Cancer risk after medical exposure to radioactive iodine in benign thyroid diseases: a meta-analysis

in Endocrine-Related Cancer

Radioiodine-131 (131I) is widely used for diagnosis and treatment of benign thyroid diseases. Observational studies have not been conclusive about the carcinogenic potential of 131I and we therefore conducted a meta-analysis. We performed a literature search till September 2011 which included 131I as a diagnostic or treatment modality (131I for treatment of thyroid cancer was excluded). Data on 64 different organ or organ group subsets comprising 22 029 exposed subjects in the therapeutic cohorts and 24 799 in the diagnostic cohorts in seven studies were included. Outcome was pooled as the relative risk (RR) using both standard and bias adjusted methods. Quality assessment was performed using a study-specific instrument. No increase in overall (RR 1.06, 95% CI: 0.94–1.19), main organ group or combined organ group (four groups known to concentrate 131I; RR 1.11, 95% CI: 0.94–1.31) risks was demonstrable. Individual organs demonstrated a higher risk for kidney (RR 1.70, 95% CI: 1.15–2.51) and thyroid (RR 1.99, 95% CI: 1.22–3.26) cancers with a strong trend for stomach cancer (RR 1.11, 95% CI: 0.92–1.33). A thyroid dose effect was seen for diagnostic doses. While there is no increase in the overall burden of cancer, an increase in risk to a few organs is seen which requires substantiation. The possible increase in thyroid cancer risk following diagnostic 131I use should no longer be of concern given that it has effectively been replaced by the use of 99mTc-pertechnetate.

Abstract

Radioiodine-131 (131I) is widely used for diagnosis and treatment of benign thyroid diseases. Observational studies have not been conclusive about the carcinogenic potential of 131I and we therefore conducted a meta-analysis. We performed a literature search till September 2011 which included 131I as a diagnostic or treatment modality (131I for treatment of thyroid cancer was excluded). Data on 64 different organ or organ group subsets comprising 22 029 exposed subjects in the therapeutic cohorts and 24 799 in the diagnostic cohorts in seven studies were included. Outcome was pooled as the relative risk (RR) using both standard and bias adjusted methods. Quality assessment was performed using a study-specific instrument. No increase in overall (RR 1.06, 95% CI: 0.94–1.19), main organ group or combined organ group (four groups known to concentrate 131I; RR 1.11, 95% CI: 0.94–1.31) risks was demonstrable. Individual organs demonstrated a higher risk for kidney (RR 1.70, 95% CI: 1.15–2.51) and thyroid (RR 1.99, 95% CI: 1.22–3.26) cancers with a strong trend for stomach cancer (RR 1.11, 95% CI: 0.92–1.33). A thyroid dose effect was seen for diagnostic doses. While there is no increase in the overall burden of cancer, an increase in risk to a few organs is seen which requires substantiation. The possible increase in thyroid cancer risk following diagnostic 131I use should no longer be of concern given that it has effectively been replaced by the use of 99mTc-pertechnetate.

Introduction

Among many radionuclides used in nuclear medicine, radioiodine-131 (131I) (physical half-life ∼8 days) plays an important role and has been used successfully for seven decades. Since its introduction in 1941 (Chapman & Evans 1946, Hertz & Roberts 1946), it is likely that millions of patients with benign thyroid disease have been administered therapeutic or diagnostic activities of 131I. The ease of its administration (Ross 2011), lack of significant adverse effects (Read et al. 2004) and low cost (Ross 2011) have prompted such widespread use even though there still remains some uncertainty in terms of its carcinogenic potential (Glinoer et al. 1987, Solomon et al. 1990, Tominaga et al. 1997, Walsh 2000, Lucignani 2007). In favour of 131I therapy, however, is the fact that the radiation dose to the thyroid is so high that in most cases a large mass of the thyroid gland is ablated and absorbed dosages to other organs are very low (ICRP 1987). It is, however, only in Graves' disease that 131I treatment achieves such high tissue doses that the thyroid epithelial cells are destroyed. With toxic nodules or toxic goitre, the attempt is to achieve lethal tissue doses in the hyperactive tissue while trying to avoid lethal doses in the ‘healthy’ tissue. So there is a possibility that we do induce carcinogenesis in the remaining gland. This is in contrast to external radiotherapy where doses are in the intermediate range to surrounding organs, resulting in sub-lethal cellular damage. Although the carcinogenic effects of moderate doses of therapeutic X-radiation and gamma radiation are fairly well described (Ron 2003), it is likely that risks after exposure to 131I are much less given that the thyroid concentrates most of the radioiodine, and doses to most organs are very small (0.04–0.14 mGy/MBq; Zanzonico 2000) and delivered over several days.

Several long-term cohort studies from groups in Sweden (Holm et al. 1991), England (Franklyn et al. 1999), Finland (Metso et al. 2007) and the US (Hoffman et al. 1982, Goldman et al. 1988) that have investigated the relationship between cancer risk following therapeutic use of 131I in benign thyroid disease provide conflicting results. For example, studies report an increase (Holm et al. 1991, Metso et al. 2007), a decrease (Franklyn et al. 1999) or no change (Hoffman et al. 1982) in overall cancer risk. Also, reports on increased breast (Metso et al. 2007), kidney (Holm et al. 1991) and thyroid (Hoffman et al. 1982, Franklyn et al. 1999) cancer risks are contradicted by other reports on breast (Hoffman et al. 1982, Goldman et al. 1988, Holm et al. 1991, Franklyn et al. 1999), kidney (Hoffman et al. 1982, Metso et al. 2007) or thyroid cancers (Holm et al. 1991, Metso et al. 2007). A recent overview of this topic by Verburg et al. (2011) suggests that the evidence is not conclusive in favour of increased risk. However, as individual studies on their own are unlikely to answer this important question, we decided to embark on a meta-analysis of the evidence available. Also, cancer is a rare outcome, so in order to estimate the incidence of cancers after exposure to 131I, long periods of follow-up and a large population is required which could be addressed by meta-analysis.

