Review of NASA's Longitudinal Study of Astronaut Health (2004)

Because the Longitudinal Study of Astronaut Health (LSAH) began in 1992, more than three decades after the first astronauts were selected, the study’s database was rapidly populated with a wealth of retrospective data. As a result, it was possible to query the data early in the study’s course. The first peer-reviewed publication appeared in 1993 (Peterson et al., 1993). It reported astronaut mortality from 1959 through 1991. This chapter summarizes the findings of that publication, two other peer-reviewed papers from later in the 1990s, and more recent analyses of morbidity and mortality that National Aeronautics and Space Administration (NASA) scientists provided to the committee in meetings held in 2003.

Hamm et al. (Hamm, 2000) updated Peterson’s figures but followed the same sample for seven more years. During that time, six members of the 295 astronaut sample died. Therefore, the numbers will vary according to which paper is quoted.

Although some published reports utilizing the LSAH are included, the committee concentrated on the organization, goals, and function of the LSAH rather than a critique of the methods and analyses, which has already passed peer review.

The space radiation environment was a major concern for NASA from the earliest days of the space program. Space travelers are exposed to radiation that is different from that to which terrestrial workers, such as those in the nuclear power or nuclear weapons industries, are subjected. A National Academy of Sciences panel was asked to develop radiation protection guidelines and identify biological responses for human exposure to space radiation (National Academy of Sciences, 1967), and NASA has maintained a database on space (and medical) radiation exposure for all astronauts since Project Mercury began in 1959. It is not surprising that the initial analysis of the LSAH data (Peterson et al., 1993)

focused not simply on mortality but also addressed the hypothesis that astronauts are at increased risk for malignant neoplasms.

The study looked at the medical records of the 195 astronauts selected between 1959 and 1991 and found that 20 deaths had occurred during the 32 years surveyed. Sixteen were due to spacecraft (8), aircraft (7), or automobile (1) accidents; 2 were due to circulatory disease; 1 was the result of a malignant neoplasm; and 1 was due to unknown causes. Standardized mortality ratios (SMR) based on the U.S. population, adjusted for age, race, gender, and calendar year, were significantly increased for all-cause deaths (SMR=181) and accidental deaths (SMR=1,346). The crude accidental death rate of 445 deaths per 100,000 person-years for the 12 occupationally related deaths was an order of magnitude greater than the 34 to 41 per 100,000 typical of the mining industry, although the SMR for all accidents was comparable to that reported for Canadian airline pilots in another study (Band, et al., 1990). The hypothesis that astronauts are at increased risk for cancer mortality compared to the U.S. population was not supported, although the relatively young age of the astronauts, the low doses of radiation during space flight, the modest interval between space flight and data analysis, and the small sample size all made statistical confirmation unlikely. Space radiation doses varied directly with mission duration (r=0.99), and average mission doses ranged from less than 0.1 milliGray (mGy) for Mercury astronauts (average mission duration of less than 1 day) to 43 mGy for Skylab astronauts (average mission duration of 57 days). The average dose for astronauts on the first 43 shuttle missions was 1.3 mGy. For each of the 13 astronaut classes examined, the average per capita dose of radiation from diagnostic medical X-rays exceeded that from space travel—in most of the earlier years by a factor of 10 or 12.

This study (Hamm et al., 1998) focused on cancer mortality in a slightly larger population of astronauts than that of the Peterson et al. study of 1993, and it included an LSAH comparison group of JSC civil servants matched to the astronaut group on age, sex, and body mass index (BMI). The observed rate of cancer mortality in each of those groups was also compared with the age- and sex-adjusted cancer mortality rates of residents of Public Health Region 6 of Texas. All 3 cancer deaths among astronauts through 1995 were in males, so the investigators chose to confine their analyses to the 210 male astronauts selected through that date. The 618 civil servant comparison participants constituted all the males in that group, and the data from the Texas general populations included only males as well.

