Objectives The number of nuclear medicine procedures has increased substantially over the past several decades, with uncertain health risks to the medical workers who perform them. We estimated risks of incidence and mortality from cancer and circulatory disease associated with performing procedures involving the use of radionuclides.
Methods From a nationwide cohort of 90 955 US radiologic technologists who completed a mailed questionnaire during 1994–1998, 22 039 reported ever performing diagnostic radionuclide procedures, brachytherapy, radioactive iodine therapy, or other radionuclide therapy. We calculated multivariable-adjusted HRs and 95% CIs for incidence (through 2003–2005) and mortality (through 2008) associated with performing these procedures.
Results Ever (versus never) performing radionuclide procedures was not associated with risks for most end points examined. However, we observed increased risks for squamous cell carcinoma of the skin (HR=1.29, 95% CI 1.01 to 1.66) with ever performing diagnostic radionuclide procedures, for myocardial infarction incidence (HR=1.37, 95% CI 1.10 to 1.70), all-cause mortality (HR=1.10, 95% CI 1.00 to 1.20) and all-cancer mortality (HR=1.20, 95% CI 1.01 to 1.43) with ever performing brachytherapy, and for mortality from all causes (HR=1.14, 95% CI 1.01 to 1.30), breast cancer (HR=2.68, 95% CI 1.10 to 6.51), and myocardial infarction (HR=1.76, 95% CI 1.02 to 3.04) with ever performing other radionuclide therapy procedures (excluding brachytherapy and radioactive iodine); increasing risks were also observed with greater frequency of performing these procedures, particularly before 1980.
Conclusions The modest health risks among radiologic technologists performing procedures using radionuclides require further examination in studies with individual dose estimates, more detailed information regarding types of procedures performed and radionuclides used, and longer follow-up.
Statistics from Altmetric.com
What this paper adds
Although the number of nuclear medicine procedures has increased substantially over the past several decades, there is little information about the risks of radiation-related cancers and circulatory diseases to the medical workers who perform them.
Radiologic technologists who ever performed procedures involving radionuclides had modestly elevated risks for some health outcomes, including squamous cell carcinoma of the skin, female breast cancer, and myocardial infarction, compared to those who never performed these procedures.
Greater frequency with which these procedures were performed, especially before 1980, was positively associated with risks of most of these outcomes.
These findings provide preliminary evidence that exposure to radiation from procedures involving radionuclides, especially before 1980, may be linked with an increased risk of certain cancers and circulatory diseases.
The field of nuclear medicine, in which radiopharmaceuticals are utilised in the diagnosis and treatment of a wide range of medical conditions, has undergone dramatic changes since its recognition as an official medical specialty in 1971.1 In the USA, the number of nuclear medicine procedures performed increased from an estimated 7 million during 1980–1982 to 18 million in 2006.2 Between the early 1970s and early 2000s, the corresponding effective doses to patients increased fivefold to sevenfold, mostly attributable to an increase in relatively high-dose diagnostic cardiac procedures.3 Such procedures have resulted in enormous medical benefits to patients. However, exposure to radiation, especially from higher dose or frequently performed procedures using radionuclides, poses potential health risks to the medical workers who perform them.
In a large prospective cohort of radiologic technologists, we evaluated risks for incidence and mortality from cancer and circulatory disease in technologists who ever versus never performed procedures involving the use of radionuclides. To the best of our knowledge, this is the first epidemiological study to prospectively assess risks for incidence and mortality from cancer and circulatory disease in medical radiation workers working with radionuclide procedures.
Overview of study
The US Radiologic Technologists (USRT) Study is an ongoing collaboration between the US National Cancer Institute, the University of Minnesota, and the American Registry of Radiologic Technologists (ARRT). The study population and methods have been described in detail previously.4 5 Briefly, a search of the ARRT records identified 146 022 radiologic technologists who were certified for at least 2 years during 1926 through 1982 and resided in any US state or territory. The cohort is followed through a combination of annual recertifications with the ARRT and linkage with the Social Security Administration database to determine vital status. Those who are found deceased or presumed deceased are linked to the National Death Index (NDI-Plus) to ascertain causes of death. The study was approved by the institutional review boards of the National Cancer Institute and the University of Minnesota, and all participants provided written informed consent.
