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.
7 Adverse Reproductive Outcomes The term "adverse reproductive outcome" includes such diverse endpoints as the inability to conceive (sterility or infertility), the premature spontaneous termination of a pregnancy (abortion), the birth of an infant with a congenital malformation or with mental or physical retardation, and the premature death of an offspring (stillbirth, neonatal, or infant death). Many host and environmental factors contribute to the origins of such outcomes. They may be caused by ma- ternally or paternally derived inherited defects, exposure to noxious environ- mental agents, including ionizing radiation, smoking, or the consumption of al- cohol, preexisting maternal illness (such as diabetes) or illness during pregnancy, and malnutrition (Bracken, 19841. This diversity of possible origins makes diffi- cult the assignment of causation in any specific instance. Of primary interest here are those adverse reproductive outcomes that may have arisen through the induction of a deleterious mutation in the paternal germ cells as a consequence of exposure to ionizing radiation. As noted in Chapter 6, this contribution to the totality of adverse reproductive outcomes is called the mutation component, and it varies substantially from one endpoint to another. Unfortunately, in most instances the precise size of this mutation component is either unknown or poorly estimated, but sufficient information is available to estimate its probable magnitude. 42
AD VERSE REPRODUCTI VE OUTCOMES 43 GENERAL REMARKS In the paragraphs that follow, the committee examines separately each of 11 adverse reproductive outcomes. This examination will include a definition of the endpoint, current estimates of its frequency, its causers), possible sources of ex- traneous variability (confounders) that may contribute to its prevalence or inci- dence, the interaction of various contributors to the occurrence of the endpoint, and the difficulties inherent in obtaining reliable estimates of prevalence or inci- dence and paternal exposure. The risk factors described primarily relate to the mother's preconceptional or gestational exposure. Although much of what Is known about the causes of these endpoints relates to maternal exposure, specific reference will be made to the status of evidence for male-mediated effects. In addition, these maternal risk factors are discussed because they are potentially strong confounders that would have to be taken into account in any study of pa- ternal radiation exposure. A confounder in this context is a variable that is causally related to the disease under study and that is associated with exposure in the study population, but that is not a consequence of this exposure (Kelsey et al., 19861. Current research on adverse reproductive outcomes has been expanded to include effects on the reproductive systems of both men and women as well as effects on the offspring. This expanded list of effects is shown in Table 4 and includes menstrual cycle changes; semen characteristics; fecundability and fer- tility; embryonic and fetal loss; any complication affecting the embryo, fetus, or mother; infant morbidity and mortality; and childhood malignancies. For ex- ample, preconceptional exposure may lead to spermatic or ovarian toxicity and germ cell mutations in either sex. These latter effects may in turn lead to infer- tili~y, spontaneous abortions, or abnormalities in the offspring. Prenatal expo- sures could lead to functional deficiencies and illnesses in the offspring that may not become apparent until childhood or even adulthood, such as developmental disorders and malignancies. In summary, the reproductive and developmental effects of environmental agents may operate through a variety of mechanisms including toxic, mutagenic, teratogenic, and carcinogenic effects. TABLE 4. Possible Reproductive Effects of Environmental Exposures . . 1. Menstrual cycle disorders 2. Hormonal changes 3. Sperm or semen abnormalities 4. Infertility 5. Single gene defects 6. Chromosomal aberrations 7. Fetal loss 8. Altered sex ratio SOURCE: Adapted from Berkowitz (1985). 9. Intrauterine growth retardation 10. Preterm birth 11. Maternal complications 12. Birth defects 13. Neonatal and infant deaths 14. Developmental deficits 15. Childhood malignancies 1 6. Other childhood diseases . This multiplicity of possible adverse reproductive effects and etiologies makes difficult the design and implementation of epidemiologic studies seeking to establish the association of a specific effect(s) with a given exposure, as well as the meaningful interpretation of a suspected association. -
44 ADVERSEREPRODUCTIVE OUTCOMES MALE-MEDIATED REPRODUCTIVE AND DEVELOPMENTAL OUTCOMES The feasibility of studying adverse reproductive outcomes among Atomic Veterans and their families hinges in large measure on the likelihood that genetic effects may be seen following exposure of the male to ionizing radiation. To assess this likelihood the committee summarized the evidence on paternal expo- sures and adverse reproduction and disease in offspring found in studies of ani- mals and humans. The role of paternal exposure in the origin of many adverse reproductive and developmental outcomes has not been investigated extensively in humans (Olshan and Faustman, 1993~. The effects of chemicals or ionizing radiation on the sperm, chromosome, and fertility have been demonstrated (Wyrobek et al., 1983; Martin et al., 1986; Geneseca et al., 1990~. The prevailing view is that exposure of the human male to chemicals and ionizing radiation is generally un- related to the occurrence of developmental endpoints such as miscarriage, birth defects, growth retardation, and cancer (Brown, 1985~. Animal studies, some repeatedly confirmed, demonstrate that paternal exposure can lead to a variety of effects, including skeletal malformations and cataracts, in the offspring. A causal connection has been established with radiation-induced mutations. There is a paucity of human data on this subject. Several potential mechanisms have been proposed to explain possible male- mediated effects on offspring. A direct effect of an agent on male germ cell DNA is the traditional explanation for the induction of some abnormalities and is of major relevance for considering the potential effects of ionizing radiation. Since the time span for male germ cells to mature into spermatozoa is only about 8-9 weeks (compared to the longevity of a lifetime for resting oocytes), the ef- fect of environmental agents must be on the spermatogonia for possible effects to be long lasting. The majority of the available data from tests in animals suggest that this mechanism is involved. More indirect mechanisms involving the