Both the astronaut group and the Johnson Space Center (JSC) comparison group had lost three members to cancer. The cancer mortality rate of the astronauts appeared increased compared to that of the comparison group (SMR =

345; 95 percent confidence interval (CI) = 66 to 756), but the apparent increase was not statistically significant. Both groups showed much lower than expected rates of death when compared to residents of Texas Public Health Region 6 (SMR for astronauts = 47; CI = 10 to 105; SMR for comparison group = 17; CI = 4 to 38). For the comparison group the difference was statistically significant. The cancer types causing the astronaut deaths were undifferentiated carcinoma of the nasopharynx, glioma, and metastatic melanoma. The comparison group fatalities were due to metastatic melanoma (2) and glioblastoma multiforme. The lack of a significant difference between the astronaut and control groups is again not unexpected, given the small number of cases, the short duration of the astronauts’ space experience (mean of 12.6 days), the low dose of space radiation (mean of 1.65 mGy), which is not significantly different from the background radiation, and the relatively young age of both groups. Additionally, the cancers found are not clearly linked to ionizing radiation. The comparisons to cancer rates in the general population are also not too surprising, given the substantially higher levels of education, income, general health, and fitness that characterize the astronaut group and their JSC comparisons. Employment itself is well known to be associated with lower mortality rates than those of the general population—“the healthy worker effect” (Fox and Collier, 1976).

The statistical analyses were confined to cancer mortality, but the report alludes to a preliminary review of the medical records that indicated that there had been at least 21 nonfatal cancer cases among the astronaut group and at least 6 cases in the comparison group. Non-melanoma skin cancers accounted for 17 of the 21 astronaut cases and 3 of the 6 comparison cases.

This study, published by Hamm et al. in Aviation, Space, and Environmental Medicine in 2000, updated mortality data still further, but it was primarily devoted to describing the study design and baseline data from the initial health evaluations of all the participants, (i.e., both comparisons and astronauts), who entered the study between 1959 and 1991. The baseline data—demographic, behavioral, and physiologic—are important indications of how closely the comparison participants might match the astronauts in initial propensity for disease.

Not surprisingly, the ages and body mass index of the comparison participants closely approximated those of their astronaut counterparts (see Table 2-1). Caucasians comprised 94 percent of the astronaut group and 90 percent of the comparison group. Women comprised approximately 11 percent of each group. Other demographic data reported were marital status (84 percent of each group were married) and education. All astronauts have at least a bachelor’s degree at selection, and 77 percent of those had a graduate degree as well. Only 36 percent of the comparison participants had an advanced degree (p = 0.001) and 6.6 percent had less than a bachelor’s degree.

T-tests were used to compare the astronaut and comparison groups on the substantial number of measures derived from baseline physical examinations and clinical laboratory tests (see Appendix B for a full list). Astronauts had significantly lower pulse rates, systolic and diastolic blood pressure, hemoglobin, and serum triglycerides, and significantly higher blood glucose. Seventy-nine percent of the astronauts had uncorrected visual acuity of 20/20 or better, and none had worse than 20/150, while only 55 percent of comparisons had 20/20 vision and 23 percent of comparisons had acuity worse than 20/150. No other statistically significant differences were identified.

The report also notes that eight comparison participants (1.6 percent) had controlled hypertension, and one had borderline hypertension at selection. Two of the comparison participants had diabetes at selection. These are disqualifying conditions for astronaut selection, so there were no cases of either hypertension or diabetes in the astronaut group at selection.

Twenty-six of the astronauts and 14 of the comparison participants had died at the time the Hamm et al. (2000) report was written. Table 2-2 shows the causes as well as relative risks.

Just as the Peterson (1993) report found, the only cause of death found to be significantly different between the two groups was injury and accidental deaths. The astronauts are clearly at a greater risk of accidental death than are the comparison participants. Eight of the accidental deaths among the astronaut group were due to two spacecraft accidents. Five deaths among this group were due to crashes of high-performance military aircraft, and three were due to accidents involving commercial or private aircraft. One astronaut died of altitude sickness and exposure to cold. One astronaut and one comparison died in car crashes. The other accidental death among the comparison population was due to a firearm.

Cancer mortality still appeared greater in the astronauts, but the apparent difference between the groups was not statistically significant.

In January 2003 and again in March 2003, NASA scientists from the Space and Life Sciences Directorate at Johnson Space Center briefed the Institute of Medicine (IOM) Committee on its analysis of LSAH data concerning overall mortality and the three clinical conditions that the committee was asked to review (cataracts, cancer, and thyroid function).

James Logan summarized the LSAH data on mortality from all causes for the committee (Logan, 2003). As in the earlier reports summarized above, overall mortality has been significantly higher for the astronaut group. Logan’s presentation in January 2003, just prior to the loss of the space shuttle Columbia and its crew of 7, reported 29 deaths among the 312 astronauts in the LSAH database and only 17 deaths among the 912 comparison participants. Accidental deaths, including 8 in spacecraft losses (3 in the Apollo fire and 5 in the Challenger explosion), accounted for 20 of the astronaut deaths (versus only 2 in the

comparison group).¹ The groups did not differ significantly in mortality from any other cause.

