Our study examined the seasonality of HT Google-related searches in Europe using GT data, with the goal to explore whether there was a seasonal characteristics of Google searches regarding HT, examine the potential impact of the countries’ geographic location on the potential seasonality, and identify possible modifiable risk factors for HT, thereby inspiring future research on the topic.

We conducted our GT data retrieval throughout May 2021, and using all the search categories, we queried GT using the search topic “Hashimoto thyroiditis.” The data were retrieved for 38 European countries, covering a 17-year period of time, from January 2004 to December 2020. A total of 36 European countries were included in the analysis, while Kosovo and Moldova were excluded due to an extremely small search volume. As we investigated the potential impact of seasonality on Google searches, the widest available search time frame of 17 years was used. A specific search topic and no search category restrictions were used to potentially capture the interest of the population of multiple European countries, as wide as possible. The search topic approach was used as it included alternative spellings and translations in other languages; this approach, when comparing European countries using different languages in their Google searches, provided a simple and uniform way to extract the data for each country. The methods were reported following suggestions from a systematic review article on the topic of the use of GT in health care research [18].

GT search data were expressed as relative search volume (RSV) normalized to range from 0% to 100% for the set search time frame and geographical location. Seasonality was assessed using a cosinor regression model from the R programming language “seasons” package. The model fits a sinusoid to the monthly input data, and its outputs include the sinusoid’s amplitude, the phase corresponding to the sinusoidal peak, and 2 P values, the smaller of which is being reported in Table 1. A P value of less than .025 was considered statistically significant. More information on the cosinor model can be found in studies by Cornelissen [26], Mei et al [27], and Wu et al [21]. Months with 0 RSV have been encoded as NA, allowed by the model.

Seasons have been defined as spring (March, April, and May), summer (June, July, and August), autumn (September, October, and November), and winter (December, January, and February).

To evaluate the potential influence of the countries’ latitudes and longitudes, the weighted population center coordinates for each country were retrieved from the Baylor University population resource [28] except for Serbia and Montenegro, where the coordinates of the capital cities were used instead, as the weighted population center coordinates for the 2 countries were not reported in the population resource.

Simple linear regression was conducted to evaluate the potential effect of latitude and longitude on seasonal amplitude and phase of the cosinor model output for each country.

All statistical analyses and data visualizations were done in R programming language (version 4.0.5; R Core Team).

The ethical committee of the University of Zagreb School of Medicine exempted this study from review.

The results of our study demonstrate statistically significant seasonality in HT-related RSV across Europe, with most seasonal peaks occurring in spring and winter months (22/30, 73.3%). Additionally, we have also observed a statistically significant impact of geographical latitude on seasonal amplitude.

A study exploring internet searches patterns regarding hypothyroidism found similar results, with more hypothyroidism-related internet searches occurring during spring globally [29]. Seasonal increases in RSV for HT may be a consequence of increased incidence or higher disease activity. Namely, the volume of disease-related internet searches is known to correlate with patients’ desire to gather more information before an appointment and to supplement information provided by the physician [30]. It is particularly interesting that in most countries, a seasonal peak occurred in spring and winter months; when viewed in the context of the statistically significant impact of geographical latitude on seasonality amplitude, this may indicate that vitamin D levels could play a role in the seasonality of HT. Significant discrepancies in vitamin D levels between northern and southern European countries have been observed, with people in northern European countries having lower levels of vitamin D, which can be associated with less sun exposure, geographical latitude, and solar zenith angle [31]. People living in higher-latitude countries receive lower yearly amounts of sunlight, which leads to a predisposition for developing vitamin D deficiency. Studies have shown that shorter days and insufficient sunlight exposure at latitudes above 40 degrees North lead to poor vitamin D synthesis in the skin [7]. Furthermore, seasonal changes in serum vitamin D levels have been implicated in the seasonality and outcomes of infectious disease [32]. Vitamin D levels are also known to have seasonal fluctuations with the lowest serum levels occurring in the late winter and early spring months [33], which seems to correlate well with the seasonal increase in the RSVs of HT in most countries. Research by Kim [34] and Jamka et al [35] highlighted the importance of vitamin D in the pathogenesis of HT. Kim’s [34] cross-sectional study showed significantly higher prevalence of vitamin D insufficiency in patients with autoimmune thyroid disease, while Jamka et al [35] demonstrated a presumed vitamin D effect on reduction in the levels of thyroid peroxidase antibodies, which has an important role in HT pathogenesis [36]. Multiple studies have demonstrated beneficial effects of vitamin D supplementation on autoimmune diseases, but most notable is the recently published study in the BMJ, which found that vitamin D supplementation (2000 IU/day) could reduce autoimmune disease rate by 22% [37]. Future research should focus on the association between vitamin D serum seasonal changes as a potential trigger of HT and vitamin D supplementation during winter and early spring months, which may prove to be an easily implemented public health measure to decrease the burden of HT.

It is important to consider some of the possible limitations of this study. First, selection bias might be present, as this study only used data pertaining to the population with internet access and those who used Google instead of other search engines. Second, medical-related searches may be performed by anyone interested in a particular medical topic and not only patients. Third, deeper analysis of individual users could not be performed due to the limitations of the available data set. Finally, the impact of potential confounding factors, such as academic cycling (ie, higher search volumes in spring during college exams), could not be excluded.

Significant seasonality of GT search volume for HT was observed in our study, with seasonal peaks in most European countries occurring during spring and winter. A significant impact of latitude on seasonality amplitude was also demonstrated. Additional studies on the topic of seasonality in HT and factors impacting it are required. If vitamin D deficiency is unequivocally proven as a contributing factor in the development of HT, vitamin D supplementation during winter and early spring months might be an easily implemented public health measure aimed at decreasing the burden of this disease.