Download - Learning Objective I
Coronary Heart Disease Risk Profile of Women with PCOS:
Collision of Evidence and Assumptions
Evelyn Talbott, Dr.P.H.Kim Sutton-Tyrrell, Dr.P.H.
Lewis H. Kuller, M.D., Dr.P.H.Jeanne V. Zborowski, Ph.D.
Dept. of EpidemiologyGraduate School of Public Health
University of Pittsburgh
Learning Objective I
To review some of the designs employed in the To review some of the designs employed in the study of Cardiovascular Disease risk factors in study of Cardiovascular Disease risk factors in women with PCOSwomen with PCOS RetrospectiveRetrospective ProspectiveProspective Historical ProspectiveHistorical Prospective
Learning Objective II
To provide an overview of the controversy in To provide an overview of the controversy in the literature to date linking PCOS and the the literature to date linking PCOS and the risk of Heart Disease in women.risk of Heart Disease in women.
Learning Objective III
To more fully understand the metabolic To more fully understand the metabolic parameters of PCOS in women in relation to parameters of PCOS in women in relation to Cardiovascular Disease risk over time andCardiovascular Disease risk over time and
to discuss future research in this area.to discuss future research in this area.
Prevalence of Polycystic Ovary Syndrome (PCOS)
PCOS is a condition that affects 5-10% of PCOS is a condition that affects 5-10% of women in the US.women in the US.
It is a heterogeneous reproductive It is a heterogeneous reproductive endocrine disorder.endocrine disorder.
Characteristics of PCOS
Chronic anovulationChronic anovulation HyperandrogenemiaHyperandrogenemia Insulin ResistanceInsulin Resistance HyperinsulinemiaHyperinsulinemia ObesityObesity Impaired Glucose Tolerance (IGT)Impaired Glucose Tolerance (IGT) Central Adiposity Central Adiposity
Etiology of Polycystic Ovary Syndrome (PCOS)
Primary dysregulation of ovarian steroidogenesis (Ehrmann, Primary dysregulation of ovarian steroidogenesis (Ehrmann, 1995); insulin post-receptor binding defect (Dunaif, 1992)1995); insulin post-receptor binding defect (Dunaif, 1992)
Multiple arrested follicles are present with increased ovarian Multiple arrested follicles are present with increased ovarian stroma (pearl necklace effect)stroma (pearl necklace effect)
Results in hypertrophied ovarian theca cells, elevated Results in hypertrophied ovarian theca cells, elevated androgens, lack of a mid-cycle estradiol surge, and disrupted androgens, lack of a mid-cycle estradiol surge, and disrupted folliculogenesis (Barnes, 1997)folliculogenesis (Barnes, 1997)
Atretic Follicle
ThickenedCapsule
From: The Menstrual Cycle (1993), Ferin, Jewelewicz, and Warren
The Polycystic Ovary
EnlargedOvary with Abundant Stroma
Differing Criteria of PCOS
ULTRASOUND BIOCHEMICAL
CLINICAL
Review of the Evidence related to CHD Risk and PCOS
Women with PCOS may be at increased risk of Women with PCOS may be at increased risk of heart disease with increased prevalence of CHD heart disease with increased prevalence of CHD risk factors: (LDL, HDL, triglycerides, risk factors: (LDL, HDL, triglycerides, hypertension and diabetes.)hypertension and diabetes.)
However, the literature to date has been However, the literature to date has been inconsistent. inconsistent.
Studies to date have been observational, largely Studies to date have been observational, largely consisting of case-control and cross sectional consisting of case-control and cross sectional designs.designs.
