We have found that diagnoses of Sjögren's syndrome and autoimmune thyroid disease are strongly associated in families with SLE. This association was found among patients with SLE with secondary Sjögren's syndrome who had a relative risk for thyroid disease of 2.3. The association was also found among SLE‐unaffected relatives of patients with SLE, where the relative risk of autoimmune thyroid disease among the 44 patients diagnosed with primary Sjögren's syndrome was 4. We found no association between the two illnesses in the unrelated control population, although the number of subjects with Sjögren's syndrome was small. It seems likely that these patients with SLE and their SLE‐unaffected relatives are especially at risk for the combination of Sjögren's syndrome and autoimmune thyroid disease, probably related in part to their predisposition to SLE.
A number of studies have found no association between primary Sjögren's syndrome and autoimmune thyroid disease8,3
. The largest of these studied 160 patients with Sjögren's syndrome and 75 healthy controls in a cross‐sectional design. A total of 36% of the patients with Sjögren's syndrome and 27% of the controls had autoimmune thyroid disease, and this difference was not significant.8
Another large study found a significant difference in that 41 (30%) of 137 patients with Sjögren's syndrome had autoimmune thyroid disease whereas only 5 (4%) of 120 controls had the same.7
These studies differed in several ways—for example, the second study included only women and the controls were not healthy normals but patients with either sciatica or osteoarthritis. In the first study, the healthycontrols were free of symptoms possibly attributable to Sjögren's syndrome,8
whereas in the second study, the controls were known to not have antinuclear antibodies, rheumatoid factor, anti‐Ro or anti‐La.7
None of these differences definitely explains the discordant conclusions. However, perhaps the most important difference is the rate of thyroid disease in the controls. In the study of D'Arbonneau et al7
that finds a difference between subjects with Sjögren's syndrome and controls, a mere 4% of the controls had autoimmune thyroid disease. These controls were women who were age matched to the patients, whose average age was 53.6 years. Thus, this is a strikingly low prevalence of autoimmune thyroid disease in a cohort of older adult women.16
We wonder whether this report found a low prevalence of thyroid disease in the controls instead of a high prevalence of Sjögren's syndrome.
Many persons may give a history of autoimmune thyroid disease without documented past disease. Thus, the strategy we used to classify subjects as having autoimmune thyroid disease might result in false positives. Subjects identifying themselves as having autoimmune thyroid disease had anti‐TPO in their sera at levels higher than those who did not thus identify themselves. There were, of course, several individuals in the autoimmune thyroid disease groups who were negative for the antibodies, but this is the expected outcome as these antibodies may disappear from the serum after many years of disease.17
Meanwhile, others in the group without clinical thyroid disease were antibody positive, which was again an expected finding.
We have relied on a self‐reported diagnosis of Sjögren's syndrome, or this condition given as a diagnosis in the medical record. In fact, the great majority of those so diagnosed or reporting this diagnosis could not be shown to meet the combined US–European Sjögren's syndrome classification criteria.14
In practical terms, this means that neither objective testing of tear or saliva production nor salivary gland biopsy was commonly found in the medical record. Patients meeting three criteria (dry eyes, dry mouth and anti‐Ro) of four needed for classification had the same rate of thyroid disease as other patients in whom Sjögren's syndrome was diagnosed. Thus, the lack of definitive diagnosis based on the commonly agreed research classification criteria is a weakness of our study. On the other hand, combinations of clinical symptoms such as dry mouth, dry eyes and sore mouth correctly classify >90% of patients with Sjögren's syndrome and normal controls.18
Thus, it is highly likely that most of our cohort had the disease.
At least 10 cohorts of patients with SLE have been examined for autoimmune thyroid disease,3
with prevalence ranging from 2.4% to 11.4%. Most of these studies did not include sex‐ and age‐matched controls. We found that the point prevalence of thyroid disease was 7.8% in these patients with SLE, which is higher than in most studies. Our patients were all from families with at least two patients with SLE. Such patients may be at higher risk of other autoimmune diseases than isolated patients with SLE. Furthermore, different ascertainment of thyroid disease from study to study is common.
Autoimmune thyroid disease is common in the general population, especially among women. In fact it is so common that the American College of Physicians recommends a case‐finding approach. By this strategy, all women >50 years of age who have not had a thyroid‐stimulating hormone measured within 5 years should undergo this test when seeing their healthcare provider, regardless of the reason for the visit.19
This recommendation was made partly because of the insidious nature of hypothyroidism. This diagnosis may be particularly difficult in patients with SLE in whom the symptoms of hypothyroidism might be attributed to SLE or its treatment. Thus, given the level of thyroid autoimmunity we have found in patients with SLE with secondary Sjögren's syndrome, it seems prudent to screen these patients periodically with a serum thyroid‐stimulating hormone. In addition, patients with primary Sjögren's syndrome who have a relative with SLE should also be screened.