Abstract
Background:
Hyperthyroidism is known to have a significant impact on quality of life (QoL), at least in the short term. The purpose of the present study was to assess QoL in patients 6–10 years after treatment for Graves' disease (GD) with radioiodine (RAI) compared to those treated with thyroidectomy or antithyroid drugs (ATD) as assessed with both thyroid-specific Thyroid-Related Patient-Reported Outcome (ThyPRO) questionnaire and general (36-item Short Form Health Status) QoL survey.
Methods:
The study evaluated 1186 GD patients in a sub-cohort from an incidence study 2003–2005 who had been treated according to routine clinical practice at seven participating centers. Patients were included if they had returned the ThyPRO (n = 975) and/or the 36-item Short Form Health Status survey questionnaire (n = 964) and informed consent at follow-up. Scores from ThyPRO were compared to scores from a general population sample (n = 712) using multiple linear regression adjusting for age and sex as well as multiple testing. Treatment-related QoL outcome for ATD, RAI, and surgery were compared, including adjustment for the number of treatments received, sex, age, and comorbidity.
Results:
Regardless of treatment modality, patients with GD had worse thyroid-related QoL 6–10 years after diagnosis compared to the general population. Patients treated with RAI had worse thyroid-related and general QoL than patients treated with ATD or thyroidectomy on the majority of QoL scales. Sensitivity analyses supported the relative negative comparative effects of RAI treatment on QoL in patients with hyperthyroidism.
Conclusions:
GD is associated with a lower QoL many years after treatment compared to the general population. In a previous small randomized controlled trial, no difference was found in patient satisfaction years after ATD, RAI, or surgery. Now, it is reported that in a large non-randomized cohort, patients who received RAI had adverse scores on ThyPRO and 36-item Short Form Health Status survey. These findings in a Swedish population are limited by comparison to normative data from Denmark, older age, and possibly a more prolonged course in those patients who received RAI, and a lack of information regarding thyroid status at the time of evaluation. The way RAI may adversely affect QoL is unknown, but since the results may be important for future considerations regarding treatment options for GD, they need to be substantiated in further studies.
Introduction
Graves' disease (GD) is a common cause of hyperthyroidism, with an incidence of 21 patients per 100,000 inhabitants in Sweden and 20–50 per 100,000 in other countries (1,2). Antibodies directed against the thyrotropin (TSH) receptor stimulate the thyroid to increase its production and release of thyroid hormones (1,3). Development of GD is associated with autoimmune disposition but also occurs sporadically (1,3). Graves' hyperthyroidism can be accompanied by extrathyroidal manifestations such as ophthalmopathy, dermopathy, and acropathy, and carries the risk of relapse as long as there is a sufficient amount of thyroid tissue left after treatment (3 –5). In addition, patients treated for Graves' hyperthyroidism run an increased risk of cardiovascular morbidity and mortality in some but not all studies (6 –8). Increased mortality after radioiodine (RAI) was thus not seen after patients were euthyroid on thyroxine (T4) replacement (7).
In recent years, more attention has been given to the potential negative impact of hyperthyroidism on patients' quality of life (QoL). Previous randomized studies of Graves' hyperthyroidism indicate that general QoL, as measured by the 36-item Short Form Health Status survey (SF-36), is similar between the three treatment modalities antithyroid drugs (ATD), RAI, or surgery (9,10). Every fifth or sixth patient, however, does not feel completely recovered within one year after treatment, and many are not yet fully recovered even after three years or more (5). There are minor deficits such as decreased vitality in GD patients three years as well as 14–21 years after treatment compared to the general population (9,10). Of particular interest for the present study, decreased general health was also observed in RAI-treated patients in a previous study (10). Patients treated with RAI also scored lower on a simple general well-being visual analog scale compared to ATD and surgery (11). Other prospective studies of shorter duration in patients with hyperthyroidism due to GD or toxic nodular goiter also indicate impairments of QoL aspects when evaluated by the SF-36 and, more recently, with the thyroid-specific Thyroid-Related Patient-Reported Outcome questionnaire (ThyPRO) (12 –15). QoL is even more impaired if the patient develops thyroid-associated orbitopathy (TAO) (16 –20).
