Systematic thyroid screening in myotonic dystrophy: link between thyroid volume and insulin resistance

Myotonic dystrophy (DM) is the most common inherited, autosomal dominant neuromuscular disorder in adults, affecting 1 out of 8000 people. This multi-systemic disease causes myotonia and muscle weakness in skeletal muscles with a risk of life-threatening cardiorespiratory disorders. The disease is very heterogeneous with regard to the age of onset, clinical manifestations, and severity. Two genetic types of DM have been described corresponding to an expansion of, respectively, cytosine thymidine guanine (CTG) and CCTG repeats in non-coding regions (3′-untranslated region) of the myotonic dystrophy protein kinase (DMPK) gene for DM1, and zinc finger protein 9 (ZnF9) gene for DM2.

The length of the (CTG)_n repeat expansion in DM1 is correlated with the severity of the disease and age of onset, defining five clinical types (congenital, infantile, juvenile, adult onset and late onset forms) [1]. The nuclear retention of mutant ribonucleic acid (RNA) alters RNA metabolism in patient’s tissues by targeting RNA-binding proteins, particularly Cytosine-Uridine-Guanine-binding protein 1 (CUGBP1) and muscle blind-like protein 1 (MBNL1). The phenotype variability is also attributed to an anticipation mechanism and possible somatic mosaicism. There is currently no curative treatment [1, 2].

Establishment of patient registries, such as DM-Scope [1‐3], helps to define the phenotype of patients. DM1 patients have an increased incidence of endocrine dysfunction, especially gonadal insufficiency, diabetes and thyroid disorders [4, 5]. Palpable thyroid gland abnormalities have been described in about 20% of DM1 patients. Moreover, an increased risk of cancer [6‐12], including thyroid cancer, was also reported in three recent observational studies (national Swedish/Danish, United Kingdom and American myotonic dystrophy patient registries) and one meta-analysis [13‐15]. However, no studies have specifically targeted a systematic biological and imaging thyroid check-up.

The purpose of this study was to determine the prevalence of thyroid disorders, especially goiters, and any correlated factors in a French cohort of DM1 patients who received the same standardized thyroid evaluation, including a systematic thyroid ultrasound (US). We secondly aimed at identifying goitrogenic factors in this population.

General, metabolic and thyroid characteristics of the whole cohort are given in Tables 1, 2 and 3, respectively.

This study is the first evaluation of the frequency of thyroid disorders in a cohort of DM1 patients using well-standardized ultrasound and biological evaluation. Sixty percent of this cohort of 115 patients had US thyroid nodules or goiter, 7.8% had PTCs and about 20% had sub-clinical thyroid dysfunction before any surgery.

This prevalence of 60% of thyroid dystrophy is close to the prevalence found by US in the general population (45 to 67%) [17], age-matched general population (42%), and in autopsy series (60%) [18, 19]. Thyroid volume did not differ between gender and age (more or less than 45 years) in this cohort.

The main interest of US scans was to identify nodules suspicious for malignancy according to the TIRADS classification, in a population in which a higher risk of cancer has been identified [12]. In this series, 17.4% (20/115) of the patients had suspicious nodules, classified as TIRADS 4 or 5, by US examination. About 22% of the performed FNAC led to undetermined cytology (Bethesda I) that seems relatively high compared to literature [20]. This might be related to the fact that FNAC was performed in nodules below 10 mm when TIRADS results were suspicious since the study was undertaken before American Thyroid Association (ATA) guidelines [21]. A little less than 14% of the whole cohort had a thyroidectomy because of suspicious FNAC or large goiter. A high prevalence of PTCs (nearly 8% of the cohort) was identified, which is consistent with the high frequency of cancers (especially thyroid cancer) already noted in four observational DM cohorts [11, 13‐15]. The overall PTC prevalence in our cohort screened routinely by US was, however, a little higher than that previously found in other DM cohort (8% vs. about 4%), which is not surprising when considering the systematic screening effect. Indeed, the increase of PTC was mainly due to an increase of micro-PTCs, which corresponds to two-thirds of the PTCs cases in our series. A high frequency of micro-PTCs has already been reported in the general population and is attributed to an improvement in diagnostic techniques [22]. Nevertheless, to get a local reference, the prevalence of micro-PTCs in operated non-DM patients in the same surgery department, studied by the same pathologists during the same period was 17% which is lower than the prevalence observed in our operated DM1 patients (6 micro-PTCs out of 16 surgeries in the DM group, e.g., 37%). Even if this comparison should be taken with caution, it suggests that DM patients may have an increased risk of micro-PTCs, in accordance with recent data demonstrating a high frequency of thyroid cancer, in DM, for yet unknown reasons.

