Rebecca C. Thurston, PhD Departments of Psychiatry, Psychology, and Epidemiology University of Pittsburgh.

Annoying Symptom or Harbinger of Disease? New Discoveries about Hot Flashes, Obesity, and Cardiovascular RiskRebecca C. Thurston, PhDDepartments of Psychiatry, Psychology, and EpidemiologyUniversity of Pittsburgh

Outline

Introduction to hot flashes Hot flashes and subclinical

cardiovascular disease Mechanisms Discussion/future directions

Hot Flashes Sensation of intense heat, sweating,

flushing

Hot flashes, night sweats (vasomotor symptoms)

Over 70% of women experience during menopausal transition

Can persist for decades

Hot Flashes Duration

0 10 20 30 40 50 60 70 800

50

100

0 8051Birth DeathFinal Menstrual Period

% U

S w

om

en

Hot Flashes Associated with pronounced

impairments quality of life:

Physical, social, emotional functioning

Sleep disruption, irritability, depressed mood, poorer cognitive function

Hot Flashes Leading cause of treatment seeking among

midlife women Findings of risk associated with hormone

therapy (HT)Most effective treatment for hot flashes

Increased interest in physiology of and nonhormonal treatments for hot flashes

Underlying physiology not well-understood

Physiology of Hot Flashes

Copyright ©2004 The Endocrine Society

Randolph, J. F. et al. J Clin Endocrinol Metab 2004;89:1555-1561

Estradiol (E2)

Follicle stimulating hormone (FSH)

Hot Flashes

Physiology of Hot Flashes

Sweating

Shivering

Thermoneutral zone

Thermoneutral zone

Tc Tc

Asymptomatic Symptomatic

Shivering

Sweating

(Freedman, 2001)

Outline

Introduction to hot flashes Hot flashes and subclinical

cardiovascular disease Mechanisms Discussion/future directions

Hot flashes long understood to have important impact on quality of life

Few medical implications?

Is a Hot Flash Just a Hot Flash? WHI & HERS: Women with hot flashes at

highest cardiovascular risk with HT use

Cardiovascular risk factors (smoking) also risk factors for hot flashes

E2 widespread cardiovascular impact

Potent vasodilator associated with hot flashes, not other types of sweating

Study Questions

Hot FlashesSubclinical

Cardiovascular Disease (CVD)

• Flow mediated dilation (FMD)

• Calcification

CV Risk Factors?

E2?

Subclinical Cardiovascular Disease (CVD) Measures

FMD: Endothelial dysfunction (lower worse)Early in CVD

Calcification: Calcified plaques in aorta and coronary arteres

Study of Women’s Health Across the Nation (SWAN)

Baseline

Yrs 4-7• FMD: Brachial

artery ultrasound• Calcification: EBT

aorta

SWAN Heart (N=557) Pittsburgh, Chicago

SWAN (N=3302)Annually:• Demographic,

Health behaviors, Affect

• Hot flashes• SBP, DBP, BMI• Blood Draw: E2,

lipids, glucose

10

B

B

B

B

2

3

4

5

6

7

8

9

1

Hot Flashes & Flow Mediated Dilation

4

6

8

10

12

None AnyHot Flashes

FM

D (

%,

M,

SD

)

B (SE) = -0.99 (0.41), p = 0.02B (SE) = -0.97 (0.44), p = 0.03B (SE) = -1.01 (0.41), p = 0.01

+ Covariates+ Covariates, E2

Age, site, race, lumen diameter, BMI, education, DBP, HT use, menopausal status, LDL, HDL, triglycerides, glucose, diabetes history, lipid med use, smoking, physical activity

Age, site, race, lumen diameter, BMI, education, DBP, HT use, HDL, LDL, triglycerides, glucose, diabetes history, lipid med use, smoking, physical activity, E2, cycle day of blood draw (Thurston et al., 2008, Circulation)

Age, site, race

*

Age, site, race, education, BMI, smoking, SBP, antidepressant use, HT, menopausal status, depressive sx, phys activity, glucose, HDL, LDL, triglycerides, diabetes hx

None Any50

60

70

80

Hot Flashes

% w

ith

Ao

rtic

Cal

cifi

cat

ion

Hot Flashes & Aortic Calcification

OR = 1.55, 1.10-2.19, p = 0.01

+ Covariates, E2OR = 1.63, 1.07-2.49, p = 0.02

+ CovariatesOR = 1.53, 1.02-2.29, p = 0.04

Age, site, race, education, BMI, smoking, SBP, antidepressant use, HT, menopausal status, depressive sx, phys activity, glucose, HDL, LDL, triglycerides, diabetes hx, cycle day, E2Age, site, race

(Thurston et al., 2008, Circulation)

*

Is a Hot Flash Just a Hot Flash?