Randomized trials have many advantages but may be of limited size and may not be fully representative of the spectra of treatment outcomes seen in the clinic. Such studies also suffer from limitations caused by bias through study inclusion criteria and possibly a more attentive monitoring. A pertinent question therefore arises: do the results from randomized trials apply to the usual and routine clinical practice when larger groups of patients are studied?
In 2003–2005, all patients with a de novo diagnosis of hyperthyroidism (n = 2916) were registered at seven centers in southern Sweden, covering 40% of the total Swedish population, to determine the incidence of GD (21). The patients were treated according to clinical practice. A follow-up study of QoL 6–10 years after diagnosis has now been performed using the validated Swedish versions of the thyroid-specific ThyPRO and the general SF-36 questionnaires (11,22). The results from the sub-cohort of adult patients with GD are presented below.
The aims of this study were to assess the long-term QoL in patients with GD who have been treated in a routine clinical setting and to examine whether the results from previous randomized trials are applicable for today's clinical practice.
Methods
A total of 2916 patients were diagnosed with hyperthyroidism throughout Sweden during 2003–2005 in an open study, and 2204 adults of these had GD (16). Of the GD group, 1976 patients were approached, of whom 245 declined to participate and 545 did not respond, leaving 1186 patients who completed the clinical survey and participated in a follow-up study 6–10 years after the original diagnosis (M ± standard deviation [SD] = 8 ± 0.9 years). Responses to the two QoL questionnaires were obtained from 975 patients (ThyPRO) and 964 patients (SF-36), along with their clinical data. All patients met the following inclusion criteria: after informed consent, the patient (i) agreed to participate in the study and (ii) had returned a completed thyroid-specific ThyPRO and/or generic SF-36 survey (23 –27). Some patients (n = 245) actively declined to participate in the study and were therefore excluded, while 545 GD patients did not respond to the study invitation at all, despite repeated efforts (non-responder group). Permission to go through all medical history records of this group was granted by the commission of ethics. Since they did not respond, ThyPRO and SF36 data for this group are not available. However, the medical history records for the non-responder group were scrutinized regarding ethnicity, treatment modality, and clinical outcome in a clinical follow-up of the study group (11). Patients with subclinical hyperthyroidism or relapse of GD were not registered or included in the incidence study 2003–2005. The patients were treated with ATD, RAI, or surgery according to usual and customary clinical practice within Sweden (11). ATD treatment was given for 12–18 months predominantly with methimazole or, less frequently, with propylthiouracil combined with levothyroxine as “block-and-replace therapy” with the goal of achieving and maintaining TSH within the normal reference range. Remission was defined as a minimum of 12 months of ATD treatment together with discontinuation of ATD and the clinical judgment that the disease was not clinically active. Recurrence of GD was defined as recurrence of hyperthyroidism after remission (11).
Validation of data
All data sheets in the two questionnaires were auto-scanned, and the data were then auto-transferred into a database using a licensed scanning software (Remark Office OMR 8©; Remark, Malvern, PA). The database was validated by cross-checking with returned QoL questionnaires, and the clinical outcome reported was validated through the case history records from randomly selected participants. Only 0.92% of the controlled data values needed correction during the validation process.
QoL questionnaires
The ThyPRO is a questionnaire that is completed by the patient measuring thyroid-related QoL of physical and mental symptoms, impaired function, well-being, and participation, as well as negative impact on overall QoL (24). It has been thoroughly validated, including cross-culturally, and consists of 85 items summarized in 13 multi-item scales and one single-item scale (22). In this study, a validated Swedish version of ThyPRO was used. In addition, a composite score summarizing the nine functions, well-being, and participation scales can be scored (28). Each item is rated on a five-point Likert scale, and items are summarized and linearly transformed to 0–100 scales, with higher scores indicating more symptoms or impact. Reference material for ThyPRO scores has been established based on the Danish general population (29). These data were obtained from a random sample with a female/male sex distribution of 4:1 drawn from the central personal registry.
