Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
5 Sex Affects Health ABSTRACT Males andfemales have different patterns of illness and different life spans, raising questions about the relative roles of biology and environment in these disparities. Dissimilar exposures, susceptibilities, and responses to initiating agents and differences in energy storage and metabolism result in variable responses to pharmacological agents and the initiation and mani- festation of diseases such as obesity, autoimmune disorders, and coronary heart disease, to name a few. Understanding the bases of these sex-based differences is important to developing new approaches to prevention, diag- nosis, and treatment. Sex should be considered as a variable in all biomedi- cal and health-related research. Studies should be designed to control for exposure, susceptibility, metabolism, physiology (cycles), and immune re- sponse variables. Males and females have different patterns of illness and different life spans, which leads to important questions about how these differences might be biologically determined. Diseases other than those of the repro- ductive system affect both sexes, often with different frequencies or presentations, or they may require different treatments. This chapter explores how these differences result from differences in exposures, sus- ceptibilities, and responses to disease-initiating agents, differences in en- ergy storage and metabolism, and disparate diagnostic and therapeutic interventions. Because it is not possible to explore all diseases, disorders, and condi- 117
118 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH lions with sex differences, the committee chose several illustrative ex- amples. The committee first briefly describes the complexities of sex dif- ferences in response to therapeutic agents and energy metabolism. The subsequent section focuses on differences in energy metabolism, obesity, and physical performance and then uses two illustrations melanoma and osteoporosis to describe sex differences. The chapter then focuses on the complexities of a normal immune response that has gone awry or that has spontaneously lost its normal immune regulation system (au- toimmune diseases). These diseases in particular demonstrate variable susceptibilities between human males and human females, as well as differential exposures to environmental factors. The chapter concludes by describing a disease whose etiology occurs from conception to the grave but that affects both sexes differently, coronary heart disease. SEX DIFFERENCES IN RESPONSE TO THERAPEUTIC AGENTS: DIAGNOSTIC AND THERAPEUTIC INTERVENTIONS Background Pharmacological agents can be used as probes to diagnose, prevent, and treat human illnesses. How much of an agent one encounters de- pends on the route of entry into the body (see Figure 5-7), absorption, distribution, metabolism, and excretion and is a major factor in the body's response, whether it is a therapeutic or an adverse response. The follow- ing discussion uses the definitions provided in Box 5-1 and the schema presented in Figure 5-1.
SEX AFFECTS HEALTH D= Drug UD = Unbound drug (free) PP = Plasma proteins T = Tissue R= Receptor Absorption - UD R Sitters) of Action 119 Sites of Tissue Storage ~ =: (3 j/ Circulation ~ UD , \ \ Excretion Metabolites ~ rev \ | Phase I Phase 11 \ + UD' Biotransformation (metabolism) ~ Excretion FIGURE 5-1 Schematic representation of the absorption, distribution, metabo- lism, and excretion of drugs. Pharmacokinetic and pharmacodynamic variables can be measured and can demonstrate differences between males and females. Pharmaco- kinetic and pharmacodynamic differences between males and females exposed to the same compound and dose do not necessarily result in different health outcomes. Outcome data that can be used, however, to establish whether differences between the sexes are clinically meaningful are sparse. The methods used to study sex differences in the effects of drugs can serve as a template for the study of the relative effects of any foreign chemical, including volatile organic chemicals. For example, most stan- dards for the human carcinogen benzene have been based on studies involving males, even though physiologically based pharmacokinetic modeling shows that females have higher levels of metabolism of ben- zene (Brown et al., 1998~. Greater appreciation of these methods will yield significant clinical data regarding the importance of sex both in drug development, prescription, and dosing and in assessments of environ- mental exposures.
20 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH In reviewing the examples described below, the committee was cog- nizant of the small numbers of subjects involved in some of the studies, the sometimes conflicting results of studies, and the significant variations observed in the results of many of the studies that purport to evaluate sex and age differences. It will be essential in future studies to define the stages of a woman's menstrual cycle, use protocols with sufficient power to detect statistically significant differences, and determine whether dem- onstrated differences deemed to be sex relevant affect clinical outcomes. Although many studies are designed to demonstrate sex differences in pharmacokinetics, few look at pharmacodynamics, and even fewer deter- mine clinical outcomes. Processes of Drug Absorption and Metabolism Absorption of Pharmacological Agents Through Different Routes of Entry Absorption of small organic molecules is usually passive, but it may involve a facilitated process or an active process that requires energy. The factors that affect the absorption of chemicals from the gastrointestinal tract are listed in Table 5-1. TABLE 5-1 Factors Affecting Absorption of Chemicals Physical characteristics Dosage form Gastrointestinal characteristics pKa Aqueous solubility Lipid solubility Disintegration Dissolution Enteric coating Controlled-release properties Presence of food and type of food in the gastrointestinal tract Rate of gastric emptying pH of the different segments of the gastrointestinal tract Intestinal microorganisms Intestinal transit time Gastrointestinal blood flow Gastrointestinal enzymes Alcohol dehydrogenase Cytochrome P450 Glucuronosyltransferase Sulfotransferase Gastrointestinal transport systems P glycoprotein Effects of other drugs
SEX AFFECTS HEALTH 121 The venous blood supply of the entire gastrointestinal tract except the rectum goes directly to the liver, where absorbed compounds may be metabolized (first-pass metabolism). This observation may be explained by the high prevalence of CYP3A4 in the upper part of the intestine and its absence from the colon and rectum. Drugs given through the rectum (suppositories) and urinary bladder do not go through significant first- pass metabolism (Buyse et al., 1998~. Interestingly, drugs applied to the vagina accumulate to a greater extent in the uterus than when adminis- tered by other routes; that is, they have a "first-uterine-pass effect" (Bulletti et al., 1997; Mizutam et al., 1995~. Gastrointestinal Tract Absorption of drugs from the stomach is affected by a variety of factors (Table 5-1~. (For a general review of the effect of gastrointestinal motility on the absorption of drugs, see Hebbard et al. [1995~.) Gastric emptying (Malagelada et al., 1993) can be measured by several techniques, of which transit times determined with a radiolabeled liquid or solid meals provide the best-validated and clinically meaningful measurements (Camilleri et al., 1998~. Differences in age, sex, body mass index, phase of menstrual cycle, and type of meal consumed lead to large inter- and intrasubject variabili- ties. The preponderance of evidence (Bennink et al., 1999; Datz et al., 1987; Gryback et al., 1996; Hermansson and Silvertsson, 1996; Hutson et al., 1989; Knight et al., 1997; Tucci et al., 1992) supports the conclusion that females empty solids more slowly than men; however, others have found that all transit variables are unaffected by sex (Madsen, 1992~. The gastric emptying of liquids has also been reported to be slower in women than men (Datz et al., 1987; Mohiuddin et al., 1999~; however, others have found no differences (Bennink et al., 1998~. In addition, young people empty their stomachs faster than elderly people do (Jane et al., 1998; Moore et al., 1983; Teff et al., 1999~. The small-bowel transit time of solids and liquids do not differ between the sexes (Horowitz et al., 1984; Madsen, 1992), but differences are seen between old and young subjects (Bennink et al., 1999; Madsen, 1992~. Slower gastric emptying in females does not likely affect the absorp- tion of most solid drugs since most absorption takes place in the small intestine. The rate of absorption of enteric coated forms is delayed (Mojaverian et al., 1987~. Drugs with narrow therapeutic indices are most likely to be harmful (Greiff and Rowbotham, 1994) or lack efficacy in persons with slow rates of gastric emptying. A slow rate of gastric empty- ing also decreases the level of absorption of alcohol. Progesterone may be responsible for the slower rate of gastric empty- ing in women (Gill et al., 1985, 1987; Hutson et al., 1989; Mathias and Clench, 1998; Riezzo et al., 1998~. Females are less sensitive than males to muscarinic blockade of stomach motility, possibly because of differences
22 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH in autonomic tone (Teff et al., 1999~. It is of interest that 80 to 90 percent of diabetic patients with gastroparesis (paralysis of the stomach) are females (Bennink et al., 1998~. Basal acid output, pH, and gastrin secretion are independent of sex (Bernard) et al., 1990; Dressman et al., 1990; Straus and Raufman, 1989) and age (Russell et al., 1993~. In patients with gastroesophageal reflux, however, the mean basal level of acid output is greater in males than females (Collen et al., 1994~. Transport systems within the gastrointestinal tract may have signifi- cant effects on drug absorption. For example, a component of the intestine and liver, P glycoprotein (Pgp), is a transmembrane efflux protein that actively transports many compounds including drugs out of cells. The livers of males express twofold larger amounts of Pgp than the livers of females (Schuetz et al., 1995~. This suggests that males transport drugs out of hepatocytes more rapidly than females, decreasing the time for biotransformation, although further studies are needed. The absorption of many drugs might be affected if a sex difference in intestinal Pgp activity also exists, and it is important to determine whether such a sex difference exists. Drug absorption rates from the rectum may also be sex dependent. Females absorbed one of the two specially prepared ondansetron suppositories (an antiemetic) differently than males (Jane et al., 1998), although this could be attributed to normal variability. The menstrual cycle has no effect on motility in the esophagus (Mohiuddin et al., 1999) or on whole-gut transport (Kamm et al., 1989~. However, evidence for a menstrual cycle effect on gastric emptying is conflicting (Gill et al., 1987; Horowitz et al., 1985; Mones et al., 1993; Parkman et al., 1996; Petring and Flachs, 1990~. Skin Gels, ointments, creams, or patches deliver small organic com- pounds through the skin (Brown and Langer, 1988; Xu and Chien, 1991~. Transit through the stratum corneum (outer skin) barrier often requires the use of absorption enhancers (Kanikkannan et al., 2000~. The level of transepidermal water loss, a measure of epidermal bar- rier permeability following injury to and recovery of the stratum cor- neum, does not differ between males and females or between Caucasians and Asians (Reed et al., 1995~. Dark skin recovers from an injury more quickly than lightly pigmented skin. In a clinical trial of transdermal clonidine for the treatment of hypertension, blood pressure reduction was independent of race, ethnicity, sex, and age (Dies et al., 1999~. In vitro, transdermal absorption of fentanyl and sufentanil (analgesics) was neither age nor sex dependent (Roy and Flynn, 1990~. Pulmonary tract Absorption of drugs via the pulmonary tract varies with breathing rate and depth (ventilation) (Gonda, 2000~. Progesterone
SEX AFFECTS HEALTH 123 may be a ventilatory stimulant. Females have greater minute ventilations and lower tidal volumes than males, and the ventilatory response to high carbon dioxide levels is greater in males (White et al., 1983~. A drug, ethionamide, given orally to healthy and ill men and women appeared in equal concentrations in alveolar cells of both sexes (Conte et al., 2000~. Women have higher ventilatory responses in the luteal phase than in the follicular phase of the menstrual cycle. Protein Binding Most small organic compounds bind to albumin or oc~-acid glycopro- tein (AAG) and less frequently to alpha, beta, and gamma globulins, lipo- proteins, or erythrocytes. AAG binds with a high affinity to basic drugs. Albumin binds to acidic drugs in a complex in which the drug readily dissociates to maintain an equilibrium between the bound and unbound (free) fractions. The unbound fraction is in equilibrium with the receptor (Anton, 1960; Shoeman and Azarnoff, 1975~. Thus, the degree of binding of drugs to plasma proteins can influence their dispositions (Gillette, 1973~. The plasma protein binding of enantiomers (mirror-image com- pounds) in racemic mixtures (containing both enantiomers) may differ, and selective binding of the enantiomers does occur (Gross et al., 1988; Waite et al., 1983~. The level of AGG, an acute-phase reactant, increases in patients with infections, cancer, and rheumatoid arthritis. Decreased albumin levels or increased AAG levels occur in patients with renal, liver, and thyroid Crohn's disease, myocardial infarction, cancer, and burns (Reidenberg and Affrime, 1973~. Increased levels of unbound drug occur in patients with uremia (Garland, 1998; Reidenberg and Affrime, 1973) and cirrhosis (Goldstein et al., 1969~. It is not known whether sex differences exist in these circumstances. In a very small study (nine women each examined through one men- strual cycle), the concentration of AGG was higher on day 4 of the men- strual cycle than on days 12, 20, and 28 (Parish and Spivey, 1991), but the study had very significant inter- and intrapatient variabilities. The level of sex hormone binding globulin has also been found to increase during the luteal phase of the menstrual cycle (Plymate et al., 1985~. The concentration of a drug in the fetus is a function of the concentra- tion in the mother, placental permeability, fetal drug clearance, and dif- ferences in the levels of protein binding between maternal and fetal plasma (Boulos et al., 1971~. The placental transfer of lipophilic drugs is good, but the transfer of hydrophilic drugs is slow. The concentration of protein is significantly lower in the fetal circula-
24 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH TABLE 5-2 Differences in Drug Concentrations Between the Mother and Fetus and Between Males and Females Mother-fetus comparisons Male-female comparisons Binding of propranolol and verapamil is lower in fetal serum Propranolol R/S enantiomer ratio is larger in maternal serum Verapamil R/S enantiomer ratio is similar in maternal and fetal serum AAG levels are higher in maternal serum Albumin levels are slightly lower in maternal serum (Belpaire et al., 1995; Notarianni, 1990) Unbound concentrations of chlordiazepoxide, diazepam, imipramine, and nitrazepam are higher in women (MacKichan, 1992) The level of the unbound fraction of S-propranolol but not R-propranolol is decreased in older women (Walle et al., 1983) lion than in the maternal circulation during early pregnancy, but the concentration in the fetus exceeds that in the mother at term. The mater- nal AAG level is very low before 16 weeks of gestation and thereafter increases at a constant rate to a fetal concentration-maternal concentra- tion ratio of 0.37 near term (Perucca and Crema, 1982~. Examples of cir- cumstances in which plasma protein variables affect drug concentrations are shown in Table 5-2. Body Composition Male-female differences in body fatness may account for the increased volumes of distribution for lipophilic drugs (such as benzodiazepines) in females (Parker, 1984; Sciore et al., 1998) and for alcohol in males (Loebstein et al., 1997; Parker, 1984; Petring and Flachs, 1990~. The level of total body water decreases with age because of a dispro- portionate decrease in intracellular water levels (Kashuba and Nafziger, 1998; Phipps et al., 1998~. It is important to adjust body composition mod- els not only for sex but also for age and body size (Kasuba and Nafziger 1998; Phipps et al., 1998~. (It is important to note, however, that some sex differences in pharmacokinetics reported in the literature are a result of differences in weights between males and females receiving the same fixed dose of the drug. Thus, pharmacokinetic parameters should be cor- rected by weight before concluding that a sex difference exists. The effects of differences in body fat, as noted above, can confound weight issues and should also be considered.)