Materials and methods

Data sources and eligibility criteria

The medical literature, till September 2011, was searched for articles reporting on the diagnostic or therapeutic use of 131I in benign thyroid diseases and the development of cancer. The references of selected articles were cross-checked for further studies and other language publications were translated where deemed necessary. Details of the search strategy and databases searched are in Supplementary Materials and methods 1, see section on supplementary data given at the end of this article. Inclusion was limited to human studies with exclusions made for 131I exposure in thyroid cancer (therapeutic or diagnostic), in nuclear accidents and reports of a non-analytic observational design.

Data abstraction and quality assessment

A data extraction table was completed for each study (see Results). A methodological quality checklist was developed based on a modification of the general checklist created by Doi & Thalib (2008) (see Supplementary Materials and methods 2, see section on supplementary data given at the end of this article). The maximum score was ten points and quality was assessed by two independent reviewers. Any discrepancies were addressed by consensus.

Statistical analysis

The primary outcome was the relative risk (RR) or the standardised incidence ratio (SIR) for those exposed to radioiodine compared with those not exposed or a standardised population respectively. The SIR was defined as the number of cases observed in the study population divided by the number of cases that would be expected in a standardised population. The RR was defined as the risk of developing a cancer within the follow-up time compared with a control group. SIR and RR were stratified by cancer site. The SIR was used as an estimate of the RR and we report the pooled RR in this meta-analysis. The latter was justified because both the prevalence of cancer and the RR for cancer is expected to be low (Chaturvedi et al. 2008). Heterogeneity was determined to be present when the calculated value of τ2 was greater than zero and/or the Q-statistic was significant at P<0.1 (Takkouche et al. 1999). Studies were pooled using the quality-effects (QE) model (Doi & Thalib 2008, Doi et al. 2011) to take advantage of both bias adjustment (Doi et al. 2011) and additional weighting by the incidence of cancers (AIHW 1998) in the population at large because incidence differs widely by cancer type, and failure to consider this would not reflect the overall burden of risk. This sort of weighting has been applied by us previously (Jamal et al. 2012) and though the latter is not possible with routine models of meta-analysis, we also report a sensitivity analysis using the conventional fixed effects model for homogenous studies and random effects model for heterogeneous studies.

Finally, a subgroup analysis was conducted by computing the ratios of SIRs (Morris & Gardner 1988; RSIR and 95% CI) in subgroups (defined by duration of follow-up, dose/activity of 131I and age at the time of exposure; see Supplementary Materials and methods 3, see section on supplementary data given at the end of this article). To check for publication bias, funnel plot asymmetry could not be used as we had few studies in each group. However, an exclusion sensitivity analysis was performed to check the impact of individual studies by excluding each study and calculating the overall effects again. All analyses were done using MetaXL version 1.2 (http://www.epigear.com).

Results

2929 unique abstracts were reviewed and the full texts of 24 studies were retrieved (Fig. 1). Seventeen of these were then excluded because they were not found to address 131I use (n=3; Berlin & Wasserman 1976, Hoffman et al. 1984, Mellemgaard et al. 1998), did not report effect sizes (RR or SIR; Munoz et al. 1978), were descriptive studies (n=7; Pochin 1960, Dobyns et al. 1974, Freitas et al. 1979, Spencer et al. 1983, Ozaki et al. 1994, Angusti et al. 2000, Listewnik et al. 2010) or had overlapping datasets (n=6; Saenger et al. 1968, Holm et al. 1980, 1988, Hall & Holm 1995a,b, Hall et al. 1996). Seven studies were finally selected for this meta-analysis (Hoffman et al. 1982, Goldman et al. 1988, Holm et al. 1991, Franklyn et al. 1999, Hahn et al. 2001, Dickman et al. 2003, Metso et al. 2007).

Figure 1
Figure 1

Flow diagram of the literature search for the meta-analysis.

Citation: Endocrine-Related Cancer 19, 5; 10.1530/ERC-12-0176

Characteristics of the included studies

We included 64 different organ or organ group subsets in five therapeutic (Hoffman et al. 1982, Goldman et al. 1988, Holm et al. 1991, Franklyn et al. 1999, Metso et al. 2007) and two diagnostic studies (Hahn et al. 2001, Dickman et al. 2003). There were 22 029 exposed subjects in the therapeutic cohorts and 24 799 in the diagnostic cohorts (see Table 1). Although the data source in the study by Dickman et al. (2003) overlapped with the study by Holm et al. (1991), the focus was diagnostic in the former study; and patients with therapeutic use of 131I were excluded. The population considered in these studies originated from the US, Sweden, Finland, UK and Germany. The mean follow-up time varied from 9.8 to 27 years. RRs were estimated in three studies (Hoffman et al. 1982, Hahn et al. 2001, Metso et al. 2007), while the other four studies reported SIRs (Goldman et al. 1988, Holm et al. 1991, Franklyn et al. 1999, Dickman et al. 2003). The quality scores of studies ranged from five to eight out of ten. All of seven studies were retrospective (see Supplementary Table 1, see section on supplementary data given at the end of this article for further details regarding each study). The diagnostic studies reported only on thyroid cancer risk.