The SF-36 is a generic extensively validated health-status measure consisting of 36 items summarized in eight subscales measuring physical, mental, and social function and well-being, ranging from 0 to 100, with higher scores indicating better health status (30). In addition, two summary scores are constructed from the eight subscales: physical and mental component summary (PCS and MCS), respectively. These are norm-scored to a mean of 50 and SD of 10 (26).
Data analyses
Patients were subdivided into three groups according to treatment received since diagnosis: (i) patients treated with ATD only, (ii) patients treated with RAI (irrespective of previous ATD treatment) but not thyroidectomy, and (iii) patients treated with subtotal or total thyroidectomy, regardless of previous ATD or RAI treatment. Mean scale scores from the above-mentioned patient groups were compared to general population norms from the Danish general population using scales for which a valid response from people without thyroid disease could be obtained (i.e., scales where the questions are not asked with attribution to thyroid disease).
Statistical methods
All statistical analyses were performed using SAS v9.4 (SAS Institute, Cary, NC). Differences in baseline characteristics and mean scale scores between the three treatment groups were evaluated with multiple linear regression using SAS PROC GLM. All estimations and tests were adjusted for age and sex. Differences were regarded statistically significant if both the overall F test of variance and the type III sum of squares test of effect of treatment had a p-value of <0.05 after controlling for multiple testing using the Hochberg Benjamini test (31). Sensitivity analyses for stability of results were performed by repeating analyses in subgroups having received only one type of treatment: (i) ATD, (ii) RAI and never ATD, or (iii) thyroid surgery and never ATD or RAI. Further, as a proxy for long-standing, recurrent, or treatment-resistant disease, the number of treatments received were also separately controlled for. Finally, to adjust for possible confounding of comorbidity, all analyses were repeated with comorbidity as a covariate. Comorbidity was registered as present or absent. Chronic autoimmune conditions that have been associated with GD and may have an impact on QoL such as diabetes type 1 or 2, vitamin B12 deficiency, rheumatic disease, celiac disease, Addison's disease, vitiligo, psoriasis, sarcoidosis, and other chronic diseases were regarded as comorbidities.
Ethical aspects
The study was approved by the Regional Ethical Review Board in Uppsala (EPN Dnr 2012/035, 2012 April 4). Informed consent was obtained from the study participants, as the study was performed according to the Declaration of Helsinki.
Results
The baseline characteristics for the participating patients, including statistically significant differences in these, are shown in Table 1. The basic characteristic for the 545 non-responder patients, where medical history records were scrutinized, showed slight deviations from the study group. In essence, patients in the non-responder group were 4.4 years younger, more often born outside Europe (26% vs. 11%), and treated with ATD (71% vs. 65%) but less often with RAI (22% vs. 27%) compared to the included GD patients (11). Otherwise, TAO and smoking habits were similar to those of the included GD patients.
Basic Characteristics of the Patient and General Population Samples
Data are shown as n (%) or median (interquartile range; Q1–Q3). Differences among groups were tested for variables not used in definition of groups or clinically implied (i.e., age, sex, and comorbidity). Patients in the RAI group may have received ATD but not thyroidectomy. Patients were allocated to the surgery group if they had received thyroid surgery, irrespective of previous ATD or RAI treatment.
p < 0.05; *** p < 0.001.
RAI, radioiodine; ATD, antithyroid drugs; na, not available.