Information about cancer and circulatory disease incidence, work history as radiologic technologists, personal diagnostic and therapeutic radiation, and cancer risk factors were ascertained in mailed questionnaires (sent in 1983–1989, 1994–1998, and 2003–2005). We attempted to validate a subset of reported cancers with pathology reports and other medical records. Self-reported but unconfirmed cancers (other than brain cancer) were included in analyses due to high positive predictive values (70–100%).
Study population and follow-up
Eligible for the current study were the 90 955 of 126 628 living technologists (72%) who completed the second survey during 1994–1998. Incidence and mortality were evaluated separately due to differences in eligibility criteria, sources of data, and length of follow-up. Incidence analyses for cancers other than non-melanoma skin cancer (NMSC) were restricted to the 63 482 technologists who responded to both the second and third surveys and did not have a prior diagnosis of cancer other than NMSC through the second survey. Technologists who reported a diagnosis of NMSC by the second survey (n=5863) were additionally excluded from analyses of basal cell carcinoma (BCC) and squamous cell carcinoma (SCC) of the skin. Circulatory disease incidence analyses were restricted to the 63 182 technologists who responded to both the second and third surveys, and did not report a prior diagnosis of circulatory disease other than hypertension by completion of the second survey. Person-time for incidence analyses was calculated from completion of the second survey to the earliest date of diagnosis of any first primary cancer other than NMSC (for other cancer incidence analyses), any first primary cancer (for NMSC incidence analyses), any circulatory disease other than hypertension (for circulatory disease incidence analyses), or completion of the third survey (2003–2005).
Cancer mortality analyses were restricted to the 84 952 technologists who responded to the second survey and were cancer-free, other than NMSC at that time. Circulatory disease mortality analyses were restricted to the 86 700 technologists who responded to the second survey, and had no prior diagnosis of circulatory disease other than hypertension at that time. Person-time for mortality analyses accrued from second survey completion until date of death or 31 December 2008, whichever came first.
Incidence and mortality outcomes were chosen a priori according to the following criteria: cancers that have been consistently associated with radiation exposure (non-CLL, BCC and cancers of the female breast, thyroid, lung, brain and colorectum),6 other cancers associated with radiation in the USRT cohort (melanoma),7 cancers of specific concern to medical radiation workers (brain cancer, SCC),8 circulatory diseases that have been associated with radiation (stroke, heart diseases (including ischaemic heart disease and, more specifically, myocardial infarction)),9–11 and a common cancer not consistently associated with radiation exposure (prostate cancer) to check for the specificity of radiation-related associations. For specific ICD-10 codes for the outcomes evaluated,12 see online supplementary table S1.Performing procedures using radionuclides was evaluated in relation to incidence and mortality for all cancers combined, female breast cancer, melanoma, prostate cancer, lung cancer, colorectal cancer, leukaemia other than chronic lymphocytic leukaemia (non-CLL), stroke and myocardial infarction. Because of low fatality from BCC, SCC and thyroid cancer, only incident outcomes were evaluated. For all causes combined, all circulatory diseases, all heart diseases, ischaemic heart disease and brain cancer, only mortality outcomes were evaluated.
In addition to other questions relating to work history, technologists were asked to report in the second survey whether they ever performed selected medical procedures and, if so, how frequently (never or rarely, monthly, weekly, or daily) they performed these procedures during three defined time periods (before 1980, 1980–1989, and after 1990). Four types of procedures involving radionuclides were queried, including diagnostic radionuclide, brachytherapy, radioactive iodine therapy and other radionuclide therapy (no examples were given in the questionnaire of procedures that fall under this ‘other’ category).
Cox proportional hazards models were used to compute HRs with 95% CIs associated with ever, compared with never, performing procedures involving radionuclides. These models used age as the time scale (to control for age), were stratified on birth cohort (<1950, 1950–1959, 1960–1969, 1970+) to control for secular trends, and were adjusted for sex and ever/never worked with all other radionuclide procedures. Missing values were modelled using a separate indicator variable. We further evaluated risks in relation to the frequency (never, monthly, weekly, daily) with which they performed these procedures before 1980, 1980–1989, and after 1990. Linear trends were assessed by assigning ordinal values to each category level and modelling these values continuously, excluding participants with missing values, and those who did not work in that time period.