transfer of toxic agents in seminal fluid and maternal exposure to agents brought home by the father have been sug- gested. Some toxicants, such as PCBs and lead, have been found in seminal fluid, and vaginal absorption of substances in semen can occur. Nevertheless, it is unclear whether the concentrations of these toxicants obtained through this mechanism would be sufficiently great to have an effect on the fetus (Hatch and Marcus, 1991~. There are some data from studies in animals linking seminal transfer of chemicals to preimplantation loss (Robaire and Hales, 1994~. There is no evidence of seminal fluid toxicants being produced as a result of radiation exposure. As noted above, evidence for the direct effects of exposure on male germ cell DNA (germ cell mutagenicity assays) from studies in experimental animals has been available for several decades. In fact, an important test in animals for germ cell mutagenicity, the specific locus test, was developed in 1951 and has been extensively used in tests involving ionizing radiation (Russell, 1951~. The
AD VERSE REPRODUCTI HE OUTCOMES 45 majority of the data from tests in animals were obtained in the evaluations of radiation and chemicals in relation to the expression of defined visible pheno- types caused by mutations at recessive loci (specific locus test), fetal loss (dominant lethal test), inherited chromosomal aberrations (heritable translocation test), and congenital anomalies (dominant skeletal and dominant cataract tests). Ionizing radiation and a small number of chemicals (compared to animal car- cinogenicity assays) have yielded positive results in these test systems. Data from tests in animals also show that exposure of males to ionizing radiation and some chemicals can produce other outcomes in the offspring such as tumors, growth retardation, and neurobehavioral effects (Olshan and Faustman, 19931. Detailed data from studies in experimental animals exposed to ionizing radiation can be found in Chapter 6. Epidemiologic associations between paternal exposure and reproductive and developmental abnormalities in offspring have been reported, although further verification is needed. Most of the studies have focused on occupational expo- sures. A brief summary of the findings from epidemiologic studies is provided here. The major epidemiologic studies relating to ionizing radiation and devel- opmental outcomes, such as the studies of the atomic bomb survivors, can be found in Chapter 6. There is evidence that a small number of agents such as ionizing radiation and dibromochloropropane (DBCP) may affect semen quality and result in reduced fertility or sterility when individuals are exposed to high doses. An increased risk of fetal loss (spontaneous abortion) has been linked with paternal occupational exposure to vinyl chloride, anesthetic gases, dibro- mochloropropane, mercury, lead, other metals, and various solvents (Savitz et al., 19941. A number of epidemiologic studies have examined the relationship between paternal occupation and birth defects (Olshan and Schnitzer, 1994~. Positive associations have been found for individuals with a variety of occupa- tions, including painters, welders, firemen, forestry and logging workers, motor vehicle operators, wood workers, farm workers, and metal workers. Individuals in these occupations have a variety of exposures, often mixed, to metals, sol- vents, pesticides, and paints. Several recent large case-control studies have reported some paternal occu- pations and exposures that may be associated with childhood cancer in the off- spring (Savitz, 1986; Savitz and Chen, 1990; O'Leary et al., 1991~. These asso- ciations include leukemia in the offspring of painters, mechanics, and machinists; childhood brain tumors in the offspring of painters, metal workers, those in elec- tronics-related occupations, and those with motor vehicle-related jobs; Wilms' tumor (childhood kidney tumors in the offspring of auto mechanics and machin- ists, welders, and painters; and neuroblastoma in the offspring of those in elec tronics-related occupations. To summarize briefly, some male-mediated environmental exposures may affect pregnancy outcome through fertilization by mutation-bearing sperm, which could lead to pregnancy loss or possibly congenital malformations or childhood cancer (Berkowitz and Marcus, 1993~. It is also possible that male- mediated effects could occur because male workers bring toxicants into the
46 ADVERSE REPRODUCTIVE OUTCOMES home environment or because toxicants are absorbed into the female genital tract via sperm or seminal fluid. The experimental animal and epidemiologic data indicate that the exposure of the male to various toxic agents may increase the risk of the full spectrum of adverse developmental endpoints from fetal loss to cancer. However, the evidence is not firm and clearly requires more study and confirmation in both laboratory and epidemiologic settings. INFERTILITY Approximately 15% of all couples of reproductive age cannot achieve a de- sired pregnancy (McClure, 1986~. The large majority of previous research has concentrated on the female component; however, half of all fertility problems may be due to male reproductive dysfunction (Swerdloff, 1985~. The role of extrinsic environmental factors on male reproduction has suggested that chemi- cal and physical exposures have an effect (Biava et al., 1978; Steeno and Pang- kahila, 1984; Bonde, 1990~. In a recent study, occupationally related exposure to electromagnetic fields was unrelated to morphology, motility, or concentration of semen (Lundsberg et al., 19951. Ionizing radiation at high enough doses can lead to temporary or permanent sterility. The doses and time course of the proc- ess in humans have been studied extensively. In general, the human data related to testicular effects are reported from accidental exposures or from males irradi- ated for therapeutic reasons, for example, as a treatment for testicular cancer. Following the administration of doses as low as 80 mGy (8 reds), a reduction in the number of spermatogonia occurs. At these low doses, the more resistant and more differentiated cells in the line may continue normal maturation. A reduc- tion in the sperm count may not be evident until 30 to 60 days have passed. The ultimate degree and duration of depletion of the sperm depend on the magnitude of the dose received. Sterility following irradiation is often a loosely applied term; it may repre- sent complete sterility, temporary subfertility, or in fact, fertility with a reduced number of sperm. Investigation of human volunteers indicates that administra- tion of 25 fractions of 150-200 mGy (15-20 red) daily may cause a decrease in the sperm count. It is of interest that in the few patients treated with radiother- apy to one testicle or to the inguinal nodes at doses of 600-2,500 mGy (60-250 red), a significant percentage were subsequently able to father children. Lush- baugh and Casarett (1976) have reviewed the literature and have found no case reports of malformed infants from parents who had received radiotherapy before the conception of the child. Five to six Gy (500 to 600 red) in a given dose will cause permanent steril- ity in most men; however, in some cases, that dose has been exceeded without causing permanent sterility. A dose of 2.5 Gy (250 red) may cause transient sterility for about 12 months. UNSCEAR (1982) has reviewed the available lit- erature and concludes that