QoL compared with the general population
As shown in Table 2, patients treated with RAI had worse thyroid-related QoL scores on all comparable scales compared to the general population. The deficits were most pronounced for the physical symptoms scales hyperthyroid, hypothyroid, and eye symptoms and for anxiety and depressivity. Patients treated with ATD or surgery had worse scores on most scales, but similar scores for tiredness and emotional susceptibility compared to the general population.
Mean (SE) ThyPRO Scale Scores Adjusted for Sex and Age Among Patients with Graves' Disease Compared to a Reference Group Randomly Selected from the Danish General Population
Patients are categorized according to treatment, as in Table 1. Scale score range 0–100, with increasing scores indicating worse health status.
Significantly (p < 0.05 with Hochberg correction) different from the general population sample.
SE, standard error; ThyPRO, Thyroid-Related Patient-Reported Outcome questionnaire.
QoL related to treatment modality
Differences in thyroid-specific QoL between the three treatment groups are presented in Table 3. As shown, patients treated with RAI had worse QoL than patients treated with ATD or surgery measured by ThyPRO. This was most pronounced for impaired sex life and tiredness and least for goiter symptoms, but compared to patients treated with ATD, the deficits were worse on all scales except cognitive complaints and overall QoL impact (Figure 1). Compared with surgery, RAI-treated patients differed most regarding tiredness, anxiety, depression, and impaired daily life. In addition, the ThyPRO-scores for RAI patients were lower on all scales except hypothyroid symptoms, eye symptoms, cognitive complaints, appearance, and overall QoL impact (Table 3). Patients treated with surgery had worse ThyPRO scores compared to patients treated with ATD on hypothyroid and eye symptoms as well as on appearance.
Mean Thyroid-Related Patient-Reported Outcome (ThyPRO) questionnaire scale scores, adjusted for sex and age among patients with Graves' disease, treated with antithyroid drugs (ATD), radioactive iodine (RAI), or surgery, as well as scores from a general population sample. Patients in the RAI group may have received ATD but not surgery. Patients were allocated to the surgery group if they had received thyroid surgery, irrespective of previous ATD or RAI treatment. Items in four scales (impaired social life, impaired daily life, impaired sex life, and appearance) are asked with attribution to thyroid disease and cannot be answered by respondents from the general population. Scale scores range from 0 to 100, with higher scores indicating worse health status.
Mean (SE) ThyPRO Scale Scores Adjusted for Sex and Age Among Patients with Graves' Disease Treated with ATD, RAI, or Surgery, as Described in Table 1
Scale scores range 0–100, with increasing scores indicating worse health status.
Significantly (p < 0.05 with Hochberg correction) different from group(s) marked by a.
Significantly worse scores are shown in bold.
With respect to general QoL, a similar pattern was found (Table 4). Patients treated with RAI had worse (lower) scores on all SF-36 scales compared to ATD, and the deficits were most pronounced for general health, role limitations, and social function. The RAI-treated group also had significantly lower general QoL measured by SF-36 except for bodily pain and the physical component summary compared to the surgically treated group.
Mean (SE) SF-36 Scale Scores Adjusted for Sex and Age Among Patients with Graves' Disease Treated with ATD alone, RAI, or surgery, as Described in Table 1
Scale scores range 0–100, with increasing scores indicating better health status.
Significantly (p < 0.05 with Hochberg's correction) different from group(s) marked by a.
Significantly worse scores are shown in bold.
SF-36, 36-item Short Form Health Status survey.
According to the sensitivity analyses, the above-described pattern was robust regarding both the number of treatments entered as covariates, which did not change the results, and, as shown in Table 5, when the analyses were restricted to patients who only received one treatment modality. For the latter, the surgery group was very small. Compared to patients receiving only ATD, patients treated with RAI had worse ThyPRO scores on eye symptoms, tiredness, anxiety, depressivity, impaired social life, and impaired daily life and on the ThyPRO composite scale, and they had worse SF-36 scores on physical function, general health, social function, mental health, and mental component summary. Compared to the small group treated only with surgery, the RAI-treated group had worse ThyPRO scores on impaired social life and worse SF-36 scores on physical function, social function, mental health, and mental component summary. Repeating analyses with adjustment for comorbidity did not alter these results.