Table 1

Characteristics of the included studies

StudyReason for exposed to 131INumber of exposed subjects in final cohortExposed groupComparison groupActivities of 131I used (mean)
Hoffman et al. (1982)Therapeutic1005Hyperthyroid women treated with 131I at Mayo clinic from 01/01/1946 to 31/12/1964Hyperthyroid women assigned to surgical treatment in Mayo clinic from 01/01/1946 to 31/12/1964392.2 MBq (equivalent to 10.6 mCi)
Goldman et al. (1988)Therapeutic607aHyperthyroid women treated with 131I at Massachusetts General Hospital Thyroid Unit from 01/01/1946 to 31/12/1964Cause- age- calendar year- time- sex-matched cancer incidence data in Connecticut from 01/01/1946 to 31/12/1964NA (approximately equal numbers of women were administered <240, 240–369 and ≥370 MBq)
Holm et al. (1991)Therapeutic10 207Hyperthyroid patients treated with 131I in seven Swedish hospitals from 1950 to 1975Age- sex- region of residence- calendar year-matched cancer incidence data in Sweden from 1958 to 1985506 MBq
Franklyn et al. (1999)Therapeutic7417Hyperthyroid patients treated with 131I in the West Midlands region of the UK from 1950 to 1991Age- sex- calendar year-matched cancer incidence data in England and Wales from 1950 to 1991308 MBq (s.d.=232 MBq)
Hahn et al. (2001)Diagnostic789Children examined for suspected thyroid disease with 131I in 10 German hospitals from 1989 to 1997Children examined for suspected thyroid disease with other methods than using 131I in ten German hospitals from 1989 to 19970.9 MBqb (interquartile range 0.4–1.5; median thyroid-absorbed dose 1 Gy)
Dickman et al. (2003)Diagnostic24 010cPatients examined with 131I in seven Swedish hospitals from 1952 to 1969Age- sex- region of residence- calendar year-matched cancer incidence data in Sweden from 1958 to 19691.6 MBqd (mean thyroid-absorbed dose 0.94 Gy)
Metso et al. (2007)Therapeutic2793Hyperthyroid patients treated with 131I in Tampere University Hospital, Finland from 01/1965 to 06/2002Age- and sex-matched group from the Population Register Centre of Finland305 MBq

Number of patients who were exposed to 131I only.

The median 131I activity administered.

Number of 131I-exposed patients who did not report previous radiotherapy to the neck and were not referred for suspicion of a thyroid tumour.

Mean activity administered to patients who did not report previous radiotherapy to the neck region and were not referred for suspicion of a thyroid tumour.

Ascertainment of cancer diagnosis

In all seven studies, cancer was the main endpoint retrospectively ascertained. Three studies (Hoffman et al. 1982, Goldman et al. 1988, Hahn et al. 2001) used either questionnaires, telephone interview, autopsy reports (if available) or medical records from the Cooperative Thyrotoxicosis Therapy Follow-up Study.

Four studies (Holm et al. 1991, Franklyn et al. 1999, Dickman et al. 2003, Metso et al. 2007) used more robust methods as data were retrieved from national cancer registries to identify cancer types from Sweden (Holm et al. 1991, Dickman et al. 2003), Finland (Metso et al. 2007) and the UK (Franklyn et al. 1999).

Quantitative synthesis of data (therapeutic use only)

Overall cancer risk

Seven main organ groups from five studies (Hoffman et al. 1982, Goldman et al. 1988, Holm et al. 1991, Franklyn et al. 1999, Metso et al. 2007; Fig. 2) suggest that there was no overall increase in cancer risk (RR 1.06, 95% CI: 0.94–1.19; Fig. 2a). Results were sensitive to exclusion of the study by Franklyn et al. (1999), exclusion of which increased the pooled risk of cancer (RR 1.13, 95% CI: 1.03–1.24; Fig. 2b). Conventional weighting methods (no population or bias weights) concurred with a RR of 1.03 (95% CI: 0.94–1.13) compared with a RR of 1.12 (95% CI: 1.04–1.20) when the study by Franklyn et al. (1999) was excluded.

Figure 2
Figure 2

The meta-analysis forest plot depicting the overall burden of cancer risk from 131I on eight major groups of organs (QE model; a) including Franklyn et al. (1999) and (b) excluding Franklyn et al. (1999).

Citation: Endocrine-Related Cancer 19, 5; 10.1530/ERC-12-0176

Organ groups that concentrate 131I

Four organ groups known to concentrate radioiodine (digestive organs and peritoneum; lip, oral cavity and pharynx; genitourinary organs and thyroid) revealed no increase in risk (RR 1.11, 95% CI: 0.94–1.31). Results were sensitive to exclusion of the study by Franklyn et al. (1999) after which risk increased (RR 1.17, 95% CI: 1.06–1.30). The latter was mainly because the study by Franklyn et al. (1999) had protective estimates for digestive and genitourinary organs. Similar results were obtained with conventional weighting (RR 1.06, 95% CI: 0.98–1.13) compared with RR 1.12 (95% CI: 1.04–1.21) after removal of the study by Franklyn et al. (1999)).

Individual organ groups

Seven organ groups (Fig. 2) from five papers did not demonstrate an increase in cancer risk (Supplementary Figure 1, see section on supplementary data given at the end of this article). Digestive and respiratory organs groups were, however, sensitive to exclusion of the study by Franklyn et al. (1999), resulting in borderline significance in favour of increased risk (lower confidence limit of 1.01–1.03, Fig. 2b). Conventional weighting methods suggest a slight difference in direction of the pooled estimate for digestive organs and peritoneum; lip, oral cavity and pharynx; and bone, connective tissues, skin and breast, although none were significant and were most likely random changes (see Supplementary Figure 2, see section on supplementary data given at the end of this article).

Individual organs

Nine individual organs (salivary glands, breast, kidney, bladder, lymphoma, leukaemia, pancreas, stomach and thyroid) reported by five studies reveal a higher risk of only kidney and thyroid cancers, while stomach cancer demonstrated a non-significant trend (Supplementary Figure 1, see section on supplementary data given at the end of this article).

The kidney cancer RR was 1.70 (95% CI: 1.15–2.51) from data in three studies. Exclusion of the study by Metso et al. (2007) decreases the significance of the kidney cancer result and puts most of the weight on the study by Holm et al. (1991).

