National, sub-national, and risk-attributed burden of thyroid cancer in Iran from 1990 to 2019

An updated exploration of the burden of thyroid cancer across a country is always required for making correct decisions. The objective of this study is to present the thyroid cancer burden and attributed burden to the high Body Mass Index (BMI) in Iran at national and sub-national levels from 1990 to 2019. The data was obtained from the GBD 2019 study estimates. To explain the pattern of changes in incidence from 1990 to 2019, decomposition analysis was conducted. Besides, the attribution of high BMI in the thyroid cancer DALYs and deaths were obtained. The age-standardized incidence rate of thyroid cancer was 1.57 (95% UI: 1.33–1.86) in 1990 and increased 131% (53–191) until 2019. The age-standardized prevalence rate of thyroid cancer was 30.19 (18.75–34.55) in 2019 which increased 164% (77–246) from 11.44 (9.38–13.85) in 1990. In 2019, the death rate, and Disability-adjusted life years of thyroid cancer was 0.49 (0.36–0.53), and 13.16 (8.93–14.62), respectively. These numbers also increased since 1990. The DALYs and deaths attributable to high BMI was 1.91 (0.95–3.11) and 0.07 (0.04–0.11), respectively. The thyroid cancer burden and high BMI attributed burden has increased from 1990 to 2019 in Iran. This study and similar studies’ results can be used for accurate resource allocation for efficient management and all potential risks’ modification for thyroid cancer with a cost-conscious view.

The decomposition analysis of thyroid cancer at the province level showed that in all provinces except one (i.e. Tehran), the largest proportion of all-ages increase in both sex incidence cases was attributable to the increase in the age-specific incidence rate. In contrast, in Tehran, the largest proportion was attributed to population aging (Supplementary Table 3). Among provinces, Sistan and Baluchistan (1156.7%) had the highest increase attributable to the increase in age-specific incidence rate, Alborz (180.8%) had the highest increase attributable to the population aging, and Hormozgan (105%) had the highest increase attributable to the population growth.  Table 1). The ASPR also increased in all provinces between these years, but the difference between the first and the last rank of ASPR among provinces was lower in 1990 (i.e. 19.94 versus 22.78). Although ASPR increased in all provinces, the extent of change varied from 15% (− 21 to 69) to 537% (146-1061) (Supplementary Table 1 and Supplementary Fig. 1). Females had higher ASPR than males in all provinces in both years. However, the uncertainty intervals of the ASPRs for females and males overlapped in some provinces. In 2019, all provinces had lower female to male ratio except one (i.e. Bushehr) compared to 1990, and the highest and lowest female to male ratio was 4.45 and 1.42. However, in 1990, it was 4.74 and 2.50 respectively (Supplementary Table 1).    Table 1 and Supplementary Fig. 1). In 1990, females had higher ASDR than males in all provinces, whereas in 2019, males had higher ASDR in five provinces. However, the significance of these differences has to be further evaluated in most of the provinces. However, in 2019, all provinces had lower female to male ratio compared to 1990. Furthermore, the highest and lowest female to male ratio in 1990 was 2.75 and 1.34, and in 2019, it was 2.05 and 0.80 respectively (Supplementary Table 1). Moreover, the age-standardized Mortality to Incidence Ratio (MIR) was higher in males in all of the provinces from 1990 to 2019 ( Supplementary Fig. 2).  Supplementary Fig. 1). In 1990, females had higher age-standardized DALYs than males in all provinces. However, in 2019, in five provinces males had higher age-standardized DALYs. Similar to previous indices, the significance of the differences between females and males are discussable. In 2019, all provinces had lower female to male ratio compared to 1990. Moreover, the highest and the lowest female to male ratio in 1990 was 2.61 and 1.26, whereas in 2019, it was 2.10 and 0.78 respectively (Supplementary Table 1).
Thyroid cancer risk attributable DALYs. The high BMI is the only risk factor attributable to thyroid cancer introduced by the GBD studies. Here, the risk-attributable burden at national level will be reported at the Table 2, and the risk attributed burden at province level are presented in Supplementary  (Fig. 5). The risk attributable DALY increased in all of the provinces from 1990 to 2019. However, the difference between the first and the last rank was lower in 1990 (i.e. 1.11 versus 1.37). Besides, the extent of change in this index ranged from 7% (− 24 to 57) to 262% . In 1990, females had higher risk attributable DALYs than males in all provinces, whereas in 2019, males had higher risk attributable DALYs in half of the provinces. However, the uncertainty intervals overlapped in all of the provinces in both years. In 2019, all provinces had lower female to male ratio compared to 1990. In addition, the highest and lowest female to male ratio in 1990 was 2.64 and 1.29, and in 2019, it was 1.61 and 0.62 respectively.