AD VERSE REPRODUCTI HE OUTCOMES (1) temporary sterility is reported to occur with single doses to the testes ranging from 1.5 to 4 Gy (150 to 400 red) and with fractionated doses of 0.1-2 Gy (10 to 200 red), and (2) permanent sterility occurs with single doses rang- ing from 5 to 9.5 Gy (500 to 950 red) and with fractionated doses of 2-6 Gy (200-600 red). 47 A person may have transient sterility for a matter of years before fertility returns. Five individuals who received doses of 2.3-3.7 Gy (230-370 red) in the Oak Ridge criticality accident were aspermic for 4 months and hypospermic for 21 months; at least one individual experienced "sterility" for several years, but the individual subsequently had a normal offspring (Andrews et al., 1980~. Ortin et al. (1990) have reviewed data on 148 boys treated for Hodgkin's disease and followed for a median of 9 years. Sexual maturation was achieved in all boys without the need for androgen replacement. Of eight boys who were treated with radiation alone, three who received a pelvic dose of 40-45 Gy (4,000~,500 red) were able to father children. Three others who received 30- 44 Gy (3,000 - ,400 red) of pelvic radiation were oligospermic. This was in contrast to an 83% incidence of absolute azospermia in boys treated with chemo- therapy and no pelvic radiation. Patients who have received whole body irradia- tion before bone marrow transplantation have been studied by Sanders et al. (1986) and Deeg et al. (1984), who report that gonadal failure occurred in almost all boys who were postpubertal at the time of irradiation. Radionuclide Irradiation Irradiation of the testes by radionuclides may result from internal or external exposure. In the case of energetic beta radiation, the germ cells may be irradi- ated by external radionuclides. Although there may be internal deposition of radionuclides such as cesium, few data on the human sperm count or sterility are available. Although some radionuclides are known to be preferentially deposited in specific organs or tissues, such as radioiodine in the thyroid or strontium in bone, no radionuclide is known that is preferentially deposited in the testes. SPONTANEOUS ABORTIONS The World Health Organization (1970) has defined spontaneous abortion as "the non-deliberate interruption of intra-uterine pregnancy before 28 weeks (LMP) in which the embryo or fetus is dead when delivered." However, in many technically developed countries 22 weeks of gestational age separates spontane- ous abortion from stillbirth (Bracken, 19841. Spontaneous abortions occur much more frequently in early pregnancy, with most thought to occur before pregnancy is recognized. Spontaneous abortion rates as high as 80% of all conceptions have been suggested, but the most widely cited rates are 30-35% (Wilcox et al., 19881. Of clinically recognized pregnan- cies, approximately 15% will spontaneously abort. Second-trimester spontane- ous abortion occurs in 1-3% of all pregnancies.
48 AD VERSE REPRODUCTI HE OUTCOMES There is considerable heterogeneity of spontaneous abortion that varies by the gestational age of pregnancy. As many as 95% of very early abortions are of congenitally malformed embryos and a large proportion of these have chromo- somal abnormalities. Of the clinically recognized pregnancies that spontane- ously abort, 65% are chromosomally abnormal and include fetuses with devel- opmental anomalies. The risk of spontaneous abortion is higher with increased gravidity, and this may relate to such maternal factors as retroversion of the uterus, fibroids, pro- lapsed uterus, cervical erosion, and uterine incompetence. A septate or bicornate uterus will also increase the risk of spontaneous abortion. There is a reoccur- rence rate of about 1.6% which is found for abortions of both euploid and ane- uploid fetuses. Repeat spontaneous abortion (sometimes called habitual abor- tion) has also been shown to relate to both very early or very late age of first menstrual cycle. Other known or suspected maternal risk factors for spontaneous abortion are maternal smoking, which primarily increases the risk of abortion of euploid fe- tuses; maternal alcohol use; occupational exposure to lead, vinyl chloride, and solvents; exposure to antineoplastic drugs; cocaine use; and heavy maternal caf- feine use. Maternal infection with malaria, rubella, rubeola, herpes virus, cy- tomegalovirus, and genital mycoplasmas has been associated with an increased risk of spontaneous abortion. Recent use of oral contraceptives or a diaphragm has been related to a decreased spontaneous abortion risk. The frequency of spontaneous abortion was studied in atomic bomb survivors, but the data were internally inconsistent (Neel and Schull, 1991~. PRETER1\I DELIVERY Preterm delivery is a major cause of neonatal morbidity and mortality. Traditionally, prematurity was defined as birth weight of less than or equal to 2,500 g. Today, a distinction is made between low birth weight births and pre- term births. Low birth weight characterizes an infant who weighs less than 2,500 g (5 pounds 8 ounces) at birth, and preterm refers to a birth that occurs at a ges- tational age of less than 37 completed weeks (<259 days). The rate of preterm delivery in the United States is approximately 10% (Berkowitz and Papiernik, 19931. The rate has varied only slightly between 9 and 10% since 1970. Established maternal risk factors for preterm birth include African-American race, single marital status, low socioeconomic status (SES), previous low birth weight or preterm delivery, multiple second-trimester abortions, cigarette smok- ing, in vitro fertilization pregnancy, and such gynecologic or obstetrical compli- cations as in utero diethylstilbestrol exposure, cervical and uterine anomalies, gestational bleeding, and placental abnormalities. In addition, urogenital infec- tions, cocaine use, and no or inadequate prenatal care are probably associated with preterm delivery. Other factors such as age, parity, and maternal weight gain appear to be weakly associated or not associated with preterm delivery.