SEX AFFECTS HEALTH Biotransformation 125 Sex has a complex effect on the pharmacokinetics of drugs metabo- lized in the liver (Harris et al., 1995; Yonkers et al., 1992~. Temazepam and oxazepam, benzodiazepines that are metabolized through conjugation, are cleared faster by males (Divoll et al., 1981; Greenblatt et al., 1980; Smith et al., 1983~. Alprazolam (Kristjansson and Thorsteinsson, 1991) and diazepam (Greenblatt et al., 1980) are metabolized via an oxidative mechanism and are cleared faster by females. Nitrazepam (Iochemsen et al., 1982) is metabolized via reduction of its nitro group, and its metabo- lism shows no sex differences. Thus, sex affects differently even drugs within the same pharmacological class and drugs with the same structures. Cytochromes P450 The cytochromes P450 (CYPs) are a superfamily of at least 17 isozymes that modulate the oxidative metabolism of drugs in the liver (Wrighton and Stevens, 1992~. CYP1, CYP2, and CYP3 are thought to be responsible for most hepatic metabolism of drugs. The CYP3A4 subfamily is the most abundant of the CYPs in the human liver and is responsible for the metabolism of cyclosporine, quinidine, erythro- mycin, dapsone, and lidocaine (Harris et al., 1995; Wing et al., 1984~. The major CYP isoform found in the human embryonic, fetal, and newborn liver is CYP3A7. The activity of CYP3A4 is very low before birth but increases rapidly at birth and reaches 50 percent of the level in adults between 6 and 12 months of age. This maturation of drug-metabolizing enzymes is the main factor for age-associated changes in nonrenal drug clearance (de Wildt et al., 1999~. Studies with tirilazad (an antioxidant) (Hurst et al., 1994), erythromy- cin (Watkins et al., 1985), and diazepam (Greenblatt et al., 1980) suggest that females have greater CYP3A4 activity, but studies with other probe drugs yield conflicting results. Drugs metabolized by CYP3A4 are exten- sively cleared by females, whereas drugs cleared by other isozymes are usually cleared faster by males (Harris et al., 1995~. The sex-specific differ- ences in CYP3A4 activity are related to estrogen and progesterone, which regulate the activity of CYP3A4 at the gene level (Harris et al., 1995~. However, the metabolism of ranitidine, also metabolized by CYPs, shows no sex difference (Abad-Santos et al., 1996~. The metabolism of some drugs eliminated through conjugation shows sex differences (Divoll et al., 1981; Greenblatt et al., 1980, 1984; Macdonald et al., 1990; Miners et al., 1983, 1984~. Excretion The kidney is the major organ of drug excretion. Drugs diffuse in their un-ionized form across the kidney glomeruli and tubules or are
26 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH secreted and reabsorbed by active tubular transport systems. Drugs are also excreted in the feces if either the drug has not been absorbed from the gastrointestinal tract or it is excreted from the liver into the bile in the intestinal tract. Reabsorption (enterohepatic circulation) may occur as the excreted drug travels through the intestines. Males have higher levels of creatinine in serum and urine and higher rates of creatinine clearance (CLCR) than females. The difference is related to the greater lean body mass of males (lames et al., 1988~. In a three-way crossover study with young and elderly men and women, the renal clear- ance of amantadine was significantly inhibited by quinine and quinidine only in the male subjects. There were no age-related effects (Gaudry et al., 1993~. The mechanism of this interaction is probably related to the differ- ential effects of quinine and quinidine on the tubular excretion rate of cations (Charney et al., 1992~. A physician would not need to consider this interaction when treating an elderly woman with quinidine for muscle cramps who developed influenza and was prescribed amantadine; a lack of attention to this drug-drug interaction in a male, however, could lead to a significant adverse reaction. CLCR is lower during the first week of menses and increases by week 4 by 20 percent. Overnight CLCR measured three times a week during 11 menstrual cycles found the median CLCR to be 7.3 percent higher during the luteal phase than during the follicular phase. Similar changes were found when intravenous chromium 51-labeled EDTA was used, a more accurate measure of the glomerular filtration rate (Paaby et al., 1987a,b). Estradiol and estriol do not affect the glomerular filtration rate, urine flow, renal plasma flow, or tubular reabsorption in humans (Christy and Shaver, 1974; Davison, 1987~. CLCR was found to be slightly increased only in the midluteal phase in another study (Phipps et al., 1998~. How- ever, the changes were attributable to changes in creatinine excretion and are not considered clinically important. It is surprising that investigators are still attempting to determine the effect of the menstrual cycle on CLCR. The renal clearance of the aminoglycoside antibiotics tobramycin (Nafziger et al., 1989) and amikacin (Matsuki et al., 1999) are not signifi- cantly altered during the menstrual cycle. In a recent review, Kashuba and Nafziger (1998) reported that most published studies have been con- ducted with small numbers of women and limited numbers of menstrual cycle phases within one menstrual cycle. In addition, studies of the effects of estrogen or progesterone on renal clearance are limited and their re- sults are contradictory. They further conclude, "there are no demonstrated clinically significant changes that occur in the absorption, distribution or elimination of drugs" during a normal menstrual cycle (Kashuba and Nafziger, 1998, p. 204~. Beierle et al. (1999) concluded that sex-related differences in the renal clearance of drugs is generally only of minor importance. However, a meta-analysis of 10 studies with 172 healthy vol-
SEX AFFECTS HEALTH 127 unteers administered an oral therapeutic dose of the antibiotic fleroxacin provided evidence that the volume of distribution/systemic availability ratio (V/F) is 20.4 liters greater in males than females and that the clear- ance/systemic availability ratio is 10.8 milliliters per minute (ml/min) greater in males than females (Reigner and Welker, 1996~. Pharmacodynamics Receptors are macromolecules on or in cells to which a drug binds to initiate its effect. Continued stimulation of a receptor may lead to its downregulation or desensitization (refractoriness). Hyperreactivity may also occur and may result from long-term administration of antagonists or the synthesis of additional receptors. Drugs may also act as substrates for enzymes or may inhibit enzymes competitively or noncompetitively. Although there is marked interindividual variation (Levy, 1998), phar- macodynamic differences between the sexes do occur. Several examples are shown in Table 5-3. (For additional examples, see Table B-1 in Appendix B.) Clinical Implications Some pharmacokinetic and pharmacodynamic sex differences may affect drug efficacy or make serious adverse events more likely. A sex difference will more likely affect drugs with narrow therapeutic indices than those with wide therapeutic indices. Adjustment of the dose or dos- ing interval of the drug may be sufficient to correct the difference, or it may be necessary to use a different drug or treatment modality. Given the amount of resources being dedicated to drug development worldwide, it is especially important to consider sex-specific issues relat- ing to clinical trials research, the effects of drugs on receptors (or sexual dimorphisms of receptors and neurotransmitters), and the sexual dimor- phism of treatment responses. Sex-related variables should be specifically and comprehensively included in all diagnostic, longitudinal, and treat- ment research studies (i.e., all clinical studies involving humans) when- ever feasible. Studies should be designed to determine the relative effects of covariables to clearly establish the contributions of sex to outcome data. Analyses should be planned a priori (even if they are secondary or exploratory analyses) that address sex-related hypotheses (i.e., they should not rely primarily on post hoc analyses). In addition, at least some diagnostic, longitudinal, and treatment studies should be powered spe- cifically to permit the appropriate analysis of sex-related variables (post hoc analyses are usually conducted with sample sizes that are too small, making it likely that type II errors [the assumption that no relationship exists when in, fact, it does] will occur). The Office of Research on
28 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH TABLE 5-3 Receptor, Enzyme, and Structural Differences Between Males and Females Sex Difference Clinical Relevance Intra-arial ocl-adrenoceptor agonist phenylephrine induces greater vasoconstriction during the luteal phase than during the follicular phase of the menstrual cycle in African-American and white females. The oc2-adrenoceptor agonist clonidine induces greater vasoconstriction in the follicular phase than in the luteal phase in white females but not African-American females (Freedman and Girgis, 2000~. Men show a significant dose-response to vasodilatation to isoproterenol; females do not. Correlation curves of blood alcohol levels and sedation; the slope of the curve is steeper in females (Ammon et al., 1996~. Angiotensin I and II infusion; negative relationship between change in heart rate and mean arterial pressure during infusion in females only. The renal vasoconstrictor response is increased in females (Gandhi et al., 1998~. The rise of D1 and D2 striatal dopamine receptors in males but not females parallels the early developmental appearance of motor symptoms of attention deficit hyperactivity disorder (Andersen and Teicher, 2000~. Blood pressure regulation. The sedative effects of ethanol are greater in women. Sex differences in cardiorenal fluid balance at the pharmacodynamic level. Possible explanation for higher attention deficit hyperactivity disorder incidence in males. NOTE: See also Appendix B. Women's Health of the National Institutes of Health (1999a,b) has made numerous recommendations in this regard. Sex Differences in Adverse Events Drugs from classes as diverse as antihistamines (terfenadine), antibi- otics (erythromycin) (Makkar et al., 1993), and antiarrhythmic drugs (d,l- sotalol) (Kuhlkamp et al., 1997; Lehmann et al., 1996) can induce a poten- tially lethal cardiac rhythm called torsades de pointes. The risk of having this drug-induced complication is far greater in females than in males. Hypokalemia, hypomagnesemia, bradycardia, and the baseline QT1 inter- 1 The QT interval on an electrocardiogram is the duration of activation and recovery of the ventricular muscle. Because QT interval varies inversely with heart rate the corrected QT, or QTc, is often used.
SEX AFFECTS HEALTH 129 vat and the degree of QT prolongation (electrocardiographic variables) increase ones susceptibility to this event (Napolitano et al., 1994~. As fe- males possess a longer average electrocardiographic QT interval-corrected heart rate (QTc) than males, a difference not found in the newborn (Stramba-Badiale et al., 1995), sex differences in cardiac ion channel func- tion account for the increased incidence in females (Rautaharju et al., 1992~. In clinical studies it was found that quinidine, a drug that induces QTc interval prolongation, also induces a greater prolongation of the QTC interval in females (Benson et al., 2000~. Furthermore, in animal models, estrogen has been found to prolong the QTC interval, affecting cardiac repolarization (Drici et al., 1996~. Recent studies with women suggest that drug-induced QT interval prolongation may be affected by the phase of the menstrual cycle (Rodriguez et al., 2001~. Studies of isolated ventricular myocytes indicate that active drugs block the delayed rectifier potassium channel (Ik) (Wesche et al., 2000~. Adjustment of dosage for body weight or surface area does not amelio- rate this problem. Recently, the U.S. Food and Drug Administration (FDA) has required label warnings for drugs that prolong the QTc interval. One drug with this effect, cisapride, was removed from the market. Although the prolongation of the QTC interval by the d-sotalol enanti- omer is said to be independent of dose and sex (Salazar et al., 1997), the small population size of the study (four males and four females, which is not sufficient to be able to draw general conclusions) exemplifies the inadequacy of the available literature on this topic. Genes associated with the long QT syndrome and sudden death have mutations that affect ionic currents involved in the control of ventricular rhythm. Only those subjects with a mutation affecting the sodium-chan- nel ionic current have sufficient prolongation of the nighttime QT interval to increase arrhythmic risk (Stramba-Badiale et al., 2000~. The sex differ- ence in QTC interval observed in adults is not seen on the 4th day of life (Stramba-Badiale et al., 1995~. Other sex-related differences in side effects have also been noted. For example, females are twice as likely as males to develop a cough while taking angiotensin-converting enzyme inhibitors (Kubota et al., 1996; Strocchi et al., 1992~. Sex Differences in Effectiveness Sex differences in the therapeutic response to the 5HT3 antagonist alosetron have been reported. FDA has recently approved alosetron for the treatment of nonconstipated irritable bowel syndrome in females. It is ineffective in males, but the reasons for this are unknown (Bardhan et al., 2000; Camilleri et al., 2000~. Another example of sex differences in the therapeutic response to a compound that acts by a serotonergic mecha-
130 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH nism is sertraline. FDA recently approved sertraline for the treatment of posttraumatic stress disorder in women. Two multicenter, placebo-con- trolled trials for the treatment of posttraumatic stress disorder demon- strated its effectiveness in females but not in males (Henney, 2000~. Differences in the rates of serotonin synthesis in the brains of female and male volunteers have been measured by positron emission tomogra- phy. This technique allows a direct measurement of the rate of serotonin synthesis in the living brain. The rate of brain serotonin synthesis was found to be 52 percent greater in male subjects than female subjects. The researchers suggest that this marked difference in rates of serotonin syn- thesis could contribute to the higher incidence in women of major unipo- lar depression and other psychopharmacologies in which serotonergic mechanisms are implicated in the pathophysiology of the disease. How- ever, a few postmortem studies found no significant sex differences in serotonin levels (Arato et al., 1991; Dean et al., 1995; Nishizawa et al., 1997~. METABOLISM, LIFESTYLE, AND PHYSICAL PERFORMANCE Male-female differences are very striking in terms of body size and composition (Bjorntorp, 1989; Legato, 1997; National Center for Health Statistics, 1987~. These sex differences are closely linked to reproductive variables (Bjorntorp, 1989; Legato, 1997~. In addition, these differences can result from a combination of bio- logical and social (lifestyle) differences, which are then manifest in vari- able rates of illness, for example, obesity and osteoporosis between males and females. The relative influence of biological factors on the develop- ment and prognosis for some diseases, such as melanoma, are sometimes difficult to sort out because of the different social practices of men and women, for example, in regard to styles of clothing, levels of sun expo- sure, and use of sunscreens. As another example, the reduced rates of osteoporosis in males are likely a combination of basic molecular and hormonal differences com- bined with a greater tendency for men to engage in physical activity and weight-bearing exercise (Damien et al., 2000~. Nevertheless, there is a limited understanding of the underlying mechanisms of these apparently fundamental sex differences and their influence on aspects of health and functioning other than reproduction (Legato, 1997~. A better understanding of how sex differences relate to physical performance variables such as strength, endurance, and overall work capacity is needed to address the question of when differential treat- ment by sex is justified.
