CARDIOVASCULAR RISK FACTORS VIEWED AS ADAPTATIONS TO PREGNANCY

CARDIOVASCULAR RISK FACTORS VIEWED AS ADAPTATIONS TO PREGNANCY: THE MATERNAL ORIGINS HYPOTHESIS Nigel Paneth MD MPH Michigan State University http://www.epi.msu.edu/faculty/paneth .htm ESPR, Prague October 7, 2007 WHY ARE CARDIOVASCULAR RISK FACTORS SO COMMON? TWO EXPLANATIONS OF THE SOURCE OF CARDIOVASCULAR RISK • Thrifty genotype (Neel 1962). Focuses on genetic evolution in relation to food scarcity. Scarcity of food throughout life favors survival of individuals with genotypes that conserve nutrients, store fat, maintain blood sugar, etc. • Thrifty phenotype (Hales and Barker 1992). Focuses on the perinatal period and the intrauterine experience. Scarcity of food prenatally favors survival of individuals with phenotypes that conserve nutrients, store fat, maintain blood sugar, etc. POPULATION-LEVEL EXPLANATIONS VS INDIVIDUAL-LEVEL EXPLANATIONS • Neither hypothesis has been primarily focused on explaining differences between populations in CVD risk factors. • The Neel hypothesis has, however, been invoked to explain rapid development of CVD risk factors in populations newly exposed to Western diets, arguing that these populations may have had generations of extreme nutritional histories. • The Barker hypothesis has had to contend with population-level contradictions to its thesis, in that populations who have large babies are often at high CVD risk, the opposite of what the Barker hypothesis argues at the individual level. For example, in the INTERSALT study, Mean BW and Mean BP across countries has a positive correlation of 0.36 AN ALTERNATIVE EXPLANATION FOR THE ORIGIN OF CARDIOVASCULAR RISK FACTORS THE MATERNAL ORIGINS HYPOTHESIS THE MATERNAL ORIGINS HYPOTHESIS • The maternal origins hypothesis has features of both the Neel and Barker hypotheses, uniting genetic evolution and the influence of the perinatal period by positing that cardiovascular risk factors originated as, and indeed, still function as, adaptations to benefit pregnancy. • It differs from the Barker hypothesis in suggesting that the evolutionary target is not the fetus and infant, but the mother. • It differs from the Neel hypothesis by focusing, not on overall genetic adaptation, but on the pregnancy experience. • It differs from both by relying for its argument on observed differences between some populations in CVD risk factors and in perinatal outcomes. • My argument is that genetic evolution for pregnancy success is the substrate for cardiovascular disease. • CVD risk factors are the price the human species pays for pregnancy success. • However, different populations pay that price in different currencies and for different evolutionary reasons. • These differences depend upon whether the population had to prioritize, in its evolutionary history, maternal survival or feto-infant survival. BALANCE OF FORCES IN PREGNANCY FETAL SURVIVAL MATERNAL SURVIVAL PREGNANCY SUCCESS MATERNAL OR FETAL? • Pregnancy success is characterized by both maternal and fetal survival • But there is competition for survival between mother and fetus – They compete for resources, especially nutrition – In humans, the combination of upright posture and large head circumference at birth makes labor and delivery more difficult than in any other mammal. – The fetus thus puts the mother’s life at risk, especially at birth • This suggests that there may possibly be separate genetic patterns that favor either maternal or fetal survival TWO TYPES OF CARDIOVASCULAR RISK FACTORS ONE SET IS “METABOLIC” • Central obesity • Dyslipidemia • Insulin resistance A SECOND SET IS “VASCULAR/HEMATOLOGIC” • Raised blood pressure • Pro-thrombotic state FETAL SURVIVAL IS ENHANCED BY GENETIC ADAPTATIONS THAN FAVOR EFFICIENT TRANSFER OF NUTRIENTS TO THE FETUS THESE ADAPTATIONS ARE THE “METABOLIC” CARDIOVASCULAR RISK FACTORS ADAPTATIONS TO ENHANCE FETAL NUTRITION IN NORMAL PREGANCY • Fat deposition, especially central fat • Elevation in lipid fractions, especially triglycerides • Increased insulin secretion and increased insulin resistance • These metabolic changes all enhance storage and transfer of nutrients to the fetus. • Pregnancy is the time in life when the metabolic CVD risk factors are good for us. • In fact, many healthy pregnant women fit several of the criteria for having the