Men accounted for 27 of the 29 (93 percent) of astronaut deaths and all 17 of the comparison group deaths. The 12 astronaut accidental deaths that were not spacecraft-related included 4 in T-38 jet trainer crashes, 4 in private plane crashes, 1 in a commercial plane crash, 1 each in car and motorcycle crashes, and 1 while mountain climbing.

Frank Cucinotta reported to the IOM committee (Cucinotta, 2003) that an optometrist who had performed annual eye examinations of astronauts for more than a decade told the LSAH staff in 1998 that he had seen numerous lens opacities among the astronauts, possibly even more frequently than in his private practice. The LSAH staff immediately initiated an investigation by the medical staff and radiation group, which resulted in the publication of a detailed analysis

by Cucinotta et al. (2001), followed by recommendations to reduce radiation exposure (see Box 1-1).

Astronauts have had eye examinations since the beginning of NASA, but at varying frequency during the early years. Prior to 1977, examinations were performed at the JSC flight clinic. Between 1977 and 1988 they were done by referral physicians in the Houston area, and the results were collected by JSC. Since 1989 they have again been performed at the JSC flight clinic. Forty-eight lens opacities have been found in 295 astronauts. No measurements of astronaut exposure to ultraviolet radiation have been made, but each astronaut except for those on the initial four Mercury flights has worn a thermoluminescent dosimeter badge while in space.

Cucinotta and his colleagues used those badge data to calculate a lens equivalent dose for space radiation. They then developed an extensive database on the exposure of these astronauts to radiation from both diagnostic medical X-rays and occupational aviation. They sorted the astronauts into high-dose and low-dose groups and computed relative hazard ratios for cataracts at age 60 and at age 65.

Table 2-3 shows a significant increase in cataract risk for astronauts in the high space lens dose group for all cataracts and nontrace cataracts. Hazard ratios using lens dose from medical X-rays alone and from aviation alone were not significant. There was a significant association between cataracts and high-inclination or lunar missions, where a much higher flux of heavy ion radiation occurs. Ninety percent of the 39 cataracts occurring after space flight were in astronauts on such missions.

It has long been suspected that exposure to solar particle events, galactic cosmic rays, and trapped protons and electrons would increase the lifelong risk of developing cataracts. Early studies of cancer patients suggested a dose threshold for cataracts of about 2 Gray, but these data from Cucinotta et al. (2001) showed that astronauts with exposures above 8 milliSieverts developed cataracts more frequently and at an earlier age than those exposed to less than 8mSv (Table 2-3). Space radiation has higher linear energy transfer than terrestrial radiation, and this study reported exposure in tissue penetration rather than surface measurement (so the absorbed dose (Gy) and the equivalent dose (mSv) would be approximately equal).

NASA has contracted with Dr. Leo Chylack, an ophthalmologist at Harvard, for a five-year follow-on study comparing the prevalence and rate of progression of cataracts in astronauts with those in a group of current and former military pilots matched to the astronauts for age and gender. Digital photography and computerized image analysis will ensure comparable, objective, and quantitative measurements for all subjects (Cucinotta, 2003).

The Committee considers this work an excellent example of the potential value of the LSAH, but it notes that the impetus for the study was the anecdotal report of more frequent cataracts by an examining doctor with no direct tie to the LSAH. When this suspicion was reported, 48 cases of lens opacification were subsequently culled from the data for 295 astronauts, and the problem was referred to the radiation and space groups at JSC for more detailed study.

Because of the known association of some cancers with radiation exposure, surveillance of astronauts for malignancies was planned from the beginning of the LSAH. Craig Fischer briefed the committee on the comparison of cancer incidence among the astronauts (Fischer, 2003), the LSAH comparison participants, and an age- and sex-matched sample of the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) database. Fourteen cases of cancer were diagnosed among the 312 astronauts followed from 1959 to the present. This is 59 percent higher than the comparison group per person/year (not statistically significant), but 46 percent lower per person/year than the SEER data (statistically significant). The distribution of the astronaut malignancy types is shown in Table 2-4. The prostate is the predominant cancer site in both the astronaut and comparison groups.

Thirty-three cases of basal and squamous cell carcinomas of the skin were excluded from the analysis of the astronaut group, and 27 were excluded from the 912 member comparison sample, an almost threefold difference in rate. A rationale offered for this deletion is that astronauts spend significant time outdoors for both training and recreation, but this is without supporting data and is inconsistent with the professed goal of assessing the overall risks of being an astronaut. The higher than threefold increase in prevalence in astronauts is statistically significant. Including these non-melanoma skin cancers with all other cancers would make the overall difference between the astronaut group and comparison group significant as well.