The thyroid cancer risk (RR: 1.99, 95% CI: 1.22–3.26) was mitigated by excluding the study by Franklyn et al. (1999), resulting in an RR of 1.58 (95% CI: 0.91–2.75). These studies had less than ten observed cases of thyroid cancer (Hoffman et al. 1982, Franklyn et al. 1999, Metso et al. 2007) and the study by Holm et al. (1991) with 18 observed cases did not demonstrate increased risk. Conventional weighting methods concurred with these results (see Supplementary Figure 2, see section on supplementary data given at the end of this article).

Finally, the only two studies, Hahn et al. (2001) and Dickman et al. (2003), on diagnostic 131I use (pooled separately) did not demonstrate an increase in thyroid cancer risk (RR=1.06, 95% CI: 0.62, 1.82). However, one of these studies was on children only (Hahn et al. 2001) while the other (Dickman et al. 2003) was predominantly undertaken on adults (only 7% were below the age of 20 years). The RR (or SIR) for thyroid cancer was 0.9 (95% CI: 0.1–5.1) in childhood exposure vs 0.91 (95% CI: 0.64–1.26) in adult exposure (excluding subjects with prior radiation therapy or referred for suspicion of a tumour). The risks in the adult study (Dickman et al. 2003) were significantly higher with prior radiation with (SIR 13.66; 95% CI: 7.06–23.87) or without (SIR 7.67; 95% CI: 3.96–13.40) referral for a thyroid tumour.

Stratification by time since exposure, age at exposure, absorbed dose and administered activity of 131I

To examine the effect of time since exposure, age at exposure and dose of 131I on cancer risk, we calculated the ratio of the SIRs (RSIR) using data from three studies that used this effect size and reported subgroups (Goldman et al. 1988, Hall et al. 1992, Dickman et al. 2003; Table 2). This was done in lieu of a subgroup meta-analysis because there was not enough data to perform a stratified analysis.

Table 2

Subgroup analysis by duration of follow-up, administered 131I activity, absorbed dose or age at exposure

StudyTypes of cancer (observed cases in group 1 vs group 2)SIR1SIR2Ratio of SIR's95% CI
Duration of follow-up
 Goldman et al. (1988)All cancer (16 vs 47)≥30 years<10 years1.330.67–2.22
Breast (5 vs 14)1.500.29–3.59
Pancreas (6 vs 3)20–29 years1.880.53–36.14
 Holm et al. (1991)All cancer≥15 years1–4 years0.980.85–1.14
Breast0.880.63–1.24
Pancreas (16 vs 6)1.590.70–6.29
Leukaemia (12 vs 8)1.040.44–3.06
Thyroid (4 vs 4)0.860.16–4.77
Parathyroid (11 vs 8)0.640.26–1.86
Stomach (35 vs 16)1.771.03–3.53
Liver, gallbladder and bile ducts (9 vs 17)0.960.45–2.60
Lung (30 vs 25)0.770.45–1.34
Kidney (14 vs 14)0.730.34–1.60
Bladder (14 vs 7)1.210.52–3.98
Brain (13 vs 8)1.400.60–4.10
 Dickman et al. (2003)aThyroid (15 vs 4)≥20 years2–5 years0.690.28–6.61
Therapeutic activity
 Goldman et al. (1988)All malignant neoplasms (48 vs 41)≥370 MBq (10 mCi)No 131I1.000.65–1.61
All digestive organs (13 vs 13)0.820.35–2.40
Pancreas (3 vs 2)0.760.10–50.07
Breast (18 vs 18)1.500.73–3.81
Brain (3 vs 1)2.500.40–82.50
All other sites (14 vs 9)0.780.36–1.74
Diagnostic-absorbed doseb
 Dickman et al. (2003)aThyroid (16 vs 5)>1.00 Gy≤0.25 Gy2.821.22–14.89
Age at exposure
 Dickman et al. (2003)aThyroid (2 vs 11)≤20 years>50 years1.020.00–3.02
 Dickman et al. (2003)aThyroid (23 vs 11)20–50 years1.060.44–2.40

Calculated from patients who did not report previous XRT to the neck region and were not referred for suspicion of a thyroid tumour.

The quality factor for all radiations emitted by 131I is assumed to be unity and thus doses in Gray are equivalent to doses in Sievert.

Time since exposure to 131I had a significant effect only on stomach cancer with the ratio of SIRs of 1.77 (95% CI: 1.03–3.53) and the >15 year SIR showing a statistically significant increase.

Thyroid dose after diagnostic and therapeutic activity administration was compared in two studies (Hahn et al. 2001, Dickman et al. 2003). The diagnostic study (Dickman et al. 2003) demonstrated that the patients with an absorbed dose >1 Gy had 2.82 times the SIR compared with patients with an absorbed dose of <0.25 Gy (ratio of SIRs 2.82, 95% CI: 1.22–14.89), although neither of the stratum-specific SIRs were significant. The therapeutic study (Goldman et al. 1988) reported that >10 mCi compared with no administration of 131I did not significantly alter SIR estimates.

Finally, there was no trend of increasing SIR of thyroid cancer as age at exposure to 131I for diagnostic use decreased (Dickman et al. 2003; RSIR=1.06; 95% CI: 0.44–2.40) when those 20–50 years old were compared with those >50 years old.

Discussion

Overall burden of cancer risk

Although case reports of several cancers (Kennedy & Fish 1959, McCormack & Sheline 1963, Bundi et al. 1977, Munoz et al. 1978, Spencer et al. 1983) after radioactive iodine treatment have raised question about the carcinogenesis of 131I, the evidence from this meta-analysis is lacking. In exposed patients, there was no increase in cancer risk overall. Individual organ groups as well as a combination of organ groups (CG) known to concentrate radioiodine (salivary glands, digestive tract and urinary tract) did not demonstrate an increased risk of cancer. Individual studies on overall (Holm et al. 1991, Metso et al. 2007) or CG (Hoffman et al. 1982) risks report a slight increase in cancer but the overall burden of risk by our population-weighted estimate as well as conventional models concurred suggesting that overall burden of risk was null.