metabolic syndrome. THESE METABOLIC ADAPTATIONS FAVOR THE FETUS, AND TO SOME EXTENT PUT THE MOTHER AT NUTRITIONAL RISK. ARE THERE PREGNANCY ADAPTATIONS THAT FAVOR MATERNAL SURVIVAL? DELIVERY HEMORRHAGE • Maternal blood loss at delivery is the most important hemorrhagic stress ordinarily encountered by the human species, with hundreds of uterine arteries/arterioles actively pumping at the moment of placental separation • In underdeveloped countries today, and probably throughout most of human history, bleeding has been the leading cause of maternal death. • Even in the US, 5% of mothers lose more than a liter of blood in delivery. (Magann EF et al: South Med J 2005;98:419-22) • It is likely that adaptations to minimize the severity of hemorrhage in labor may be important determinants of genes controlling two processes critical for minimizing blood loss vasoconstriction and thrombosis DOMINANCE OF THE PRO-COAGULANT, VASOCONSTRICTIVE STATE IN PREGNANCY • ↑ thrombin, PAI, thromboxane, factors VII, VIII, X • Thromboxane/prostacyclin balance favors vasoconstriction • Greatly enhanced risk of thrombotic disorders in pregnancy (20-fold) • These adaptations seem likely to have evolved to reduce the risk of delivery blood loss COULD PRE-ECLAMPSIA PROTECT AGAINST DELIVERY BLEEDING? • Blood pressure normally drops as pregnancy progresses because of decreased venous tone and increased capacitance. This is partly due to fetally-induced remodeling of maternal placental vasculature to increase blood flow to the fetus. • Pre-eclampsia (PE) may be an exaggerated version of ordinary pregnancy adaptions to reduce the risk of delivery bleeding since the primary lesion in PE is reduced cytotrophoblast invasion of uterine spiral arteries. • In PE, the mother seems to be trying to prevent the baby from so remodeling the uterine vasculature that she risks dying in labor from bleeding. TWO TYPES OF CVD RISK FACTORS AND PREGNANCY • The “metabolic” type of CVD risk factor (obesity, hyperlipemia, insulin-resistance) favors fetal and infant survival • The “vascular” type of CVD risk factor (hypertension, hypercoagulability) favors maternal survival in pregnancy and delivery. • By advantaging one member of the pregnancy dyad, each pattern favorable to one poses risks to the other. – Large fetal size enhances fetal survival but takes nutrients from the mother, and probably puts her at higher risk for dying from blood loss in labor. – More vasoconstriction and thrombosis may reduce risk of the mother dying in labor but can fetal growth and fetal survival. THE BALANCE BETWEEN THESE TWO ADAPTATIONS MAY BE REFLECTED IN POPULATION BIRTHWEIGHT DISTRIBUTIONS • In most populations a balance is seen between these two survival forces, but some populations appear to have evolved so that one or the other pattern is somewhat more dominant. • Large birthweights in the face of nutritional scarcity are a sign of a tilt toward fetal survival. Small birthweights and tendencies to preterm delivery are a sign of a tilt towards maternal survival IS THERE EVIDENCE OF A LINK BETWEEN A POPULATION’S CVD RISK PROFILE AND ITS BIRTHWEIGHT DISTRIBUTION? THE MEXICAN-AMERICAN PERINATAL PARADOX • Even under adverse socio-economic circumstances, Mexican-Americans in the US have relatively large babies who experience low infant mortality • More precisely, this is a Meso-American paradox, because it applies also to all Native American populations in the US. 30 25 PERCENT LOW BIRTHWEIGHT (<2,500g) BY ETHNIC GROUP (US - 2004) AND PERCENT POVERTY (CENSUS, 2005) 25.3 24.3 20.6 20 15 10.3 10 6.4 6.8 7.1 8.2 7.5 7.9 13.4 5 0 MEXICANAMERICAN ALL HISPANIC NON-HISPANIC WHITE AMERICAN INDIAN ASIAN BLACK AFRICAN-AMERICAN PERINATAL CHARACTERISTICS • Relatively high neonatal mortality • Lower mean birthweight, and higher frequency of low birthweight. • Lower mean gestational age, and higher frequency of preterm delivery • Relatively favorable neonatal survival for a given birthweight or gestational age below the mean. • Relatively unfavorable neonatal survival for birthweights and gestational ages above the mean • In mothers, relatively high risks of pre-eclampsia BIRTHWEIGHT IN WEST AFRICAN AND US MESOAMERICAN POPULATIONS:PRELIMINARY QUANTITATIVE REVIEW OF THE LITERATURE • Thanks to Crystal Pirtle MPH, PhD candidate in epidemiology at MSU • She identified 32 studies of West African birthweights and 20 studies of US Meso-American birthweights • In 18 of the West African studies, sample size and mean BW were available. 