Graves' Disease: Prospective Up to 10‐Year Outcome for 100 Consecutive Individuals at a Single Centre

Graves' disease is named after Robert J. Graves, who described the condition in 1835 1. The prevalence of Graves' disease is approximately 1% in the population as a whole: the incidence is 20–30 new cases/100,000/year 2. Both genetic and environment factors, such as familial clustering, negative life experiences, high iodine intake, and smoking, predispose to Graves' disease being clinically manifest 3. Graves' disease remains a challenging condition to manage in the 21st Century, with no new pharmacotherapies for many years. However, new options for treatment utilising immunotherapy approaches have potential 4. In order to study the trends of thyroid-stimulating hormone (TSH), free thyroxine (FT4), and free liothyronine (FT3) with Graves' disease, we undertook an evaluation of outcomes for 100 people diagnosed with Graves' disease, for up to a 10-year period after diagnosis. We also organised the data into different groups according to biographical status and what therapy the individual underwent. All individuals were seen in the endocrinology outpatient clinic as consecutive presentations at a single centre in the United Kingdom. Individuals were first seen between 2012 and 2014. The Electronic Patient Record (EPR) was accessed to collect data which was then collated in a Microsoft Excel spreadsheet. Biographic data (age and gender) were recorded. The results for those who underwent thyroid surgery, received radioactive iodine treatment, diagnosis of thyroid eye disease, together with deaths and relapses, were compiled. Patients remained under the care of the Endocrinology Clinic, and data were gathered through the clinical follow-up. For up to 10 years, data were collected for TSH, FT4, and FT3 levels at baseline (0 month), 3, 6, 12 months, as well as 2, 5, and 10 years. The data were collected at Outpatient Clinic appointments in the endocrinology centre. Follow-up was ascertained through the local electronic patient record. Outcomes were reported at approximately 10 years from initial assessment where available, or otherwise at the most recent documented follow-up within the local healthcare system. Not all patients had tests done at these specific intervals. Patients were mostly still resident within the catchment area throughout the follow-up period, but some were lost to follow-up, mostly due to non-attendance at outpatient appointments. TSH receptor antibody (TSHRAb) levels were also recorded along with any recheck of TSHRAb levels. FT4, FT3, and TSH were measured on the Siemens Atellica Immunoassay platform (Manchester, UK) with TSHRAb measured on the Dynex DS2 2-Plate ELISA Processing System (Chantilly, USA). Continuous variables were summarised as mean (standard deviation) or median (interquartile range), and categorical variables as counts and percentages. Kernel density plots were used to display the distribution of FT4 values at baseline and at the most recent follow-up. Logistic regression was used to assess the association between baseline TSH receptor antibody level and relapse from unadjusted and age- and sex-adjusted models. Analyses were performed using IBM SPSS Statistics version 21 (Chicago, USA). Data processing and figure generation were performed in Python 3.11 using Pandas, NumPy, Matplotlib, and Seaborn. Of the 100 consecutive patients, 15 were male and 85 were female. The median age at diagnosis was 39.5 years (Interquartile Range or IQR: 31–52) with median 43 years (IQR: 31–61) for males and 39 years (IQR: 31–51) for females. The youngest female at time of diagnosis was 18 years old, while the youngest affected male was 43. The oldest patients were a 73-year-old male and 79-year-old female at diagnosis. There was a markedly skewed distribution of FT4 at baseline with a median FT4 of 29.3 pmol/L falling to a median 16.2 pmol/L (IQR: 14.2–19.6 pmol/L) at 10 years or (last available test result), transitioning to an approximately Gaussian distribution. The levels of TSH were suppressed at baseline, with a median value of 5.0 mU/L). The number of missing results was as follows: 3 months (n = 20), 6 months (n = 23), 12 months (n = 25), 18 months (n = 26), 2 years (n = 30), 5 years (n = 30), and 10 years (n = 46). There was no difference in the demographic profile (by age and sex) between those for whom 10-year follow-up data was available vs. those with data only available up to an earlier point in time. The baseline levels of TSHRAb (Reference Range: 0.0–0.9 IU/L) were available for all 100 individuals and showed a median level of 7 U/mL (6 IU/mL for males and 7 IU/mL for females). Overall, the females were more likely to have dispersed values compared to males. Subsequent TSHRAb titre was available for 34 individuals and showed median levels lower at 3.3 IU/mL for females (n = 29) with the median time of rechecking at 2 years (range: 1 month to 10 years). Eighty-seven individuals received carbimazole in the first year. Eight people were started on propylthiouracil, three did not receive any medical treatment while two received both propylthiouracil and carbimazole in that time. At 10 years after first assessment or the most recent follow-up point (Figure 1), seven of the individuals had died (three males (20% of the total males) and four females (4.7% of the total females)). Of the original group of 100 people, 28 of those still alive were still on medication at last follow-up (four or 26% for males and 24 or 28.2% for females). Two of the females were on propylthiouracil, while the rest of the individuals (n = 26) were on carbimazole at the time of last follow-up. At last follow-up, 39 individuals were euthyroid post-pharmacotherapy alone, that is, not on pharmacotherapy. Twenty-two remained on pharmacotherapy alone, 13 had received radioactive iodine, 12 had undergone thyroid surgery, while smaller groups required combined approaches (pharmacotherapy plus radioactive iodine in four, pharmacotherapy plus surgery in two, and radioactive iodine plus surgery in one). The development of thyroid eye disease showed consistent proportions in both sexes with four (26.7%) males affected and 21 females (24.7%) affected. Overall, 37 individuals relapsed, nine of these being post-radioactive iodine. For those who relapsed the median age at diagnosis was 34 years (IQR: 30–53) while the median age at diagnosis for those who did not relapse was 42.0 years (IQR: 32–52). Thirteen (35%) of the 37 individuals who relapsed in relation to thyrotoxicosis were still on anti-thyroid medications at the 