Thyroid

Lack of an overall increase in burden of cancer risk is consistent with the low levels of organ doses expected from 131I exposure that remain far from the reported dangerous threshold of between 100 and 1000 mGy (Kloos 2011). This level of exposure is reached, after 131I administration, within the thyroid, but is thought to be less important because the thyroid is ablated and tissue does not remain for carcinogenesis to ensue. Also, while a significant risk of thyroid cancer has been observed after administration of therapeutic X-irradiation with doses as high as 60 Gy in childhood (de Vathaire et al. 1988, Tucker et al. 1991), this has not been observed in several large cohort studies after use of a similar radiation-absorbed dose of 131I for treatment of hyperthyroidism (Dobyns et al. 1974, Holm et al. 1980, 1991, Goldman et al. 1988, Angusti et al. 2000, Metso et al. 2007).

The two studies that demonstrate an increase in the risk of thyroid cancer (Hoffman et al. 1982, Franklyn et al. 1999) differ from the other studies. The study by Hoffman et al. (1982) reports an RR 9.1 but this result was based on three cases in the 131I group and one case in the control (surgical) group (Hoffman et al. 1982). There was a tendency toward increased risk for women <40 years old and decreased risk with increasing age at exposure (Hoffman et al. 1982). This result could have been biased by medical screening following treatment of hyperthyroidism, patient selection bias, choice of surgically treated patients as control groups and limited observation time. In the case of the study by Franklyn et al. (1999) data on cancer incidence (for patients and controls) were drawn from different sources, so bias was likely. Franklyn et al. (1999) compared site-specific data on cancer incidence from follow-up with data from England and Wales in preference to comparison with incidence from the study region because regional control data were incomplete and about 20% of patients had moved from the region before study completion. Indeed, after exclusion of the study by Franklyn et al. (1999) our pooled estimate for thyroid cancer (RR 1.99; 95% CI: 1.22–3.26) declines to non-significance (RR 1.58; 95% CI: 0.91–2.75) and this is contrary to the protective trend seen for other cancers.

Overall, there is the problem of confounding by indication in studies that report increased risk as thyroid cancer is far more common in those with nodular thyroid disease (as compared with Graves'; Angusti et al. 2000). Also thyroid cancer, if associated with Graves' disease, is found more commonly in surgically treated patients than in patients after 131I therapy (Behar et al. 1986). Indeed, under 20-year-old individuals who did not presumably have nodular disease and who received 131I for Graves' disease, followed-up for 36 years, revealed no cases of thyroid cancer (Read et al. 2004). Also, stratification by duration of follow-up (Holm et al. 1991) did not result in a significant increase in the SIR for thyroid cancer.

Diagnostic 131I use had no increase in risk demonstrable, and this concurs with several studies of diagnostic 131I use in children or adolescents (Holm et al. 1988, Hall et al. 1996, Hahn et al. 2001, Dickman et al. 2003). There was also no increase in SIR when stratification by duration of follow-up or age at exposure was done (Dickman et al. 2003). It has been estimated that for an adult, assuming a 15% uptake, a thyroid scan with 1.85–7.4 MBq 131I orally exposes the thyroid to 388–1554 mGy (SNM 1999) and the study by Dickman et al. (2003) demonstrated an increased SIR when the dose was >1000 mGy as compared with <250 mGy. This might be of concern given that this was an absorbed dose to a non-ablated thyroid gland. Nowadays, however, it is much more common to use 75–370 MBq 99mTc-pertechnetate i.v., which results in a much lower thyroid exposure, with the highest exposure being to the upper large intestine to the order of 5–23 mGy.(SNM 1999). While there is little evidence to suggest that the use of diagnostic 131I in adults is carcinogenic, additional data are needed to clarify the risks associated with childhood medical exposure (Ron 2003).

Other organs

Doses to other organs are quite low and an excess of cancers has never been detected for doses below 100 mGy (Kloos 2011), and after repeated X-ray examinations, an excess of cancer is reported only for cumulative doses >500 mGy (Vaiserman 2010). Suit et al. have reviewed secondary carcinogenesis after radiation exposure and conclude that the factorial decrease in risk of radiation-induced secondary cancer would be greatest for reduction in dose levels below 2 Gy. This dose is not reached in any organ (other than the thyroid) after 131I therapy (Suit et al. 2007) and doses to other organs (apart from the stomach) parallel doses received from common medical imaging procedures, ranging up to 20 mGy (highest for chest/abdominal CT and coronary angiography).

The highest non-thyroidal exposure is reported to be the stomach at about 0.41 mGy/MBq (Zanzonico 2000), the total dose averaging 250 mGy (Edmonds & Smith 1986, Huysmans et al. 1996). Within the gastrointestinal system, the stomach is most exposed, since ingested 131I is completely absorbed from the stomach into blood and does not pass through the other compartments of the gastrointestinal tract (Zanzonico 2000). Three studies (excluding Franklyn et al. (1999)) concur in terms of increasing trend for stomach cancer (Hoffman et al. 1982, Holm et al. 1991, Metso et al. 2007) even though the meta-analysis of neither the digestive tract cancer on the whole nor the stomach cancer revealed increased risk. The strongest effect was with the study by Metso et al. (2007) (stomach cancer had 30 cases in the exposed group and an RR 1.76). In addition, stratification by duration of follow-up revealed significant increases in the SIR for stomach cancer (Holm et al. 1991). Data on the dependence of risk on radiation dose for stomach carcinogenesis in patients treated with radiation demonstrate that the slope (of the regression line fitted through the risk values as a function of dose) for the stomach is significantly positive (Suit et al. 2007) and it might therefore be prudent to consider caution especially in patients with a strong family history of gastric carcinoma (Meyer 1994).