The relevance of routine thyroid ultrasound in DM has however to be discussed. Indeed, most cancers were micro-PTC with a good prognosis and low risk of recurrence [21]. In our series, five cancers had not been suspected clinically but were only found on US examination. The final diagnosis was two cases of micro-PTCs and one low-risk macro-PTC, and a delayed diagnosis probably would not have changed the patient’s prognosis. Nevertheless, out of the nine PTCs, four micro-PTCs were multifocal, with one of the four having a capsular breach (pT3) and two macro-PTCs were of intermediate risk. Thyroid US screening is non-invasive and not very expensive. Therefore, on one hand, routine thyroid evaluation seems justified in order to not miss any cancers. On the other hand, an over-diagnosis of PTC, which usually has a good prognosis and progresses slowly, could lead to over-treatment with all its implied risks [23] (e.g., general anesthesia and cervicotomy), especially if not done in a reference center. Most of these patients have comorbidities and are considered high-risk in terms of cardiorespiratory complications, although no serious postoperative complications were reported in our group. Follow-up study is needed to determine if the evolution pattern of PTCs in DM patients and also the prognosis are similar to general population. This will help to justify a systematic screening, in this population, now considered as a high-risk population for cancer.

Our secondary aims were to determine whether any associated factors to goiter in this DM population, could guide the US thyroid screening. The classical factors of goitrogenesis, such as family history of thyroid disease and smoking, were not overtly involved, despite it must be noticed that one third of the population was smokers [24]. Female gender was more frequent in the whole cohort but was similar between the UsGoiter and UsNon-Goiter groups (63.6% vs. 60.6%). Nevertheless, childbearing was more frequent in the UsGoiter group, and the thyroid volume was associated with parity, as usually reported [25]. Patients with PTC were, however, more often male but the difference of PTCs frequency between male and female was not statistically significant in accordance with a recent meta-analysis [12]. However, a more severe phenotype in males has already been recognized in DM1 patients [3], and if the difference of number of CTG repeats between sex was not significant, it tended to be slightly higher in men as compared to females.

The most significant parameters associated with thyroid volume were BMI and glucose metabolism. This relationship has been emphasized in the literature, both in obese people [17, 26] and in normal-weight, hyperinsulinemic, lipodystrophic patients [27]. Indeed, half of our DM patients were overweight, 39% were glucose intolerant or diabetic, and 25% were receiving statins; this confirms the frequency of insulin resistance confirmed by higher 120-min OGTT glucose and insulin levels [28]. This high prevalence of metabolic abnormalities has been demonstrated in other DM cohorts [29], with a range between 14.6 and 21.1% in 1856 patients, with no difference between genders. Insulin is a growth factor and could therefore favor thyroid growth, as demonstrated in a recent age-matched control study in DM1 [30].

