Send letter to journal:
Re: Response to Chris J Kalman

June 8th 2009

We are writing to respond to Dr.Kalman in regard to our article entitled “Maternal Occupational Exposure and Risk of Spontaneous Abortion in Veterinary Practice”.

Regarding the risk relating to radiation, we presented data separately for those with 1-4 x-rays a week in Table 2. Readers are able to use these analyses rather than those in Table 3 if they prefer. The results for exposure to radiation as a continuous variable are also presented in the text for readers who require more detail (results section, page 3, column 1, line 16).

In regard to the way data were presented, particularly regarding exposure to waste anaesthetic gases, we suggest that this is a different situation to the dose-response variable describing x-rays a week. We categorized the variable surgery/gas scavenging using both two and three categories, but in the table we used three categories in order to disentangle the effects of surgery and waste anaesthetic gases. The results for the variable surgery/gas scavenger using two categories were reported in the text (results section, page 3, column 1, line 6).

This study, as documented in the abstract, only reports an increased risk of spontaneous abortion in veterinarians who performed more than five radiographic examinations per week. However, we suggest that it would be wise to advise veterinarians of the possible risk due to radiation even at low doses, particularly for young females. It has long been established that x-rays can have harmful effects, including sterility, miscarriage, mutagenesis and carcinogenesis and, while empirical evidence of harm is difficult to measure at low doses, the theories of radiation and cancer biology have to led to widespread adoption of the ALARA principle (as low as reasonably achievable).

Yours sincerely,

Adeleh Shirangi on behalf of all authors

Send letter to journal:
Re: Response to authors

Dear Sir

I am grateful for the speedy response from Dr Shirangi and her colleagues to my original letter. I regret the delay in my further response.

Importantly, the authors justify the combination of the ‘no exposure’ and ‘one to 5 films per week’ categories (ie the elimination of an exposure category) on the identification that the crude relative risk data did not indicate any adverse outcome for the taking of one to 5 films. I would, of course, agree that had there been such an adverse outcome, it would have been inappropriate to carry out this combination. The question however centres around the justification of this action when, in fact, there is demonstrated beneficial outcome which reaches statistical significance. There is, of course, significant publication around considerations of low dose radiation effects, including hormesis, so that lack of biological plausibility would not appear to offer such justification.

In addition, it appears that this treatment of the data is not standardised in the paper. It is noted, for example, that in the category dealing with surgery/gas scavenging, the crude data is in 3 exposure categories, and there is no adverse outcome between the first 2 categories. Notwithstanding this, there appears to have been no move to amalgamate these 2 groups in the multiple logistic regression.

I think an important role of Occupational Physicians is to be able to comment on and advise workers in relation to research. This appears a difficult task to do in relation to a study which shows the greatest risk of spontaneous abortion to be in those who are not subject to any x-ray exposure, with a statistically significant reduction in risk for those taking one to 5 films but, nevertheless, produces an abstract and conclusion based around an apparently demonstrated radiation risk and advice to warn workers.

Yours sincerely

Dr Chris Kalman MBChB FRCP (Glas) FFOM Director/Consultant Occupational Physician

Send letter to journal:
Re: Response to Chris J Kalman

We are writing to respond to Dr.Kalman in regard to our recent article entitled Maternal Occupational Exposure and Risk of Spontaneous Abortion in Veterinary Practice.

Dr. Kalman believes that he has picked up an anomaly in the presentation of data on radiation. We would like to reassure readers of the journal that there is no anomaly in the presentation of data and all presented results including radiation are correct. The presented results for radiation in Table 2 are the crude risks. The presented results in abstract are derived from Table 3 which indicates adjusted relative risks. In Table 1 the radiation exposure has been categorized into three levels of exposures (No exposure, 1-5 films/w, >5 films/w). The crude relative risk for radiation exposure from Table 1 indicated that taking 1 to 5 films did not have any adverse outcome which has been mentioned in the results section, paragraph 4, column 2, page 2. Therefore, in Table 3 to estimate adjusted relative risks using logistic regression, we combined the first two exposure categories (No exposure and 1-5 films/w) as the reference group (less than 5 films/w) and (more than 5 films/w) as the exposure variable. A similar grouping has been used in another study in female veterinarians with similar results (Reference No.5). So we have not eliminated the unexposed group from logistic regression analysis as postulated by Dr. Kalman. Furthermore, in Table 1, if we combine the first two exposure categories in one group (as the reference group), we will also see an increased risk of spontaneous abortion in those who exposed to radiation for more than 5 films per week compared with those who performed 5 or less films including unexposed group (crudeRR:1.38, 95%CI:0.94, 2.04 ).

Dr. Kalman also raises the issue that there are no measurements presented for the dose of radiation. Our study was a questionnaire-based study, and unfortunately we were not funded to obtain actual measurements. In the absence of measurements, we used the self-report of the number of X -rays taken as a proxy for dose. In some ways, this may have been a better measurement of the number of X-rays undertaken at the time of pregnancy as it would be virtually impossible to obtain measurements for what may have been a different job many years previously.

Ionising radiation at high doses is historically known to be embryotoxic and teratogenic. Therefore, we suggest that it is always wise to continue to inform female (and male) veterinarians of the potential risks in this profession.

Send letter to journal:
Re: Maternal Occupational Exposure and Risk of Spontaneous Abortion in Veterinary Practice

Shirangi and her colleagues set out to examine the relationship between occupational exposures and spontaneous abortion in female veterinarians. One exposure examined, is in relation to the use of x- rays, where the authors report in the abstract that veterinarians who reported performing more than 5 radiographic examinations per week had a statistically significant elevated risk of spontaneous abortion compared with those performing 5 or less (OR 1.82, 95% CL 1.17:2.82). This leads onto a conclusion that female veterinarians should be fully informed of the possible reproductive effects of ionising radiation.

Looking at the text of the article, the authors reference that spontaneous abortion has been linked to occupational exposure to radiation, although their consideration is limited to studies of female veterinarians and there is no mention of the large amount of occupational data on radiation workers, or radiobiology in terms of deterministic or stochastic effects. In relation to the study design, there is no consideration of measured radiation dose or radiation protection practice, with consideration limited to the crude division of 3 classes of practice: those not exposed to x-rays at all, those indicating between one and 5 films per week, and those undertaking more than 5 films.

Interestingly, although not commented upon in the text, it appears that there is a statistically significant reduction in relative risk from those not exposed at all, for those undertaking one to 5 films (0.57 (0.40 to 0.79)). The risk for over 5 films is also less than that for no exposure, although this does not reach to statistical significance. Given a situation where it is clear that there is a higher crude risk of spontaneous abortion amongst the unexposed than the exposed, one much question why further analysis in terms of the logistic regression was undertaken, and how such further analysis revealed the stated result. The first question must be one for the authors, although the answer to the second appears to be based on the elimination of the unexposed group from the subsequent analysis. Table 3 indicates only 2 exposure categories, namely 5 or less films, or over 5 films per week.

Those providing Occupational Health support to radiation workers in the UK were in the forefront of establishing ethical standards for the provision of prepublication information to workers on the results of studies which they participated in. Shirangi and her colleagues do not report their arrangements for providing such information. It is however difficult to understand how a cohort of workers of the intelligence and training of female veterinarians would not have picked up this apparent anomaly in the data at worker briefing sessions.