Kathleen McMonigal briefed the IOM committee on NASA’s discovery of the adverse effects of iodinization of space shuttle drinking water (McMonigal, 2003; McMonigal et al., 2000). Although the role of the LSAH in detecting these thyroid function abnormalities was indirect, it is relevant to the discussion of LSAH design and execution.

In 1990 a female astronaut was diagnosed as hypothyroid by her flight surgeon. In the course of their discussion, the surgeon noted that he had seen several such cases in astronauts, but they were young men, in which hypothyroidism is rather unusual. Daily thyroid hormone replacement therapy was prescribed for the astronaut, and she subsequently flew two more shuttle missions without difficulty. The thyroid problem resurfaced in 1997 when Dr. McMonigal was monitoring a terrestrial project in which life support systems were being tested by 4 astronauts in isolation for 30, 60, and 91 days. She was told that there might be a problem with the iodine concentrations in the water and was asked to check the thyroid function of one of the astronauts in who had participated in the previous 60-day test. He had started with a high normal thyroid stimulating hormone (TSH) level and subsequently became hypothyroid during the test.

The physician discovered that the iodine levels in the test subjects’ drinking water was 5 milligrams per liter and the concentration of iodine in the water on the previous shuttle missions had ranged from 3 to 4 mg per liter. The decision was made to test the new group of 4 euthyroid astronauts at 30 days, and if thyroid function abnormalities were found, reduce the iodine concentration. The recommended daily allowance (RDA) for iodine is about 0.15 mg, but these subjects were ingesting more than 10 to 20 mg in drinking water alone. The danger of ingesting such doses has been known for many years. The Wolff-Chaikoff effect (high doses of iodine block the organification of thyroglobulin) was described more than half a century ago (Wolff et al., 1949), and the Jod-Basedow effect (potential hyperthyroidism after iodine administration) more than a century ago. All four astronauts in the test showed marked elevations of TSH. Anion exchange resin filters were then installed at the tap, which lowered the iodine concentration to 0.25 mg per liter (approximately a 16-fold reduction), and thyroid function measures returned to normal during the period of observation. Two of the eight test subjects studied showed functional abnormalities, one hypothyroid and one hyperthyroid, which returned to normal in about a year.

The episode described above prompted a review of the LSAH database for evidence of thyroid abnormalities in the astronaut population, because high concentrations of iodine had been the bacteriocide of choice on U.S. space vehicles to that point. Analysis of postflight TSH levels indicated that a transient but statistically significant elevation over preflight levels on the day of return was common. By the next annual physical examination, TSH values had typically

returned to preflight levels. Mean preflight TSH level for all 134 astronauts for whom data were available at the time (1998) was 2.84 microInternational units per ml (µIU/ml) (SD=2.24). Upon return, mean TSH was 3.43 µIU/ml (SD=2.59). SKYLAB astronauts (n=9) remained in orbit an average of 57 days, and they had much higher TSH levels immediately upon return (8.44+/− 6.94) and three days after return (8.56 +/− 3.70), although the levels eventually returned to preflight levels (4.89 +/− 3..65). U.S. astronauts who flew on the Russian space station MIR for periods of 2-6 months (n=6) showed no significant increase in TSH upon return. Mir’s drinking water was treated with silver nitrate rather than iodine. Postflight elevations in TSH are also not seen in data from the 79 astronauts who have flown on the space shuttle after iodine levels in the drinking water were reduced to 0.25 mg/liter in 1998.

The LSAH data were also searched for eight ICD codes related to clinical thyroid disease. Thirty-nine cases were uncovered by this 1998 search. Nine male astronauts (3.8 percent of all male astronauts) and 3 female astronauts (8.3 percent of all female astronauts) were found to have clinical thyroid disease, along with 18 male (2.5 percent) and 9 female (7.8 percent) comparison subjects. The overall odds ratio comparing astronaut incidence to comparison group incidence was 1.39 (95 percent CI=0.69 to 2.78). For males only, the odds ratio was 1.56 (CI=0.67 to 3.63). For females only, the odds ratio was 1.07 (CI = 0.27 to 4.18). None of these comparisons was statistically significant, nor were further tests for association of disease with estimated iodine intake during mission, gender, or role (pilot or mission specialist).

This page intentionally left blank.