Mean (SE) ThyPRO and SF-36 Scale Scores Adjusted for Sex and Age Among Patients with Graves' Disease Having Received Only One Treatment Modality: ATD, RAI, or Surgery
Scale scores range 0–100 for ThyPRO and SF-36 subscales, whereas SF-36 component summaries are scored with a mean of 50 and standard deviation of 10. For ThyPRO, higher scores indicate more symptoms/impact, whereas for SF-36, higher scores indicate better health status.
Significantly (p < 0.05) different from group(s) marked by a.
Significantly worse scores are shown in bold.
Discussion
The main purpose of the present study was to assess both long-term thyroid-specific and general QoL after treatment for GD in a large cohort of consecutively enrolled patients treated according to usual and customary clinical practice in Sweden.
The patients treated for GD had reduced QoL compared to the general population. Regardless of treatment modality, GD patients had worse QoL on the disease-specific ThyPRO compared to the general population sample, although this was to a varying extent. This is in line with a previous randomized, prospective, long-term study (14–21 years of follow-up) of 179 patients treated for GD where decreased vital and mental aspects of QoL were observed compared to the general population (10). The present study, based on 6–10 years of follow-up in a patient group more than five times larger supports the observation that patients treated for GD have worse thyroid-specific QoL than the general population.
There may be several possible explanations for the QoL impairment observed in GD patients. In all treatment groups, patients have more eye symptoms than the general population, which may be explained by persistent or even undiagnosed mild orbitopathy that may contribute to the lower QoL for the GD group. After a diagnosis of hyperthyroidism, patients have a significantly higher risk of being diagnosed with both somatic and psychiatric disease, which could explain some of the QoL impairments at long-term follow-up (6). Being diagnosed with a disease may change the patients' health perception, which could influence their perceived QoL and thereby the response to QoL assessments. In fact, indications of such mechanisms have been identified among patients with thyroid diseases (32). Could our choice of reference data for the general population partly explain the deficit in QoL for the GD patients? ThyPRO data from the general population in Denmark were used as the reference population, since no ThyPRO general population data from Sweden are available (29). The two Nordic countries, however, are very similar in life-style, overall standard of living, and health, and the countries have a similar welfare system (33). Therefore, the Danish norm data are regarded as a valid comparator. In addition, previous research has shown essentially identical levels of health-related QoL in Denmark and Sweden, and a cross-cultural validation of ThyPRO found minimal differences between these two populations (22,34).
The three different treatment modalities were also compared, and it was discovered that RAI-treated patients had significantly worse QoL outcome compared to patients who received ATD or surgery when evaluated with both the thyroid-specific as well as general QoL survey. The most immediate explanation would be older age and excess comorbidity in the RAI-treated group. However, the result still stands when the analysis was adjusted for these parameters. According to Cohen et al., all identified differences were small, but clinically significant (i.e., with effect sizes 0.2–0.5, corresponding to a difference in ThyPRO 0–100 scores of 4–10 points) (35). Thus, large sample sizes are required for identification of such differences. The RAI group, apart from individuals who received RAI as the primary choice of treatment, also consisted of patients with one or more recurrences of GD but where other treatments were regarded unfeasible. In any event, the latter patients presumably had a more extended and complicated disease course, which may also negatively affect QoL in the long term. The patients treated with RAI may have a more negative perception of their QoL due to this more protracted treatment course. RAI treatment causes thyroid antigen exposure, which further causes an autoimmune response that is clearly demonstrated by prolonged elevation of TSH receptor antibodies (TRAb) after therapy (5,36). Whether TRAb or unidentified thyroid-related autoimmune components per se may have a negative long-term impact on QoL is largely unknown, but a negative impact of TRAb on general physical well-being has been observed (37).