Thyroid cancer in SDI levels of Iran. The overall age-standardized rate of incidence, prevalence, deaths, and DALYs of thyroid cancer increased from 1990 to 2019 in almost all provinces in Iran from all five SDI levels (Fig. 6) Ir a n (I s la m ic R e p u b lic o f)

A lb o rz A rd e b il B u s h e h r C h a h a r M a h a a l a n d B a k h ti a ri
E a s t A z a rb a y e ja n F a rs G ila n G o le s ta n H a m a d a n H o rm o z g a n Il a m Is fa h a n K e rm a n K e rm a n s h a h K h o ra s a n − e − R a z a v i K h u z e s ta n K o h g ilu y e h a n d B o y e r− A h m a d K u rd is ta n L o re s ta n M a rk a z i M a z a n d a ra n N o rt h K h o ra s a n Q a z v in Q o m S e m n a n S is ta n a n d B a lu c h is ta n S o u th K h o ra s a n T e h ra n W e s t A z a rb a y e ja n Y a z d Z a n ja n Ir a n (I s la m ic R e p u b lic o f)

A lb o rz A rd e b il B u s h e h r C h a h a r M a h a a l a n d B a k h ti a ri E a s t A z a rb a y e ja n F a rs G ila n G o le s ta n H a m a d a n H o rm
o z g a n Il a m Is fa h a n K e rm a n K e rm a n s h a h K h o ra s a n − e − R a z a v i K h u z e s ta n K o h g ilu y e h a n d B o y e r− A h m a d K u rd is ta n L o re s ta n M a rk a z i M a z a n d a ra n N o rt h K h o ra s a n Q a z v in Q o m S e m n a n S is ta n a n d B a lu c h is ta n S o u th K h o ra s a n T e h ra n W e s t A z a rb a y e ja n Y a z d Z a n ja n

Discussion
In this study, we have shown that although variability existed among provinces, the general burden of thyroid cancer increased from 1990 to 2019 in Iran. Moreover, the difference between provinces was higher, and female to male ratios were lower in 2019 comparing to 1990. The high BMI attributable DALYs also increased between these two years. As a matter of fact, based on the uncertainty intervals, the ASIR and the ASPR of thyroid cancer was higher in 2019 in Iran, whereas the ASDR, DALY, and the high BMI attributable DALY did not seem to be significantly different between these two years. This study's results also suggest that SDI in a province is not significantly associated with the thyroid cancer burden in 2019. Indeed, some provinces with different SDIs had similar ASIRs. Hence, SDI had no effect on Thyroid cancer epidemiologic measures in Iran in 2019. It is worth noting that the observed ranges of disparity in all of the indices among the provinces had overlapping uncertainty intervals. The implication of this finding is that there might not be significant differences in the burden of thyroid cancer among provinces, which needs to be evaluated in future studies.
One explanation in support of the disparities in the incidence rate among different provinces might be differences in the healthcare access and quality. The more aggressively a nodule is approached due to higher healthcare access, the more will the case finding be. In addition, it is believed that one of the factors affecting the increasing incidence and prevalence of thyroid cancer in Iran is the increasing number of skilled endocrinologists and radiologists 8 . Hence, the sub-clinical disease is more likely to be diagnosed [23][24][25] . However, since the availability of the diagnostic modalities might be affected by the SDI in a province [26][27][28] , an increase in the thyroid cancer ASIR in low SDI provinces, including Sistan and Baluchistan might not be solely due to the increased diagnostics. Besides, the decrease in the ASDR in four provinces, including Tehran and other two provinces nearby Tehran from 1990 to 2019, might be reflecting the inequities in the treatment of thyroid cancer across the country due to inequities in healthcare quality, possibly in favor of the capital. Nevertheless, based on the uncertainty intervals, the change from 1990 to 2019 might only be significant in Sistan and Baluchistan and only in the ASIR index. This province is a relatively deprived area in Iran and the increase in the ASIR in the recent years might be mostly due to the better diagnostics. Furthermore, the age structure of a population affects the incidence, prevalence, deaths, and DALYs of a disease. Specially, since Iran is a country in the Eastern Mediterranean Region of the WHO with a dynamic age composition 29 , the effect of the population age structure is strong in the findings of this study.
