Early X-ray workers: an effort to assess their numbers, risk, and most common (skin) affliction

Objective To assess quantitatively the number of early X-ray workers, their risk of becoming a radiation victim, and their most common radiation-induced (skin) disease. Methods Information on professional life and occupational disease was retrieved from the Ehrenbuch, a book of honour containing biographies of 404 radiation victims, as well as member and congress lists of the German and US radiological societies, obituaries, books, articles, and the Internet. Results The estimated numbers of X-ray users in a medical setting in the US increased from about 300 to 600 in 1900–1903, in Germany from about 700 to 1200 during 1905–1908. The risk for a beginning user eventually to die from radiation was 1–2 % in these years, but up to 10–25 % in 1896. Data on 198 victims of fatal radiation-induced skin disease were collected. The incidence of the various stages of skin afflictions with a fatal outcome was characterized by very wide distributions. Conclusions After 1896, the radiation risk decreased very fast at first and more slowly thereafter to nearly zero in 1935. Many victims became quite old, partly because of the slower progress of tissue reactions at lower radiation doses, partly because of the success of often multiple surgical interventions. Main messages US and German X-ray users amounted to several hundreds to thousand in 1900–1908. The risk eventually to die from radiation was about 1–2 % during 1900–1908. After 1896, this risk decreased from >10 % to nearly zero in 1935. The incidence of subsequent stages of skin harm varied strongly in time. X-ray victims could become quite old, dependent on radiation dose and surgery. Electronic supplementary material The online version of this article (doi:10.1007/s13244-015-0457-2) contains supplementary material, which is available to authorized users.

a Lifespan shortening in France is somewhat larger than in Germany and USA (both p=0.03) b Standard deviation of the paired differences In 43 of the 182 cases the actual lifespan exceeded the projected value.
In conclusion, the average lifespan of a radiation victim was shortened by about ten years compared to the lifespan of a double in the general population. With respect to a no radiation using peer this difference might be somewhat larger.

Appendix 2 Estimation of the number of early X-ray users in the US and Germany
A complication in our application of the Chapman model is that the individuals belong to different professional groups with potentially different risks and different likelihoods of being a member of a radiological society, or of being present at one of its meetings.
To investigate the consequences of this we will use the mathematically simpler Lincoln-Petersen formalism, which should be sufficiently representative for the similar Chapman modification. As indicated before, nearly all X-ray users fall into one of three professional categories, that of medical doctors (md), suppliers, (sup), or technicians (tech). We will consider the situation in which the suppliers and technicians have a risk and a representation within societies or congresses which differ from those of medical doctors. Suppose the risk to become a martyr at some time is for a medical doctor f, for a supplier αf and for a technician βf, respectively. Assume further that the probability that a doctor will be a society member or Notice that Eq. A1 and Eq. A2 give the same result if γ = δ = 1, independent of the values of α and β. This of course reflects the basic condition for the validity of the model. But when γ and δ are not equal to 1, the outcome of Eq. C1 depends on all four parameters α, β, γ and δ. Equations A1 and A2 can be used to find the difference between the Lincoln-Petersen estimate and the true value for various combinations of α, β, γ and δ.
We will illustrate the calculation of the number of X-ray users for the US. The total number of (future) victims (N E ) living in each of the years 1900-1903 was estimated from biographical data (from the Ehrenbuch and other sources). For the US in 71% of the cases time data were available, and the so obtained (too low) estimate was corrected for missing data by scaling with (1/0.71), assuming the missing data had the same temporal distribution as the rest (may result in non-integers). The other two required numbers (N SE and N ST ) were obtained by counting in member lists.
The Chapman equations are ('var' stands for variance): The results are shown in Table A2 and Fig. A1; those for Germany in Fig. A2.
It is difficult to confirm our estimates, but a rough consistency check on the number of Xray workers in the US in 1903 is possible. Pitkin stated in that year "about one-third of the prominent operators and instrument dealers have hands which have been more or less severely injured" [36], and Hesse in 1911 reported that about 25% of those who developed a malignancy died [25]. Assuming that 100% of the 'röntgen-hands' developed cancer in the long run, the number of persons to die from the X-ray users working in 1903 is estimated as 609* ⅓ *1.0 *0.25 =51. The number of future victims actually at work in 1903 amounted to 48.
The good agreement must be partly fortuitous, as the uncertainties are considerable. For instance, the 'one third' of Pitkin is probably an educated guess, and the assumed fractions of 100% and 25% may be off, the first possibly being too high, the second too low, as Hesse had only a short period for observation, or too high if later improvements in surgical interventions prevailed.   Table A3.  1901 1902 1903 1904 1905 1906 1907 1908 1909 Persons