9 studies had N > 1,000. • In 9 of the Meso-American studies, sample size and mean BW were available. 6 studies had N > 1,000. MEAN BIRTHWEIGHT IN WEST AFRICAN AND MESO-AMERICAN POPULATIONS 4,000 3,500 3,000 2,500 2,000 1,500 1,000 500 0 All studies: All studies: All studies unweighted weighted > 1,000: weighted 3,636 3,014 3,570 2,989 3,596 2,982 West African MesoAmerican THE METABOLIC SYNDROME IN MESO-AMERICAN POPULATIONS • Most Meso-American populations are at high risk of the first three components of the metabolic syndrome – abdominal obesity, dyslipidemia and insulin resistance. 50% of Pima Indians have type II diabetes by the age of 50 • • The key pregnancy complication in MesoAmerican populations is gestational diabetes However, Meso-Americans are not at especially high risk of the fourth component, hypertension US AGE-ADJUSTED DIABETES AND IMPAIRED FASTING GLUCOSE BY ETHNIC GROUP (NHANES, 1999-2002) Source: Diabetes Care 2006;29:1263-8 35 31.6 30 26.1 25 20 15 10 5 3 0 MEXICAN-AMERICAN NON-HISPANIC WHITE BLACK 17.7 2.7 3.6 US AGE-ADJUSTED HYPERTENSION PREVALENCE IN HISPANIC 30 AND NON-HISPANIC WHITES (NHANES, 1999-2002) Source: MMWR 1/14/2005;54:7-9 27.4 25 25.1 20 15 10 5 0 ALL HISPANIC NON-HISPANIC WHITE MESO-AMERICANS HAVE BABIES WITH LARGE BIRTHWEIGHTS, AND HAVE AN INCREASED RISK OF THE METABOLIC, BUT NOT THE VASCULAR CVD RISK FACTORS. THIS IS A FETAL SURVIVAL PATTERN CARDIOVASCULAR RISK IN AFRICANAMERICANS • Unlike Meso-Americans, the most distinctive cardiovascular risk factor in African-Americans is hypertension • Insulin resistance and abdominal obesity occur, but less commonly than in MesoAmericans • Dyslipidemia is less severe; BMI-adjusted HDL-C levels are actually higher in African-Americans than in whites 45 40 35 30 25 20 15 10 5 0 US AGE-ADJUSTED HYPERTENSION PREVALENCE BY ETHNIC GROUP (NHANES, 1999-2002) Source: MMWR 1/14/2005;54:7-9 40.5 27.4 25.1 ALL HISPANIC NON-HISPANIC WHITE BLACK PREGNANCY ADAPTATIONS IN AFRICAN-AMERICANS • Higher rates of pre-eclampsia indicate a tendency towards vasoconstriction • Shorter mean gestation and slower fetal growth will also produce less delivery bleeding. • A predisposition to early delivery may also protect the mother from experiencing fullblown eclampsia. • This pattern is a maternal survival pattern WHY SHOULD PEOPLE OF WEST AFRICAN ORIGIN NEED SPECIAL PROTECTION FROM DELIVERY BLEEDING? • Possibly because of a high prevalence of anemia in pregnancy due to malaria infection. • Malaria is known to have altered the gene pool to produce the sickle-cell trait • Delivery bleeding is likely to especially threaten maternal survival in anemic women. SPECULATION ON THE KIND OF ENVIRONMENT THAT PRODUCES EACH SET OF PREGNANCY ADAPTATIONS • Dry climates with low rates of chronic infection but intermittent food supplies will produce the metabolic, fetoprotective adaptation pattern. • Humid climates with high rates of anemiaproducing chronic infections will produce the vascular, maternoprotective adaptation pattern. BIRTHWEIGHT AND CARDIOVASCULAR RISK ARE THUS GENETICALLY LINKED Populations with unusually large babies are often at risk for obesity, diabetes and hyperlipemia, but not particularly for raised blood pressure. I conclude that the genes for metabolic risk factors evolved to make the babies large. • Populations with unusually small babies are often at risk for hypertension. I conclude that the genes for hypertension evolved to protect the mother from excessive bleeding at delivery IMPLICATIONS FOR CARE AND PREVENTION This hypothesis is only designed to understand how population predispositions to different components of the cardiovascular risk profile may have developed as a result of perinatal survival pressures. It does not minimize in any way our continuing public health need to find environmental ways to prevent and ameliorate the human consequences of our shared evolutionary history, particularly in populations at risk. BUT THERE IS ONE DIFFICULTY • If populations with low birthweights are susceptible to hypertension then that can explain the fetal origins observation of an inverse association between birthweight and blood pressure in later