYTOKINES

drive increases in both

A third factor may
ANGIOTENSINOGEN
visceral fat and BP
HPA axis activation

New Perspectives on Obesity

BMI and the obesity paradox
How good is BMI in diagnosing obesity?
Implications of central obesity
Does sleeping less make us fat?
Management of the high risk obese
patient

2011 MFMER | slide44

Pretest questions:
Who are these innovators
who have transformed
society? What, if anything, do
they have to do with the
obesity epidemic?

Increase of Melatonin and Induction of Sleep by Darkness

Kripke DF, 2012

We are chronically sleep deprived

- BFB

McKnight-Eily, JAMA 2008; Krueger Am. J. Epidemiol 2009; Sleep in America 2002; Aguiar
National Bureau of Economic Research 2006

Sleep Deprivation Induced by the

Information Highway
(2011 Sleep in America Poll from the NSF)

Two thirds of Gen Y (19 to 29 yrs) and one

third of Gen X use a social networking site

before bed

People who text before bed (half of Gen Y)

are more likely to be sleepy and drive
while drowsy

7 of 10 Brits are tweeting, poking, surfing

and writing on walls at night instead of

going to sleep (Travelodge Survey/Graham
Jones)

Epidemiology of sleep duration

Prevalence of sleep duration 7
hrs/day by state

2011 MFMER | slide51

Centers for Disease Control and Prevention,

Morb Mortal Wkly Rep. 2016

Short sleep
duration and
obesity
Epidemiologic
evidence

1984

1987

NHANES I
Gangswisch, Sleep 2005;
Patel, Am J Epidemiology
2006

1992

RATINGS OF HUNGER AND APPETITE

7.5

HUNGER

5.5

AFTER 2 DAYS OF
4-H BEDTIME

3.5

AFTER 2 DAYS OF
10-H BEDTIME

52

GLOBAL
APPETITE

42
32
22
9

HUNGER (cms)
GLOBAL APPETITE (cms)

11

13 15 17 19
CLOCK TIME

6.0 0.5
39.7 3.0

21

7.2 0.4
47.7 3.4

p level
<0.01
0.010

% change
+24%
+23%

Spiegel et al, Ann Int Med, 141 (2004), 846-850

Sleep Deprivation and Energy Balance

Design

SCREEN
1 d/1 n

1 week
home
actigraphy

ACCLIMATION
3 d/3n

EXPERIMENTAL
8 d/8 n

RECOVERY
4 d/3 n

Simple randomization stratified by gender to sleeping ad lib

or reduction in time in bed by 1/3

Subjects awoken every day at 6 am but bedtime based on

randomization

Typical time in bed: 10:00 pm 6:00 am = 8 h

Deprived time in bed: 12:40 am 6:00 am = 5.3 h

Sleep Deprivation and Energy

Balance

Mayo Sleep Monitoring System (MSMS)

Custom-modified sleep/wake monitoring system
based on Compumedics Siesta 802

64 channels, 512 Hz, 16 bit resolution

Mayo 10/20 electroencephalogram (EEG)

Electro-oculogram (EOG)
Electromyogram (EMG)
Electrocardiogram (ECG)
Extensible, i.e., easily allows addition of more
channels

Specialized electrode placement to allow safe,

23 hour/day monitoring

telemetry allowing real-time monitoring

<1 kg, wearable, unobtrusive

Sleep Deprivation and Energy

Balance
Outcomes caloric intake

main outcome caloric intake per day

(kcal)

free access to food 24/7

CRU Metabolic kitchen, SMH

cafeteria, snack basket, outside food

daily log by dieticians

Sleep Deprivation and Energy

Balance
Outcomes energy expenditure

six sensors (left panel):

four inclinometers (I)
two triaxial accelerometers (A)
14 axes of data are binned and stored every
half-second on two data loggers.
Specially designed undergarments (right
panel)
Levine, Science 1999; Levine, Science 2005; Levine, Acta Physiol Scand 2005

Sleep Deprivation and Energy Balance

Results caloric intake
P<0.01

Calvin et al, Chest 2013

P=NS

Sleep Deprivation and Energy Balance

Results caloric intake per hour
Was the increased caloric intake due to more time available to
eat?
P=0.01
P=0.02

Calvin et al, Chest 2013

Sleep Deprivation and Energy Balance

Results energy expenditure
P=NS

Estimated maximal (95%) energy change <50 kcal/day

Calvin et al, Chest 2013

Sleep Deprivation and Energy Balance

Conclusions

Modest sleep restriction increased

caloric intake by 549 kcal/day

Activity energy expenditure did not

change, total energy expenditure
likely changes minimally

2011 MFMER | slide62

Pretest questions:
Who are these innovators
who have transformed
society? What, if anything, do
they have to do with the
obesity epidemic?

Tim Berners-Lee
(html and www)
2011 MFMER | slide-64

New Perspectives on Obesity

BMI and the obesity paradox
How good is BMI in diagnosing obesity?
Implications of central obesity
Does sleeping less make us fat?
Management of the high risk obese
patient

How are we doing with high risk

obesity?

Recognition, Diagnosis and

Management of Obesity after MI
n = 627 randomly selected patients discharged
after MI, between 1/1/01 and 12/31/02

Five teaching hospitals

Overweight = BMI 25 kg/m2
Obese = BMI 30 kg/m2
Markedly obese = BMI 40
kg/m2
(Lopez-Jimenez et al, Int J Obesity)

Results
Mean BMI = 31 13 kg/m2 - documented in 14%
83% overweight, 55% obese, 8% morbidly obese
For BMI 30, only 20% had obesity documented
as PMH, current medical problem or final
diagnosis

Weight loss described as part of plan at

discharge in 7% of overweight and 9% of obese

post-MI patients

Results

New Perspectives on Obesity

BMI and the obesity paradox
How good is BMI in diagnosing obesity?
Implications of central obesity
Does sleeping less make us fat?
Management of the high risk obese
patient

CanJohnnycomeoutandeat?

Acknowledgements
Francisco Lopez-Jimenez, M.D.
Prachi Singh, Ph.D.
Fatima H. Sert-Kuniyoshi, Ph.D.
Thais Coutinho
Abel Romero-Corral, M.D.
Diane E. Davison, R.N.
Michael D. Jensen, M.D.
Andy Calvin, M.D.
Naima Covassin, Ph.D.
Virend K. Somers, M.D., Ph.D.
Funding: National Institutes of Health
American Heart Association

Thank you