The following comments are offered on the basis that, although the arguments, assumptions and/or conclusions contained within may be flawed, in part or in whole, they may promote further discussion or deliberations that may have scientific value. All opinions are my own and do not reflect the views of any organisation or society with which I may be,or have been, affiliated.
These comments relate to Section 4.7 Eye
In Paragraph (678) the draft text assert that there is a compelling “weight of evidence” to support a lowering of the recommended acute threshold for the purposes of radiation protection to a nominal value of 500 mSv.
The assumption that this is prudent, based on the epidemiological studies that have been reviewed, may be premature for the following reasons, which are outlined below by type of study.
1 Infant Exposures
The Hall et al. (1999) study relates to infants of ages 0 – 18 months, with a mean age of 5 months and a median age of 4 months at the time of first treatment. Augusteyn 2007 [1] has developed an empirical formula for the increase in the (wet) weight of the lens of the eye, from which the rate of increase at any age may be found via differentiation. If the Augusteyn formula is valid then the rate of (wet) weight increase in the infant lens may be between 40 – 50 times greater than in the adult. If there are approximately linear relationships between rate of growth, rate of epithelial cell division, and tissue sensitivity, the Hall et al. study implies that the relative risk for an “adult” (actually age >4 years) per Gy would not be significantly different from unity, and would support the status quo.
2 Radiologic Technicians
The Chodick et al. (2008) paper has an impressive statistical methodology (and number of subjects), but is limited by the following:
i) the conclusion of an ERR/Gy of 2.0 (2sigma interval: -0.7, 4.7) is not statistically significant
ii) in deriving the ERR/Gy the radiation associated variable “age at radiotherapy to the head/neck” is adjusted for, but other radiotherapy variables are not
iii) the paper quotes references the Salisbury Eye Study which “suggested that the positive predictive values of self reported cataracts were 76 percent, i.e. the expectation is that only 76% of reported cataracts are actual cataracts. The claimed result may therefore be an artefact of subjects with higher lifetime occupational dose having a greater probability of falsely self-reporting a cataract.
iv) there was no question put to the subjects of the study on the number of occurrences of traumatic ocular injuries. Note that given the n = 35,705, and the composition of the group, it is not unreasonable to assume that of the order of 5000 – 10000 of these would have experienced one or more incidents of domestic violence prior to 2004.
v) estimating uv exposure by means of mean residential ultraviolet exposure is unlikely to provide even a qualitative distinction between differently exposed groups. Questions relating to average time spent outdoors, and the use of (sun)glasses, solariums, and/or hats would have provided better discrimination.
vi) Positive self report (“Yes”) of hypercholesterolemia (>240 mg/dl) was positively linked with cataract (with a HR of 1.49 for “Yes” respondents vs. “Never” respondants following multivariate analysis). However the self-reported rate at baseline appears to be low, allowing for age, and does not take into account the potential for subjects who may have subsequently been diagnosed with hypercholesterolemia between 1984 and 2004
3 A-bomb survivor data
The problem with A-bomb data (e.g. Nakashima et al. 2006) is that they only consider UV exposure from an occupational/lifetime perspective. They do not explicitly consider UV exposure coincident with the x/gamma/neutron exposure during an atomic bombing. A rough estimate assuming 5% of the energy of a 20kT nuclear weapon is converted into UV (A&B) and that the proportion of UVA to UVB matches that of an ideal black body radiator at 4000K suggests that a person at 2km from the epicentre may be exposed to 7.5E+05 J/m2 UVA and 7.5E+04 J/m2 UVB, within the first 0.25 seconds, if the epicentre was within their field of view. These values are greater than typical annual accumulated exposures (in Japan) by factors of approximately 2.5 for UVA and 15 for UVB. At such intensities, even indirect (reflected only) exposures may be significant.
I’d therefore expect a synergistic (detrimental) effect to occur from simultaneous UV and ionising radiation exposure, with the UV exposure causing damage directly, and enhancing the damage from ionising radiations by interfering with damage repair mechanisms within the eye.