SEX AFFECTS HEALTH 131 Differences in Body Composition Differences in body composition are mediated by sex hormones and sex differences in behavior and tend to diminish after the fertile adult years. It is thought that these differences emerge around puberty; how- ever, sex differences in body composition have also been observed well before the onset of puberty (Taylor et al., 1997~. Body fat is an important component of energy balance, and there is clearly a relationship between systems of energy regulation and reproduction (including puberty and menarche) in various animal species and humans, but this relationship is not as simplistic as was once thought (Caprio et al., 2001; Schneider et al., 2000; also discussed in Chapter 3~. Mammary fat and gluteal-femoral fat are preserved in females under conditions of starvation and are preferen- tially mobilized during lactation. Food intake varies during the menstrual cycle, and hormonal changes during pregnancy may increase a female's appetite. Pregnancy causes weight gain over and above that from the fetus and placenta, and a return to the prepartum weight requires weight loss. Menopause is associated with a shift toward relatively more fat as well as toward the deposition of more fat in the abdominal region (Poehlman and Tchernof, 1998~. Adult males are, on average, taller than adult females and generally have a greater proportion of muscle and a lower percentage of body fat than females of the same weight and height. The greater muscle mass of males is associated with greater physical strength. Males also typically have relatively more abdominal fat and less gluteal-femoral fat than fe- males. Energy Metabolism and Body Composition Foods and beverages contain substances that can be oxidized to en- ergy (heat, measured in calories) as fuel for vital processes and physical work; these substances include fat (~9 kilocalories per gram), carbohy- drate (~4 kilocalories per gram), and protein (~4 kilocalories per gram) (Goran, 2000~. The ethanol in alcoholic beverages provides calories (~7 kilocalories per gram) but is also a toxin. Individual variation in the effi- ciency of energy utilization (the amount of heat produced per unit of food energy consumed) is in part genetically determined but can be influenced by behavioral factors such as physical training, cigarette smoking, and alcohol use (Collins et al., 1994; Goran, 2000; Suter et al., 1994~. Body weight reflects the weights of bone, visceral organs, skeletal muscle, adipose tissue (fat, both intra-abdominal and subcutaneous), blood, and other body fluids (Heymsfield et al., 1997~. The adipose tissue component has the greatest variability. Daily energy requirements de- pend primarily on basal metabolic needs, that is, the amount of energy
132 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH needed to support basic functions, such as during sleep or at rest. The basal or resting energy expenditure is determined primarily by body size and body composition (i.e., the proportions of fat and lean tissue). Energy expenditure during energy utilization is higher in muscle than in fat tis- sue. Thus, at the same weight and height, a fatter individual expends less energy than one with relatively more lean tissue. Other contributors to total energy expenditure include the small amount of energy needed to metabolize food and the variable amount of energy expended during physical activity. Energy intake in excess of energy requirements is stored preferentially and primarily as fat (Figure 5-2~. Sex differences in energy requirements derive from sex differences in body size, body composition, and activity levels (Goran, 2000~. Larger individuals (e.g., males) have higher energy needs. Estimated energy re- quirements for males and females are similar before puberty (i.e., before the occurrence of sex hormone-driven changes in body size and fatness) (National Research Council, 1989~. After puberty, energy requirements are higher for males. Compared with a man of the same weight and height, a woman has less lean tissue and, therefore, a lower basal metabo- lism and lower energy expenditure per unit of work. Obesity The interest in obesity has increased in recent years because of an alarming increase in the prevalence of obesity in both adults and children (Flegal et al., 1998; Mokdad et al., 1999; Troiano and Flegal, 1998) and because type II diabetes (adult-onset, non-insulin-dependent diabetes BODY FAT | Energy | Intake Human organism: The Biological and Behavioral Interface 1 Energy | | Expenditure | FIGURE 5-2 The major affecters of body fat. Source: Bouchard (1992). Reprint- ed, with permission, from C. Bouchard. 1992. Genetic aspects of human obesity. In: Obesity. P. Bjorntorp and B. N. Brodoff, eds. Philadelphia: J. B. Lippincott Company. Copyright 1992 by l. B. Lippincott Company).
SEX AFFECTS HEALTH 133 mellitus), as one clear manifestation of the associated metabolic aberra- tions, is also increasing in parallel in both adults and children (Mokdad et al., 2000; Rosenbloom et al., 1999~. Simultaneously, advances in genetics and molecular biology have begun to provide new insights into mecha- nisms of appetite regulation, energy metabolism, and obesity causation, with the promise of effective pharmacological approaches to the treat- ment of obesity in the foreseeable future. Taken together, these develop- ments increase both the urgency of and the potential for the identification of the biological factors that underlie obesity and that interact with the plethora of identified sociocultural and behavioral contributors to obesity (World Health Organization, 1998a). The following questions remain unresolved, however. · Does the biological predisposition of women to develop fat stores also predispose women to become obese under conditions of high levels of food availability? · Are sex-specific approaches needed in obesity treatment? · Are the systems of energy regulation in females more primed for weight gain or weight retention relative to those in males? If so, under what circumstances is this tendency expressed or suppressed? · Given that many obesity-related health risks are specifically linked to abdominal fat, is a given level of overall obesity less problematic for females than for males? · Can the study of sex differences in energy balance facilitate under- standing of the etiology of obesity? These questions become particularly pressing in light of the current epi- demic of obesity within both the United States and worldwide (Mokdad et al., 1999; World Health Organization, 1998a). The development of obesity probably reflects a natural response to an overabundance of food energy and limitations of requirements for physi- cal activity (Hill and Peters, 1998; lames, 1995), conditions that overwhelm the physiological capability to maintain energy balance. There are, never- theless, numerous scientific questions related to variations in individual susceptibility to these conditions, physiological aberrations that may result from a chronic overabundance of food and physical activity limita- tions (e.g., changes in set points for energy regulation), and ways to lever- age compensatory responses through genetic or drug therapy. Sex differ- ences in obesity, along with age and ethnic variations in overall levels of obesity and in the sex ratio of obesity, provide interesting leads that can be used to address these questions. Commonly used definitions of obesity are based on relative weight for height but do not account for either fatness or fat patterning. These definitions may therefore overestimate fatness in males relative to that in
34 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH females. To the extent that obesity-related health risks are tied to abdomi- nal obesity (Bjorntorp, 1996), definitions of obesity based only on body size will overestimate the health risks associated with obesity in females (who have relatively less abdominal fat) in comparison the health risks associated with obesity in males (Laws et al., 1997~. Obesity prevalence data also indicate that the tendency of females to be more obese than males differs according to ethnic, socioeconomic, and environmental cir- cumstances (Table 5-4~. This makes the question of sex differences in obesity more complicated; that is, what is the underlying predisposition to obesity in females, and under what circumstances is it expressed or not expressed? Possible explanations for a greater prevalence of obesity in females include the following: (1) overconsumption, that is, eating behaviors that predispose females to consume too much food in relation to energy needs, possibly including physiologically determined disorders of appetite regu- lation; (2) metabolic efficiency, for example, physiological factors that predispose females to store relatively more consumed energy at any given level of intake; (3) low energy expenditure, that is, possible behavioral sex differences in the ability to offset energy intake through routine or leisure time physical activity; and (4) less success in voluntary weight control because of either behavioral of physiological factors. These explanations are complementary and may combine to increase the predisposition of females to a positive energy balance. From an epide- miological perspective, exploration of these possibilities might include comparisons by ethnicity and socioeconomic status to determine whether one or more are particularly applicable or inapplicable to subgroups of females. For example, assuming that the underlying biological sex differ- ences in obesity determinants are relatively similar, the higher level of obesity in non-Hispanic African-American females compared with that in non-Hispanic white females might be due to higher levels of occurrence of the behavioral risk factors implied above in the African-American fe- male population (overeating, low levels of activity, less effective weight control) (Kumanyika, 1998~. Sex differences in energy metabolism are not observed in all species, but when differences are observed, females are more obese (Hoyenga and Hoyenga, 1982~. A convincing scenario can be constructed in which dif- ferential evolutionary pressures on males and females would lead fe- males to be smaller (i.e., shorter) than males and females to have less muscle but to be more efficient users of energy (Hoyenga and Hoyenga, 1982), as follows. In mammals, sex differences in body size appear to occur in relation to the degree of differentiation in reproductive roles. Larger size confers a greater advantage with respect to the reproductive roles usually held by males (e.g., competition for territory, defense of the group, and competition for females). The greater muscle mass in human
SEX AFFECTS HEALTH TABLE 5-4 Obesity Prevalence Data for Selected U.S. Adults, by Sex 135 Percent Group Males Females Female:Male Non-Hispanic whites, 1988-94, age 220a Non-Hispanic whites, 1988-91, age 225, <grade 12 educations Non-Hispanic whites, 1988-91, age 225, grade 12 educations Non-Hispanic whites, 1988-91, age 225, >grade 12 educations Non-Hispanic African Americans, 1988-94 age 220a Non-Hispanic African Americans, 1988-91 age 225, <grade 12 educations Non-Hispanic African Americans, 1988-91 age 225, grade 12 educations Non-Hispanic African Americans, 1988-91, age 225, >grade 12 educations Mexican Americans, 1988-94, age 220 b Mexican Americans, 1988-91, age 225, <grade 12 educations Mexican Americans, 1988-91, age 225, grade 12 educations Mexican Americans, 1988-91, age 225, >12 grade education b Puerto Ricans, 1982-84, age 20-74b Cuban Americans 1982-84, age 20-74b American Indians in OK, 1994-96b (questionnaire) American Indians in NM and AZ, 1994-96b (questionnaire) American Indians in WA and OR, 1994-96b (questionnaire) American Indians in ND and SD, 1994-96b (questionnaire) 46 Alaska Natives, 1996b (questionnaire) 41 Asian Americans, ages 18-59C (questionnaire) 57 Samoans in Manu'a 56 Samoans in Oahub 75 20 39 36 30 21 29 29 36 21 39 43 40 25 29 38 33 33 23 39 38 30 37 55 51 49 33 53 40 1.2 1.0 1.1 1.0 1.8 1.8 1.8 1.4 1.6 l.4 0.9 47 37 1.5 34 36 34 43 47 30 38 77 80 1.2 1.2 0.9 1.3 1.1 a Obesity was defined as a body mass index of 230 kilograms per square meter. b Obesitv was defined as a body mass index of >27.8 kilograms ner sauare meter for J J (A 1 1 males and a body mass index of 2 27.3 kilograms per square meter for females. c Obesity was defined as a body mass index of 225 kilograms per square meter. SOURCES: Lauderdale and Rathouz (2000~; National Heart, Lung, and Blood Institute, National Institutes of Health (1998~; and Will et al. (1999~.