Year
When not all X-ray users in the reference region were a society member the extrapolation will be too low as an estimate of the number of X-ray users in the country; on the other hand, a higher density of users in the reference area than the country average would lead to an overestimation.
Especially in Berlin the density of society members seems to be very high (compare the two 'Fraction' rows). Considering the possible (but unknown) magnitudes of these counteracting effects, the extrapolations look compatible with the Chapman estimates, but no more can be said.   is for all scenarios the (identical) derivative of the parallel lines in Fig. A6 (left) times the risk to become a martyr (100*0.02=2). Without the arguments given above, scenario L+ might have come to mind when noting the correspondence between the initial peak in Fig. 3 and that in Fig.   A6 (right, red). Rejecting scenario L+, the sharp drop in Fig. 3 can only be explained if the risk decreased over time. Independent testimony for an initially very fast reduction of risk, as Figs. 3 and 4 in the main article suggest, comes from data on burns in patients and operators collected by Codman (Fig. A7). In his 1902 article he writes "The main reasons for such a decrease have been the bitter teachings of experience and the fact that the introduction of better apparatus has done away with long exposures and the close approximation of the tube."

Fig. A7
Cases of X-ray burns in patients and operators with known date of occurrence (n=86) collected by Codman [37] If our interpretation of Fig. 3 is correct, it seems that in 1896 or 1897 a large part of the critical (high) exposure must have taken place in a rather short time immediately after starting Xray work, because if the initial risk had stayed high for several more years, more persons who started in 1898 and 1899 would have become a victim. In later years, say after 1900, on the basis of Fig. 3 little can be said about the time that must have passed before critical harm was done.
From the biographies it appeared that in many cases the exposure could have lasted many years, but with a view to the gradual improvement of shielding it is likely that the first years contributed most to the cumulative dose.