Other subclinical CVD measures?

Intima Media Thickness (IMT)

IMT: Thickness of medial and intimal layers of carotid artery Most well-validated and

widely-used measure of subclinical CVD

Study Questions

Hot Flashes Intima media thickness (IMT)

CV Risk Factors?

E2?

Study of Women’s Health Across the Nation (SWAN)

Baseline

Baseline Yrs 4-7• IMT: Carotid artery

ultrasound

SWAN Heart (N=557) Pittsburgh, Chicago

SWAN (N=3302)Annually:• Demographic,

Health behaviors, Affect

• Hot flashes• SBP, DBP, BMI• Blood Draw: E2,

lipids, glucose

10

B

B

B

B

2

3

4

5

6

7

8

9 F

F

F

F

1

Follow up Yrs 6-9• IMT

Cross Sectional Association between Hot Flashes and IMT

age, site, race, education, BMI, smoking status, SBP, HDL, LDL, triglycerides, glucose, diabetes status/meds, CVD status/meds, HT use, menopausal statusage, site, race, education, BMI, smoking status, SBP, HDL, LDL, triglycerides, glucose, diabetes status/meds, CVD status/meds, HT use, E2, cycle day of blood draw

+E2B (SE) = 0.03 (0.01), p = 0.03B (SE) = 0.03 (0.01), p = 0.02

(Thurston et al., 2011, Menopause)

*

Association between Hot Flashes Across Visits and IMT

B (SE) = 0.03 (0.01), p = 0.03B (SE) = 0.02 (0.01), p = 0.04 +E2

age, site, race, education, BMI, smoking status, SBP, HDL, LDL, triglycerides, glucose, diabetes status/meds, CVD status/meds, HT use, menopausal statusage, site, race, education, BMI, smoking status, SBP, HDL, LDL, triglycerides, glucose, diabetes status/meds, CVD status/meds, HT use, E2, cycle day of blood draw

(Thurston et al., 2011, Menopause)

*

Relation between hot flashes and IMT by obesity status

Study Visits with Hot Flashes

None One Both0.6

0.62

0.64

0.66

0.68

0.7

0.72

0.74

0.76

NormalOverweightObeseIM

T (

mm

)

Hot flashes *BMI p<0.01

Hot Flashes and Subclinical CVD Women with hot flashes had higher

subclinical CVD (FMD, calcification, IMT) Persist controlling for CVD risk factors, E2 Most pronounced with high BMI Hot flashes mark something more? Consider role of the vasculature in hot

flashes Mechanisms?

Outline

Introduction to hot flashes Hot flashes and subclinical

cardiovascular disease Mechanisms Discussion/future directions

A Note about Measurement Epidemiologic studies use

questionnaire measures of hot flashes Crude, memory and reporting influences

Physiologic, diary measures of hot flashes Data in “real time”

More precise

Insight into reporting influences

Physiologic Measurement of Hot Flashes

Hot Flash Diary Occurrence Severity Bothersome Location on body Aura Emotions Health behaviors…

“False Positive” Hot Flash Reporting

Physiologic

Reported (Diary)

Yes No

Yes 347 208

No 394 --

Psychological Factors Associated with False Positive Hot Flashes

Low Medium High0

0.1

0.2

0.3

0.4

0.5

0.6

0.7Depressive sxState anxietyTrait anxiety

Level of Negative Affect

Fa

lse

po

sit

ive

re

po

rtin

g r

ate

(Thurston et al., 2005, Psychosom Med)

**

†

† p < 0.1* p < 0.05

Emotional Antecedents of “False Positive” Hot Flashes

0

0.5

1

1.5

2

OR

Fa

lse

Po

sit

ive

Ho

t F

las

h

Fru

stration

Sa

dn

es

s

Stre

ss

Tire

d

Ha

pp

y

Re

laxe

d

In

Co

ntro

l

*

*

* p < 0.05(Thurston et al., 2005, Psychosom Med)