It is also worth noting that the observed trend of the rising thyroid cancer incidence was more significant between the years 1990 and 2010 compared to the years between 2010 and 2019. Although the duration between 1990 and 2010 was longer, this difference between these two time intervals might be due to the revision of the American Thyroid Association's guidelines on thyroid cancer diagnosis and treatment and other countries developing similar guidelines in 2010 30 . However, according to the GBD studies' data source, there are limited number of data sources for thyroid cancer burden in Iran after 2010 31 . Hence, most of the data between 2010 and 2019 is based on estimations. Therefore, more precise data needs to be acquired from future field studies.
The observed pattern and the increasing trend of thyroid cancer incidence might result from various factors. The improvement of medical techniques has occurred in diagnostics and it might have resulted in a higher rate of diagnosis of subtle malignancies 32 . Accordingly, the diagnosis of subtle malignancies leads to overdiagnosis which is estimated to account for over 80% of female thyroid cancer in countries including South Korea, Belarus,    (per 100,000) Incidence    33 . Similarly, in Spain, the increased incidence of thyroid cancer is mostly attributed to the higher diagnosis rate 34 . However, in comparison with global rates, the increase in the incidence rate of thyroid cancer in Iran was relatively lower 35 . This finding might be explained by a relatively lower rate of overdiagnosis in Iran. In fact, although the increased ASIR and DALYs in other countries, including South Korea 36 were mostly due to the higher diagnosis rates by cancer screening programs, this scenario does not apply to Iran because currently there are no thyroid cancer screening programs in the country.
Since there are various burdens of thyroid cancer in the world, it might be a matter of the cost-effectiveness of thyroid cancer screening programs in different regions 37,38 . Even in South Korea, the necessity of these screenings remains controversial 39 . Thus, the cost-effectiveness of thyroid cancer screening programs in all countries including Iran needs future studies to be evaluated. Nevertheless, the role of increased diagnosis must not overshadow the role of other potential risk factors of thyroid cancer. Besides, advanced diagnostics are expensive and therefore, the increasing burden must not be solely attributed to overdiagnosis 40 . For example, in China, the exposure to endocrine-disrupting chemicals such as polychlorinated biphenyls (PCBs), asbestos, pesticides, and polybrominated diphenyl ethers (PBDEs) have been suspected to play role in the increasing incidence 41,42 .
Another study in the Czech Republic also reported the role of other factors such as chemical exposures and iodine deficiency in the increasing burden of thyroid cancer 43 . Previously in Iran, radiation exposure and underlying benign thyroid disease have been found to be associated with the thyroid cancer incidence 44,45 . Further studies are required to assess the role of these factors more accurately in the increasing burden of thyroid cancer in Iran.
It is worth noting that despite the increasing incidence rates, the lower difference in death rates between 1990 and 2019 might be due to better clinical management of thyroid cancer in recent years 42 . The evidence in support of this assertion is that although the ASIR of thyroid cancer increased from 1990 to 2019, the age-standardized MIR has decreased in all provinces ( Supplementary Fig. 1). However, even though Iranian health care system have been evidently successful to some extents, it is currently facing new challenges regarding the non-communicable disease and this may affect the performance of managing these disease including thyroid cancer 46 . Hence, more detailed analysis needs to be done in Iran at subnational level using indices measuring the quality of care 47 .
Finally, the increased burden attributable to the high BMI in Iran suggests that Iranian population is failing to manage this risk factor 48 . To explain the thyroid cancer burden disparities, disparities in other risk factors of thyroid cancer in different provinces needs to be assessed as well [11][12][13][14][15][16] . Accordingly, East Azarbayejan and Sistan and Baluchistan, which were among the provinces with the highest ASDRs and DALYs, had the lowest urinary iodine concentration in 2014 15 . Moreover, Sistan and Baluchistan had the highest increase in ASIR, ASPR, DALYs, and ASDRs of thyroid cancer. On the other hand, Kurdistan was among the provinces with the highest urinary iodine concentration and the lowest ASIR, ASPR, ASDR, and DALYs. However, these are descriptive interpretations and more precise analyses are needed to evaluate the significance conclude that the aforementioned risk factors are correlated with the findings of this study.