It could be argued that the thyroid hormone status after treatment at the time of follow-up may have influenced the patients' QoL assessments. For example, patients treated with ATD are more likely to be euthyroid in the long run compared to subjects who had RAI or surgical thyroid ablation and thus require levothyroxine therapy. In an earlier study, an association was seen between low T3 concentrations and a low general health score and an inverse relation between TRAb and the PCS of the SF-36 (37). On the other hand, no association was observed between T4 or TSH levels and general QoL or between TSH and the response to the question “Do you feel well now?” at the three- to four-year follow-up after treatment (37). It is therefore unlikely that the worse QoL score for the RAI group in the present study is related to patients' thyroid hormone levels. In addition, no consistent thyroid hormone or antibody parameters were available at follow-up in the present study. The RAI group had an average age of 53 years at registration in 2003–2005 and is now 6–10 years older, but age has been accounted for in the sensitivity analysis. Older age is therefore an unlikely explanation for worse QoL after RAI treatment.
Treatment-related difference in QoL outcome is a new and interesting observation of clinical relevance, and it differs from previous studies (10,20). At three-year follow-up of GD patients randomized to ATD, RAI, or surgical treatment, no significant differences were found, as evaluated with a clinical questionnaire developed for the purpose in collaboration with an expert in the QoL field at the Swedish National Center of Health Statistics in Sweden (9). Likewise, differences related to treatment modality were not found when the SF-36 instrument was used in a follow-up of the same study population 14–21 years after treatment, or in a subsequent randomized trial between the ATD and RAI of 308 subjects (10,18). The major difference between the earlier observation of no treatment-related differences in QoL and the present study is the larger sample size in the present study and better instruments for QoL measurements, including both the thyroid-specific ThyPRO and general SF36 QoL questionnaire. It is also possible that the open study design with patients treated as part of clinical routine may better reflect real-life results compared to a randomized study of limited size. In the present study, the RAI group also had worse QoL when only subjects who had undergone one kind of treatment were included in the analysis, and when sex and age were adjusted for in the analysis. In this way, confounding of the data by change of treatment modality and relapses was ruled out. This strengthens the validity of the observation. Patients treated with RAI had more comorbidities (44%) than patients treated only with ATD or with surgery (29% vs. 33%, respectively). However, adjusting analyses for comorbidity did not alter the results or change the conclusions, confirming the robustness of the present findings.
RAI treatment has been used since the 1940s and is the preferred choice of treatment for an index patient with GD in North America, and less preferentially, it is also commonly used in Europe and throughout the rest of the world (38,39). RAI treatment is reliable and safe, but it is associated with an increased risk for development of TAO, although this may to a large extent be prevented with glucocorticoid prophylaxis in patients with milder TAO activity (39 –44). The risk of TAO associated with RAI may contribute to its diminished use as the preferred treatment for an index case of Graves' hyperthyroidism, as has been observed over the last decades in the United States and Europe (38,39). Increased cardiovascular morbidity and mortality, predominantly cerebrovascular events and atrial fibrillation, especially during the first year after treatment with RAI but also in the long term, has also been reported, but this seems to be a consequence of hyperthyroidism per se rather than the previous treatment with RAI (8,45,46). Hyperthyroidism also leads to increased bone turnover and loss of bone mass. An increased risk of fracture after hyperthyroidism is well known, although not all studies have shown this (47 –49). Registry studies have shown an association between previous RAI treatment and increased fracture risk (50). Taken together, there are several reported possible consequences of previous RAI treatment, some of which may have relevance for the QoL impairments observed in the present study and which may support an indirect link between RAI treatment and reduced QoL. Increased risk of extrathyroidal cancer after RAI for benign thyroid diseases has been reported in some but not all studies (51 –56). However, the cancer incidence after treatment for GD is extremely small (≤1%), and solid cancers usually first develop >10–15 years after radiation exposure. It is therefore unlikely that reduced QoL in RAI-treated GD patients can be explained by secondary cancers at this 6- to 10-year follow-up.