136 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH males would be consistent with this. Greater metabolic efficiency would confer resistance to starvation in times of limited food supplies or of cycles of feast and famine. This would be more advantageous for the females of species in which the main role of females is not only to bear the offspring but also to rear them until they are independent and would particularly apply in cases of a long period of dependency of the off- spring. Low levels of heat production per calorie consumed would pro- mote starvation resistance but would predispose an individual to store fat during periods with increased food supplies. The link between reproduc- tive function and energy balance implies that progesterone and possibly estrogen affect processes related to food intake, weight gain, heat produc- tion, and heat loss. Some such effects can apparently be identified in the perinatal period (Hoyenga and Hoyenga, 1982~. Current approaches to the study of sex differences in energy metabo- lism and obesity are compatible with the general view that there are physiologically mediated differences in energy regulation, storage, and utilization. Many differences appear to be mediated by male-female dif- ferences in fat patterning, for example, the amounts, types, and metabolic characteristics of fat in the abdominal and gluteal-femoral regions (Bjorntorp, 1989~. Mechanisms under study include hormonal influences on appetite regulation (Hassink et al., 1996; Kennedy et al., 1997; Roca et al., 1999), energy expenditure (Nicklas et al., 1997), gastric emptying time (Gryback et al., 1996), resting metabolic rate (Carpenter et al., 1998; Weyer et al., 1999), use of fatty acids as energy sources and rate of mobilization of fatty acids (Fletchner-Mors et al., 1999; Laws et al., 1997; Lonnqvist et al., 1997; Sumner et al., 1999), and changes in body composition and en- ergy metabolism associated with menopause (Poelhman and Tchernof, 1998) or puberty (Molnar and Schulz, 1997~. Females are less likely to mobilize fat from certain adipose tissue stores (Laws et al., 1997; Lonnqvist et al., 1997; Sumner et al., 1999), al- though this varies with age and ethnicity. The gluteal-femoral fat depots in females are much larger than those in males and have higher lipopro- tein lipase activity, which promotes the uptake of lipid and greater oc- adrenergic activity than p-adrenergic activity, which promotes the reten- tion of fat in the cells. Changes in the lipid-accumulating patterns of cells in this fat depot are observed with pregnancy (increased) or menopause (decreased), consistent with the theory that this aspect of female fat accu- mulation and fat metabolism has a reproduction-related function (Bjorntorp, 1989~. In addition, within the abdominal area, females demon- strate more lipolytic activity in subcutaneous fat, whereas visceral fat is more active in males. These differences do not apply after menopause. The finding that age- and sex-related differences are mediated primar- ily by changes in the energy density rather than the volume of foods consumed (Marti-Henneberg et al., 1999) implies sex differences in appe-
SEX AFFECTS HEALTH 137 tile regulation. Studies of sex differences in levels of circulating leptin a hormone involved in appetite regulation also suggest sex differences in appetite regulation. Higher leptin levels occur in females at all levels of body weight, with a steeper gradient of increase in leptin levels occurring with obesity in females. However, an increase in abdominal fat levels and insulin resistance associated with leptin levels was observed only in males. Leptin levels were always higher in obese girls than in obese boys. These studies suggest that girls are relatively resistant to the appetite-suppress- ing effects of leptin. The appetite-regulating effects of serotonin (5-OH- tryptophan) may also differ between the sexes (Roca et al., 1999~. Physical Performance Studies of physical performance and energy metabolism during ex- ercise and physical work suggest intriguing sex differences that may be qualitatively different from the sex differences in energy regulation that are observed under resting conditions (Bjorntorp, 1989~. From an evolu- tionary perspective, in keeping with sex role differentiation, males may be more adapted than females for brief spurts of intense energy expendi- ture, whereas females may be more adapted than males for sustained but less intense energy expenditure (Hoyenga and Hoyenga, 1982~. For ex- ample, under conditions of moderate exercise, females preferentially uti- lize fatty acids, sparing muscle glycogen reserves and permitting sus- tained performance (Bjorntorp, 1989; Tale and Holtz,1998~. This difference could, however, be related to physical training rather than underlying sex differences. In the same vein, sex differences in the ratios of different types of muscle fiber are consistent with a greater potential for oxidative metabolism in the muscles of females. Again, however, these differences may reflect sex differences in muscle morphology that are influenced by behavior (through differences in activity and exercise habits) instead of intrinsic differences. In addition, some part of the sexual dimorphism in muscle fiber composition appears to be mediated by sex hormones; for example, it is not observed after menopause (Bjorntorp, 1989~. The Institute of Medicine Committee on Military Nutrition Research (1992, 1998) described some major policy implications potentially associ- ated with sex differences in body weight and composition and their rela- tionship to physical performance. With respect to the armed services, a conflict was identified between the standards of body composition re- quired for women to achieve an appearance goal and those necessary for performance of many types of military tasks (Institute of Medicine, 1992~. Specifically, among heavy women, greater lean body mass and upper body strength confer advantages for physical fitness and endurance, but higher weights are considered a disadvantage with respect to appearance (that is, the proper "military bearing," which is associated with leanness).
138 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH The Committee on Military Nutrition Research observed that both the standards for percent body fat used for eligibility and the methods used for assessment of body composition and evaluation of physical per- formance varied considerably among the different branches of the mili- tary, suggesting a need for evidence on which to base a sound consensus. The 1998 Committee on Military Nutrition Research (Institute of Medi- cine, 1998) cited the earlier recommendation of the 1992 Committee (In- stitute of Medicine, 1992) that "body composition standards be based on considerations of task performance and health and be validated with regard to the ethnic diversity of the military" (Institute of Medicine, 1998, p. 2~. The need to improve the scientific basis for sex-specific criteria for body size and composition for military performance is evident (Institute of Medicine, 1998~. The fundamental questions are whether and under what circumstances men and women can be held to similar performance standards and what health or reproductive benefits or risks accrue to women under these circumstances. Bone Metabolism and Osteoporosis Osteoporosis is a disorder of low bone mass, microarchitectural de- generation, and bone weakness that leads to fracture. The most common sites of osteoporotic fracture are the thoracic vertebrae and femoral neck (hip). Fifteen percent of all Caucasian women in the United States and 35 percent of women in the United States over age 65 have osteoporosis; one of every two Caucasian women will suffer an osteoporotic fracture in her lifetime (American Journal of Medicine, 1993), a lifetime risk similar to the combined risk of developing breast, endometrial, and ovarian cancer (Holbrook et al., 1985~. The rate of mortality within the first year after fracture is 15 to 20 percent; less than one-third of fracture patients return to their prior functional status. The bones of humans reach their peak mass in the third or fourth decade of life. Thereafter, men lose bone density at a slow, steady pace (0.3 to 0.5 percent per year). Women lose bone density at this same pace until menopause, at which time they lose 2 to 3 percent per year for approximately 10 years and then resume a rate of loss comparable to that in men (Khosla et al., 1999~. Osteoporosis results when the rate of bone resorption by osteoclasts outstrips the rate of bone formation by osteo- blasts (Manolagas, 2000~. Bone is composed of a honeycomb-like structure (trabecular bone in vertebral bodies and the femoral neck) and dense bone (cortical bone in the outside tube-like structures of long bones). Trabecular bone has a greater surface area per gram of mineral and is more likely to be subjected to osteoclastic degradation than cortical bone. Trabecular struts, like the
SEX AFFECTS HEALTH 139 architectural cross-struts that support bridges, give strength to bone. In osteoporosis, these struts erode through, weakening bone strength dis- proportionately to the amount of calcium loss. Many factors contribute to bone health. Regarding the whole organ- ism, both body size and frailty are important factors in clinical osteoporo- sis: large people have large bones and are less likely than small people to suffer a fracture. Heavy people have subcutaneous fat to absorb trauma in falls and are therefore less likely than thin people to suffer a fracture if they fall. Old people have relatively slow reflexes and frequent gait dis- turbances and are thus more likely than young people to fall. Behavioral habits contribute to bone strength: weight-bearing exer- cise creates strong bone structure; loss of weight bearing, for example, during space travel or bed rest, weakens bone (Turner, 1999~. Extreme exercise resulting in amenorrhea causes bone loss (Drinkwater et al., 1984; Hobart and Smucker, 2000~. Caloric repletion, however, restores the menses and protects the bones (Warren and Stiehl, 1999~. Smokers are at high risk for the development of osteoporosis (Kato et al., 2000~. Hormones are also important to bones. Both estrogen and testoster- one are critical for achievement and maintenance of peak bone mass. A deficiency of either hormone in both men and women or a deficiency of growth hormone reduces bone mass (Braidman et al., 2000~. Rapid bone turnover can result in osteoporosis (Garnero et al., 1996; Ross and Knowlton, 1998~. Bone resorption follows a circadian cycle, being maxi- mal at night. Among the various life events that affect bones, pregnancy and lacta- tion divert calcium from the mother's bones to the baby; if the mother does not replete her calcium, she will suffer a net bone loss (Black et al., 2000; Horst et al., 1997~. Genetic factors also affect bones. In controversial studies, different alleles of the vitamin D receptor make individuals' susceptibilities to os- teoporosis different. Mutations of the collagen I gene, as in osteogenesis imperfects, lead to extremely weak bones (Heegaard et al., 2000: McGuigan et al., 2000~. Various environmental factors affect bones. Dietary calcium and vita- min D (from diet or sun exposure) are required for bone formation, and calcium and vitamin D deficiencies during childhood cause rickets. Cal- cium requirements differ by age and sex: infants should obtain 400 to 600 mg of calcium per day, children should obtain 800 mg per day, adoles- cents should obtain 1,200 to 1,500 mg per day, premenopausal women should obtain 1,000 mg per day, and postmenopausal women should obtain 1,000 to 1,500 mg per day. Men under age 65 should obtain 1,000 mg of calcium per day, and men over age 65 should obtain 1,500 mg per day. Many pharmaceutical agents cause bone loss. Corticosteroids, for in-
140 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH stance, hinder the actions of osteoblasts, promote renal calcium loss, and inhibit the absorption of dietary calcium. Other drugs that cause bone loss are heparin, thyroxin, and several anticonvulsants, while diuretics pre- vent calcium loss. The major risk factors for osteoporosis are positive family history, weight less than 127 pounds and current tobacco use. Lesser risk factors are white race, female sex, age more than 65 years, postmenopausal sta- tus, low levels of calcium intake, alcoholism, sedentary life style, and chronic illness (Huopio et al., 2000~. Bone strength thus has genetic, hor- monal, life stage, life event, behavioral, and environmental components. Those factors that contribute most to the predominance of osteoporosis among women are estrogen loss at menopause, lower levels of exercise among women, lower levels of sun exposure among women, and preg- nancy. Exposures and Different Patterns of Melanoma Occurrence and Survival The different rates of melanoma between men and women serve as an example of the complex interaction of biology, exposures, and social fac- tors in the manifestation and progression of disease. Melanoma, a malignancy of melanocytes (pigment cells), which are primarily found in the skin, is a public health concern for several reasons: · It is largely preventable; two-thirds of cases are attributable to sun exposure (Gilchrest et al., 1999~. · Most deaths should be avoidable, as most lesions are easily recog- nized early on, when excision is still curative (Piepkorn, 2000~. · The incidence of melanoma and the rate of mortality from mela- noma are increasing more rapidly than those for almost any other malig- nancy (Koh et al., 1995; National Cancer Institute, 1986; Piepkorn, 2000~. · It is the most common fatal malignancy in young adults (Kosary et al., 1996~. The incidence of melanoma is slightly higher among women than men, but the rate of mortality from melanoma is higher among men (Stidham et al., 1994; Tsao et al., 1998~. Both sex and gender differences appear to play important roles in the risk for melanoma. Most melanomas can be attributed to ultraviolet (UV) light-induced mutations in key regulatory genes in melanocytes; UV light-induced cu- taneous immune suppression may also contribute to melanoma (Piepkorn, 2000~. Risk factors include fair skin, red or blonde hair, easy sunburning (specifically, a history of sunburns during childhood), a large number of nevi (benign proliferations of pigment cells, commonly called "moles"),