10-year or most recent follow-up point with all 13 of these individuals being on carbimazole in the range of 5–15 mg per day. Relapse rate was higher in those people who were treated with anti-thyroid medications only and for those who had a higher baseline TSHRab (univariate logistic regression: hazard ratio 1.6 (95% confidence interval 1.4–1.8) for likelihood of relapse over 10 years vs. baseline TSHAab). The rate of relapse of thyrotoxicosis was twice in females compared with males, but the number of males overall was small at 15 so statistical comparison is not meaningful. Thirty-four of the 37 individuals who relapsed were females with males 3 in number. Median TSH levels were similar at baseline for both individuals who relapsed and who did not but the individuals who did not relapse showed consistently higher TSH levels for the first 18 months, reaching a median value of 1.2 mU/L (IQR: 0.07–2.9 mU/L) compared with individuals who relapsed, in whom median TSH levels slightly increased over a similar period to 0.3 mU/L (IQR: 0.1–2.9 mU/L). Baseline TRAb level showed a modest positive correlation with baseline FT4 (Spearman's rho = 0.284, p = 0.004). No significant correlations were observed between baseline TRAb and FT4 at 3, 6-, 12-, 18-, or 24-month follow-up. In summary, we here describe the prospective outcomes for 100 consecutive patients presenting with Graves' disease. There have been a number of studies that have looked at the longer-term outcome for patients with Graves' disease as summarised in a 2020 review 3. Our paper goes out to a timeline of up to 10 years following initial diagnosis and presentation. The fact that 28% of patients (some following surgery or radioactive iodine treatment) remained on anti-thyroid medication at last follow-up and that only 39% were euthyroid following pharmacotherapy alone, highlights the importance of patients with Graves' disease undergoing long term TFT checks and those on active treatment remaining under specialist endocrinology follow-up in the longer term (this could be with shared care)—but also is indicative of the relevance of new therapeutic approaches, including immunotherapy 4, 5. This proportion of people with Graves' disease still on pharmacotherapy does not reflect any specific departmental treatment policy. The percentage of people remaining euthyroid off treatment at up to 10-year follow-up accords with the seminal 1961 paper of Drury, who reported that approximately 30% of Graves' patients remain in remission vs. the higher 50% figure that is often quoted 6. Only about one third of patients remained in remission off anti-thyroid treatment, with a further one third requiring definitive treatment with thyroidectomy or radioactive iodine, with significant implications for ongoing care costs. Individuals who did not relapse showed consistently higher TSH levels for the first 18 months, while in individuals who relapsed there was no difference in FT4 levels vs. those who did not relapse. This suggests that ensuring adequate suppression of excessive thyroid hormone production (with TSH consequently higher) is important in relation to stabilising the trajectory of thyroid hormone profile in people treated for Graves' disease and points to the importance of interval monitoring of TSHRAb titre in people with Graves' disease 4. However, reduction in TSHRAb antibody levels does not necessarily correlate with treatment response either to pharmacotherapy or radioactive iodine 4. Our finding that baseline TSHRAb was associated with a higher relapse rate for those on anti-thyroid treatment alone, is in keeping with previous studies 7 with repeated measurements of TSHRAb increasing the precision of relapse prediction 8. It should also be stated that although serial TSHRAb measurements are used in clinical practice and have been reported in prior studies to aid relapse risk stratification, the present service-based dataset was not designed to assess whether repeat TSHRAb measurement improves outcomes or clinical decision-making. In relation to limitations, we accept that we have looked at a relatively small sample of people. Also, not all patients had tests done at the specific intervals. We agree that missing data at prespecified follow-up timepoints is a limitation of this service-based evaluation and reduces power for longitudinal inference. Follow-up testing was undertaken as part of routine clinical care rather than at protocol-defined intervals, resulting in incomplete biochemical data at several interim timepoints. We accept that the regression analysis should be considered exploratory in nature. However, this longer-term follow-up study, up to 10 years post-diagnosis in a single centre, provides an important insight regarding the trajectory of thyroid hormone status in people with Graves' disease. The high proportion of deaths in men is in keeping with the elevated cardiovascular mortality previously reported by Okosieme et al. 9 but the number of males is small. It is recognised that early and effective control of thyrotoxicosis among patients with Graves' disease is associated with improved survival. Rapid and sustained control of hyperthyroidism should be prioritised in the management of Graves' disease 9. In this regard, we report that a proportion of people still require anti-thyroid treatment many years after diagnosis. The relapse rates in relation to various modes of treatment are comparable to previous surveys 10. This report, over an up to 10-year time scale, highlights the importance of the creation of national long-term registers of people with Graves' disease with long-term monitoring of thyroid function, so that longer-term outcomes can be fully understood from a biochemical and immunological perspective and that the outcomes of new therapies evaluated, with pooling of data from thyroid clinics in diverse settings. Muhammad Muneeb Ibad and Adrian Heald wrote the manuscript. Natalie Gallant, Sarah Jamil, Muhammad Muneeb Ibad, Sangeeth Veluchamy, and Mrinalini Srichandran collected the data and continued to the analysis led by Sangeeth Veluchamy. Waseem Majeed, Suhani Bahl, Lakdasa Premawardhana, Colin Dayan, Onyebuchi Okosieme, Peter Taylor, and Rupinder Kochhar provided clinical context. All authors contributed to and have approved the final version of the manuscript. The authors have nothing to report. The study was approved by our local ethics board reference SRH 2025: 13. The authors declare no conflicts of interest. The data that support the findings of this study are available from the corresponding author upon reasonable request.