ADVERSE REPRODUCTIVE OUTCOMES 49 Still, the association remains inconclusive for other factors such as low prepreg- nancy weight, physical activity, and psychological stress (Berkowitz and Pa- piernik, 19931. The frequency of congenital malformations is higher in preterm than in full-term pregnancies (Placek, 1977; Hartikainen-Sorri and Sorri, 1989~. Standing for long periods of time at work (Teitelman et al., 1990) and a history of asthma and bronchodilator use during pregnancy have also been implicated as risk factors for preterm delivery (Doucette and Bracken, 1993~. A short interval between pregnancies has recently been reported to be a possible risk factor for delivery of a preterm, low birth weight infant and, potentially, as an important explanatory factor for the racial disparity in adverse pregnancy outcomes (Rawlings et al., 19951. Neither paternal nor maternal exposure to low dose radiation is considered to be a risk factor for preterm birth. There are reports that the number of pre- term births in the seventh to eighth month were increased in contaminated areas shortly after the Chernobyl nuclear power plant accident in 1986 (IAEA, 1991~. However, it is not clear to what extent this may have been due to such factors as maternal nutrition, stress, cigarette smoking, and alcohol consumption. No ef- fect of parental irradiation on mean birth weight could be demonstrated in survi- vors of the atomic bombs at Hiroshima and Nagasaki (Neel and Schull, 1991~. STILLBIRTHS AND NEONATAL DEATHS (PERINATAL DEATHS) Stillbirth is widely accepted to be a fetal death occurring at 28 weeks of gestation or later (from the last menstrual period ELMP]~. Because both stillbirth and early neonatal death (under 7 days of age) share in a number of causes, which are themselves somewhat different from causes of death in older neonates, the term "perinatal death" is often used to refer to both of them. In 1977, the World Health Organization recommended including all infants weighing at least 500 g or, if the birth weight is unavailable, infants delivered at 22 or more weeks of gestation as stillbirths. Occasionally, all deaths of infants delivered at 16 weeks or more of gestation are included as stillbirths. Perinatal (neonatal) mortality rate is typically measured as the number of deaths occurring within the first 28 days of life per 100,000 live births. Neonatal mortality has declined substantially in the past 50 years. In the United States in 1992, neonatal death rates were 537.5 per 100,000 live births; however this dif- fered significantly among black and white infants (rates of 1,083.1 per 100,000 live births and 434.6 per 100,000 live births, respectively) (DHHS, 1995~. The higher perinatal (neonatal) mortality rate (PMR) is seen in lower socio- economic (SES) groups, but it is difficult to disentangle the behavioral, social, and environmental factors that correlate with socioeconomic class. PMR is consistently higher in unmarried women and women who smoke, and PMR de- creases with parity. PMR increases with maternal age and variety of maternal diseases; including diabetes, blood group incompatibility, uterine infections, renal disease, and preeclampsia. Pregnancy complications include placenta pre
so ADVERSE REPRODUCTIVE OUTCOMES via and abruptio, pelvic anomalies, malpresentation, prolonged duration of labor, uterine rupture, umbilical cord complications, and fetal complications. These maternal factors are among those that should be taken into account in evaluating paternal influences on PMR. Results from the atomic bomb survivors show that high doses had no effects on PMR (Neel and Schull, 1991), so no effect would be expected at the much lower doses received by Atomic Veterans. BIRTH DEFECTS Conventionally, a birth defect (a congenital malformation) implies a failure of proper or normal morphologic development. Such failures can vary greatly in the threat that they pose to an individual's physical and mental integrity and survival. As a result, it is common to divide malformations into major and minor malformations although there is no absolute dividing line between these two classes of defects. Major birth defects include those failures of normal devel- opment that are incompatible with life (anencephaly, for example), that are seri- ously life-threatening (such as many congenital heart defects), or that materially compromise the individual's ability to function effectively in the society of which he or she is a member (cleft palate). All other birth defects are construed as minor. Birth defects, whether major or minor, are often further classified in two ways: (1) malformations arising from anomalies of a gene or a chromosome and (2) developmental disorganization occurring in an embryo or fetus with a nor- mal genotype. Human Chromosome Abnormalities Four subtypes of chromosomal abnormalities can be considered on the basis of whether the affected cell is somatic or germinal and whether the anomaly is structural or numeric. Many chromosomal abnormalities are incompatible with embryonic or fetal survival and are never expressed in the human phenotype at delivery, although some (e.g., trisomy 16) may be relatively common in early spontaneous abortion. Table 5 shows the prevalence of the more frequently seen chromoso- mal abnormalities in individuals born to parents exposed to ionizing radiation at Hiroshima and Nagasaki and to controls. The data were tabulated by Hook (1984) and were collected by Awa (for a recent summary of the findings see Awa et al., 1987, 19889. The offspring of exposed parents (in = 8,322) and control parents (in = 7,976) who were delivered between May 1946 and December 1984, are being studied at about age 13 years. This late age at examination screens out many of the chromosomal anomalies that might incur early mortality, and thus, the data provide reliable estimates on only two types of chromosomal abnormalities,
ADVERSEREPRODUCTIVE OUTCOMES 51 namely, sex chromosome aneuploids and balanced structural rearrangements, which confer relatively little threat to survival among liveborn infants. The data presented do not suggest any major differences according to exposure. How- ever, when the frequency of sex chromosome aneuploidy is examined in the context of the combined parental gonadal dose, there is a small but statistically nonsignificant increase in these anomalies with increasing parental dose (Neel et al., 1990~. Developmental Abnormalities Developmental abnormalities can affect any organ, and as a result, there are several hundred potential diagnostic categories. Early fetal lethality reduces the number of malformations seen at birth. The most frequent developmental mal- formations and their prevalence in those under age 1 year are given in Table 6 (Bracken, 1983~. The most commonly observed malformations at birth are of the ventricular septum of the heart; the majority of malformations occur at rates of <2 per 1,000 live births (Merlin, 19639. Overall, in some 3% of newborns a major congenital anomaly will be diagnosed at birth, and an additional 3-5% will be diagnosed with a major congenital anomaly in the first 10 years of life. If the Atomic Veterans each fathered an average of two or three children, the total would be about 500,000 offspring. The rates for birth defects given in Ta- ble 6 can be used to estimate the numbers of birth defects that would be seen among these children of Atomic Veterans in the absence of any radiation ef- fects. For example, spine bifida is routinely observed in 1.4 in 1,000 babies, which would be 700 among the infants in 500,000 offspring of Atomic Veter- ans. Among the offspring of Atomic Veterans there would be 5,000 infants with heart defects, and nearly 600 would be diagnosed with cancer during childhood. In total, one would expect to find 15,000 infants with major malformations among the children of Atomic Veterans if the rates among those children were the same as those in the general population. A subcategory of malformation, sometimes called Reformational, are those in organ systems that developed normally but that incurred a secondary deform- ity. Usually, this is due to intrauterine molding because of oligohydramnios and physical constraint (examples of these are some types of talipes, congenital hip dislocation, and midline cleft palate).