Although the significance of the differences were debatable, the variability was also observed in different sexes. In each year from 1990 to 2019, the female gender had more ASIRs and ASDRs. Even though the reason behind the higher burden among the females in thyroid cancer remains unknown, there are some possible explanations. One explanation is that there are identical hormonal signaling pathways playing role in thyroid and breast cancer, both of which having the highest incidence in women 49,50 . Another explanation is that the female gender is associated with higher usage of diagnostic medical imaging modalities, especially after the age of 45 51 . However, in 2019, female to male ratio was lower in most of the provinces in case of thyroid cancer burden. Indeed, the high BMI attributable DALY was higher in males in half of the provinces in 2019. Thus, it might be deductible that the high BMI in males is resulting in the convergence of the thyroid cancer burden in females and males in Iran.
However, in the case of deaths, the higher rates among women was mostly due to the higher incidence. Following this theory, the age-standardized mortality to incidence ratio was calculated in all provinces of the country. Generally, the age-standardized mortality to incidence ratio was higher in males in the provinces of Iran during all years from 1990 to 2019. However, the significance of this difference needs to be evaluated as well. Indeed, a previous study (i.e., in 2009) had indicated that the survival rate was not significantly different between the two genders in Iran 52 . However, this study's finding was consistent with the previous studies in the U.S, indicating that the survival was lower in the male population 53 . The differences in the health care seeking attitudes among men and women are probably the explanation behind these findings 54 .
There were also disparities among different age groups in the thyroid cancer incidence rate. The highest incidence rate of thyroid cancer in 2019 among men was in the 55 to 59 age group, whereas among women, it was in the 65 to 69 age group. In contrast, in 1990, the highest incidence rate was in the 75 to 79 age group among men and women. This finding might be due to the earlier diagnosis of thyroid cancer in recent years. However, this result was in contrast with the findings of the studies in the U.S community 55 . The results also indicate that thyroid cancer rarely affects children under the age of 15 in Iran 8 .
This study's results are in agreement with the World Health Organization warning against thyroid cancer in the world 56 . The 0.049% (0.044-0.053) of total DALYs in 2019 was due to thyroid cancer 57 , however, only 0.003% of the cancer research budget is allocated to thyroid cancer in the world 58 . Therefore, resource allocation for this cancer research, subsequent management, and improving healthcare quality might need to be revised across the globe. Plus, the world is failing in lifestyle improvement 10 , and consequently leading to the increased distribution of risk factors (e.g., high BMI). Thus, risk modification as a strategy in managing the thyroid cancer burden remains challenging. However, it has to be elucidated that what makes provinces with low burden of thyroid cancer (e.g. Zanjan) different from the others and modify the affecting factors in other provinces.
Similar to the other GBD studies, the limitations of this study are mostly due to the data availability 2 . Despite the fact that GBD study relies on modeling to construct estimates, sufficient primary data is not available from every location due to weak registry systems and this lack of data might challenge the accuracy of estimations. Accordingly, representing the UI around estimates remains challenging because of the sparse data. Therefore, www.nature.com/scientificreports/ despite improvements in precision of the estimation models, there is always a need for better and more primary data collection. Another limitation of this study was considering high BMI as the only risk factor for thyroid cancer. Finding more causal connections between risks and outcomes would result in better explanation of the observed patterns and spotting proper intervention points. To accomplish this objective, other studies including Mendelian randomization studies could be used in the meta-regression 10,59 . Furthermore, thyroid cancer is not limited to a single histology and different histological types have different risk factors and outcomes. Hence, disaggregation of the thyroid cancer burden estimation by histology would provide more comprehensive insights towards this cancer 60 .
On the other hand, providing sub-national data, decomposition analysis, analyzing the effect of SDI on thyroid cancer burden, calculating MIR and attributed burden to a risk factor are among the strengths of this article.
In conclusion, this study illustrates that the thyroid cancer burden and high BMI attributed burden has increased from 1990 to 2019 in Iran. This study and similar studies' results would help in the health systems appropriate resource allocation for managing the thyroid cancer burden and reduce the disparities in this regard. However, this resource allocation must have a cost-conscious view to avoid overmanaging and must be able to enhance the awareness of the medical community about overdiagnosis. In addition, the exposure to risk factors of this cancer including high BMI and the other abovementioned factors need to be reduced as much as possible, especially among the high-risk groups.