Patients treated with RAI and surgery reported more eye symptoms compared to the ATD-treated group. The reason for this is not clear but may in some cases be due to development of TAO associated with RAI treatment, as reported earlier (41,42,44). Another possible explanation for elevated eye scores in surgically treated patients may be that thyroid ablation with total thyroidectomy was often preferred in patients with a more complicated clinical course such as persistent or relapse of GD, as well as long-standing active TAO, to reduce the amount of thyroid tissue and thereby reduce antigen expression.
Strengths of the study include that the results are based on the largest cohort to date of patients who have been treated for GD according to usual clinical practice at the participating centers within Sweden. All patients with GD were registered in the incidence study during 2003–2005, and they were all recently invited to participate in the present study. Therefore, the validity of the results is not biased due to narrow study inclusion criteria often used in randomized trials. Furthermore, both a validated thyroid-specific and a validated general QoL survey (SF36) were applied to assess the long-term QoL impact in patients treated for GD. The data are both derived from patient reports of comorbidity and validated through the individual medical history record. The follow-up was performed long after the patients were diagnosed and treated for GD (2003–2005), and therefore reflect the long-term QoL rather than just the immediate effects of treatment.
Limitations of the study include that no difference in patient satisfaction was shown years after ATD, RAI, or surgery in the small randomized controlled trial (9). It is now reported that in a large non-randomized cohort, patients who received RAI had adverse scores on the ThyPRO and SF-36. These findings in a Swedish population are limited by comparison to normative data from Denmark, older age, and possibly a more prolonged course in those patients who received RAI, as well as a lack of information regarding thyroid status at the time of evaluation. The follow-up was performed 6–10 years after treatment for hyperthyroidism, which also may be a weakness, as it can make it difficult to answer questions with attribution to thyroid disease, that is, the patients may find it difficult to judge whether their symptoms at present are related to their previous thyroid disease. Another possible weakness is the multicenter design of the study, since there may have been local variations in the clinical routine, despite the fact that each center adheres to a Swedish treatment consensus (57). The lack of randomization to ATD, RAI, or surgery may also imply a limitation when the effectiveness of the three treatment modalities to restore QoL at long-term follow-up is compared. The response rate was 51% (975/1916) of the total GD cohort at follow-up. Whether QoL in the non-responder group is different from the included GD group is unknown. They could have declined participation because they were doing well or because they did not feel well. In addition, a different ethnic background may affect willingness to participate.
Finally, slightly more patients provided a sufficient response to ThyPRO compared to SF-36. The content of a disease-specific instrument is often experienced as more relevant by the patients, which previously has been shown to increase the willingness to complete a questionnaire. This is considered the most likely explanation for this difference in response frequency (58).
In summary, patients treated for Graves' hyperthyroidism had worse QoL 6–10 years after diagnosis compared to the general population. Patients treated with RAI had more impaired long-term thyroid-specific and generic QoL compared to patients treated with ATD and surgery. Given the large number of patients who develop GD, the new observation of long-term negative impact of RAI on QoL needs further substantiation due to its important clinical consequences.
Footnotes
Acknowledgments
This work was performed with grants from the ALF-agreement in Västra Götaland County, Sweden, and from the Swedish Medical Society, Agnes and Knut Mørk's Foundation, and Else and Mogens Wedell-Wedellsborg's Foundation. M.L. was supported by grants from the Faculty of Medicine at Lund University and the Skåne Research Foundation. We thank Associate Professor Anthony Firek, University of California Riverside School of Medicine, for constructive criticism of the text.
Author Disclosure Statement
T.W. holds the license for ThyPRO and has worked as consultant for AbbVie, Inc. No competing financial interests exist for the remaining authors.