life. • But how can this hypothesis be compatible with the observation that low birthweight is linked with isulin resistance/diabetes in later life? THE FETAL INSULIN HYPOTHESIS (Hattersley AT et al: Nature Genetics 1998; Lancet 1999; Weedon MN et al Diabetes 2005; Am J Hum Gen 2006) • Fetal growth is stimulated by fetal insulin secretion, which in turn is stimulated by maternal hyperglycemia • Glucokinase is an enzyme that serves as a glucose sensor, regulating insulin secretion • Some mutations in glucokinase-regulating genes produce insulin resistance. These mutations produce higher birthweights (as much as 500 grams) when the mother has them, because the mother’s insulin resistance and hyperglycemia stimulates insulin secretion in the fetus. • The same mutations produce lower birthweights (as much as 600 grams) when the fetus has them because the fetus is resistant to insulin and to its growth-promoting properties. • Since glucokinase mutations are common (30% of the population), the observed relationship between low birthweight and later insulin resistance may be explained. THANK YOU VERY MUCH I WELCOME QUESTIONS – ESPECIALLY THOSE THAT TEST THIS HYPOTHESIS FURTHER EVIDENCE FOR A GENETIC LINK BETWEEN BIRTHWEIGHT AND CVD RISK FACTORS Cai et al Human Genetics 2007;121:737-744 – – – Sample of 1,030 Hispanic families in Texas Genetic correlations of BW with CVD risk factors ranged from 0.30 – 0.59 “Chromosome 10q22 harbors genes influencing both birth weight and childhood body size and cardiovascular disease risk in Hispanic children” PERINATAL EPIDEMIOLOGY A discipline that uses epidemiologic approaches to investigate human health phenomena occurring during pregnancy and infancy either as outcomes of interest or as exposures that may lead to adverse health states in later life. PERINATAL EPIDEMIOLOGY TRAINING PROGRAM AT MSU (T-32) TRAINING IN PERINATAL EPIDEMIOLOGY AT MSU • Program funded by NICHD in May 2005 • The only T-32 training program in the nation focused solely on perinatal epidemiology • Support restricted by NIH rules to US citizens/green card holders • We have two post-doctoral positions per year • Support is for two years • Accepting applications for 2007-8 until May 2007 • Would love to have a pediatrician in the program! • If interested, email cv to paneth@msu.edu CONCERNS RAISED ABOUT THE BARKER HYPOTHESIS 1. 2. 3. 4. 5. 6. 7. Hypothesis is moving target Conflict of hypothesis with population data Famine studies by and large unsupportive Interventions to raise birthweight not generally successful Failure to address confounding Multiple and selective comparisons Weak and inconsistent effects 8. Inappropriate adjustments KEY CRITICAL PAPERS • Elford J, Whincup P, Shaper AG: Int J Epidemiol 1991; 20:833-844 and J Epidemiol and Comm Health 1992; 46:18 • Joseph KS, Kramer M: Epidemiol Reviews 1996;18:158-172 • Paneth N, Ahmed F, Stein AD: J Hypertension 1996;14 (suppl) S121-129 • Huxley R, Neil A. Collins R: Lancet 2002;360:659-665 • Huxley R, Owen CG, Whincup P et al: JAMA 2004; 292:2755-64 • Tu, Y-K, West R, Ellison GTH et al: Am J Epidemiol 2005; 161: 27-32 Hattersley AT and Tooke JE Lancet 1999; 353: 178992 Common variants in the glucokinase gene • Several cohorts including ALSPAC • Polymorphism at position –30 • Prevalence of 30% • A allele associated with increase in fasting blood glucose of 0.075mmol/L in pregnancy Weedon et al Diabetes 2005; 54: 576581 Hattersley AT and Tooke JE Lancet 1999; 353: 1789-92 EFFECT OF A ALLELE IN GCK IN MOTHERS ON BIRTHWEIGHT (NO EFFECT OF ALLELE IN CHILD) Weedon et al Diabetes 2005; 54: 576-581 SAVING MOTHER OR SAVING BABY? For each set of genes that have evolved because of pregnancy and perinatal pressures, one must consider whether they are designed for optimal fetal or for optimal maternal survival, which may be in competition. • In Meso-Americans, genes seem to be favored that protect the fetus, by transferring nutrients and producing big babies. The price paid is diabetes and obesity. • In African-Americans, genes seem to be favored that protect the mother, by producing a smaller, less mature baby, and a tendency to preeclampsia. The price paid is hypertension and somewhat lower neonatal survival. • These two populations are best seen as illustrating processes that are universal and found to some degree in all populations.