SEX AFFECTS HEALTH 141 and a family history of melanoma (Piepkorn, 2000~. Melanomas are statis- tically associated with intermittent sun exposure and occur on body sites intermittently exposed to the sun (Bentham and Aase, 1996; Nelemans et al., 1993~. In men the midback is the most common site, and in women the posterior calves are the most common site (Piepkorn, 2000~. Because mel- anocytes are present in approximately equal numbers and have the same distribution over the body in both sexes, it is difficult to invoke a biologi- cal explanation for these differences in the locations of occurrence of mela- noma; instead, the differences are attributed to clothing styles. The star- tling increase in the incidence of melanoma over the past 70 years, from a 1 in 1,500 to a 1 in 75 lifetime risk for Americans (Koh et al., 1995; Piepkorn, 2000), may reflect the increasing popularity of revealing bathing suits and other casual attire. The possibility that hormones influence melanoma has long been a subject of debate. The darkness of skin pigmentation at some body sites is affected by estrogenic hormones (Abdel-Malek, 1998~. For example, preg- nancy is associated with melasma (a reticulated facial pigmentation also known as the "mask of pregnancy"), the linea nigra (hyperpigmentation extending from the umbilicus to the pubic area along the midline), and darkening of the areolae (Abdel-Malek, 1998~. Several studies suggest an adverse effect of pregnancy on the prognosis of melanoma (Piepkorn, 2000), and melanoma cells have been reported to express estrogen recep- tors (Piepkorn, 2000~. Melanoma is also influenced by immune factors. For example, melanoma is responsive to adjuvant immunotherapy (Piepkorn, 2000~. Altered immune status during pregnancy may contrib- ute to the spread of melanoma. Men age 50 or older have a striking excess rate of mortality due to melanoma compared with that for women (Tsao et al., 1998~. It is unclear whether this reflects biological or behavioral differences. It is known, however, that most melanomas are first suspected by women, whether the lesions are on themselves or on their spouses, and that women then arrange for physician examination (Koh et al., 1992~. Such observations suggest ways in which both sex and gender figure in the incidence and prognosis of melanoma. Sun exposure histories be- tween men and women differ because of occupational and recreational sun exposures, clothing styles, and willingness to apply sunscreens. These gender-specific issues may influence who develops melanoma and at what site. Gender-specific influences can then be compounded by sex- and gender-neutral biological factors. For example, midback lesions (most common in men) carry a worse prognosis than lesions on other body sites (Piepkorn, 2000~. Once a melanoma has developed, hormonal differences may influence the probability of disease spread; and differences in body awareness and social priorities may influence how quickly medical atten- tion is sought, with a large consequent influence on the prognosis. Thus,
42 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH a better understanding of the contributions of sex and gender differences to melanoma may have an enormous effect on the incidence and progno- sis of this devastating malignancy. SEX DIFFERENCES IN AUTOIMMUNE CONDITIONS Because certain rheumatic, hepatic, and thyroid autoimmune diseases are predominant in females but other autoimmune diseases are not, the fact of autoimmunity alone does not explain the sex differences in au- toimmune disease incidence. In humans, exposure and other extrinsic factors explain most sex differences in the incidence of infectious dis- eases. If infections induce autoimmune diseases, differences in exposure may likely explain the sex differences. Background Most mammals respond to infections with a combination of innate (inflammatory) and adaptive (immune responses (Medzhitov and laneway, 2000~. Infectious agents include viruses, priors, bacteria, myco- bacteria, fungi, and parasites. The innate response engulfs, walls off, and, when appropriate, kills the invader with toxic cell products. The innate immune response recruits and in part directs the cells of the adaptive immune response. Adaptive immunity recruits and engages memory cells and their products to assist inflammatory cells. Autoimmunity occurs when the adaptive immune system attacks normal tissue. Autoimmunity may result from a normal immune response to an invading organism gone awry or from the loss of normal immune regulation. The levels and types of models used to study infection and inflamma- tion are listed in Box 5-2. Each can be tested by use of various challenges: spontaneous or induced infection, vaccination, and response to common environmental stimuli. Spontaneous illness constitutes another form of test. Sex Differences in Types of Exposure and Portals of Entry Portals of entry into the intact body include the skin; eyes; mouth, gastrointestinal tract, and rectum; nasal passages and lungs; and, in fe- males, the vagina. The urethra is less commonly an entry point, but it may be important in venereal diseases. Exposure may occur through direct penetration, as from a knife wound or transmission of a parasite through an insect bite; indirect penetration, such as by radiation; inhalation of a gas, aerosol, or organism; percutaneous absorption (through the skin); ingestion; or absorption from a mucosal surface. Sex differences in types of exposure and portals of entry are rarely
SEX AFFECTS HEALTH 143 studied. Hypothetically, for behavioral and social reasons, males may experience more penetrating trauma and may inhale higher levels of toxic (industrial) materials. In general, a male's skin may be exposed to more (industrial) toxins, and female's skin may be exposed to more toxins in detergents and cosmetics. In the course of work or relaxation, males may be more likely to place more unusual materials (building nails, tobacco pipes) in their mouths. In addition, males and females may have different diets. Females have a mucosal surface (vagina) absent from males and encounter products and agents (tampons, semen, medical instruments, douches) through vaginal insertion that males do not. Furthermore, in females the cervical barrier between the internal and the external environ- ments is transiently broken during menstruation. Finally, sexual practices present different types of exposure for males and females. Normal Processes Innate and Adaptive Immunities This section reviews differences in innate and adaptive immunities between females and males. Gonadal hormones partly control these nor- mal defense systems. The literature on the nonhormonal effects of sex on mechanisms of innate and adaptive immunities is sparse, however. Adaptive immunity varies markedly by sex; innate immunity varies
44 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH less. Mature young females mount more vigorous immune responses than others. Whole-organism elements of inflammation and immunity include cycling, hormones, growth and nutrition, life stages, and life events (dis- cussed below). Of these, only hormones have been extensively studied. Chronobiological events occur over short intervals (e.g., brain waves and heart beat), days (sleep cycle), weeks (menstrual cycle), or years (sea- sonal cycles) (Young, 2000~. Male and female chronobiologies, that is, menstrual or estrous cycles, should thus be considered separately from the associated hormonal changes. Both sleep deprivation and jet lag, ex- amples of chronobiological events, are immunosuppressive (Ishida et al., 1999~. Hypothetically, a chronobiological effect on immunity may occur in menstruating females or may render a fertile female vulnerable at cer- tain times of the month. Studies that have compared exogenously cycled and noncycled castrated animals have not been done. Leukocytes and their products constitute the innate immune system. Leukocyte function, which is assessed by cellular synthesis of degradative enzymes (Kuslys et al., 1996), oxidative metabolism (Garcia-Duran et al., 1999), and adherence and phagocytosis (Ito et al., 1995; Tosefsson et al., 1992; Miller, 1999; Mondal and Rai, 1999; Spitzer, 1999; Spitzer and Zhang, 1996a,b; St. Pierre Schneider et al., 1999), is modulated by estrogen. Some estrogen-induced changes increase the levels of functioning of leukocytes in females; others decrease their level of functioning. Overall changes are small, and sex differences in levels of leukocyte functioning probably do not affect human illness. The adaptive immune response includes activation and suppression of T and B lymphocytes, macrophages, and dendritic cells; secretion of their cytokine products; production of immunoglobulin antibodies; and activation of the complement and coagulation systems. The adaptive immune response of females, as measured by determi- nation of the level of cell proliferation or immunoglobulin levels, is more vigorous than that of males. The sex differences are relatively small. Com- parable differences are seen between Caucasians and African Americans and between young and old individuals. Persons with chronic inflamma- tory illnesses have activated immune systems. The implication of the dif- ferences between males and females is unknown. As a rule, estrogenic hormones upregulate and androgenic hormones downregulate the cellular effecters of human adaptive immunity: lym- phocytes, macrophages, and dendritic cells (Ahmed and Talal, 1999; Kanda et al., 1999~. The adaptive immune response varies during the menstrual cycle. However, most experiments on immune cells examine specific questions (e.g., does estradiol upregulate expression of a certain substance without considering physiological age, menstrual cycle, or other
SEX AFFECTS HEALTH TABLE 5-5 Sex Differences in Immunocytes as Determined in Representative Recent Studies 145 Human Estrogen stimulates mononuclear cell nitric oxide synthase (Stefano et al., 1999) Estrogen stimulates macrophage inflammatory mediators (D'Agostino et al., 1999) Lymphocyte glutathione S-transferases are not correlated with sex (Van Lieshout et al., 1998) Stress and depression downregulate hypopituitary-pituitary-adrenal axis (Olff, 1999) CD4/CD8 ratios are higher in Guinean girls (Lisse et al., 1997) Estrogen decreases apoptosis of mononuclear cells of menstruating women (Evans et al., 1997) Estrogen inhibits monocyte interleukin 1 production (Morishita et al., 1999) Animal Estrogen downregulates monocyte adhesion and migration in females (rabbits) (Nathan et al., 1999) Estrogen receptor is nonfunctional in male thymocytes (mice) (Kohen et al., 1998) Lymph nodes from myelin basic protein-immunized male mice are less encephalitogenic than those from female mice (Kim and Voskuhl, 1999) Intracerebroventricular interleukin 1 beta upregulates progesterone and prolactin in females but not males (rats) (Turnbull and Rivier, 1995) Mononuclear cell binding to atheromas is lower in hypercholesterolemic females (rabbits) (Holm et al., 1998) Estrogen upregulates Bc1-2 and blocks tolerance (mice) (Bynoe et al., 2000) variables (Lockshin, 1999~. Reviews are readily available (Cannon and St. Pierre, 1997; Cutolo et al., 1995; Draca, 1995; Gaillard and Spinedi, 1998; Kammer and Tsokos, 1998; Marchetti et al., 1998; Wilder, 1995~. Represen- tative recent data are displayed in Table 5-5. The hypothalamic-pituitary-adrenal-gonadal axis, which exerts im- portant control on the adaptive immune system, differs between males and females. Exercise, stress, and depression all downregulate immune function (Irwin, 1999; Nehlsen-Cannarella et al., 1997~. Since each of these differs between the sexes, sex differences in the resultant illness may occur. Immunization Sex-specific studies of responses to immunizations with vaccines show intriguing serological differences (differences in circulating anti- body levels) but few clinical differences between females and males (Table 5-6~. Sex differences in serological response are not generalizable among organisms. Adverse systemic reactions to immunization, particularly ar- thritis, are more common in females.
146 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH TABLE 5-6 Sex Differences in Immunization Reference Disease Finding Trollfors (1997) Singh and Datta (1997) Singh et al. (1999) Pertussis Measles Measles Atabani et al. (2000), Measles Forthal et al. (1995) Sankilampi et al. (1997) Nichol et al. (1996) Influenza Govaert et al. (1993) Influenza Cardell et al. (1999), Hepatitis B Havlichek et al. (1997) Chiaramonte et al. (1996) Hepatitis B Chen et al. (1997) Hepatitis A Benjamin et al. (1992) Rubella Rebiere and Galy-Eyraud Mumps (1995) Pneumococcal infection No sex difference No sex difference in antibody response Girls have higher case fatality rates (because of a lack of immunization) Antibody-dependent cellular cytotoxicity is lower in females, but neutralizing antibody levels are equal in both sexes Elderly women have lower antibody levels More systemic adverse reactions in females More local adverse reactions in females . Seroconverslon rates are equivalent in males and females Antibody titers are higher in young women Antibody titers are higher in females Arthritis is 3.5 times more . · . common in girls No sex difference in risk of developing vaccination . . . menmglUs Pregnancy Estradiol levels increase 100-fold and estriol levels increase 1,000-fold during human pregnancy. Within these ranges, estrogens upregulate im- mune functions, but clinically evident immunological changes during pregnancy are small. Pregnancy-specific proteins suppress lymphocyte function. Changes differ at different stages of pregnancy, with no appar- ent general pattern. Cutaneous and humoral immune responses to spe- cific microbial antigens are selectively depressed, as are leukocyte chemo- taxis and adhesion. Overall, during pregnancy the immune system deviates markedly from that during the nonpregnant state; the long-term effects, if any, of this deviation on women's biology or health are un- known. Specific infections, such as those caused by the measles virus, appear to be particularly virulent in pregnant women.
SEX AFFECTS HEALTH 147 The striking sex difference in many autoimmune disorders is incom- pletely understood. A new aspect under study is the role of fetal cells transferred to the maternal blood (microchimerism) (Bianchi, 2000) and their persistence postpartum, in some instances for decades, in the patho- genesis of scleroderma (also called systemic sclerosis, a connective tissue disorder that leads to fibrosis of skin and internal organs). In women with scleroderma who had given birth to at least one son before disease onset, male DNA was detected more often and in larger amounts in their blood than in their normal sisters who had given birth to one or more sons (Artlett et al., 1998; Nelson, 1998; Nelson et al., 1998~. Persistent micro- chimerism of maternal lymphocytes in the circulation of offspring occurs as well, but its relationship, if any, to autoimmune disorders, is not yet known. The implications of the association between autoimmune disor- ders and fetal microchimerism are unknown. The findings, however, do indicate a profound biological difference between men and women that may be relevant to sex ratios of disease. Abnormal Processes Infectious Disease General Principles Males do not differ from females in terms of their responses to infections, regardless of whether the invading organism is a virus, bacterium, mycobacterium, or parasite. In experiments with ani- mals in controlled settings, males are more susceptible to parasites, fungi, bacteria, and viruses (Klein, 2000), probably because of hormonal effects. Most sex differences in humans, however, are caused by differences in exposures (a societal level effect) instead of differences between males and females at the individual, organ, or cell level. The following are ex- amples of sex differences caused by different exposures: · Exposure. Sex differences in attack rates occur with kuru (a disease caused by a prior), in which, for cultural reasons, only females eat the infected tissue; tuberculosis, when it is contracted in prison or homeless shelters; and infection with human immunodeficiency virus (HIV) when it is transmitted by male homosexual intercourse or needle sharing. · Portal of entry. Differences in genitourinary anatomies and local immune responses cause different clinical phenotypes of gonorrhea and herpes genitalia. · Organism load. Sex differences in organism loads are especially rel- evant in HIV infection; in part, these differences are also a function of portal of entry. · Receptors. In experimental coxsackievirus myocarditis, male mouse hearts contain more viral receptors and receive higher viral loads.