Among other individual organs, we did find a higher risk of kidney cancer (RR 1.70). In terms of urinary cancers, others have found no increased risk to the bladder (Hoffman et al. 1982, Holm et al. 1991, Franklyn et al. 1999), though two studies (Holm et al. 1991, Metso et al. 2007) concur on renal cancer. Our pooled estimate for individual urinary organs suggests that the increased risk is mainly to the kidney. When we examined strata by duration of follow-up (Holm et al. 1991), the ratio of SIRs was not different from unity for kidney cancer. However, there were 90 cases of kidney cancer reported in three studies (Hoffman et al. 1982, Holm et al. 1991, Metso et al. 2007) and this raises concern. Fortunately, data from patients treated with radiation suggest that the RR (radiotherapy to general population) as a function of dose up to 15 Gy for the kidney is non-significantly positive and the curve for bladder is flat (Suit et al. 2007). Both the latter observation and the smaller dose expected to this organ are reassuring, but caution must still be exercised based on our results. Apart from kidney, thyroid and stomach, only risk of breast cancer has been suggested to be increased by one study (Metso et al. 2007) which is contrary to this meta-analysis and earlier studies (Hoffman & McConahey 1983, Goldman et al. 1988, Holm et al. 1991).

Conclusion

The thyroid, kidney and stomach are the only three organs that remain under question in terms of increased cancer risk and need further investigation. While there is no demonstrable increased burden of risk overall after 131I administration, the risks of thyroid, kidney and stomach cancers need to be kept in mind when individualising therapy. For example, it may be prudent to consider limiting use of ablative 131I to patients without risk factors for the cancers in question. In the case of the thyroid, we could limit ablative use to diffuse toxic goitres or alternatively make efforts to follow-up stringently if administered for treatment of nodular disease. In addition, persons with a history of atrophic gastritis, intestinal metaplasia or dysplasia and gastric ulcers as well as a genetic predisposition to gastric cancer might also best be excluded. In the same way, subjects with acquired or genetic cystic diseases of the kidney, chronic hepatitis C infection, analgesic abuse, or who have received cytotoxic chemotherapy in childhood might also best avoid therapeutic 131I. Fortunately, diagnostic 131I has now been replaced by 99mTc-pertechnetate, so this is no longer of concern.

Supplementary data

This is linked to the online version of the paper at http://dx.doi.org/10.1530/ERC-12-0176.

Declaration of interest

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.

Funding

This research did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector.

References

  • AIHW 1998 Cancer in Australia 1998: incidence and mortality data for 1998. Australian Insitutes of Health and Welfare

  • AngustiTCodegoneAPelleritoRFaveroA2000Thyroid cancer prevalence after radioiodine treatment of hyperthyroidism. Journal of Nuclear Medicine4110061009.

    • Search Google Scholar
    • Export Citation
  • BeharRArganiniMTain-ChengW1986Graves' disease and thyroid cancer. Surgery10011211127.

  • BerlinNIWassermanLR1976The association between systemically administered radioisotopes and subsequent malignant disease. Cancer3710971101. (doi:10.1002/1097-0142(197602)37:2+<1097::AID-CNCR2820370818>3.0.CO;2-8).

    • Search Google Scholar
    • Export Citation
  • BundiRSScottJSHalnanKE1977Chronic myeloid leukaemia following radioiodine therapy for carcinoma thyroid. British Journal of Radiology506164. (doi:10.1259/0007-1285-50-589-61).

    • Search Google Scholar
    • Export Citation
  • ChapmanEMEvansRD1946The treatment of hyperthyroidism with radioactive iodine. Journal of the American Medical Association1318691. (doi:10.1001/jama.1946.02870190010003).

    • Search Google Scholar
    • Export Citation
  • ChaturvediAKMbulaiteyeSMEngelsEA2008Underestimation of relative risks by standardized incidence ratios for aids-related cancers. Annals of Epidemiology18230234. (doi:10.1016/j.annepidem.2007.10.005).

    • Search Google Scholar
    • Export Citation
  • DickmanPWHolmLELundellGBoiceJDHallP2003Thyroid cancer risk after thyroid examination with I-131: a population-based cohort study in Sweden. International Journal of Cancer106580587. (doi:10.1002/ijc.11258).

    • Search Google Scholar
    • Export Citation
  • DobynsBMShelineGEWorkmanJB1974Malignant and benign neoplasms of the thyroid in patients treated for hyperthyroidism: a report of the cooperative thyrotoxicosis therapy. Follow up study. Journal of Clinical Endocrinology and Metabolism38976998. (doi:10.1210/jcem-38-6-976).

    • Search Google Scholar
    • Export Citation
  • DoiSAThalibL2008A quality-effects model for meta-analysis. Epidemiology1994100. (doi:10.1097/EDE.0b013e31815c24e7).

  • DoiSABarendregtJJMozurkewichEL2011Meta-analysis of heterogeneous clinical trials: an empirical example. Contemporary Clinical Trials32288298. (doi:10.1016/j.cct.2010.12.006).

    • Search Google Scholar
    • Export Citation
  • EdmondsCJSmithT1986The long-term hazards of the treatment of thyroid cancer with radioiodine. British Journal of Radiology594551. (doi:10.1259/0007-1285-59-697-45).

    • Search Google Scholar
    • Export Citation
  • FranklynJAMaisonneuvePSheppardMBetteridgeJBoyleP1999Cancer incidence and mortality after radioiodine treatment for hyperthyroidism: a population-based cohort study. Lancet35321112115. (doi:10.1016/S0140-6736(98)12295-X).

    • Search Google Scholar
    • Export Citation
  • FreitasJESwansonDPGrossMDSissonJC1979Iodine-131: optimal therapy for hyperthyroidism in children and adolescents?Journal of Nuclear Medicine20847850.

    • Search Google Scholar
    • Export Citation
  • GlinoerDHeschDLagasseRLaurbergP1987The management of hyperthyroidism due to Graves' disease in Europe in 1986. Results of an international survey. Acta Endocrinologica. Supplementum285323.