In addition, although the adult lens is effective in blocking 300nm – 400nm UV, children under 5 years old have a window of transmittance with a peak (of 5% transmittance) at 320nm, which reduces to 0.1% by age 22 (Boettner and Wolter 1962 [2]). In fig 5 of their paper, Nakashima et al. present a graph of “Odds ratio per Sv for posterior sub-capsular opacity (PS) vs. age at exposure. The interesting feature of this is that the odds ratio for the group who were in their 20s at the time of the bombings is close to unity, but increases as the mean age of the group at exposure decreases. Now the growth rate of the lens of the eye for children >5 years old is postulated to be equal to that of the adult (in contrast to most other tissues). Assuming again that growth rate is linearly related to tissue sensitivity, if the age related PSC opacity curve were due only to exposure to ionising radiation, I’d expect that the odds ratio should be fairly equal between the 10-20 and 20-30 age groups, with a spike in the under 10 group. The downward age trend may be indicative of a uv mediated effect, related to the age related uv transmission characteristics of the lens described above.
Nakasima et al. also report “For in utero [at the time of bombing] subjects we observed no significant dose-response relationship in any lens change, probably because of the small number of subjects”. A possible alternative suggestion may be that their coincident UV exposure was nil.
Minamoto et al. (in press) [3], demonstrate that in addition to the intercity (Hiroshima vs Nagasaki) odds-ratio differences relating to cataract, there is also a difference in the location of cortical cataracts, with Nagasaki residents being more prone to cataracts in the inferior nasal portion of the lens. They conclude that this may be explained by differences in solar UV exposure. A possible alternative explanation may be related to the height at which the explosions occurred – the Hiroshima explosion was at ~600m altitude compared to 500m altitude at Nagasaki. A Nagasaki resident was therefore more likely to have experienced the UV exposure from the blast in their field of view. In either case, the difference in cataract location supports a UV influence in A-bomb survivor cataract incidence.
4 Chernobyl Cleanup Workers
The Worgul et al. (2007) is less thorough in it’s consideration of potential confounding variables, than the Hall et al. study: Neither hypercholesterolemia nor hypertension are considered, and these were found to have statistically significant hazard ratios in the Hall study. Smoking is considered, but without a measure of relative use. Alcohol use was not considered (although not found to be significant in the Hall study, there may be significant differences in “range of use”, between a predominantly female north American cohort and a predominantly male eastern European one).
However, potentially the largest confounding variable is the effect of the cumulative radiation exposure on the general health of the individuals, physical, mental, and (in its broadest sense) spiritual. With increasing dose, it is reasonable to assume an increasing frequency of physical health problems that may contribute either directly or indirectly to cataractogenesis; and not unreasonable to assume that an increasing frequency of mental health issues may also indirectly contribute.
4 Pilots
The Raffnsson et al. (2005) data relates to pilots (as a status) versus controls (Rekyavik residents). The calculations of the adjusted odds ratios relating to occupational doses lack clarity.
Corrections for UV were applied, with regard to sun bathing habits. The authors state “The UV exposure of pilots on board aircraft is minimal, according to measurements in the cockpit”. This relates to previous work by Diffey & Roscoe 1990 [4], in which they used “UV sensitive” film to record cockpit doses.
I’ve not seen the full Diffey paper, but note that “polysulphone” was used as “UV sensitive film” in a similar study in 1988. Polysulphone is UV-B sensitive, with a peak response between 297nm – 305 nm Beyond 305 nm there is a linear reduction in sensitivity to about 4% at 330nm. Hence, if this was used by Diffey, his study provides no measure of pilot UV-A exposure.
I understand that aeroplane windows are (or were) typically of composite acrylic composition, giving excellent UV-B absorption, but significantly poorer UV-A absorption (unless coated). Pilot occupational UV-A exposures may therefore contribute to an increased risk of cataract formation, particularly nuclear cataracts, since the majority of incident UV-A is absorbed in the lens itself.
The case vs control odds ratio is therefore not surprising, and may not require or indicate any significant contribution from ionising radiations.
TJR
Additional References
[1] Augusteyn R C. Growth of the human eye lens. Molecular Vision. 2007; 13: 252–257.
[2] Boettner E A and Wolter J R. Transmission of the ocular media. Invest. Ophthalmol. Vis. Sci. December 1962 vol. 1 no. 6 776-783
[3] Atsushi Minamoto, , Kazuo Neriishi, , and Eiji Nakashima. UV radiation may explain intercity difference for cataract in A-bomb survivors. Journal of Photochemistry and Photobiology B: Biology, In Press (as of 25/02/2011)
[4] Diffey BL, Roscoe AH. Exposure to solar ultraviolet radiation in flight. Aviat Space Environ Med 61:1032–1035; 1990.