cd o · - cd ~ - ce o a) o x c~ a' s~ ct o 4~ o m U: - Ct · _ . _ ~_ ._ a' C~ C~ P~ t,V - o ,~ ~ ~0 O ~ ~ O _ Ct ~ c: ·_ O Ce O O ~ Z O C~ m O ~ .- V) e ~o "S ~ =i= ~ O C~ ._ _ Ct £ ¢ ~ ~r~ ~ ~ . .. . . . _ C~ oo o o ~o° ~ . -, ~-, ~ oo . . .. . . . _ C~ o- o o ~o° .. . . o ~ oC~ o CM ~ o o . . . . . . . o ~ o o o o ~o o o ~ ~ . . . . - , _ o ~ o o C~ C) C~ C) ._ .- C~ _ _ e ~_ ~C C) s~ o _ ~ ~Ct Ct ,= _ - ~ ~ ~ ~ ~ o £ x ° ~ ~ vo' v, ~ ~ 50 - Ct c~s tC P~ - a' ~ X X ~ ~ ° V' s: C ~ ~ ~ x ~ e X '~ ~ ° e e ~
~ - oo ~ ~ ~ ~ cr~ ~ 0 o o - o o ~o o c~ ~ ~ ~ ~ o ~ ~ o ~ ~ ~ c~ ~c~ ~ ~ ~ - ~ o o o o o -o o ~ ~c~ ~ ~ c~ - )~ ~- ~ ~ o o c~ o o ~o o cN u~ ~ oo ~ - ~ oC ~ ~ =} -} ~ -~ ~ o'^ =~-^ o o o o o c~o o ~ ~- - (~] ~ ~ - ~ . .. . .- O O O O O -O 0 ~ ~r,_ e c) - ~ V -~ c) ) , ct - ~ ~oo `~ n c~ ; ~, ~ O D ~ ~ =~ ~- E c: E 7 r_ :^ O cn Ct ._ Ct X O V: C~ C) ._ X V' C~ V, . :~ C~ _ ~ C~ ._ rl ~ O ,L, ~ 5 c' · ~ (~ u _^ 0 5 $~ O O _ -S O ct _ X 5: cs ~ v' v' cn O ~ v, c ~c~ - ~ oo cr~ - o o o . - o ~i o
54 AD VERSE REPRODUCTIVE OUTCOMES The range of risk factors studied for their effects on developmental malfor- mations is too large to summarize in any detail here. Several maternal infections (e.g., rubella, cytomegalovirus, toxoplasma, and herpes virus) have been associ- ated with an increased risk of malformations and may account for 2% of all mal- formations. Maternal diabetes is linked to a range of malformations. Numerous environmental substances have been studied (usually, inadequately) but with inconclusive results. Lead at relatively high doses is a behavioral teratogen, but it is not related to physical malformations. Mercury, especially in organic forms (methyl mercury), has some physical teratogenic properties, but it is primarily a neurobehavioral teratogen. Maternal alcohol and cocaine use are related to a range of malformations at moderate levels of exposure. Cigarettes and caffeine do not appear to be teratogenic at their usual levels of intake. Only a few phar- maceutical agents have clearly been shown to be teratogenic, including diethyl- stilbestrol, thalidomide, phenytoin, and some other antiepileptic drugs. Warfarin and some cytotoxic drugs have also been linked to developmental anomalies. Nutritional deficiency, particularly folate deficiencies, increases the frequency of developmental anomalies. These maternal exposures should be excluded in considering the birth defects attributable to paternal exposures. There was no evidence of an increased risk of congenital malformations following paternal exposure to ionizing radiation or in children of atomic bomb survivors (Neel and Schull, 19911. MATERNAL ILLNESSES The committee has interpreted periparturient diseases of the mother as dis- orders that can cause maternal and neonatal morbidity. Numerous chronic and pregnancy-related disorders in the mother may affect the mother's and neonate's health and well-being. These include infections such as cytomegalovirus, herpes simplex virus, and hepatitis; maternal cardiovascular disease including congeni- tal and rheumatic heart diseases; hypertensive disorders; pulmonary disorders such as pneumonia and asthma; diabetes mellitus, including gestational diabetes; hyper- and hypothyroidism; maternal adrenal gland disorders such as Cushing syndrome and Addison disease; collagen vascular disease such as systemic lupus erythematosus; renal diseases such as glomerulonephritis and nephritis; placenta previa, abruptio placenta, and retained placenta; hematologic disease including idiopathic thromobocytopenic purpura and Rh incompatibility; myasthenia gra- vis; and epilepsy and other necrologic diseases. None of these diagnoses, how- ever, are known to be related to irradiation of either parent.