    • Search Google Scholar
    • Export Citation
  • GoldmanMBMaloofFMonsonRRAschengrauACooperDSRidgwayEC1988Radioactive iodine therapy and breast cancer. A follow-up study of hyperthyroid women. American Journal of Epidemiology127969980.

    • Search Google Scholar
    • Export Citation
  • HahnKSchnell-InderstPGroscheBHolmLE2001Thyroid cancer after diagnostic administration of iodine-131 in childhood. Radiation Research1566170. (doi:10.1667/0033-7587(2001)156[0061:TCADAO]2.0.CO;2).

    • Search Google Scholar
    • Export Citation
  • HallPHolmLE1995aCancer in iodine-131 exposed patients. Journal of Endocrinological Investigation18147149.

  • Hall P & Holm LE 1995b Cancer incidence and mortality after I-131 therapy for hyperthyroidism. In Radiation and Public Perception: Benefits and Risks. Eds JP Young & RS Yalow RS. Washington: Amer Chemical Soc

  • HallPBoiceJDJrBergGBjelkengrenGEricssonUBHallquistALidbergMLundellGMattssonATennvallJ1992Leukaemia incidence after iodine-131 exposure. Lancet34014.

    • Search Google Scholar
    • Export Citation
  • HallPMattssonABoiceJD1996Thyroid cancer after diagnostic administration of iodine-131. Radiation Research1458692. (doi:10.2307/3579200).

    • Search Google Scholar
    • Export Citation
  • HertzSRobertsA1946Radioactive iodine in the study of thyroid physiology; the use of radioactive iodine therapy in hyperthyroidism. Journal of the American Medical Association1318186. (doi:10.1001/jama.1946.02870190005002).

    • Search Google Scholar
    • Export Citation
  • HoffmanDAMcConaheyWM1983Breast cancer following iodine-131 therapy for hyperthyroidism. Journal of the National Cancer Institute706367.

    • Search Google Scholar
    • Export Citation
  • HoffmanDAMcConaheyWMFraumeniJFJrKurlandLT1982Cancer incidence following treatment of hyperthyroidism. International Journal of Epidemiology11218224. (doi:10.1093/ije/11.3.218).

    • Search Google Scholar
    • Export Citation
  • HoffmanDAMcConaheyWMBrintonLAFraumeniJFJr1984Breast cancer in hypothyroid women using thyroid supplements. Journal of the American Medical Association251616619. (doi:10.1001/jama.1984.03340290030015).

    • Search Google Scholar
    • Export Citation
  • HolmLEDahlqvistIIsraelssonALundellG1980Malignant thyroid tumors after iodine-131 therapy: a retrospective cohort study. New England Journal of Medicine303188191. (doi:10.1056/NEJM198007243030404).

    • Search Google Scholar
    • Export Citation
  • HolmLEWiklundKELundellGEBergmanNABjelkengrenGCederquistESEricssonUBLarssonLGLidbergMELindbergRS1988Thyroid cancer after diagnostic doses of iodine-131: a retrospective cohort study. Journal of the National Cancer Institute8011321138. (doi:10.1093/jnci/80.14.1132).

    • Search Google Scholar
    • Export Citation
  • HolmLEHallPWiklundKLundellGBergGBjelkengrenGCederquistEEricssonUBHallquistALarssonLG1991Cancer risk after iodine-131 therapy for hyperthyroidism. Journal of the National Cancer Institute8310721077. (doi:10.1093/jnci/83.15.1072).

    • Search Google Scholar
    • Export Citation
  • HuysmansDBuijsWvan de VenMTvan den BroekWJKloppenborgPWHermusACorstensFH1996Dosimetry and risk estimates of radioiodine therapy for large, multinodular goiters. Journal of Nuclear Medicine3720722079.

    • Search Google Scholar
    • Export Citation
  • ICRPRadiation dose to patients from radiopharmaceuticalsAnnals of the ICRP1819871171. (doi:10.1016/0146-6453(87)90003-0).

  • JamalMHDoiSARousseauMEdwardsMRaoCBarendregtJJSnellLMeterissianS2012Systematic review and meta-analysis of the effect of North American working hours restrictions on mortality and morbidity in surgical patients. British Journal of Surgery99336344. (doi:10.1002/bjs.8657).

    • Search Google Scholar
    • Export Citation
  • KennedyWMFishRG1959Acute granulocytic leukemia after radioactive iodine therapy for hyperthyroidism. New England Journal of Medicine2607677. (doi:10.1056/NEJM195901082600207).

    • Search Google Scholar
    • Export Citation
  • KloosRT2011Survey of radioiodine therapy safety practices highlights the need for user-friendly recommendations. Thyroid219799. (doi:10.1089/thy.2010.2102.com).

    • Search Google Scholar
    • Export Citation
  • ListewnikMHBirkenfeldBChosiaMElblBNiedziałkowskaKSawrymowiczM2010The occurrence of malignant thyroid lesions in patients after radioiodine treatment due to benign thyroid diseases. Endokrynologia Polska61454457.

    • Search Google Scholar
    • Export Citation
  • LucignaniG2007Long-term risks in hyperthyroid patients treated with radioiodine: is there anything new?European Journal of Nuclear Medicine and Molecular Imaging3415041509. (doi:10.1007/s00259-007-0509-z).

    • Search Google Scholar
    • Export Citation
  • McCormackKRShelineGE1963Leukemia after radioiodine therapy for hyperthyroidism. California Medicine98207209.

  • MellemgaardAFromGJorgensenTJohansenCOlsenJHPerrildH1998Cancer risk in individuals with benign thyroid disorders. Thyroid8751754. (doi:10.1089/thy.1998.8.751).

    • Search Google Scholar
    • Export Citation
  • MetsoSAuvinenAHuhtalaHSalmiJOksalaHJaatinenP2007Increased cancer incidence after radioiodine treatment for hyperthyroidism. Cancer10919721979. (doi:10.1002/cncr.22635).