148 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH TABLE 5-7 Sex Differences in Viral Infections in Humans Reference Disease Finding Suligoi (1997) HIV infection Swanson et al. (1995) Sullivan et al. (1999) Herpes Rubella Benn et al. (1997) Measles Dollimore et al. (1997) Measles Females are at higher risk of infection, but there is no difference in the rate of progression to AIDS Males have higher-risk behavior Male/female incidence ratio = 2, likely because of lower immunization rates for boys Vitamin A affects the antibody concentration in males more than in females Same incidence in boys and girls, higher fatality rate in girls · Social response. Clinical acknowledgment of illness, particularly in the face of stigma, may differ between males and females, as may compli- ance with therapy. Tuberculosis and leprosy in developing countries are examples. · Response to therapy. Sex differences in response to therapy are dis- cussed below in terms of drug metabolism. Virus In humans, sex differences in viral illnesses are primarily due to differences in behavior, such as vaccination rates or exposure (Table 5-7~. The apparent higher rate of fatality from measles in girls is not explained, and the different effects of vitamin A on antibody titers in males and females are controversial. In animals, sex differences of virus effects occur in both directions. Bacteria Bacterial diseases affect males and females approximately equally. Even in the conversion of acute to chronic Lyme disease (an illness that closely resembles autoimmune disease), the incidence and severity in males and females are similar (Pena and Strickland, 1999~. Mycobacteria, Fungi, and Parasites Males animals are slightly more susceptible to infection with mycobacteria, fungi, and parasites (Klein, 2000~. In humans, sex-specific rates of infection with mycobacteria, fungi, and parasites are approximately equal; most differences can be explained by differences in exposures (Table 5-8~. Leprosy and Chagas' disease induce lupus autoantibodies and tuber- culosis induces rheumatoid factor, but these diseases do not induce clini- cal autoimmune disease in humans. Data are scant, but infected males
SEX AFFECTS HEALTH TABLE 5-8 Sex Differences in Mycobacterial, Fungal, and Parasitic Diseases 149 Reference Disease Finding Holmes et al. (1998) Tuberculosis Hudelson (1996) Tuberculosis Freire et al. (1998) Leprosy Disease rates are higher in males; young women progress from infection to disease more frequently Females have more exposure risk, are less compliant with therapy Ninety-four percent of antineutrophil cytoplasmic antibody-positive patients are male Rao et al. (1996) Leprosy Longer delay in identifying skin changes in women Corredor Arjona Trypanosomiasis, No sex differences in positivity by enzyme- et al. (1999) Chagas' disease linked immunosorbent assay Albarracin-Veizaga Chagas' disease Females are more frequently seropositive et al. (1999) Michael et al. (1996) Filariasis More prevalent in males Rogier et al. (1999) Malaria Incidences are the same in boys and girls and females do not differ in their autoimmune responses to these diseases. Autoimmune Disease Definition, Classification, and Female/Male Ratios In autoimmunity the immune response is directed against host antigens instead of foreign invaders. The host antigens are either localized, as in thyroid and skin diseases, or ubiquitous, as in lupus. Autoimmunity characterizes the pro- totypical diseases whose occurrences differ by sex. Autoimmune diseases pose the central question for the study of such sex differences: what mechanisms explain discrepancies in disease occurrences by sex? Explain- ing sexual dimorphisms in autoimmune diseases will likely bring to light heretofore unknown important biological differences between females and males. Autoimmunity is defined to occur when an antibody binds to or re- acts with an autoantigen (an extract of a normal tissue). Cell-mediated mechanisms may participate in autoimmunity (Draca, 1995; Marchetti et al., 1998; Olff, 1999; Wilder, 1995~. Other causes of autoimmunity include immunization and passive transfer of antibodies in animal models or participation of the major histocompatibility complex (MHC). There is no consensus definition, however. The different definitions and classifica- tions partly explain why different diseases are named autoimmune dis- eases in standard medical texts. Most authorities agree that thyroid and
150 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH TABLE 5-9 Female/Male Ratios Associated with Common Autoimmune Diseases Disease Female/Male Ratio Hashimoto thyroiditis Primary biliary cirrhosis Chronic active hepatitis Graves' hyperthyroidism Systemic lupus erythematosusa Scleroderma Rheumatoid arthritis Idiopathic thrombocytopenic purpuraa Multiple sclerosis Autoimmune hemolytic anemia Pemphigus Type I diabetesa Pernicious anemia Ankylosing spondylitis Goodpasture nephritis/pneumonitis 0 9 8 7 6 3 2.5 2 2 2 1 1 1 0.3 0.2 NOTE: Not all diseases are predominant in females. aAge specific. rheumatic diseases are autoimmune diseases; they differ about inflamma- tory bowel disease, multiple sclerosis, some skin diseases, and juvenile- onset diabetes. Some autoimmune diseases are strikingly predominant in females, others are not predominant in either sex, and still others are predominant in males. Table 5-9 lists several such diseases. The female/male ratios vary 50-fold. Predominance in females applies to some, but not all, of these diseases. The label "predominant in females" is commonly used for human illness and refers to sex differences in incidence. By contrast, in parallel diseases in experimental animals, the term often refers to differences in disease severity. In human autoimmune diseases, severity is similar in females and males. Environmental Causes of Autoimmunity The likelihood that environ- mental factors (toxins and infections) induce autoimmune disease is sup- ported by the diseases and circumstances listed in Table 5-10. In several exogenously induced mimics of autoimmune disease, sex differences in disease occurrence are caused by exposure differences. These diseases are described here. Drug-induced lupus (Yung et al., 1997) and toxin-induced sclero- derma-like disease (Abaitua Borda et al., 1998; Shulman, 1990) loosely resemble but are not identical to idiopathic autoimmune diseases, sug-
SEX AFFECTS HEALTH TABLE 5-10 Autoimmune Diseases in Which Environmental Triggers Are Prominent Agent Disease Sex Toxin Infection Scleroderma-like disease M, Fa Drug-induced lupus M Cryoglobulinemia Chronic Lyme disease Fogo selvagem B27 spondyloarthropathy M Subacute bacterial endocarditis Acute rheumatic fever F NOTE: M, males; F. females; -, no sex predominance. aIndustrial causes (polyvinyl chloride, mining) are predominant in males; toxic product causes (cooking oil, tryptophan) are pre- dominant in females. 151 Besting that exogenous agents cause idiopathic autoimmune disease. More males than females take drugs that induce lupus (male predomi- nant), and more males are exposed to silica inducers of scleroderma-like disease (male predominant). In Spain it was found that more females were exposed to the contaminated cooking oil that causes a scleroderma- like illness (female predominant). More females than males took contami- nated L-tryptophan, a putatively natural antidepressant; the resulting epidemic of eosinophilia-myalgia syndrome was predominant in females. Chronic Lyme disease is a self-perpetuating autoimmune illness that is initiated by but that does not require the persistence of live Borrelia burgdorferi organisms (Carlson et al., 2000~. Its incidence has no sex differ- ence, yet it closely resembles rheumatoid arthritis, whose incidence does have a sex difference. Fogo selvagem, or Brazilian endemic pemphigus foliaceus, is transmitted by a black fly bite and is presumed to be caused by an infectious agent passed at the time of the bite. Fogo selvagem shows no sex preference (Hans-Filho et al., 1999), but spontaneous pemphigus foliaceus, which occurs elsewhere in the world, is predominant in fe- males. Contact dermatitis may be predominant in females, but differen- tial exposure to allergens is the likely cause (Kwangsukstith and Maibach, 1995~. The possibility that infection induces rheumatic autoimmune disease is widely (Miller et al., 2000; Montgomery et al., 1999) but not universally (Ringrose, 1999) accepted. It is unknown how infections induce different diseases by sex (other than by different exposure rates). Potential male- female differences in the processing of infecting organisms, differences in vulnerable periods, or differences in threshold immune responses still apply.
52 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH Hormonal Causes of Autoimmunity, Including Life Events Case re- ports of clinical exacerbation or remission of autoimmune diseases after castration or hormone treatment suggest that gonadal hormone modula- tion plays an important role in disease severity in individuals but consti- tutes weak evidence for sexual dimorphisms in disease incidences in populations (Lahita, 1999~. Studies of the effects of postmenopausal estro- gen or oral contraceptive therapy on autoimmune disease incidence most often show that such therapy has little effect (Petri and Robinson, 1997~. Synoviocyte estrogen receptors may be target organs in rheumatoid ar- thritis (Castagnetta et al., 1999), a possible explanation for the predomi- nance of this illness in females. However, chronic Lyme disease causes a similar joint inflammation but is not predominant in females, and ankylosing spondylitis also causes a similar joint inflammation, but it is predominant in males. Androgens have no apparent role in ankylosing spondylitis (Giltay et al., 1999~. Although experimental feminization wors- ens autoimmune diseases in animal models and experimental masculin- ization ameliorates autoimmune diseases in animal models, variations in both severity and incidence are found. Rheumatoid arthritis goes into remission during pregnancy, contra- dicting the theory of estrogen-enhanced immunological activity. The remission of rheumatoid arthritis is likely due to a human leukocyte anti- gen (HLA) mismatch between mother and fetus rather than to pregnancy- associated hormones (Nelson et al., 1993; Ostensen, 1999~. Multiple scle- rosis also goes into remission during pregnancy (Confavreux et al., 1998~. Although it is often cited that pregnancy induces the flare-up of lupus, lupus in fact does not worsen or worsens only slightly during pregnancy (Lockshin, 1993~. Estrogen replacement therapy, oral contraceptives, and ovulation induction do not worsen lupus (Guballa et al., 2000~. Insight can be gained through the study of rheumatoid arthritis and multiple sclerosis through the study of pregnancy and events during ges- tation as well as during the postpartum period. In these two diseases, clinical symptoms frequently lessen substantially and can even abate dur- ing the third trimester of pregnancy, but the diseases flare soon after delivery. At the cell level gonadal hormones modulate the immune response. Why and how (if it influences disease incidence at all) this effect influ- ences disease incidence is unclear. Estrogen could play a permissive role, allowing survival of forbidden autoimmune clones. A threshold mecha- nism, that is, a specific level of estrogen at a vulnerable time, could ex- plain the increase in incidence, but no such threshold has been postu- lated, tested, or demonstrated in humans. Hormones may also influence the frequency of autoimmune disease in males and females in ways that are independent of the immune system. It is possible that sex differences in the endothelium are critical for disease initiation. A still undiscovered
SEX AFFECTS HEALTH 153 sex difference related to, for example, ovulation- or menstruation-related cytokines, apoptosis of nonimmunological cells, or the presence of vagi- nal flora or the immune response to vaginal flora may be responsible for the different disease experiences of the two sexes. Genetic Causes of Autoimmunity Evidence supporting the concept of genetic control of autoimmunity consists of studies with families and twins, the HLA types associated with specific illnesses, the identification of genes that enhance disease susceptibility, and transgenic experiments in which illness is induced in experimental animals (Seldin et al., 1990~. Evidence of this type is particularly strong for spondyloarthropathy (Taurog et al., 1999), rheumatoid arthritis, and lupus. HLA types by themselves do not explain the sexual dimorphism of the genetic causes of autoimmunity (Chen et al., 1999; Greenbaum et al., 2000), although sex differences in HLA-associated disease expression may occur (Lambert et al., 2000~. For each haplotype associated with a sexually dimorphic autoimmune disease, there is another haplotype associated with an autoimmune disease that is not sexually dimorphic. For instance, HLA B27 is associated with both spondyloarthropathy (female/male ra- tio, 0.3) and uveitis (female/male ratio, 1.0), whereas DR3 is associated with Graves' disease (female/male ratio, 7.0), systemic lupus erythemato- sus (female/male ratio, 6.0), and myasthenia gravis (females/male ratio, 1.0~. With the exception of the CD40 ligand, few putative autoimmune markers identified to date are on the X or Y chromosome. No conclusive evidence for imprinting or X-chromosome inactivation differences exists for autoimmune diseases. The X chromosome has no role in human ankylosing spondylitis (Hoyle et al., 2000~. Non-MHC genes may be rel- evant. In a mouse model of diabetes, mutation of a tissue or a develop- mental stage-specific proteasome product shows sex differences (Hayashi and Faustman, 1999~. The sexual dimorphism of T-cell trafficking may be due to sex-determined cell surface markers or might be secondary to genomic or nongenomic effects. Life Stage Causes of Autoimmunity Most diseases that are predomi- nant in females cluster in the young-adult years, whereas autoimmune diseases that affect younger or older patients are more evenly divided between the sexes (Table 5-11~. Characteristics of young adulthood that may explain the predomi- nance of a disease in females include the chronobiological effects of men- strual cycles, gonadal hormones, threshold effects, vascular responses, immune responses, vaginal flora, and other as yet unknown variables. Very little experimental work has considered these topics.
154 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH TABLE 5-11 Peak Ages of Various Autoimmune Diseases Disease Female/Male Ratio Age (year) Hashimoto thyroiditis Primary biliary cirrhosis Chronic active hepatitis Graves' hyperthyroidism Takayasu's arteritis Systemic lupus Idiopathic thrombocytopenic purpuraa 3 Myasthenia gravisa Scleroderma Rheumatoid arthritisb Dermatomyositisa Poststreptococcal glomerulonephritis Multiple sclerosis 2 Idiopathic thrombocytopenic purpuraa 1 Henoch-Schonlein purpuraa 1 Insulin-dependent diabetes mellitus Ulcerative colitis Henoch-Schonlein purpuraa Dermatomyositisa Pernicious anemia Rapidly progressive glomerulonephritis Myasthenia gravisa Bullous pemphigoid Giant cell arteritis Immunoglobulin A nephropathy B27 spondyloarthropathy Amyotrophic lateral sclerosis Goodpasture nephritis/pneumonitis NOTE: Ages that signify "young adulthood" are indicated in boldface. Higher female/male ratios cluster in young adulthood, as do very low female/male ratios. aTwo age peaks are listed. bSome studies list an older age range. SOURCE: Beeson (1994~. 0 9 8 6 6 3 3 2.5 2.3 2 1 0.5 0.4 0.3 0.3 0.2 30-50 30-75 30-50 30-50 10-30 15-45 20-30 20-30 30-50 20-40 1-15 5-15 20-35 2-5 2-5 2-15 15-40 30-65 40-60 40-80 50-70 50-80 60-75 50-90 10-30 20-50 40-70 15-35 Animal Models of Autoimmunity Animal models of autoimmune dis- ease give mixed messages about the causes of sex differences in autoim- munity. Table 5-12 displays some relevant data for three animal models of human autoimmune diseases, two of which are predominant in fe- males and one of which is dominant in males. Some information is dated and can be challenged by new technology. Animal models of autoimmune disease use immunization, inbreed- ing, and transgenic and gene knockout methods. In the model of thyroidi-
SEX AFFECTS HEALTH 155 tis, different strains of mice and rats are variably susceptible, implying strong genetic control of this disease. Only young adult mice and rats were studied, however. In one rat strain, susceptibility was predominant in females. On the basis of backcross experiments, the X chromosome determines susceptibility. In mice, estrogen enhanced the antithyroid an- tibody titer (a marker of disease) but not thyroiditis itself. The severity of induced thyroiditis was also found to vary with diet. Thus, genetic, X- chromosome, hormonal, and extrinsic factors may influence the occur- rence of thyroiditis. In mouse models of spontaneous lupus, the incidence and severity of lupus are predominant in NZB x NZW (F1) strain females, whereas they are sex neutral in MRL lpr/lpr mice and are predominant in males of the BXSB strain. In animal models, spontaneous autoimmune lupus develops in young adulthood, implying that maturation or cumulative damage is required for disease expression. At maturation, susceptible mouse strains have more numerous and more avid estrogen receptors on lymphoid and uterine tissues than nonsusceptible strains, an explanation of strain sus- ceptibility differences by strain but not of susceptibility differences by sex. Castration and replacement experiments demonstrate estrogen en- hancement and testosterone suppression of spontaneous disease. When animals are raised in a germfree environment, however, neither the phe- notype nor the disease incidence changes, except that females tend to have higher autoantibody levels than males. Raising animals in a germ- free, antigenfree environment ameliorates disease. Males and females in germfree environments are affected equally. Gene knockout experiments give conflicting results. Of the two studies listed, one shows a markedly worse occurrence of glomerulonephritis in females, whereas the other shows equal incidences of glomerulonephritis in both sexes. Thus, in ex- perimental lupus, genetic, hormonal, life stage, and environmental fac- tors all appear to be relevant and sex differences remain unexplained. The human HLA B27 gene transgenically expressed in rats induces a phenotype with features of psoriasis and ankylosing spondylitis. In a germfree environment, the spondylitis does not occur but the psoriasis does. Introduction of specific gastrointestinal pathogens to the germfree animal induces the spondylitis. The predominance of psoriasis and spondylitis in males is true of this model, as it is of the disease in humans. Genitourinary anatomy, sex hormones, immune response, and unknown factors are possible explanations. In summary, animal models suggest that autoimmune diseases have specific genetic, hormonal, life stage, and environmental causes. The hu- man sex differences are not reproduced in many of the animal models, but no attempt has been made to understand why.