Click for larger picture
PCOS and Cardiovascular Risk PCOS is associated with an increased prevalence of PCOS is associated with an increased prevalence of
cardiovascular disease risk factors (Wild, 1988; cardiovascular disease risk factors (Wild, 1988; Talbott, 1995)Talbott, 1995)
In both lean and obese PCOS cases exhibited > fasting In both lean and obese PCOS cases exhibited > fasting insulin levels (Slowinska-Srzednicka, 1991)insulin levels (Slowinska-Srzednicka, 1991)
PCOS women have evidence of metabolic PCOS women have evidence of metabolic derangements similar to the Metabolic Cardiovascular derangements similar to the Metabolic Cardiovascular Syndrome (Syndrome X) [dyslipidemia (Syndrome (Syndrome X) [dyslipidemia ( LDLc, LDLc, HDLc, HDLc, triglycerides), hypertension, and Type II triglycerides), hypertension, and Type II diabetes]diabetes]
Literature continued
Conversely, other investigators have noted Conversely, other investigators have noted little difference in LDLc levels in thin versus little difference in LDLc levels in thin versus obese PCOS cases and controls (Norman l992, obese PCOS cases and controls (Norman l992, Conway, l995).Conway, l995).
Earlier studies had small sample sizes and did Earlier studies had small sample sizes and did not control for potential confoundersnot control for potential confounders
Summary of Studies Linking Adverse Lipid Profiles in Women with PCOS
Click for larger picture
Hyperinsulinemia, Insulin Resistance, and Impaired Hemostatic Function
Hyperinsulinemia and insulin resistance, as observed in PCOS, Hyperinsulinemia and insulin resistance, as observed in PCOS, have been associated with increased CVD risk (Haffner, 1992; have been associated with increased CVD risk (Haffner, 1992; Savage, 1996; Reaven, 1998)Savage, 1996; Reaven, 1998)
Recent evidence suggests hyperinsulinemia in PCOS correlates Recent evidence suggests hyperinsulinemia in PCOS correlates with increased CVD risk independent of obesity (Mather, 2000) with increased CVD risk independent of obesity (Mather, 2000)
Increased risk may be a direct effect of elevated insulin or an Increased risk may be a direct effect of elevated insulin or an indirect effect, through the influence of insulin on other indirect effect, through the influence of insulin on other metabolic processes (e.g. thrombosis, fibrinolysis) metabolic processes (e.g. thrombosis, fibrinolysis)
Whether this translates into a true “ increased risk” of CHD is Whether this translates into a true “ increased risk” of CHD is not knownnot known
CHD Mortality -Previous Literature
To date, one historical prospective study of To date, one historical prospective study of 768 PCOS women (McKeigue et al) 768 PCOS women (McKeigue et al) undergoing wedge resection in the l930-1979 undergoing wedge resection in the l930-1979 has been conducted. There was no increased has been conducted. There was no increased risk of cardiovascular disease noted in this risk of cardiovascular disease noted in this population.population.
Types of Cohort Studies
Click for larger picture
Pierpoint et al., J Clinical Epi, 51(7), 1998
Click for larger picture
Brief Literature Review
In an earlier study of sudden cardiac death among In an earlier study of sudden cardiac death among women aged 25-64 in Allegheny County (Talbott ’89), we women aged 25-64 in Allegheny County (Talbott ’89), we noted an > rate of childlessness in ever married cases noted an > rate of childlessness in ever married cases compared to controls (N=80 cases, 80 controls). compared to controls (N=80 cases, 80 controls).
Detailed reproductive histories,however, were not Detailed reproductive histories,however, were not available. available.
The question of whether a reproductive endocrine The question of whether a reproductive endocrine abnormality such as PCOS was somehow linked to abnormality such as PCOS was somehow linked to sudden cardiac death was intriguing.sudden cardiac death was intriguing.
Risk of CHD in PCOS (Phase I l992-94) (Talbott, Guzick, Berga et. al.)
PCOS was a common enough diagnosis in medical PCOS was a common enough diagnosis in medical practices that a large-scale study was possible.practices that a large-scale study was possible.
The specific aim of our first study was to provide The specific aim of our first study was to provide information on whether women with PCO have a information on whether women with PCO have a greater incidence of known risk factors for CHD greater incidence of known risk factors for CHD compared to age matched controls.compared to age matched controls.