    • Search Google Scholar
    • Export Citation
  • MeyerMA1994Re: cancer risk after iodine-131 therapy for hyperthyroidism. Journal of the National Cancer Institute8610261027. (doi:10.1093/jnci/86.13.1026).

    • Search Google Scholar
    • Export Citation
  • MorrisJAGardnerMJ1988Calculating confidence intervals for relative risks (odds ratios) and standardised ratios and rates. BMJ29613131316. (doi:10.1136/bmj.296.6632.1313).

    • Search Google Scholar
    • Export Citation
  • MunozJMGormanCAElvebackLRWentzJR1978Incidence of malignant neoplasms of all types in patients with Graves' disease. Archives of Internal Medicine138944947. (doi:10.1001/archinte.1978.03630310034015).

    • Search Google Scholar
    • Export Citation
  • OzakiOItoKMimuraTSuginoKKitamuraYIwabuchiHKawanoM1994Thyroid carcinoma after radioactive iodine therapy for Graves' disease. World Journal of Surgery18518521. (doi:10.1007/BF00353753).

    • Search Google Scholar
    • Export Citation
  • PochinEE1960Leukaemia following radioiodine treatment of thyrotoxicosis. BMJ215451550. (doi:10.1136/bmj.2.5212.1545).

  • ReadCHTanseyMJMendaY2004A 36-year retrospective analysis of the efficacy and safety of radioactive iodine in treating young Graves' patients. Journal of Clinical Endocrinology and Metabolism8942294233. (doi:10.1210/jc.2003-031223).

    • Search Google Scholar
    • Export Citation
  • RonE2003Cancer risks from medical radiation. Health Physics854759. (doi:10.1097/00004032-200307000-00011).

  • RossDS2011Radioiodine therapy for hyperthyroidism. New England Journal of Medicine364542550. (doi:10.1056/NEJMct1007101).

  • SaengerELThomaGETompkinsEA1968Incidence of leukemia following treatment of hyperthyroidism. Journal of the American Medical Association205855862. (doi:10.1001/jama.1968.03140380059014).

    • Search Google Scholar
    • Export Citation
  • SNM 1999 Society of Nuclear Medicine Procedure Guideline for Thyroid Scintigraphy (online). Society of Nuclear Medicine. Available: http://interactive.snm.org/docs/pg_ch05_0403.pdf (accessed 17/04/2012)

  • SolomonBGlinoerDLagasseRWartofskyL1990Current trends in the management of Graves' disease. Journal of Clinical Endocrinology and Metabolism7015181524. (doi:10.1210/jcem-70-6-1518).

    • Search Google Scholar
    • Export Citation
  • SpencerRPChapmanCNRaoH1983Thyroid carcinoma after radioiodide therapy for hyperthyroidism. Analysis based on age, latency, and administered dose of I-131. Clinical Nuclear Medicine8216219. (doi:10.1097/00003072-198305000-00007).

    • Search Google Scholar
    • Export Citation
  • SuitHGoldbergSNiemierkoAAncukiewiczMHallEGoiteinMWongWPaganettiH2007Secondary carcinogenesis in patients treated with radiation: a review of data on radiation-induced cancers in human, non-human primate, canine and rodent subjects. Radiation Research1671242. (doi:10.1667/RR0527.1).

    • Search Google Scholar
    • Export Citation
  • TakkoucheBCadarso-SuarezCSpiegelmanD1999Evaluation of old and new tests of heterogeneity in epidemiologic meta-analysis. American Journal of Epidemiology150206215. (doi:10.1093/oxfordjournals.aje.a009981).

    • Search Google Scholar
    • Export Citation
  • TominagaTYokoyamaNNagatakiSChoBYKohCSChenJLShiY1997International differences in approaches to 131I therapy for Graves' disease: case selection and restrictions recommended to patients in Japan, Korea, and China. Thyroid7217220. (doi:10.1089/thy.1997.7.217).

    • Search Google Scholar
    • Export Citation
  • TuckerMAJonesPHBoiceJDJrRobisonLLStoneBJStovallMJenkinRDLubinJHBaumESSiegelSE1991Therapeutic radiation at a young age is linked to secondary thyroid cancer. The late effects study group. Cancer Research5128852888.

    • Search Google Scholar
    • Export Citation
  • VaisermanAM2010Radiation hormesis: historical perspective and implications for low-dose cancer risk assessment. Dose–response: a Publication of International Hormesis Society8172191.

    • Search Google Scholar
    • Export Citation
  • de VathaireFFrancoisPSchweisguthOOberlinOLe MoniqueG1988Irradiated neuroblastoma in childhood as potential risk factor for subsequent thyroid tumour. Lancet2455doi:10.1016/S0140-6736(88)90445-X).

    • Search Google Scholar
    • Export Citation
  • VerburgFALusterMLassmannMReinersC2011(131)I therapy in patients with benign thyroid disease does not conclusively lead to a higher risk of subsequent malignancies. Nuklearmedizin. Nuclear Medicine509399.

    • Search Google Scholar
    • Export Citation
  • WalshJP2000Management of Graves' disease in Australia. Australian and New Zealand Journal of Medicine30559566. (doi:10.1111/j.1445-5994.2000.tb00856.x).

    • Search Google Scholar
    • Export Citation
  • ZanzonicoPB2000Age-dependent thyroid absorbed doses for radiobiologically significant radioisotopes of iodine. Health Physics786067. (doi:10.1097/00004032-200001000-00010).

    • Search Google Scholar
    • Export Citation

If the inline PDF is not rendering correctly, you can download the PDF file here.

 

      Society for Endocrinology

Related Articles

Article Information

Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 2972 2759 102
PDF Downloads 358 320 33

Altmetrics

Figures

Cited By

PubMed

Google Scholar