AD VERSE REPRODUCTI HE OUTCOMES Table 6. Prevalence per 1,000 Live Births of Congenital Malfonnations by Diagnostic Group from Five Major Hospitals in Connecticut, 1974-1976 and Expected Prevalence at Birth for a Population of 500,000 NewbomsU Diagnostic Group Number per 1,000 Expected Number (All Diagnosed) Among 500,000 Births All malformations 57.0128,505 All neoplasms 1.16580 Hemangiomas and lymphangioma 2.061,030 Strabismus 0.87435 Heart block fibrillation, tachycardia 0.33165 Inguinal hernia plus obstruction 7.683,840 Anencephaly 0.41205 Spina bifida 1.40700 Hydrocephaly 1.40700 Eye anomalies 0.54270 Common truncus 0.21105 Transposition of great vessels 0.99495 Tetralogy of Fallot 0.70350 Ventricular septal defect 6.073,035 Atrial septal defect 0.50250 Heart valve 1.73865 Other heart anomalies 0.87435 Total heart anomalies 11.075,535 Patent ductus arteriosus 2.191,095 Coarctation of aorta 0.62310 Single umbilical artery 0.78390 Cleft lip palate 1.86930 Pyloric stenosis 8.014,005 Tracheal-esophageal fistula 0.45225 Other digestive system anomalies 1.69845 Undescended testicles 0.62310 Hypospadias 1.40700 Congenital hydrocele 1.57785 Other genital organ anomalies 1.03515 All talipes 3.301,650 Polysyndactyly 2.191,095 Limb reduction 0.66330 Congenital dislocation of hip 0.99495 Lower limb anomalies 0.21105 Skull and face anomalies 0.62310 Other muscular-skeletal anomalies 1.16580 Skin anomalies 2.621,310 Down syndrome 1.36680 Other autosome anomalies 0.58290 Other unspecified multiple anomalies 1.28640 55 " Data are based on malformations found among 24,224 live births in five Connecticut Hospitals over a 24-month period. SOURCE: Modified from Bracken (1983).
56 AD VERSE REPRODUCTI VE OUTCOMES ALTERED SEX RATIO When studies of the Hiroshima-Nagasaki atomic bomb survivors and their offspring began, it was believed that a person's gender was determined in a simple way. Individuals inheriting an X chromosome from their father were destined to be females, whereas those individuals who inherited a Y chromo- some from their father would be males. Thus, females would have two X chro- mosomes and males would have only one. These notions suggested, in turn, that when mutations in genes on the X chromosome induced by ionizing radiation are incompatible with survival (are lethal), their expression would be manifested differently in the two genders. More specifically, since a father normally transmits his single X chromosome to his daughters, if a lethal mutation were present on the X chromosome in the father's sperm, it could find expression only in his daughters. On the other hand, since mothers transmit their X chromo- somes equally to their sons and daughters, a lethal mutation might be expressed in either sex. If the mutation was dominant, the two sexes would be affected equally often; however, if the mutation was recessive, since the male has only one X chromosome, it would invariably manifest itself in males, but in females manifestation of the new mutant would occur only if the second X chromosome fortuitously carried a functionally similar gene. From this rationale the likelihood of a mutation increases as the dose in- creases. If the father were exposed, more female embryos would be lost, and the sex ratio (males/females) at birth would rise in proportion to dose. If the mother was exposed, more male embryos would be lost, and the sex ratio at birth would fall in proportion to dose. If both parents were exposed? the resulting sex ratio -or proportion of male births would be related to the individual parental doses and the frequency of dominant versus recessive lethal mutations. As can be seen, this theory of sex determination made fairly specific predictions that could be compared with the actual observations that were accumulating in survivors of the Hiroshima and Nagasaki atomic bombs. When the data from the years 1948 through 1953 were examined, it ap- peared that the proportion of male births, in fact, declined with dose when the mother was exposed and increased, albeit modestly, with increasing paternal dose (Schull and Neel, 1958~. The rate of change with dose was not, however, statistically significant, although it was in the direction predicted by the theory outlined above. For this reason, when the clinical phase of the studies ended, data on the sex ratio continued to be collected on the supposition that the rate of change might become statistically significant with further information. To this end, observations of the frequency of male births were continued through 1966. However, when these additional observations were analyzed, the results did not support the earlier findings; indeed, the modest changes seen were opposite those predicted by theory (Schull et al., 1966~. Today, the earlier understanding of sex determination is known to have been overly simplistic. First, it did not take into account the occurrence of X chromo
AD VERSE REPRODUCTI HE OUTCOMES 57 some aneuploids, as in the Klinefelter and Turner syndromes, which make the predictions less precise. The first of these aneuploids was discovered in 1959 (in unrelated studies), and soon thereafter many others were identified. As a result of these discoveries, it is known that it is possible for some females to have only one X chromosome (or even as many as five X chromosomes), and for some males to have two or more X chromosomes. Moreover, these individuals with abnormal numbers of X chromosomes are more frequent in most populations than one would expect following exposure to ionizing radiation, at least at low doses. Second, it is now known that in females only one of the two X chromo- somes within a cell is functionally active. This inactivation of one of the X chromosomes makes the prediction of the behavior of a potentially lethal gene on the X chromosome more difficult, particularly if the inactivation is not ran- dom (and it does not appear to be random). Given these developments, the sim- ple, early arguments are much less compelling, and prediction of the effects of lethal mutations on the proportion of male births more tenuous. Thus, the sex ratio is no longer considered useful in estimating of genetic risks following ex . . . . . posure to Ionizing radiation. MORTALITY AMONG THE CHILDREN OF EXPOSED PARENTS The largest study by far of the effects of parental exposure to ionizing ra- diation on mortality among their children conceived subsequent to irradiation is the study of the offspring of the atomic bomb survivors. That study involves the surveillance of approximately 72,000 children born alive between May 1946 and December 1984. It includes individuals whose parents were exposed at a wide variety of ages and doses. Surveillance of a cohort of this size is possible in Ja- pan because of the existence of a unique record resource. Since the latter part of the 19th century, the Japanese government has maintained an obligatory system of household censuses, known as the koseki. These censuses, which are under the jurisdiction of the Japanese Ministry of Justice, incorporate information on all events that affect the composition of a family, such as birth, death, adoption, and marriage. It is thus possible to determine the life status of an individual wherever that individual may reside in Japan if the location of the individual's household record is known. Customarily, on a cyclic basis, the koseki of the cohort are inspected anew to identify deaths that may have occurred since the last review cycle. If an indi- vidual has died, it is possible to obtain a copy of the death certificate and deter- mine the stated primary cause of death and secondary contributors through the regional health centers of the Ministry of Health and Welfare. Follow-up is vir- tually complete; the vital status of more than 99% of individuals in the cohort can be determined. On those rare occasions when vital status cannot be deter- mined, it is usually because of the migration of the individual to another country.
