Part 1. Cancer Production by Prenatal Exposure to Ionizing Radiation
In an important communication in Lancet, 1988, Yoshimoto, Kato, and Schull reported as follows (p.665):
"This study examines the risk of cancer (incidence) over 40 years among the in-utero exposed survivors of the atomic bombing of Hiroshima and Nagasaki, and adds eight years of follow-up to a previous report confined to mortality. Only two cases of childhood cancer were observed among these survivors in the first 14 years of life; both had been heavily exposed. Subsequent cancers have all been of the adult type. Not only did the observed cancers occur earlier in the 0.30+Gy [30+ rads] dose group than in the 0 Gy dose group but also the incidence continues to increase, and the crude cumulative incidence rate, 40 years after the A-bombing, is 3.9-fold greater in the 0.30+ Gy group."
This finding of a significant elevation in cancer rate in those irradiated in-utero is important. However, the total incidence of cancers is numerically small, 18 cases in all, with 16 of the 18 occurring in adults. This suggests, according to the authors, that:
"These results, when viewed in the perspective of fetus doses, suggest that susceptibility to radiation-induced cancers is higher in prenatally than in postnatally exposed survivors (at least those exposed as adults). However, definitive conclusions must await further follow-up studies" (p.665).
It is very early in this study. By 1984 (which was the closing date of the Yoshimoto study), those exposed in-utero were just about forty years old. The cancers are largely yet to come. There were three breast-cancers in the series of 18 cancers in toto, and all three of them were in the ostensibly unexposed category. In view of the random fluctuations of small numbers (and three cases make a severe "small numbers problem"), and in view of what is already known about age-years 0-9 from larger parts of the same study (see Chapter 3, for example; also Gofman 1990), the distribution of the three breast-cancer cases in Yoshimoto 1988 is almost certainly due to pure chance, with no biological meaning. We consider it highly reasonable to assume that breast-cancer sensitivity is about the same for irradiation in-utero as it is for irradiation at 0-9 years of age. Of course, we will continue to observe the on-going follow-up, of the in-utero cohort of A-bomb survivors, to ascertain whether the assumption is strenghthened or weakened by continued observation.Part 2. Proportion of Infants Receiving X-Ray Doses in-Utero (U.S.A)
The most common reason that pregnant women receive x-ray examinations which expose their infants, is to determine whether or not the women will be able to deliver the infants vaginally. The x-ray examinations sometimes take place during labor itself. According to Kevin Kelly and co-workers (1975):
"Roentgenographic evaluation of the relative sizes of the fetal head and maternal pelvis has been used clinically almost since the advent of medical radiography. The technique was considerably refined by Colcher and Sussman in 1944. Further refinements and variations have been instituted since that time."
Pelvimetry was a routine topic in medical textbooks of the 1930s and 1940s. For example, from Christopher's Textbook of Surgery, Third Edition (1942), we quote Dr. James T. Case, Professor of Radiology at Northwestern University Medical School (p.1635):
"Roentgen measurements of the pelvic diameters can be taken without any special apparatus; the ordinary x-ray equipment of a hospital should be satisfactory. Space here does not permit description of the technic of pelvic and fetal mensuration (see the author's description in Curtis: Obstetrics and Gynecology, Phila., W.B. Saunders Co, 1932, vol.3, p.762)."
(a) Frequency of In-Utero Irradiation: MacMahon's Study
Dr. Brian MacMahon (1962) published a paper entitled "Prenatal X-Ray Exposure and Childhood Cancer." Our major interest here is in his evaluation of the frequency of prenatal exposure to x-rays.
The study-population consisted of 734,243 children born in, and discharged alive from, any of 37 large maternity hospitals in the northeast United States in the years 1947-1954. These hospitals were located in the nine states comprising the Northeast Region of the United States as defined by the U.S. Bureau of the Census. All but three of the hospitals were situated in Massachusetts, Rhode Island, Connecticut, or New York City. The frequency of intrauterine x-ray exposure in the population was estimated in MacMahon 1962 by review of the records of a 1 percent systematic sample.
MacMahon acknowledged that births in these hospitals were not representative of the general population of births in the area, since the included hospitals were not rigorously representative. For example, a special effort was made to obtain the collaboration of three hospitals in which the use of x-ray pelvimetry was believed to have been high, and in general, large hospitals were approached. We shall deal with the bias of this selection-process in a moment.
The MacMahon study concerns x-ray examinations of the maternal abdomen and pelvis in which the fetus received essentially direct whole-body exposure. The majority of such exams are pelvimetry examinations, but flat-plate examinations for twins, placentography and series-studies of the intestinal tract or urinary tract are also included.