A Note about Measurement

Mood and emotions can impact hot flashes

When using self-report measures only, consider the role of emotion

Best to have physiologic + diary measures (laboratory/clinical studies)

Hot Flashes and Autonomic Nervous System

Etiology of hot flashes: Role of autonomic nervous system speculatedSympathetic, parasympathetic (vagus)

Reduced parasympathetic (vagal) control of heart rate linked to elevated CVD risk

High frequency heart rate variability(HF-HRV) index of cardiac vagal control

Study Question

Hot Flashes Cardiac vagal control (HF-HRV)

PMBC FLASHES Study

• 30 peri and postmenopausal women, aged 40-60, >4 hot flashes/day, no HT, SSRI/SNRIs

• Lab procedures to induce hot flashes

• Continuous ECG (HF-HRV) and sternal skin conductance

• Hot flashes physiologically measured and self-reported

Observation30 min

Stress 5 min

Heat30 min

Cold Pressor1 min

Data Reduction & Analysis

Spectral analysis of heart rate time series Linear mixed models

Minutes during flash compared to non-flash pre and post flash periods

Covariates: Age, task, race, menopausal status, education, smoking, anxiety, BMI, diabetes, use of cardiovascular meds/HTN, physical activity

FlashPre-flash Post-flash

Hot Flash

Data Reduction & Analysis

Minute -10 Minute +10

Reduced Cardiac Vagal Control During Hot Flashes: Laboratory

* p < 0.05 vs. minute zero

Pre-Flash Period

Post-Flash Period

Flash Period

(Thurston et al. 2010, Menopause)

Flash Period

Ambulatory Study: During Daily Life• 42 peri and postmenopausal women,

aged 40-60, >4 hot flashes/day, no HT, SSRI/SNRIs, or CV medications

• Wore ambulatory monitor for 24 hourso Hot flashes o ECG o Respiration

• Hot flashes physiologically measured and self-reported

Reduced Cardiac Vagal Control During Hot Flashes: Ambulatory (24 hrs)

Pre-flash Post-flashp<0.0001 p<0.0001

Minutes surrounding hot flash

HF

-HR

V (

lnm

sec2 )

(Thurston et al., 2012,

Menopause)

Hot Flash

Autonomic nervous system and hot flashes Reduced HF-HRV during hot flashes

Laboratory and ambulatory settings

Insight into etiology of hot flashesReproductive hormonal

Thermoregulatory

Autonomic nervous system?

Mechanism linking hot flashes to CVD risk?

In SWAN, hot flashes associated with elevated subclinical CVD

Mechanisms?Autonomic nervous system (Thurston et al., 2010,

Thurston et al., 2012)

Inflammation/hemostasis (Thurston et al., 2011)

Hot Flashes and CVD risk

Other mechanisms: Inflammation/hemostasis

Inflammation/hemostasis and hot flashes?

Regulated in part by vascular endothelium

One study: IL-8 elevated among women with hot flashes (Yasui et al., 2006)

Sensitive to reproductive hormones

CV risk factors?

Study Questions

Hot Flashes

E2?

Inflammation/ hemostasis

CRP, PAI-1, Factor VIIc, TPA-antigen, fibrinogen

B

Study of Women’s Health Across the Nation (SWAN)

SWAN (N = 3302)Annually:• Demographic, Health behaviors,

Medications/Health status, Affect• Hot flashes, night sweats• SBP, DBP, BMI• Blood Draw: E2, lipids, glucose

7

6

5

4

2

3

4

5

6

7

8

9

1

3

1

B

Inflammatory/hemostatic markers:• CRP, PAI-1, and tPA-ag• Fibrinogen and FVIIc

10

0 1 3 4 5 6 75

6

7

8

9

10

None 1-5 Days 6+ Days

SWAN Visit

TP

A-a

nti

gen

(lo

g)

Hot Flashes and TPA-antigen

Hot flashes in past two weeksCovariates: education, menopausal status, alcohol, parity smoking, exercise, affect, BMI, CV meds, diabetes/insulin, steroids, pain med, antidepressants

p<0.001

(Thurston et al., 2011)

0 1 3 5 7100

105

110

115

120

125

130

135

140

None 1-5 Days 6+ Days

SWAN Visit

Fac

tor

VIIc

(lo

g)

Hot Flashes and Factor VIIc p<0.01

Hot flashes in past two weeksCovariates: education, menopausal status, alcohol, parity smoking, exercise, affect, BMI, CV meds, diabetes/insulin, steroids, pain med, antidepressants

(Thurston et al., 2011)

In SWAN, hot flashes associated with elevated subclinical CVD

Mechanisms?Autonomic nervous system (Thurston et al., 2010,

Thurston et al., 2012)

Inflammation/hemostasis (Thurston et al., 2011)

Lipids (Thurston et al., 2012)

Hot Flashes and CVD risk

What about lipids?