58 ADVERSE REPRODUCTIVE OUTCOMES Insofar as malignant tumors are concerned, the mortality data are supplemented with information in the tumor registries maintained by the Medical Associations of Hiroshima and Nagasaki with the assistance of the Radiation Effects Research Foundation. These registries date from 1957 (Hiroshima) and 1958 (Nagasaki). The causes of death or incident cases are routinely coded by using the current International Classification of Disease and the International Classification of Cancer of 1976. The first summarization of the mortality data occurred in 1966, and since then other publications have appeared, the most recent being in 1991 (Yoshimoto et al., 1991~. At that time, the age of the average living member of the cohort was 28.8 years, and some 80% had completed their nineteenth year of life. The average dose received by the parents, father and mother combined, was 430 mSv (43 rem) on the basis of a neutron of 20 radiobiological effectiveness (RBE). It should be noted that the assumption of an RBE of 20 is not critical to what follows since the results described here are not materially affected by the choice of any other RBE between 1 and 20. Deaths were divided into those at- tributable to cancer and those attributable to other causes. Emphasis in the analysis of these data was placed on the 67,586 liveborn children when the pa- rental doses could be computed by using the DS86 system (Shimizu et al., 1992) of dosimetry; however, a second analysis was based on all 72,228 children. That analysis used imputed doses received by the parents of the 4,642 children for whom direct calculation of the parental DS86 dose was not possible. The two analyses did not differ appreciably in their results. Briefly, the findings were as follows. With regard to cancer, among the 26,894 liveborn children whose parents received DS86 doses of 10 mSv (1 rem) or more, there were 40 cases of cancer (1.5 per 1,000 births) and 14 instances of benign or unspecified tumors (0.5 per 1,000 births), and among the 40,692 chil- dren whose parents were not exposed or who were exposed to less than 10 mSv Sv (1 rem), there were 75 cases of cancer (1.8 per 1,000 births) and, again, 14 cases of benign or unspecified tumors (0.3 per 1,000 births) (Yoshimoto et al., 1991; see also Yoshimoto et al., 1990~. Statistical analyses revealed no signifi- cant increase in the frequency of cancer deaths (cases) with increasing parental dose either for all cancers combined or for leukemia, the most common of the childhood cancers, which were considered separately. This was also true when the data were restricted to those individuals whose fathers alone were exposed. However, examination of the data suggested that only 3-5% of the tumors of childhood that were observed are associated with an inherited genetic predispo- sition that would be expected to exhibit an altered frequency if the parental mu- tation rate were increased. In other words, the mutational component was small. With regard to causes of death other than cancer, 3,709 deaths occurred among the 67,586 individuals in the cohort with known parental DS86 doses (54.9 per 1,000 births). Again, the frequency of such deaths did not increase in a statistically significant manner with increasing parental dose. This was also true
AD VERSE REPRODUCTI HE OUTCOMES 59 for four major disease categories, namely, infectious and parasitic diseases, dis- eases of the respiratory system, diseases of the digestive system, and certain conditions originating in the perinatal period. However, if the data on all dis- eases except neoplasms are taken at face value, there is a small but nonstatisti- cally significant excess relative risk with increasing dose. If this excess relative risk of 0.038 per Sv (per 100 rem) for death by age 20 is accepted as real and not ascribable to chance, the increased risk of death per 0.01 Sv (per 1 rem) is 0.00038. To the extent that these values are applicable to the U.S. population, they imply that in 1983 when the probability of a liveborn child dying by age 20 in the general U.S. population was 0.0200, if the father had been exposed to 100 mSv (10 rem) this probability would be about 0.0238 (National Center for Health Statistics, 1986~. Put somewhat differently, in 1983, when two individu- als in every 100 live births would have been expected to die by their twentieth year of life, if the father had been exposed to 100 mSv (10 rem), the number of expected deaths would be about 2.4. This increase is too small to be within the resolving power of present epidemiologic studies. CANCER AND LEUKEMIA IN PARTICULAR U.S., Japanese, Russian, and Italian scientists have reported transgenera- tional cancer in the mouse or rat after exposure of the parent either in utero or before mating to carcinogenic chemicals or ionizing radiation (reviewed by To- matis, 19941. Several years ago a report in the British media of an excess num- ber of cases of childhood leukemia in the village of Seascale near the Sellafield nuclear facility in West Cumbria, prompted a more careful case-control study to ascertain whether this alleged excess could be explained. The findings of the resulting study by Martin Gardner and colleagues supported the earlier allegation (Gardner et al., 1990 a, b; Gardner, 1992~. These investigators suggested that the excess was due to the relatively high doses received occupationally by the fa- thers of the patients. More specifically, they argued that paternal preconception exposure, in particular, exposure in the six months immediately prior to the con- ception of a child, induced mutations in sperm that resulted in the offspring de- veloping leukemia. Their findings stimulated a number of other studies aimed at confirming, if possible, their findings. Little or no support for the hypothesis that paternal exposure leads to an ex- cess risk of childhood leukemia has been found in studies carried out in the United States (Jablon et al., 1991), France (Hill and Laplanche, 1990), Germany (Michaelis et al., 1992), and Canada (McLaughlin et al., 19931. Moreover, the findings among children at Sellafield were significantly at variance with those among the children of the atomic bomb survivors (Little, 1990, 1991, 19921. A further complication in the acceptance of the findings and hypothesis of Gardner and colleagues has been the finding of similarly raised levels of leukemia around potential nuclear facilities in Britain (Cook-Mozaffari et al., 1989) and Germany
