Does radioactive iodine treatment in hyperthyroidism increase risk for cancer?
Yes. Ionizing radiation is an established human carcinogen and is the most well-established cause of cancer in humans, apart from smoking.
Radioactive iodine therapy targets the thyroid gland, but it also inadvertently exposes other organs and tissues to low to moderate doses of radiation.
The Cooperative Thyrotoxicosis Therapy Follow-up Study (CTTFUS), initiated in 1946 when radioactive iodine therapy (RAI) was first introduced into practice, is the oldest and largest cohort study of patients treated for hyperthyroidism. About two-thirds of patients were treated with RAI with or without surgery or drugs and have been followed ever since. A well-known previous analysis of this data evaluated cancer mortality rates overall and by various characteristics. Those results found no clear evidence linking RAI therapy to risk for cancer-related death and concluded that RAI appeared to be a safe therapy for hyperthyroidism.
However, despite the large sample size and long follow-up, the study was likely underpowered to detect modest treatment-related effects. The choice to use an external comparison group made these results susceptible to confounding. There was no formal radiation risk assessment looking at cancer risk per unit organ doses, apart from hematopoietic malignancy-related deaths. Dose estimates at the time used overly simplistic assumptions.
However, in a 24-year extension study, we conducted a long-term effort to reconstruct organ tissue doses for RAI-treated patients. With updated organ and tissue dosimetry estimates, we were able to provide, for the first time, reliable assessments of radiation-related risks.
We saw a statistically significant 6% increased risk for all solid cancers with a 100 milligray (mGy) stomach dose, including a 12% increased risk for female breast cancer with a 100 mGy breast dose. For all solid cancers combined, we found that dose-response follows a linear pattern, whether you look at it continuously or categorically.
In the cohort, median administered dose was 8 mCi, a common dose at the time. Based on the observed slope of this relationship, for the current levels of administered RAI dose — between 10 mCi and 15 mCi — I would estimate an RR for solid cancer mortality between 1.09 and 1.15, or a 9% to 15% increase in risk. For a 10 mCi to 15 mCi dose range, there would be between 20 and 30 excess solid cancer deaths across the lifetime for every 1,000 patients treated at this dose level. These estimates are based on our dose-response model and current mortality rates in the U.S. population.
Given the biological plausibility of these findings, as well as the consistency of our estimates for solid cancer and breast cancer with other high-quality epidemiologic studies, including the large Life Span Study of Japanese atomic bomb survivors, it is reasonable to conclude that RAI therapy for hyperthyroidism leads to at least a small increased risk for cancer in non-thyroid organs and tissues.
- Kitahara CM, et al. JAMA Intern Med. 2019;doi:10.1001/jamainternmed.2019.0981.
Cari Kitahara, PhD, is an investigator in the division of cancer epidemiology and genetics, radiation epidemiology branch, of the NIH’s National Cancer Institute.
No. Current expert opinion holds that RAI is safe for hyperthyroidism therapy with respect to cancer risk.
This is reflected in the most recent American Thyroid Association hyperthyroidism guideline and also in a 2019 statement from the Society for Endocrinology and the British Thyroid Association.
Studies in this field are based on 100 years of cohort data, retrospective and quite heterogenous. Some studies look at incidence; others look at mortality. Importantly, comparator groups for RAI patients vary across studies. Almost all studies in this field lack data on important cancer confounding factors, such as reproductive history. This heterogeneity probably explains why some results are mixed; however, the preponderance of the data is reassuringly negative, and there is no clear pattern of increased cancer risk at the doses used for hyperthyroidism.
The important work with the CTTFUS cohort by Kitahara and colleagues has led to some controversy. There are reasons to take caution when interpreting small magnitude results — the potential for multiple testing errors, registry errors and residual confounding due to antithyroid drug questions. In this CTTFUS database, there are no data on some important cancer-causing confounders, potentially biasing the findings. What if smokers, for example, experienced worse Graves’ disease and received more RAI treatment?
The study design itself is controversial in that only the RAI-dosed group was presented. If the severity of hyperthyroidism is a risk factor in itself for cancer, and if it affects dosing of RAI, as it may in some hospitals, then perhaps the RAI dose may not be the sole cause of downstream cancers. Data from a cohort not exposed to RAI would provide meaningful context. From the clinician’s point of view, you do not get to choose between RAI and nothing. You are choosing between RAI and drugs or surgery, or some combination.
What would move the needle? We need a new study and a new database. We could take advantage of modern electronic medical records with better tracking of antithyroid drugs. Critically, we need to have control of factors like smoking, BMI over time, and reproductive health factors, looking at incidence and mortality. Long follow-up is needed, and we might need to look at control of thyroid status after treatment.
How do we handle the isolated positive signals for cancer if the needle is not fully moved? There are no consistent patterns across the data. Isolated findings cannot simply be aggregated. We need more definitive, consistent data to change the standard of care. In my view, the “yes” advocates have the burden of proof.
Brian W. Kim, MD, is chief of the division of endocrinology and director of the thyroid cancer program at Rush University Medical Center.
Click here to read the cover story.