Well known CV risk factor

Some research suggestive of adverse lipid profile among women with hot flashes

CV risk factors?

Study Questions

Hot Flashes

E2?

Adverse lipid profile?

LDL, HDL, Triglycerides, ApoB, ApoA1

B

7

6

5

4

2

3

4

5

6

7

8

9

1

3

1

B

Lipids:• LDL, HDL, triglycerides, ApoB, ApoA1

10

Study of Women’s Health Across the Nation (SWAN)

SWAN (N = 3302)Annually:• Demographic, Health behaviors,

Medications/Health status, Affect• Hot flashes, night sweats• SBP, DBP, BMI• Blood Draw: E2, FSH, glucose

Hot Flashes and LDL Cholesterol

0 1 3 4 5 6 7100

105

110

115

120

125

130

None 1-5 Days 6+ Days

SWAN Visit

LD

L, m

g/d

L

Hot flashes in past two weeksCovariates: age, site, race, education, menopausal status, alcohol use, physical activity, smoking, anxiety, BMI, CVD status/medication, lipid lowering medication

p<0.001

(Thurston et al., 2012)

Hot Flashes and ApoB

0 1 3 4 5 6 795

100

105

110

115

120

None 1-5 Days 6+ Days

SWAN Visit

Ap

oB

, m

g/D

l

p<0.0001

Hot flashes in past two weeksCovariates: age, site, race, education, menopausal status, alcohol use, physical activity, smoking, anxiety, BMI, CVD status/medication, lipid lowering medication

(Thurston et al., 2012)

Hot Flashes and Triglycerides

0 1 3 4 5 6 780

90

100

110

120

130

140

150

160

None 1-5 Days 6+ Days

SWAN Visit

Tri

gly

ceri

des

, mg

/Dl p<0.0001

Hot flashes in past two weeksCovariates: age, site, race, education, menopausal status, alcohol use, physical activity, smoking, anxiety, BMI, CVD status/medication, lipid lowering medication

(Thurston et al., 2012)

In SWAN, hot flashes associated with elevated subclinical CVD

Mechanisms?Autonomic nervous system (Thurston et al., 2010,

Thurston et al., 2012)

Inflammation/hemostasis (Thurston et al., 2011)

Lipids (Thurston et al., 2012)

Hot Flashes and CVD risk

Outline

Introduction to hot flashes Hot flashes and subclinical

cardiovascular disease Mechanisms Discussion/future directions

Hot Flashes and CVD risk Hot flashes associated with elevated

subclinical CVDMultiple mechanisms

Subtypes of hot flashes? Synergize with other CV risk factors? Physiologic, diary measures of hot

flashes?

Next Steps: CVD Risk and Hot Flashes New study designed to address CVD risk

and hot flashes R01HL105647: N=300 with and without

hot flashes, 5 years Detailed physiologic, psychological

mechanisms Physiologic, diary hot flash measures

Implications? Better understand physiology of hot

flashes

Midlife marker of CVD risk?

Aggressive risk factor reduction among women with hot flashes?

Improve health of midlife women

SWAN has grant support from the NIH, DHHS, through the NIA,

NINR, NHLBI, ORWH (NR004061; AG012505, AG012535, AG012531, AG012539, AG012546, AG012553, AG012554, AG012495, HL065581, HL06551)

Thurston: K23AG029216University of Pittsburgh Institute

on Aging

The content of this presentation is solely the responsibility of the authors and does not

necessarily represent the official views of the NIA, NINR, ORWH or the NIH.

Karen Matthews, PhD Kim Sutton-Tyrrell, DrPH Rachel Hess, MD, MSc

Samar El Khoudary, PhD Faith Selzer, PhD

Susan Everson-Rose, PhD, MPH

Ellen Gold, PhD Imke Janssen, PhD Lynda Powell, PhD Israel Christie, PhD

Carolyn Crandall, MD, MS Barbara Sternfeld, PhD

Acknowledgements

Thank you!

Questions?