Covariates that were considered a priori to be risk factors for specific outcomes were evaluated as potential confounders (see online supplementary table S1 for a list of covariates examined by outcome). Data on alcohol consumption, cigarette smoking, body mass index, age at menopause, age at first childbirth, number of live births, family history of breast cancer, skin tone, hair colour and eye colour, were obtained from the second mailed survey. Education background was obtained from the third questionnaire.
Formal tests did not indicate a lack of convergence of the Cox models or departure from the assumption that hazards were proportional over the age timescale.
All statistical tests were two-sided and were performed using the SAS V.9.2 statistical software (Cary, North Carolina, USA).
Table 1 compares selected demographic and work history characteristics for the 90 955 radiologic technologists who were eligible for any of the incidence or mortality analyses by ever/never performed diagnostic radionuclide, brachytherapy, radioactive iodine therapy, or other radionuclide therapy procedures. The majority (77%) of participants were women. A total of 22 039 technologists (24% of the 90 955 respondents) reported ever working with procedures involving radionuclides, including 14 270 (16%) with diagnostic radionuclide, 6809 (7%) with brachytherapy, 11 153 (12%) with radioactive iodine therapy, and 5390 (6%) with other radionuclide therapy procedures. The distributions among technologists who ever and never performed any specific radionuclide procedures were generally similar according to the characteristics in table 1. Exceptions include somewhat higher percentages of male (28% vs 22%) and college-educated technologists (35% vs 29%) and technologists who ever performed fluoroscopically guided procedures (32% vs 23%) among those who ever versus never performed these procedures. Additionally, technologists born in the 1940s were somewhat more likely to have ever versus never performed brachytherapy (39% vs 33%), administered radioactive iodine (41% vs 32%), and conducted other radionuclide therapy procedures (41% vs 33%).
Because additional adjustment for the potential confounders (see online supplementary tables S1 and S2), as well as ever (versus never) performing fluoroscopically guided interventional procedures, had little influence on the HRs (<10% change in log-transformed HRs for specific radionuclide procedures after inclusion in models), results presented were based on minimally adjusted models (ie, adjusted only for sex and work with other nuclear medicine procedures and stratified by birth cohort). We found no associations between ever working with diagnostic radionuclide procedures and all-cancer incidence, total mortality, or mortality from specific causes; however, we observed a 29% increased risk for SCC incidence (HR=1.29, 95% CI 1.01 to 1.66) (table 2). Ever working with brachytherapy procedures was associated higher myocardial infarction incidence (HR=1.37, 95% CI 1.10 to 1.70), total mortality (HR=1.10, 95% CI 1.00 to 1.20), and mortality from all cancers (HR=1.20, 95% CI 1.01 to 1.43), but no associations were observed for the other outcomes examined.
Apart from an inverse association for female breast cancer mortality (HR=0.40, 95% CI 0.17 to 0.94), based on 10 deaths, no associations were observed for ever working with radioactive iodine therapy procedures.
While we observed no association between ever working with other radionuclide therapy and cancer or circulatory disease incidence, risks of increased mortality were observed for all causes (HR=1.14, 95% CI 1.01 to 1.30), female breast cancer (HR=2.68, 95% CI 1.10 to 6.51) and myocardial infarction (HR=1.76, 95% CI 1.02 to 3.04). Ever working with any radionuclide procedure was associated with increased all-cause mortality (HR=1.13, 95% CI 1.06 to 1.19) and lung cancer mortality (HR=1.37, 95% CI 1.10 to 1.70).
For the outcomes that were significantly associated with ever (versus never) performing radionuclide procedures, we further investigated risks associated with frequency of performing these procedures (never, monthly, weekly, daily) in three time periods (before 1980, 1980–1989, after 1990) (see online supplementary table S3). Positive associations were observed for technologists who more frequently performed the procedures before 1980 for SCC incidence with diagnostic radionuclide procedures, for all-cause mortality, all-cancer (except NMSC) mortality, and myocardial infarction incidence with brachytherapy, and for all-cause, female breast cancer, and myocardial infarction mortality with other radionuclide therapy procedures. All-cause and myocardial infarction mortality increased significantly with frequency-performed other radionuclide therapy procedures in 1980–1989. No associations were observed with frequency-performed radionuclide procedures in 1990 onwards.