TSH levels were lower in the UsGoiter group compared with the UsNon-Goiter group, in contradiction with what would have been expected due to the goitrogenic effect of TSH. Unexpectedly, thyroid volume was inversely correlated with TSH and FT4 levels. This unforeseen result could be explained by the relative autonomy of some goiters or pituitary dysfunction related to DM1, but also suggests other promoting factors. Iodine deficiency has nearly vanished since the systematic supplementation of salt [31] in France and so was not specifically studied in our routinely screened cohort. Previous studies [32, 33] reported an association between vitamin D levels and (CTG)_n expansion size or anti-TPO antibodies that we did not observed in our cohort. However, we observed an inverse correlation between thyroid volume and vitamin D level. Such a correlation has never been previously described. Nevertheless, vitamin D is involved in cell differentiation and apoptosis, and a lack of vitamin D could favor goiter and/or PTC [34]. Compared to age-matched controls, DM patients have lower vitamin D levels [30], a fact we confirm since 76.5% of our population had vitamin D deficiency. This low level of vitamin D could be explained by the physical disability potentially resulting in less outdoor exposure and a higher BMI. Indeed, in our study, BMI was positively correlated to the number of CTG repeats and was inversely correlated to vitamin D deficiency, which is in accordance with some studies [35]. This supports an indirect mechanism linking vitamin D levels and thyroid volume. Even if vitamin D supplementation is recommended, no interventional studies have demonstrated that it could reduce the prevalence of goiter in this population, whatever the direct or indirect mechanisms.

Surprisingly, no correlation between thyroid abnormalities and CTG repeats was observed in our patients, despite the association between the number of CTG repeats and disease severity [1]. This is probably related to the fact that the half of our cohort corresponded to an adult form with a number of CTG repeats between 280 and 1000. Repeats size is only modestly associated with disease severity in adult-onset DM1, whose repeats size span a wide spectrum, at difference with forms associated to less than 100 or more than 1000 CTG repeats. However the median number of CTG_(n) repeat tended to be higher in patients who developed PTCs compared to others patients which is another argument for a higher risk related to the disease and not only a screening effect. The mechanisms leading to a higher risk of cancer in DM1 patients have not been described yet. A direct role in carcinogenesis of the abnormal RNA processing or/and the metabolic syndrome observed in these patients may contribute to increased cancer risk, as well as insulin resistance. The presence of genetic mosaicism and variable expression of CTG in thyroid tissue [36] may explain that some patients with high number of CTG repeats detected in leukocytes do not developed thyroid abnormalities.

Finally, the answer to the question “Should we screen DM1 patients by US thyroid?” remains unclear. The fact that these PTC were frequent (8% of patients), occurred at any age, were multifocal in 4/9 cases, with one third being pT3 or of intermediate risk argue for a systematic US screening. However, we did not find a higher prevalence of US goiter and nodules in this DM1 cohort compared to US-screened general population. Nevertheless, the limited size of our cohort, the fact that this cohort corresponds to a limited sample of the DM1 population followed in a reference center (probably only one third), the effect of genetic screening in pre-symptomatic patients in family inquiries, the lack of matched control group introduce selection biases. Moreover the good prognosis of most thyroid cancers identified in our series does not argue for a systematic screening. Therefore, a targeted screening in case of palpable nodule, especially in male overweight or diabetic people, seems more appropriate. If surgery is required, it should be performed by an experienced anesthesiologist and surgical team, after cardiorespiratory work-up, with a reinforced post-operative monitoring.

In this cohort of DM1 patients, we observed a high prevalence of PTCs that may not be only due to the effect of a systematic screening, in accordance with the high risk of cancer described recently in DM patients. In addition, we show that these PTC were mainly micro-PTC and that overweight and glucose intolerance were the main factors associated with higher thyroid volume. Considering the good prognosis of PTCs in general population, targeted US screening in case of clinical abnormality and/or in overweight (BMI ≥ 25) or “metabolic” patients seem currently more appropriate than systematic US screening. Further studies are needed to understand the mechanism leading to cancer in DM1 patients and to study the evolution of these cancers compared to the general population. The link with insulin resistance could open new perspectives even in general population.