156 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH TABLE 5-12 Animal Models of Human Autoimmune Diseases Model Method Animal Female/Male Ratio Gene Thyroiditis Immunization Mouse F = M Inbred st with thyroglobulin (MHC) Immunization with Rat F > M Inbred st thyroglobulin polymer MHC), chromo (Lillehc 1981) Lupus Spontaneous disease Mouse F > M Inbred st MHC, c meet; 0 mmunl are rele Spontaneous disease Mouse F = M Inbred st MHC, meet; 0 immunl are rele
SEX AFFECTS HEALTH 157 ale Ratio Gene Hormone Life Stage Environment Inbred strains Estrogen enhances Young adults Modified by diet (MHC) antibody level but tested (Bhatia et al., not clinical disease 1996) in susceptible strains (Okayasu et al., 1981) Inbred strains, Castration; estrogen Young adults polygenic (not replacement increases tested MHC), X severity and incidence chromosome in males (Lillehoj et al., 1981) Inbred strains, Castration; estrogen Disease Modified by diet MHC, comple- replacement increases develops in meet; other severity and incidence young adulthood immune genes in males (Ahmed and (all strains); are relevant Talal, 1999) in the susceptible strain the estrogen receptor is more numerous and has a higher affinity at maturity (Dhaher et al., 2000) Inbred strains MHC, comple- ment; other immune genes are relevant Disease develops Germfree; no in young difference in adulthood incidence or severity between males and females or conventionally raised controls (Unni et al., 1975); germfree, ant i g en fre e males but not females have lower lymph node weights than conven- tional controls; the rate of glo- merulonephritis . . IS . .ess in germ- free, an ti g en fre e animals (Maldonado et al., 1999) continued on next page
158 TABLE 5-12 Continued EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH Model Method Animal Female/Male Ratio Gene Gene knockout Mouse F > M p21, antil slightly females glower in fema et al., 2 Gene knockout Mouse F = M DNase I, worse i (Napir~ Spondylopathy Transgenic Rat MHC NOTE: In human thyroiditis the female/male ratio is 9, in lupus the female/male ratio is 6, and in spondyloarthropathy the female/male ratio is 0.3. F. female; M, male. Summary Gonadal hormones can modulate the adaptive immune response, but hormone effects alone are unlikely to explain the excess incidence of au- toimmune illness in females. If gonadal hormones play a role, they must do so through a threshold or permissive mechanism. Although suscepti- bility to autoimmune illness is regulated by genetic background, no ge- netic mechanism that would explain sexual dimorphism has yet been postulated. The epidemiological risk factors for the sex-discrepant au- toimmune diseases young age and female sex are consistent with the differential exposures to causative agents between males and females, the existence of vulnerable periods in females, and threshold effects rather than cell- or organ-level differences in the biologies of females and males. A long period of latency between exposure and clinical disease is pos- sible, complicating the search for etiologies that differ by sex. A SPECTRUM OF SEX DIFFERENCES ACROSS A DISEASE: CORONARY HEART DISEASE Coronary heart disease begins in utero, evolves through childhood, and emerges in middle and old age as a devastating and crippling prob- lem. Plaques of cholesterol and other cellular materials are deposited in
SEX AFFECTS HEALTH 159 ale Ratio Gene Hormone Life Stage Environment p21, antibody slightly worse in females; severe glomerulonephritis in females (Balomenos et al., 2000) DNase I, antibody slightly worse in females (Napirei et al., 2000) MHC Disease develops in young adulthood Disease develops in young adulthood Disease develops Germfree; In young adulthood intestinal bacteria are required for phenotype; more frequent in males (Taurog et al., 1999) the inner lining of the coronary arteries and over time compromise the flow of blood, causing cell and organ death myocardial infarction. Coronary heart disease is a major cause of death in the United States (National Center for Health Statistics, 1999~. In general, women manifest symptoms 10 to 20 years later than men (National Center for Health Sta- tistics, 1999) and have a higher prevalence of primary risk factors (Becker et al., 1994; Greenland et al., 1991; Steingart et al., 1991~. Men, however, die at an earlier age. Advances in knowledge about heart disease have led to steady de- clines in the rates of mortality from heart disease. Yet, many questions remain, and none is more compelling than the differences between the sexes. Sex Differences in Development of Coronary Heart Disease The etiologies of coronary heart disease encompass the environment, genetics, age, and lifestyle. Environment The most important environmental agents influencing coronary heart disease are diet, drugs, airborne toxins, and, possibly, infectious agents;
160 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH and the effects of these agents may be synergistic. For example, overeat- ing or obesity and a sedentary lifestyle promote high blood pressure, high cholesterol, and diabetes, which are major risk factors for coronary heart disease in both males and females. Yet, susceptibilities and responses vary by sex. Genetics Dyslipidemias are among the strongest genetic contributors to coro- nary heart disease (Goldstein et al., 1973a,b; Hazzard et al., 1973~. Familial hypercholesterolemia (FH) is a low-density lipoprotein (LDL) cholesterol disorder caused by a genetic defect of the LDL receptor on cells. It is autosomal dominant and is more severe in homozygotes because of a gene dosage effect (Goldstein et al., 1995~. Homozygotic defects occur in about 1 in 1 million individuals, and such individuals present with cholesterol levels that range from 650 to 1,000 milligrams per deciliter (mg/dl); heterozygotic defects occur in about 1 in 500 individuals, and such individuals present with cholesterol levels that range from 350 to 550 mg/dl (Goldstein et al., 1995~. Homozy- gotes develop atherosclerosis, which leads to coronary heart disease and death usually before age 20. Sex differences are undocumented in ho- mozygotes but do occur among heterozygotes; heterozygous males die about 10 years earlier than heterozygous females (Table 5-13~. In the FH heterozygote population, the age differences in incidence between the sexes are comparable to those for the general population. Age Sex differences in heart disease mortality occur over the life span TABLE 5-13 Estimated Risk of Symptoms of Coronary Heart Disease and Death from Myocardial Infarction in Heterozygotes at Different Ages Percent Male Heterozygotes Female Heterozygotes Age (years) Symptoms Deaths Symptoms Deaths 40 20 3 0 50 45 25 20 2 60 75 50 45 15 70 80 75 30 SOURCE: Goldstein et al. (1995, p. 1986).
SEX AFFECTS HEALTH 7,000 1 6,000 - 5,000 - Q I, 4,000- ct a' o a' ~ 2,000- z 3,000 - 1 ,000 - o 161 Men rot=_ 35-44 45-54 55-64 Age 65-74 75-84 285 FIGURE 5-3 Death rates for diseases of the heart by age and sex, 1995-1997. Source: National Center for Health Statistics (1999~. between women and men (Figure 5-3~. The Barker hypothesis postulates that the genesis of coronary heart disease may begin in utero, perhaps in response to malnutrition and stress (Barker, 2000; see also Chapter 3~. The fetus may adapt to the lack of critical nutrients or an excess of maternal stress hormones by permanently altering metabolic, endocrine, and car- diovascular systems in such a manner as to promote atherosclerosis later in life (Barker, 1999~. Several longitudinal studies have demonstrated sex differences in coronary heart disease. The Framingham Heart Study, initiated in 1948 (Dawber et al., 1951), demonstrated that cardiovascular differences exist between the sexes. For example, it established the classical risk factor concept for coronary heart disease. This landmark study showed that a raised serum total cholesterol level, high blood pressure (systolic and diastolic), and smoking increase the risk of developing coronary heart disease in men and women in a graded fashion. Women develop coro- nary heart disease about 10 years later than men and women's risk is smaller. The Bogalusa Heart Study followed African-American and Caucasian children over time, beginning in the 1960s, demonstrating that sex differ- ences in risks for coronary heart disease begin at an early age. The longitudinal Tromso Heart Study (Norway) (Stensland-Bugge et
162 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH al., 2000) reexamined 3,000 middle-aged women and men in Norway 15 years after an initial examination. Hypertension, total cholesterol levels, high-density lipoprotein (HDL) cholesterol levels, and body mass index independently predicted increased carotid intimal thickness in both women and men. However, triglyceride levels were an independent risk factor in women but not men, whereas physical activity and smoking were independent risk factors in men but not women. The National Health and Nutrition Examination Survey is a federal epidemiological follow-up study that has collected national data on U.S. residents since the 1960s and that provides evidence of the incidence and prevalence of a variety of health indicators by age, race, and sex (National Center for Health Statistics, 2000b). These data again show differences in cardiovascular disease incidence between the sexes and among racial and ethnic groups. For example, the age-adjusted risk for incident coronary heart disease is higher in African-American women ages 20 to 54 years than in Caucasian women of the same age and lower in African-American men than in Caucasian men of the same age (Gillum et al., 1997~. In terms of risk factors, men have higher prevalences of hypertension, cigarette smoking (until age 65), and excess weight, whereas women have higher prevalences of elevated serum cholesterol levels and obesity (Na- tional Center for Health Statistics, 2000a). Regarding the age-specific prevalences of hypertension and serum cholesterol levels greater than 240 mg/dl, men have a higher prevalence than women until about their late 40s and early 50s. After that, the prevalence is higher in women. These studies underscore the fact that coronary heart disease begins early in life, that it continues across the life span, and that sex differences exist. What they do not demonstrate is why such differences exist. Cigarette Smoking Many airborne toxins that are absorbed through the respiratory sys- tem predispose an individual to the development of coronary heart dis- ease. One of the most pervasive and lethal of these agents is cigarette smoke. Cigarette smoking is more frequent among men than women, espe- cially in African Americans (National Center for Health Statistics, 1999) (Table 5-14~. Data from the Centers for Disease Control and Prevention indicate that both men and women who smoke have three times the risk of dying from heart disease as individuals who do not smoke. In women, smoking (Baron et al., 1988; Michnovicz et al., 1986) promotes susceptibility to early menopause and to a number of diseases, including coronary heart disease.
SEX AFFECTS HEALTH 30 - 25 - o . _ In ~ 20 - Q o Q o C' 15 - 10 - 5 - 0 - 163 TABLE 5-14 Smoking Prevalence by Race and Sex, 1998 Percent African American Caucasian Males 29 26 Females 21 23 SOURCE: National Center for Health Statistics (1999~. Males Females White African American White, African American, non-Hispanic non-Hispanic FIGURE 5-4 Age-adjusted high serum cholesterol levels (> 240 mg/dl) among individuals ages 20 to 74 years, by sex and race, 1988-1994. Source: National Center for Health Statistics (1999~. Cholesterol Sex differences exist in the levels of lipoprotein subfractions of choles- terol: LDL cholesterol, HDL cholesterol, and very-low-density lipopro- tein (VLDL) cholesterol (Figure 5-4 and 5-5~. Before menopause, women have higher HDL cholesterol ("good cho- lesterol") levels and lower LDL cholesterol ("bad cholesterol") levels. During the pert- and postmenopausal periods, however, LDL cholesterol levels rise and HDL cholesterol levels drop. Rates of death from heart disease rise with age in both men and women, but the rate of ascent rises more sharply in women as menopause ensues.
64 60 - 50 - o . _ ~ 40 - Q o o ~ 30 - a' C' a' 20 - 10 - O- EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH Males Females 20-34 35-44 45-54 55-64 65-74 Age (years) FIGURE 5-5 Non-age-adjusted high serum cholesterol levels (>240 mg/dl) by sex and age. Source: National Center for Health Statistics (1999~. Estrogen affords women a protective advantage against coronary heart disease before menopause. 17-~-Estradiol increases HDL choles- terol levels and decreases LDL cholesterol levels, stimulates nitric oxide, and inhibits vasoconstricting factors (Collins, 2000~. These factors may operate independently or synergistically (Collins, 2000~. Estrogen also prevents calcium influx through ion channels in the membranes of vascu- lar smooth muscle cells, preventing vessel contraction. During perimenopause estrogen levels begin to decline and continue to do so through menopause, and HDL cholesterol levels fall as LDL cholesterol levels rise. Hepatic LDL cholesterol receptor activity may ex- plain these changes (Semenkovich and Ostlund, 1987~. Another explana- tion for the increased incidence of cardiovascular disease after meno- pause is that the LDL cholesterol particles may become denser and therefore less protective (Campos et al., 1988; Haffner et al., 1993~. High triglyceride levels present a greater risk to women than to men. HDL cholesterol levels may be a better predictor than LDL cholesterol levels of coronary heart disease risk in women (Gordon et al., 1989; lacobs et al., 1990~.