On the basis of a relatively short follow-up from completion of the second survey (1994–1998) to completion of the third survey (2003–2005) for incident outcomes (median=9 years) and to the end of 2008 for mortality outcomes (median=13 years), results from this study suggest some potential modest adverse health effects associated with performing radionuclide procedures, particularly for three types of outcomes: SCC of the skin in technologists performing diagnostic radionuclide procedures, female breast cancer in technologists performing other radionuclide therapy procedures (this category excludes brachytherapy and radioactive iodine therapy), and myocardial infarction in technologists performing brachytherapy procedures and other radionuclide therapy procedures. All-cause mortality was also associated with ever performing brachytherapy, and mortality from all causes and all cancers (except NMSC) were related to ever performing other radionuclide procedures. Greater frequency with which these procedures were performed, especially before 1980, was positively associated with risks of most of these outcomes.
Few studies have evaluated health risks associated with repeated, chronic exposure to low doses of radiation, as experienced by radiologic technologists and other medical workers exposed to radiation. Although studies of nuclear medicine workers indicate that annual equivalent doses are well below the maximum recommended levels for individuals occupationally exposed to ionising radiation,13–16 risks associated with regular exposure to radiation at these lower levels over a prolonged period of time remain unclear. Prospective studies of occupational exposure to radiation among medical workers, including radiologists and radiologic technologists, have shown elevated incidence or mortality risks for leukaemia,17–22 lymphoma,19 cancers of the skin,19 20 pancreas,20 lung,19 20 breast18 23 24 and thyroid,25 plus ischaemic heart disease and cerebrovascular disease,18 particularly for those who worked in the earliest decades when doses to medical radiation workers were generally highest. Results from studies of general radiologists and radiologic technologists may not be generalisable to technologists who perform procedures in nuclear medicine, or other procedures involving radionuclides, and may receive higher cumulative radiation exposure owing to the wide range in photon energies of the radioisotopes used and the limited effectiveness of lead aprons in protecting against radiation from procedures involving higher energy photons, such as positron emission tomography.26 The increased risks of mortality from breast cancer and myocardial infarction associated with other radionuclide therapy procedures are largely consistent with studies of other medical workers, as well as studies of Japanese atomic bomb survivors that are based on a single acute exposure.11 27 An increased risk of circulatory disease was similarly observed in the National Registry for Radiation Workers but not the 15-Country Study of nuclear workers.28 29
However, the inconsistencies in our findings for breast cancer and circulatory diseases warrant some discussion. Although a relatively strong increased risk of breast cancer mortality was observed for technologists who ever performed other radionuclide therapy, no increased risks were observed for diagnostic radionuclide procedures, and radioactive iodine therapy was associated with an unexpected reduced risk of breast cancer mortality. Similarly, no increased risks of circulatory diseases were found in technologists who ever versus never performed diagnostic radionuclide or radioactive iodine therapy procedures, which would be expected to yield higher doses to the technologist compared with brachytherapy or other radionuclide therapy procedures. Although additional adjustment for a wide range of breast cancer and circulatory disease risk factors had a negligible influence on the results (see online supplementary table S2), it is possible that residual confounding by some unknown or unmeasured factor could have biased the results in the positive or negative direction.
While data on the radiation-related risks of specific types of NMSC are limited, studies of Japanese atomic bomb survivors and patients treated for tinea capitis have shown associations with BCC but not SCC of the skin.27 30 31 Our finding of an increased risk for SCC of the skin, but not BCC, related to ever performing diagnostic radionuclide procedures was supported by an observed increasing risk with increasing frequency with which these procedures were performed, especially before 1980. While similar excesses were not found in technologists who performed brachytherapy, radioiodine therapy, and other radionuclide therapy procedures, diagnostic procedures are generally performed with greater frequency, and technologists performing these procedures may have higher cumulative doses.