Methods
Using both historical and current records Using both historical and current records from the Reproductive Endocrine Clinic at from the Reproductive Endocrine Clinic at Magee Women’s our goal was to identify 300 Magee Women’s our goal was to identify 300 women with PCO and match them with 300 women with PCO and match them with 300 neighborhood controls.neighborhood controls.
Recruitment continued
Review of records of Dr. David Archer Review of records of Dr. David Archer (1972-1986) /Drs. David Guzick and Sarah (1972-1986) /Drs. David Guzick and Sarah Berga (2800 records, 10% met PCOS Berga (2800 records, 10% met PCOS criteria)criteria)
Definition of PCOSDefinition of PCOS :chronic anovulation :chronic anovulation and 1) evidence of clinical (and 1) evidence of clinical (hirsutism) hirsutism) or or biochemicalbiochemical (Total T > 2.0 nmol/L) (Total T > 2.0 nmol/L) hyperandrogenism or 2) LH:FSH > 2.0hyperandrogenism or 2) LH:FSH > 2.0
Research Methodology
Clinic VisitClinic Visit Questionnaire dataQuestionnaire data: demographics, lifestyle : demographics, lifestyle
assessment, medical, reproductive, and menstrual assessment, medical, reproductive, and menstrual historieshistories
AnthropometricsAnthropometrics (height, weight, waist and hip (height, weight, waist and hip circumferences)circumferences)
Blood pressure assessmentBlood pressure assessment Lipids ( total cholesterol, triglycerides, LDLc, HDLc)Lipids ( total cholesterol, triglycerides, LDLc, HDLc)
Fasting glucose and insulinFasting glucose and insulin
Results
Hormone levels PCOS cases vs controlsHormone levels PCOS cases vs controls Blood lipid and insulin levels cases vs Blood lipid and insulin levels cases vs
controlscontrols AnthropometricsAnthropometrics Decided to look at age relationshipsDecided to look at age relationships
Selected Demographic and Reproductive Factors in PCOS Cases and Controls
VARIABLES
Age
BMI
HDLT
LDLc
Insulin
Triglycerides
WHR
CASES (n=244)
35.3 +7.4
29.9 +7.95
51.2 +14.74
119.9 +31.8
23.3 +17.8
123.6 +88.7
0.83 +0.13
CONTROLS (n=244)
36.7 +7.7
26.6 +6.77*
56.1 +14.43*
112.0 +32.6*
13.6 +8.7*
87.3 +63.1*
0.76 +0.07*
* p < 0.01** p < 0.05
CHD Risk Factors by Age in PCOS
Table 3. DISTRIBUTION OF CHD RISK FACTORS IN PCOS CASES AND CONTROLS BY AGE- 19-34 years
19-24
25-34
Cases N=24
Controls
N=21
Cases N=87
Controls
N=73 Body Mass Index
27.1 (6.4)
25.7 (7.5)
28.7 (7.1)
25.9 (6.5)*
Waist/Hip Ratio
.76 (.06)
.73 (.06)
.82 (.10)
.77 (.07)
Total Cholesterol HDLT (mg/dL) Calc LDL (mg/dL)
186.5 (29.1) 58.5 (16.6) 111.1 (31.4)
163.2 (37.9)* 55.0 (13.9) 91.1 (39.2)
192.6 (36.8) 50.1 (14.5) 120.1 (37.0)
181.1 (38.2) 54.9 (13.8)* 110.1 (33.3)*
Triglycerides (mg/dL) *
Median
83.5 (36.1)
82.5
85.5 (69.3)
73.0
114.3 (81.0)
97.0
77.4 (42.9)
62.5 ** Insulin (μU/L)
22.0 (14.5)
16.8 (12.3)
21.9 (21.8)
13.2 (10.1)**
Systolic (BP) (mm Hg)
108.9 (9.2)
106.7 (10.2)
109.9 (11.3)
107.6 (9.9)
Diastolic BP (mm Hg)
70.6 (7.9)
67.6 (5.9)
71.0 (9.0)
69.8 (7.3)
*p<.05 **p<.01
Adapted from the Journal of Clinical Epidemiology vol 51, No 5, 1998, Pages 415-422.