SEX AFFECTS HEALTH 165 As noted above, endogenous estrogen present in premenopausal women appears to have a cardioprotective effect. Observational studies of postmenopausal women taking hormone replacement therapy, and experimental studies with animals, demonstrate that unopposed estrogen has a strong effect in preventing atherosclerosis and its clinical sequelae (Barrett-Connor, 1998a; Barrett-Connor and Grady, 1998~. These studies however have not been supported in two large clinical trials of estrogen combined with progestin in women with established coronary disease (Hulley et al., 1998; Herrington et al., 2000~. Other clinical trials including women both with and without established coronary disease are currently in progress. As a result, the issue of whether hormone replacement therapy has cardioprotective effects must be considered unresolved. Hypertension The causes of hypertension are unknown (Williams, 1991), yet studies suggest multiple factors, polygenetic and environmental. Essential hyper- tension is the most common form, and it affects a substantial portion of the U.S. population. Risk factors for hypertension include age, sex, smok- ing, diet, elevated blood cholesterol levels, obesity, diabetes, sedentary lifestyle, and family history. Hypertension is both a risk factor for coronary heart disease and a disease itself. As a result of hypertension, the heart wall thickens and its function declines. There are differences in the association of hypertension and coronary heart disease by sex (Fiebach et al., 1989; Kannel et al., 1976; Sigurdsson et al., 1984) and race (Johnson et al., 1986~. Men have higher blood pressure levels than women (National Center for Health Statistics, 2000a, p. 245~. Height differences as well as differ- ences in mass contribute to differences in blood pressure between men and women. The genesis of hypertension in adulthood often occurs in childhood, and weight is a greater predictor for hypertension in girls than in boys (Cook et al., 1997~. Blood pressure is higher during the follicular phase than during the luteal phase of the menstrual cycle in both normotensive and hyperten- sive women (Dunne et al., 1991~. African-American women respond to stress with a higher diastolic pressure and higher plasma epinephrine levels during the follicular phase than during the luteal phase, but Cauca- sian women (and men of all ethnicities) do not exhibit any significant change in blood pressure over the course of a month (Ahwal et al., 1997; Mills et al., 1996~. These results have been disputed by investigators (Litschauer et al., 1998), who believe that the differences may be caused by an interaction of sex and task characteristics (cause of stress). The angiotensin-converting enzyme deletion-insertion polymorphism appears to be associated with systemic hypertension in Caucasian men
166 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH (O'Donnell et al., 1998~. In Japanese men but not Japanese women there is a similar association between angiotensin-converting enzyme gene poly- morphism and hypertension (Higaki et al., 2000~. Two hypertension-associated conditions are sex specific: preeclamp- sia and hypertension of pregnancy. Both conditions are characterized by increasing levels of hypertension as pregnancy progresses and are most common in the last trimester. Both are serious and can be fatal. Whether either of these conditions progresses to coronary heart disease is not un- derstood, although two studies have shown that such a relationship exists (Croft and Hannaford, 1989; Rosenberg et al., 1983~. Diabetes Mellitus Diabetes mellitus is a risk factor for coronary heart disease and is an example of the sex differences in the risk for coronary heart disease. Pre- menopausal women, who are not typically at risk for coronary heart dis- ease, are at risk if they have type I (juvenile) or type II (adult-onset, non- insulin-dependent) diabetes mellitus (Figure 5-6~. Rates of mortality from coronary heart disease are two to four times greater in diabetic men than nondiabetic men and three to seven times greater in diabetic women than nondiabetic women (Barrett-Connor and Wingard, 1983; Kannel and Abbott, 1987; Manson et al., 1991; Pan et al., 1986~. Diabetes may negate estrogen receptor binding and thereby miti- gate the positive estrogenic effect (Ruderman and Haudenschild, 1984~. O 70- o 60- <,, 50- <,, 40- 30 - ~a 20- 0 10- o ,$ ~ ,,~ Diabetic Men ,, ~ , ___-~ =~~~ ~ Diabetic Women Men ~ ,, ~ ~_-~~~ ~ ~ _ ·Women 0-3 4-7 8-11 12-15 Duration of follow-up (years) 16-19 20-23 FIGURE 5-6 Mortality from coronary heart disease and diabetes in men and women ages 25 to 64. Source: Krolewski et al. (1991~. Reprinted, with permission, from A. S. Krolewski, l. H. Warram, P. Valsania, B. C. Martin, L. M. Laffel, and A. R. Christlieb. 1991. Evolving natural history of coronary artery disease in diabetes mellitus. American journal of Medicine 90:56S-61S. Copyright 1991 by Elsevier Sci- ence Ltd.
SEX AFFECTS HEALTH 167 Diabetes of pregnancy may represent a risk for the development of non-insulin-dependent diabetes mellitus after pregnancy. Current evi- dence suggests that gestational diabetes is a risk factor for coronary heart disease (Mestman, 1988; O'Sullivan, 1984; Stowers, 1984~. Differences in Presentation of Coronary Heart Disease The manifestations of coronary heart disease vary in presentation and intensity between women and men (Table 5-15~. Coronary heart disease presents in women 10 to 15 years later than it does in men. Women often have comorbidities, such as congestive heart failure, hypertension, diabetes, and others. Diabetic women are particularly vul- nerable to complications after a myocardial infarction (Greenland et al., 1991~. More men present with myocardial infarction as the initial manifesta- tion of the disease, but the event is more often fatal in women (Greenland et al., 1991; Kannel and Abbott, 1987; Lerner and Kannel, 1986; Murabito et al., 1993; Wenger, 1985~. Women admitted with acute myocardial in- farction are more likely, on average, to be older than men being admitted and therefore have more severe coronary artery disease. However, in the Framingham study, nearly 66 percent of sudden deaths due to coronary heart disease in women occurred in those with no previous symptoms of disease (Mosca et al., 1999~. TABLE 5-15 Complications of Acute Myocardial Infarctions, by Sex Percent Females Males Complication (N = 1,524) (N = 4,315) P value Mechanical complications Angina in hospital 8.2 8.8 NSa Congestive heart failure on admission 26.8 24.4 0.02 Cardiogenic shock 11.1 7.4 <0.002 Arrhythmic complications Sinus tachycardia 3.2 3.1 NS Supraventricular arrhythmia 12.1 13.0 NS Atrioventricular block 12.6 10.0 NS Ventricular tachycardia 12.9 19.0 <0.005 Ventricular fibrillation 5.8 7.4 0.11 Cardiac arrest 9.8 6.6 <0.0005 aNS, not significant. SOURCE: Greenland et al. (1991~.
168 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH Among individuals with myocardial infarction, men more often present with ventricular tachycardia and women more often present with cardiogenic shock and cardiac arrest (Greenland et al., 1991; Milner et al., 1999). Although many women present with "classical" signs and symptoms, some do not. Women are less likely to present with severe chest pain than men and at the time of diagnosis are more likely to be experiencing con- gestive heart failure (events that may be age rather than sex related). Sex Differences in Treatment of Coronary Heart Disease (Myocardial Infarction) Women do not fare as well as men after a myocardial infarction for the following reasons: · Women with myocardial infarctions are older. · Women have more "silent" myocardial infarctions. · Men have larger collateral circulations. After a myocardial infarction, women younger than 65 years of age are more than twice as likely to die as men of the same age (Vaccarino et al. 1999~. Possible explanations include the following: -Diabetes, heart failure, and stroke are more prevalent in younger women. die. -Plaque erosions are more common in premenopausal women who -Arterial narrowing is less and reactive platelets levels are higher in younger women. -Women are less likely to be given effective interventions, such as aspirin, beta-blockers, and thrombolytic agents. Differences also extend to the use of different diagnostic and thera- peutic procedures, in that women have fewer diagnostic procedures (O'Farrel et al., 2000; Shaw et al., 1994; Steingart et al., 1991; Vaccarino et al., 2001; Wong et al., 2001~. A great deal has been posited about whether women are discriminated against or whether mitigating factors account for the differences in assessment of women for coronary heart disease. Certainly, tests for coronary heart disease are more frequently conducted for men. · Women have less obstructive disease at earlier ages than men, and thus, the noninvasive tests have less predictive value for women (Wenger, 1994~.
SEX AFFECTS HEALTH 169 · Subclinical disease, which occurs in younger women, is two to three times more likely to be myocardial infarction or stroke (Kuller et al., 1995~. · Women have a smaller coronary artery size (lumen). This reduces the possibility of angiography or angioplasty and bypass surgery and thus of diagnosis and a better outcome (Sheifer et al., 2000~. However, the technology is gradually changing to accommodate this need. A 1991 study examined such differences in Massachusetts and Mary- land (Ayanian and Epstein, 1991~. The findings confirm that women are less likely than men to undergo diagnostic and therapeutic procedures (Table 5-16~. Men were between 15 and 45 percent more likely to undergo selected procedures. The investigators (Ayanian and Epstein, 1991) sug- gest several reasons for such discrepancies: · physician perception of the severity of the disease in men versus women; · physician perception of the risks and efficacies of diagnostic and therapeutic procedures between men and women (women have higher rates of mortality after the use of procedures); · higher rates of admission for women with an absence of true coro- nary heart disease versus the rates for women with ischemic symptoms; · patient's perceptions and preferences (women may be more will- ing to adhere to a lifestyle of medications and limitations than to face surgery); and · bias in health care delivery. TABLE 5-16 Male: Female Odds Ratios for Use of Diagnostic Procedures for Coronary Heart Disease Mean (Range) Odds Ratio Massachusetts Disease Angiography Maryland Revascularization Angiography Revascularization Any coronary 1.28 heart disease (1.22-1.35) Myocardial 1.39 infarction (1.23-1.58) (1.10-1.51) 1.45 (1.35-1.55) 1.31 1.15 (1.08-1.22) 1.29 (1.10-1.51) 1.27 (1.16-1.40) 1.40 (1.02-1.91) NOTE: Odds ratios were calculated by multiple logistic regression, with adjustment for principal diagnosis, age, secondary diagnosis of congestive heart failure or diabetes melli- tus, race, and insurance status. SOURCE: Ayanian and Epstein (1991).
170 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH Racial and ethnic differences exist between the sexes. A surveillance of hospital admissions for myocardial infarction and of in-hospital and out-of-hospital deaths due to coronary heart disease between 1987 and 1994, showed that Caucasian men had the greatest average annual de- crease in rates of mortality from coronary heart disease and myocardial infarction, followed by Caucasian women, African-American women, and African-American men. No differences in rates of hospitalization for a first myocardial infarction were evident between men and women over the time studied. Furthermore, over the time period studied, the rate of reinfarction decreased and the rate of survival increased (Rosamond et al., 1998~. Summary Sex differences in the development, recognition, and treatment of coronary heart disease exist across the life span. There is mounting evi- dence that these differences are not solely related to hormones. Scientists and clinicians have just begun to appreciate and address these differences by studying the earliest stages of development and the subsequent effects of the internal and external environments. Research to date, however, has posed as many questions as it has answered. FINDINGS AND RECOMMENDATIONS Findings Many diseases affect both sexes, and the diseases often have different frequencies or presentations in males and females; therefore, different preventive, diagnostic, and treatment approaches may be required for males and females. Exposures, susceptibilities, responses to initiating agents, energy metabolism, genetic predisposition, and responses to thera- peutic agents are important factors in understanding how each sex re- sponds to insult, injury, disease progression, and treatment (Figure 5-7) Compounds that may be differentially internalized by males and fe- males include . . poses; foods for energy and nutrients, including vitamins and minerals; drugs for diagnostic, prophylactic, therapeutic, or recreational pur- · environmental compounds to which exposure is either purposeful (preservatives in foods) or inadvertent (pollutants or secondhand ciga- rette smoke); and · microorganisms that act as pathogens or commensals.
SEX AFFECTS HEALTH Physical: Sounds, light, heat, vibrations, gravity, radiation Chemical: Drugs, food and supplements, environmental compounds Infectious: Bacteria, mycobacteria, viruses, fungi, parasites - ~_ \ / Entry Gastrointestinal tract, respiratory tract, skin, eyes, urogenital tract, parenteral routes, transplacental :~v~> 1 ,\W Responses Genetic, molecular, cellular, organ, organ system, whole organism (dependent or independent of environment) Factors That Affect Responses Genotype, growth and development, life stage, hormone cycles, pregnancy, chronobiology, prior exposures and responses, current health status FIGURE 5-7 External agents. 17 ,/ These compounds can enter the body via the placenta, gastrointesti- nal tract, respiratory tract, eyes, urogenital tract, the transdermal route, or the parenteral route. Portals of entry can differ between the sexes, affect- ing types and incidences of disease. In addition, individuals are exposed to environmental factors such as sensory stimuli (sounds, light, heat), forces (vibration, gravity), and radiation.
72 EXPLORING THE BIOLOGICAL CONTRIBUTIONS TO HUMAN HEALTH Exposures and responses to exogenous agents may be influenced by growth and development; other aging processes; reproductive events; body size and composition; cumulative exposures, prior responses, and current health status; and genotype. Some exogenous factors (e.g., vitamins) are essential, whereas others are of no apparent consequence and some (e.g., cigarette smoke) are det- rimental. Some factors are beneficial at low doses but harmful at high doses (e.g., ultraviolet radiation). It is essential to understand whether males and females respond differently to these various substances, whether they do so at various stages in their life spans, and if so, how and with what implications. A continuing challenge is the identification of sex differences in health and illness that are methodologically robust, hereafter referred to (pro- vocatively) as "true" sex differences. Different biological, lifestyle, and social structural contexts for males and females can lead to spurious infer- ences about sex. Even the most methodologically valid comparisons are potentially susceptible to biased interpretations, intentional or not. Exposure, susceptibility, responses to initiating agents, energy me- tabolism, and responses to therapeutic agents matter and must be consid- ered in any consideration of sex differences in health and disease. How- ever, in many instances these differences in disease manifestations and health outcomes cannot be explained by the obvious anatomical or sex hormone differences between males and females. In some instances, soci- etal practices and beliefs, independent of biological sex, account for the differences, but in other instances, the differences remain unexplained. The following recommendation is presented as a result of these consider- ations. Recommendation RECOMMENDATION 6: Monitor sex differences and similarities for all human diseases that affect both sexes. Investigators should · consider sex as a biological variable in all biomedical and health- related research; and design studies that will -control for exposure, susceptibility, metabolism, physiology (cycles), and immune response variables; -consider how ethical concerns (e.g., risk of fetal injury) con- strain study designs and affect outcomes; and -detect sex differences across the life span. Also see Recommendation 3 (in Chapter 3) for a discussion of the need to mine cross-species information.