This was the first study of radiologic technologists to evaluate the relationship between performing procedures involving radionuclides and risks for major health outcomes. A key strength of the study was the prospective design, which minimised the potential for recall and selection biases. Because this study collected information on work history prior to cancer and circulatory disease incidence and mortality, any bias related to inaccuracy in self-reported work history would be non-differential by outcome, and would be expected to, if anything, bias the results toward the null as opposed to away from the null. Compared with other studies of medical radiation workers, the large number of female participants provided a unique opportunity to assess breast cancer risks. Finally, the detailed covariate data allowed for control of potential confounders.
The major limitation of this study was the lack of estimated or measured doses from the specific procedures examined. We had limited information from the second survey on the specific types of procedures performed and other factors related to the cumulative radiation exposure from these procedures during the years in which the participants trained and worked as radiologic technologists. The self-reported work history information was not validated, and our findings were not adjusted for exposure from other radiation sources apart from performance of fluoroscopically guided interventional procedures, which had a negligible influence on the results. The preliminary findings from this study require confirmation in studies with more detailed information on the specific procedures performed, radionuclides used, and radiation protection practices employed that could be used to estimate cumulative doses to the technologists, or in studies with objective measures of exposure (ie, from film badges), which would enable quantitative estimation of dose–response risks.
Another limitation of the study is the reliance on self-reported diagnoses of cancer and circulatory disease. However, radiologic technologists, like other medical workers, are expected to have a greater awareness of their own health status and to provide accurate medical information, compared with the general population. Previously, we reported that 89.2% of all cancers and 99.4% of breast cancers that were reported on the second survey were medically confirmed during a review of medical records.5 However, under-ascertainment of incident cancer and circulatory disease cases is a concern, as some of the participants who failed to respond to follow-up surveys may have died from their disease before having the opportunity to report their diagnosis. To address this concern, we conducted a validation study comparing self-reported incident cancers to cancers identified through linkage with four state cancer registries for a subset (approximately 25%) of the cohort that resides in those states.32 We found that self-report failed to identify 12% of cancers diagnosed between 1990 and 1998, but that the proportion of breast cancers missed was only 2.6%, and the proportion of prostate and colon cancers missed were each <5%. Ascertainment was lower for lung cancer (16%) and melanoma (22%). For outcomes associated with relatively low survival, we evaluated mortality risks in addition to, or in lieu of, incidence. While adjustment for some of the major outcome-specific risk factors did not appreciably alter the radiation-related results, it is possible that residual confounding by these or other unmeasured factors could explain some of the observed results. The lack of associations in the current study for leukaemia and cancers occurring in radiosensitive sites, including the brain, thyroid and colon, may reflect measurement error, the limited range of exposure, or inadequate follow-up time. The findings from this study should be interpreted cautiously, especially for those cancer outcomes for which we have few cases. Also, in view of the total number of associations examined, some of the findings could be due to chance alone.
In this study of US radiologic technologists certified before 1980, ever versus never performing procedures involving radionuclides was not associated with the majority of the cancer and circulatory disease outcomes that we examined. However, we found some evidence that performing these procedures was associated with modest elevations in mortality from all causes, all cancers other than NMSC, female breast cancer, lung cancer and myocardial infarction, and incidence of SCC of the skin and myocardial infarction; however, these findings should be interpreted with caution given the small number of cases for many of the outcomes of interest, among the other limitations of this study. Because per-procedure radiation doses and demand for many of these procedures continue to increase, the preliminary findings from the current study should motivate additional studies on the health risks to nuclear medicine technologists and other medical radiation workers who perform procedures involving radionuclides. Ideally, such studies will have longer follow-up and include more detailed exposure assessment, capturing the specific types of procedures and tasks performed, radioisotopes and protection practices used, and length of contact with and proximity to patients,26 33 as well as objective measures of dose, to enable quantitative estimation of radiation dose–response risks.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
- Data supplement 1 - Online tables
Funding This research was funded by the intramural programme of the Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of Health, Department of Health and Human Services, USA.
Competing interests None declared.
Ethics approval Institutional Review Boards of the National Cancer Institute and the University of Minnesota.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.