Click for larger picture
CHD Risk Factors by Age in PCOS (cont.)
Table 3. DISTRIBUTION OF CHD RISK FACTORS IN PCOS CASES AND CONTROLS BY AGE- 35+ years
35-44
45+
Cases N=106
Controls N=105
Cases N=27
Controls
N=45 Body Mass Index
31.3 (7.8)
26.9 (6.9)**
30.8 (10.3)
27.1 (6.3)
Waist/Hip Ratio
0.8 (.01)
0.8 (.08)
.89 (0.3)
.76 (0.1)*
Total Cholesterol HDLT (mg/dL) Calc LDL (mg/dL)
200.3 (30.6) 49.9 (14.3) 123.1 (29.0)
185.0 (33.5)** 54.7 (12.7)** 112.2 (30.9)**
198.4 (33.8) 54.1 (13.9) 117.0 (27.6)
203.3 (32.3) 62.4 (17.2)* 121.5 (29.3)
Triglycerides (mg/dL) *
Median
132.3 (90.8)
109.5
86.5 (50.3) *
73.0 **
149.4 (117.2)
111.0
93.5 (91.4)
67.0 * Insulin (μU/L)
24.4 (19.7)
15.2 (11.6)
22.0 (17.0)
12.6 (6.8)**
Systolic (BP) (mm Hg)
116.4 (15.3)
110.6 (13.4)**
119.7 (21.2)
115.5 (15.8)
Diastolic BP (mm Hg)
75.0 (11.2)
70.7 (8.5)**
72.7 (13.1)
73.2 (8.9)
*p<.05 **p<.01
Adapted from the Journal of Clinical Epidemiology vol 51, No 5, 1998, Pages 415-422.
Click for larger picture
Age Relationships
Risk profile by 10 year age groupsRisk profile by 10 year age groups LDL levels: Women with PCOS had sig > LDL levels: Women with PCOS had sig >
LDL levels at ages less than 35 yearsLDL levels at ages less than 35 years Insulin was higher in each age group than Insulin was higher in each age group than
controlscontrols BP was significantly greater than controls in BP was significantly greater than controls in
the 35 to 44 group onlythe 35 to 44 group only
Control for BMI
PCOS cases <40 yrs. showed significantly PCOS cases <40 yrs. showed significantly higher LDL levels than controls and this higher LDL levels than controls and this appears to be independent of BMIappears to be independent of BMI
simple regression line of risk factors by age for simple regression line of risk factors by age for cases and controls.cases and controls.
Predictors of CHD Risk Factors in PCOSTable 2. Estimated Parameters from Multiple Linear Regression: Predictors of
CHD Risk Factors
Independent
Variable
HDLT
HDL2
CHOL
LDL
SBP
DBP
Trig1
PCOS2
Age Hormone Status
Insulin BMI
Smoking
-2.512
-.025 .050
-.159* -.437* -.474
-2.158*
.031 1.029 -.054* -.238* -.523
9.121*
.857* -3.687
.094 .653*
-1.833
6.980*
.539* -2.82
-.02 .761*
-1.867
.763
.394* -1.041 .143* .799*
1.650*
.520
.174* -.468 .103* .514* .952*
.197* .011* -.046 .010* .016* .037
1 Triglyceride levels were log transformed
*p<.05
Summary
Younger women less than 40 with PCOS Younger women less than 40 with PCOS exhibit adverse lipid profilesexhibit adverse lipid profiles
Approximately 50% of PCOS are insulin Approximately 50% of PCOS are insulin resistant as demonstrated by a GIR <5resistant as demonstrated by a GIR <5