60 ADVERSE REPRODUCTIVE OUTCOMES (Michaelis et al., 19921. This illustrates anew the difficulties inherent in attempt- ing to interpret any cluster of cases. Nonetheless, the issue remains contentious, and at least one other multi-institutional case-control study has reported findings that suggest an increased risk of infant leukemia among the children of fathers exposed to diagnostic X rays prior to the conception of the child (Shu et al., 1994~. That study, however, could not confirm the effect of maternal exposure prior to conception that had been reported by Bithell and Stewart (1975), nor did its design permit the assessment of the dose-response relationship for exposed fathers. IMMUNE DEFICIENCY Various genetic abnormalities cause severe immunologic deficiencies in the offspring, and these deficiencies predispose them to infection and non-Hodgkin's lymphoma. Among these diseases are congenital agarnmaglobulinemia, ataxia- telangiectasia, Wiskott-Aldrich syndrome, Bloom syndrome, and Chediak- Higashi syndrome (Seibel et al., 19931. An excess of such syndromes, infections, or lymphomas has not been observed as causes of mortality in the F. generation among survivors of the · atomic bomb blasts in Hiroshima and Nagasaki (Yoshimoto et al., 19911. NEUROLOGIC DEFICIT, INCLUDING MENTAL RETARDATION Definitions of the terms "necrologic deficit" and "mental retardation" vary widely. This is perhaps inevitable when there is a continuum of intellectual or necrologic potentials. Still, it complicates the comparison and integration of the findings from different studies when the definition has not been the same among studies. For example, among the survivors exposed in utero to the atomic bombing of Hiroshima and Nagasaki, an individual was considered to be se- verely mentally retarded if he or she was unable to form coherent sentences, to perform simple arithmetic tasks, or to manage his or her own affairs or was insti- tutionalized. The World Health Organization restricts the term "severe mental retardation" to those individuals with an IQ of less than 50; individuals with IQs in the range of 50 to 70 are described as mildly retarded. Diagnosis of mental subnormality may rest on the clinical experience of the examining physician, on structured tests of mental performance, or commonly, on both. The frequency of the diagnosis depends on the severity of the handicap and the age, at least through the school years, at which the individual is exam- ined (Gesell and Amatruda, 19751. Subtle limitations of mental performance that are readily recognizable in the pubertal child may be very difficult to diag- nose in a newborn. This difference is well illustrated by the findings of the Col- laborative Perinatal Project conducted by the National Institute of Neurological
AD VERSE REPRODUCTI VE OUTCOMES 61 Diseases and Stroke. That study, which began on January 1, 1959, and ceased registration in December 1965, involved 15 university-affiliated medical centers and led to the study of 55,908 pregnancies (DHEW, 1973~. Although the study had a number of objectives, one of the major aims, as initially stated, was to "determine the relationship between factors in the perinatal environment and the continuum of human reproductive failure, with particular reference to the central nervous system for: (a) early manifestation of deficits (infancy and early child- hood), and (b) later manifestations of deficits (5 to 15 years)." Although this study remains one of the better sources of information on the frequency of congenital malformations, including necrologic deficits, available for the population of the United States, it also illustrates that, since a necrologic handicap is not a single disease entity and the potential causes of a deficit are many, it can be extremely difficult to recognize the cause in a specific instance It is known that inherited gene and chromosomal defects, as in Down syndrome or the fragile X syndrome, can lead to the occurrence of mental retardation, but attempts to estimate the contribution that genetic factors make to the overall fre- quency of mental retardation have generally been unsatisfactory. Most instances of mental retardation are idiopathic; that is, they have no clear identifiable cause. The situation is different, however, with respect to the exposure of a preg- nant woman to noxious agents, particularly early in her pregnancy, when unto- ward effects on the mental development of the child have been demonstrated Among those agents for which the evidence is most persuasive are ionizing ra- diation, methyl-mercury, lead, and alcohol. The effects of exposure in these instances can manifest themselves in a variety of ways: a possible increase in the frequency of overt mental retardation, a loss in performance on standard intelli- gence tests, and poorer performance in school. The findings, particularly with respect to ionizing radiation, have been reviewed in some detail in the 1993 re- port of the United Nations' Scientific Committee on the Effects of Atomic Ra- diation (UNSCEAR, 19931. These effects are not genetic in origin, although the possibility that individuals may differ in their sensitivities to noxious agents for genetic reasons cannot be rejected. The effects stem from the exposure of the developing embryo or fetus itself to the noxious agent at vulnerable stages in its development. Unfortunately, little evidence dealing with the risk of mental retardation among infants conceived following parental exposure to ionizing radiation is available. Clinical studies of the children of the atomic bomb survivors con- ceived after the exposure did not extend beyond the 10th month following birth and thus provide limited information. The data that are available from the study of atomic bomb survivors fail to demonstrate an increased risk of any form of mental retardation, including an increased risk of Down syndrome.