Number per year
Year

Appendix 4 On the exposure of early X-ray workers
To obtain an impression of the dose the X-ray victims received, it is worthwhile to look Gy (temporary) to 7 Gy (permanent), dry shedding of the skin at 14 Gy, wet at 18 Gy, and ulceration at about 24 Gy. Such reactions are seen after hours to several weeks. Recovery depends on the dose and the size of the exposed area, and can last for months or even more than a year. Many of the earliest pioneers suffered such acute X-ray 'burns' in all these degrees of seriousness, for instance in self-experiments and demonstrations [37,38]. Very severe burns left the skin more vulnerable to further exposure. A slight complication here is that the unfiltered Xrays used by the pioneers were very soft so that their penetration was low and the dose in the skin possibly not uniform. So the surface dose may have been even higher than the ICRP dose for the same effect. Today noticeable acute tissue reactions in radiology are limited to patients undergoing very extended or suboptimally performed interventional procedures.
Of more importance in the present context, however, are the long term effects that usually take (much) more than a year to become manifest and which may include skin cancer. Again according to ICRP 118, persons exposed to an acute dose of 10.5 Gy have a risk of 1% to suffer from skin atrophy, after a dose of 17 Gy the risk has risen to 50%. Thus a single severe radiation accident can already have fatal consequences in the long term by (deterministic) tissue reactions.
Corresponding doses by protracted exposure are 40 Gy (1%) and 69 Gy (50%). We assume that these latter ICRP values, determined from radiotherapy treatments of 30 fractions, are applicable to chronic exposure as might have existed in the case of the pioneers. Atrophy may be accompanied or followed, especially at higher doses, by various other distressing skin conditions, often called 'chronic dermatitis' [39]. The hand showing chronic dermatitis is known as a 'röntgen hand'. The ICRP values show how high (40 Gy-70 Gy and possibly higher) the skin dose of many (chronically) exposed X-ray victims must have been. Higher doses led to more serious afflictions and to a faster appearance [40]. Skin irritating chemicals used for developing photographic plates or films may have worsened damage caused by radiation. Effects on deeper lying organs were much less frequently noticed, among others because the X-rays had low penetrating power.
ICRP 118 doesn't specify the full spectrum of late effects, probably because they no longer occur in a radiation protection context. We therefore cite from Cole's description of the more severe forms of chronic radiodermatitis [41]. Pigmentation ('tanning') is often seen as a first signal of exposure, already after low doses. After higher doses, however, pigmentation may completely and permanently disappear, as may the hair and dermal appendages like sebaceous and sweat glands. Often telangiectasia develops between a few weeks and about a year but it may progress for years. Injuries to the nails like longitudinal ridges and brittleness, or (often temporary) loss of nails are seen. Keratoses, raghades, callous patches and ulcers may form.
Excruciating and untreatable pain is another debilitating consequence. Chronic dermatitis that remains unhealed is believed always to lead to cancer [27,42]. An impressive description of his experience with X-rays is given by (the later martyr) Pitkin in a poetic but in the end grim story [36]. About pain he writes: "For a description of the pain and suffering, hyperaesthesia and paraesthesia, no language, sacred or profane, is adequate." In order to understand how early radiologists incurred the above mentioned skin damaging doses of X-rays one should look at 1.) their knowledge, 2.) their experimental circumstances and 3.) their behaviour. Several books provide information on these topics [6-15], but also two articles by Ratkóczy should be mentioned [43,44].
1.) Scientific knowledge of X-rays hardly increased for a long time after Röntgen's original publication of December 1895. Moreover, the available knowledge was not helpful in understanding the X-rays' effect on tissues [19]. Notwithstanding this limitation a lot of practical experience was acquired. 'Burns' (acute dermatitis) were observed within months in 1896 [45].
Severe cases, even with necrosis (or gangrene), followed in 1897 [38]. By 1902 it was also known that cancer could be induced [46]. These observations did not directly lead to appropriate measures to limit exposure. One important reason may have been the disagreement on the cause of the biological effects. The majority of users believed for quite some time, at least to after 1900, that electrical effects originating from the electrically activated tube and wires were responsible for the harm (an excellent discussion is given by Serwer [19], Chapter 2). This believe led to placing earthed metal foils between tube and user or patient. Although the metal attenuated and filtered the X-rays to some extent, it gave no adequate protection. Several other causes were considered as well, idiosyncrasy was believed to play an important role as several operators had never contracted burns. Static high voltage generators were first (end1896) claimed not to 'burn' [47], later for several years believed to be more safe than induction coils [38]. These false beliefs caused unnecessary exposure. Interestingly, even after conclusive proof had been given that X-rays were the causing agent of the damage, the radiological community did not accept this immediately, as it was perceived as against longstanding experience [19]. All this probably delayed the development of protection by shielding from X-rays by a heavy material like lead, which was to become a key factor for more safety.