The systematic sample comprised 7,346 live births, in which 104 occurred in multiple births and 7,242 in single pregnancies. X-ray of the maternal abdomen or pelvis was recorded in 32 (30.8%) of the multiple and 770 (10.6%) of the single pregnancies. In his Table 2, MacMahon lists 9.9% of all the female children as x-rayed, out of 3,570 cases observed. This is the percentage of special interest to us. For personal reasons, some readers may be interested in MacMahon's finding that first births showed a much higher rate of x-raying than did second and later births. Moreover, this finding has important implications for arranging proper control groups in health studies of in-utero radiation.
MacMahon obtained no estimates of in-utero doses in this series of cases. For doses, we must search elsewhere (Part 3).
(b) Frequency of In-Utero Irradiation: Kelly's Study
Kevin Kelly and colleagues (1975) addressed the question of "The Utilization and Efficacy of Pelvimetry." They reported the following (p.66):
"This study analyzed clinical information from 67,078 single deliveries of 1,000 grams or greater from 16 hospitals [in the years 1969 and 1970]. Pelvimetries were performed during 6.9 percent (4,599) of these deliveries ..." Later (p.68), they reported that pelvimetry accounted for 72% of the pelvic and abdominal x-ray procedures in their study. Thus, the rate of prenatal irradiation was (6.9% / 0.72), or 9.6% --- in remarkably close agreement with the MacMahon findings for a period some 20 years earlier. Like MacMahon, Kelly et al noted that their study-population was not a perfectly random sample, due to some geographic, economic, and racial factors. Also, all of the hospitals providing data were training institutions.
Our Adjustment Downward in the Frequency of X-Raying
For Dr. MacMahon's purposes (comparing frequency of prenatal X-raying in children having a malignancy, versus frequency in children having no malignancy), seeking out a few hospitals with a high frequency of pelvimetry is not a serious matter. For our purposes, such bias is undesirable. In order to avoid overestimating the frequency of prenatal irradiation, we shall simply reduce the total frequency of x-raying to 75% of the total observed by Dr. MacMahon. This should take care of the possible consequence of MacMahon's selection-process. So:
(9.9%) x (0.75) = 7.4 % of live births had pelvimetry or abdominal radiation.Part 3. The Question of Radiation Dose to the Fetus in Such Studies
Robert Berman and Benjamin Sonnenblick (1957) have addressed the issue of doses received by the fetus and female pelvis (1957, p. 4):
"The purpose of this communication is to record actual measurements in roentgens of radiation dosage directed to the depths of the female pelvis during the exposure of films for x-ray pelvimetry and hysterosalpingography." Measurements were provided for a total of 28 women, 13 of whom had pelvimetry films taken at or near 38 weeks of gestation.
The total intrapelvic dose of x-ray radiation, as measured in the posterior vaginal fornix in 10 patients who were exposed for the full exam (4 pelvic views), had an average value of 2.9 R, with 6 patients receiving less than 3 R and only one receiving more than 4 R. The total range of doses was 2.1 to 4.4 R.
We shall use the Berman-Sonnenblick estimates for dose, and the MacMahon estimates (adjusted downward by us) for frequency of the examination.Part 4. Preparation of the Dose Estimate for the Master Table
- Item 1: In the Master Table we have an entry (in Column A) of 905,213 female infants in the 0-1 age-year group, nationwide. This number is approximately the number of female live-births in the average year of 1920-1960.
- Item 2: We shall use the reduced value from the MacMahon studies as the frequency of examinations irradiating the fetus (pelvimetries plus abdominal examinations). That value is 7.4 % of live-births per year.
- Item 3: Number of infants who received the radiation in the 905,213 live-births = (905,213 births) x (0.074) = 67,000, rounded off, per year.
- Item 4: Person-rads per year = (67,000 persons) x (Radiation Dose). For dose, we shall evaluate the two extremes of Berman and Sonnenblick: 2.1 and 4.4 Roentgens. Since this is medical x-radiation, and since these are depth doses measured in Roentgens, we use the approximation here that person-Roentgens are equivalent to person-rads.
Person-rads = 67,000 persons x 2.1 rads = 140,700 person-rads.
Person-rads = 67,000 persons x 4.4 rads = 294,800 person-rads.
The other approximation we make here is that the smaller component in the MacMahon frequency (the abdominal radiation) delivers about the same fetal dose as the dose delivered by pelvimetry.
- Item 5: We wish never to overestimate radiation dose, so we shall accept the lowest of the person-rad values obtained in Item 4, namely 140,700 person-rads.
- Item 6: For the Population Exposure in the Master Table, we need to distribute the person-rads into the entire 905,213 persons:
Population Exposure (in-utero) = 140,700 person-rads / 905,213 persons
= 0.16 rad per fetus.
This conservative value, reduced in Item 2 and reduced again in Item 5, is entered into Column K of the Master Table. For this singular entry, we can use the table's first row, because (as noted in Part 1) we will use the same conversion-factor, from dose to breast-cancer, as we use for the year after birth until the first birthday.
In view of Kelly's comment (Part 2), we can assume pelvimetry was in common use for the entire 1920-1960 period.
# # # # #