2.) Initially in 1896 the X-ray tube was completely bare, radiating in all directions, although it was later known that behind the anti-cathode the level of radiation was lower [38].
The primary X-ray beam produced a lot of secondary radiation by scattering from surrounding walls and other objects, which also exposed the photographic plate, fluoroscope, operator and patient. For improvement of image quality soon a diaphragm was introduced to limit the contribution of secondary radiation to the image. This also provided some protection for operator and patient. The introduction of the compression diaphragm (1903) still further reduced scatter, giving better image quality and a little more protection. Considerable reduction of the skin dose was achieved by using X-ray beam filters, e.g. of leather [48] and later aluminium [49]. To give an example, an Al-filter of 0.5 mm thickness reduces the air kerma of a 50 kV X-ray beam coming from a tungsten anti-cathode, with a 45 degree angle, only filtered by 1 mm glass of the tube wall and at 20 cm distance, from 2.67 mGy/mAs to 1.33 mGy/mAs, thus by a factor of 2.0.
For 2 mm glass this reduction is 1.5 (=1.13/0.77) (estimates made with PCXMC, STUK-Radiation and Nuclear Safety Authority, Helsinki, Finland). The anode material at the time was platinum, with atomic number Z=78, while our PCXMC-estimates are for tungsten (Z=74). The X-ray output scales approximately with Z, thus was slightly higher than our estimates above. The filtering turned out also to improve image quality.
Rollins, a very prudent and ingenious dentist, started from 1898 onwards to place the tube in a box covered with white lead paint (white lead was an isolator, a metal, e.g. Pb, could not be brought close to the tube as it might facilitate discharges between anode and cathode) [50,51]. was an X-ray tube made of leaded glass, except for a small window of normal glass for the exit of the X-rays. Later lead sheet was used more abundantly in the form of screens, between patient and observer, at the sides of the patient forming a box intercepting secondary rays, in movable shields and in the walls of cubicles for observers [44]. Lead foil or leaded rubber was used to shield the patient outside the X-ray field of view. 3.) Because the source of the X-rays, the gas tube, was notoriously unstable, X-ray systems in 1896 must have required endless experiments to establish conditions that gave useable X-rays. As no harm was expected from exposure to X-rays, protection was not an issue.
The worker would often stay close to the functioning X-ray system as can be seen on many early photographs. Burns were initially considered as no more dangerous than ordinary burns by heat or the sun. To gauge the quality of the radiation coming from the tube, the operator would use his hand as test object in front of the fluoroscope and close to the tube, while the other (also bare) hand held the fluoroscope (this was a second tube test, next to visual observation of the gas discharge). The whole body was in the primary beam. Fluoroscopy of patients was also a high exposure activity, as the operator was standing in the rays transmitted through the patient and in the scattered radiation coming from everywhere. And at moments where he was not in the patient's shadow he was exposed to the full beam. But in a discussion in 1903 Kassabian remarked that the number of dangerous fluoroscopies had decreased in comparison with the first years [36]. Assisting patients (e.g. children) during the making of radiographs, or holding the photographic plate, were other risky endeavours. Innumerous demonstrations with X-rays formed still another occasion for high doses because often the operator's body was used as test object. This could be for visitors at fairs or in shops, but also for clients in the case of manufacturers and salesmen of X-ray systems. Also in the production process of X-ray tubes the exposure could be very high, as there was a lengthy evacuation in an oven during which the functioning tube had to be observed. The heat served to drive residual gases out of the glass wall and metal parts. Some improvement in the stability of the gas tube was obtained by various forms of automatic vacuum regulation (1896 and later). In 1904 a phantom ('osteoscope') that could be used as a substitute for a hand in gauging radiation quality was produced by Max Kohl from Chemnitz (Germany) [53].
The reasons X-rays after 1903 still caused skin cancer fatalities (about 37% of all skin cancer deaths) are probably multiple. In arbitrary order: lack of education, economic reasons as both safety bringing hardware and a lower efficiency by prudent behaviour had its costs, falling into or continuing unsafe practices as radiation was not felt, ordinary slackness, large workload, extensive fluoroscopy, extreme professional zeal. The latter is a recurring characteristic in the (often somewhat hagiographic) biographies. Although the number of X-ray victims is large, fortunately it involved only a relatively small fraction of all workers.