Article Text
PDF
Abstract
Objectives To describe clinical and laboratory characteristics and outcomes in patients with antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) and thyroid disease (TD). We also aimed to calculate incidence and identify predictors of TD in two large cohorts of patients with AAV.
Methods The study comprised 644 patients with AAV in a population-based cohort from southern Sweden (n=325) and a cohort from a specialised vasculitis centre in Cambridge, UK (n=319). Diagnosis and classification of AAV and TD were confirmed by medical record review. Person-years (PY) of follow-up were calculated from AAV diagnosis to the earliest of TD, death or the end of study. Cox-regression analysis was employed to study predictors of TD.
Results At AAV diagnosis, 100 individuals (15.5%, 77 females) had TD, 59 had myeloperoxidase (MPO)-ANCA+ and 34 had proteinase-3 (PR3)-ANCA+. Patients with TD tended to have lower C reactive protein, lower haemoglobin and fewer constitutional symptoms. Survival and renal survival was greater in those patients with AAV with pre-existing TD. During 4522 PY of follow-up, a further 29 subjects developed TD, yielding an incidence rate of 641/100 000 PY. No analysed factor predicted de novo TD in AAV. The prevalence of TD among patients with AAV in southern Sweden was 18%.
Conclusion TD is a common comorbidity in AAV, affecting nearly one in five. While TD diagnosis is more common in females and MPO-ANCA+, these factors do not predict de novo TD after initiation of AAV treatment, necessitating monitoring of all patients with AAV with respect to this comorbidity.
- vasculitis
- epidemiology
- incidence
- prevalence
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information. Raw data are protected by confidentiality laws in Sweden and cannot be shared. All data relevant to the study are included in the article. Please contact the corresponding author.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Patients with antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) suffer higher rates of morbidities, including thyroid disease (TD), compared with the general population.
WHAT THIS STUDY ADDS
Incidence rate of TD is higher post AAV diagnosis than in the background population and is similar in males and females.
The majority of cases occur in the 5 years following AAV diagnosis.
Eighteen percent of individuals with AAV were found to live with TD.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Prospective studies are needed to understand the temporal relationship between AAV and TD. This will increase our understanding of the co-occurrence of these diseases and shed light on a common or overlapping causative pathway and suggest appropriate therapies.
Introduction
Antineutrophil cytoplasmic antibody (ANCA)-associated vasculitides (AAV), including granulomatosis with polyangiitis (GPA), microscopic polyangiitis (MPA) and eosinophilic granulomatosis with polyangiitis (EGPA) affect small vessels.1 They are characterised by vasculitis and granulomatous manifestations in kidneys, lungs, nervous system and gastrointestinal tract. Current therapy for AAV consists of induction treatment with cyclophosphamide, rituximab and high-dose glucocorticosteroids, alone or in combination, and subsequent maintenance with azathioprine, methotrexate or rituximab.2 All treatments are associated with side effects and risk for long-term morbidity.3 4 Patients with AAV are at increased risk of developing comorbidities such as stroke,5 severe infections,4 cancer6 and venous thromboembolisms.7 These comorbidities may be secondary to active inflammation, treatment or both.3
Autoimmunity is a common cause of thyroid disease (TD), which has high prevalence worldwide and is generally successfully treated with well-tolerated therapy but can have serious adverse effects if left untreated.8–11 The vasculitides are comparatively rare, nonetheless an association with TD is reported, and there may be shared aetiologies. Aspects of a flare in AAV, such as fatigue or weight loss, could be rooted in endocrine dysfunction.12 Studies have revealed increased presence of TD in AAV, both prior to13 14 and following the AAV diagnosis.15–19 While antithyroid drugs such as propylthiouracil (PTU) or methimazole (MMI) can trigger myeloperoxidase (MPO)-AAV, several reports have directly linked TD to AAV.20–24
We aimed to investigate aspects of TD in patients diagnosed with AAV using cohorts from Sweden25 26 and the UK.27 Study objectives were to (1) describe clinical and laboratory characteristics and outcome in patients diagnosed with TD prior to and after diagnosis of AAV, (2) determine the incidence rate and prevalence of TD in AAV, (3) identify possible predictors of TD in AAV and (4) characterise patient outcomes.
Methods
Patients
Southern Sweden
A population-based cohort of patients with AAV in a defined geographical area in Skåne, the southernmost region of Sweden, has been continuously updated since 1997 to include all incident cases. This study was based on 325 cases of AAV diagnosed from 1997 through 2016. Case retrieval and ascertainment has been previously described.25 26 Patients were followed from AAV diagnosis to death or the conclusion of study on 1 January 2019. The area is deemed to be iodine sufficient, and Sweden has recommended iodine salt supplementation.10 28
Cambridge, UK
The hospital-based cohort of patients with AAV treated and followed at the Vasculitis and Lupus Clinic, Addenbrooke’s Hospital, Cambridge, UK consisted of 319 patients diagnosed with AAV from 1992 through 2019. Data for this retrospective cohort have previously been used to describe pulmonary involvement in AAV with data collected during 2014.27 In 2019, a new search was conducted among those with AAV at the hospital to identify individuals diagnosed with TD. Patients were followed from diagnosis of AAV to death or the conclusion of the study, 1 October 2019. The UK is considered to be iodine sufficient.10 29
Confirmation of AAV diagnosis
In both cohorts, diagnosis of small vessel vasculitis was confirmed by case record review. Patients had been diagnosed with small vessel vasculitis by organ biopsy or, alternatively, by surrogate markers for granulomatous or vasculitic disease.30 We classified the disease as one of three AAV phenotypes (GPA, MPA, EGPA) using the algorithm of the European Medicines Agency.30
Case identification and ascertainment of thyroid disease
TD was defined as hypothyroidism, non-toxic goitre, thyrotoxicosis (hyperthyroidism), thyroiditis or any combination. Medical records indicating TD needed to be supported by laboratory results, prescribed medication, histological and/or radiological findings. To differentiate from non-thyroidal illness and to increase validity, TD status and medication were reviewed 12 months after the original diagnosis. Medical records before and after AAV diagnosis were reviewed to ascertain the temporal relationship between AAV and TD. Cases of TD that were recognised during the investigation confirming vasculitis were included as incident cases after AAV diagnosis, as they were judged to have developed in conjunction or after the diagnosis.
Southern Sweden
The patients in the AAV cohort with a history of thyroid diagnosis (International Classifications of Disease, Version 10 (ICD-10) codes E00–E07 (disorders of the thyroid gland)) were identified using the Skåne Healthcare Register, a comprehensive administrative register recording information of healthcare visits and diagnostic codes in Skåne since 1998.31 32 The diagnosis of TD was subsequently verified via individual case record review. Digital case records (2002 and later) of patients not assigned a TD code were screened for TD using a text string search for keywords ‘thyroid’, ‘goitre’ and ‘thyroxine’ as well as by review of all digital thyroid laboratory records. If any of these were indicative of TD, the diagnosis was confirmed by individual case record review.
Cambridge, UK
Search parameters applied to patients attending the specialised Vasculitis and Lupus Clinic at Addenbrooke’s Hospital, Cambridge, UK included assignment of an ICD-10 code indicating any TD (E00–E07), thyroid replacement medication or a text string in patient case records referring to any TD. Patients treated with anti-CD52 antibody (alemtuzumab) (n=5), which is associated with an increased risk of thyroid disorder,33 were excluded from the analysis.
Data collection
Demographic and clinical data of all patients with confirmed AAV from the time of diagnosis were collected retrospectively. Laboratory data consisted of complete blood profile, inflammation parameters, serum creatinine level and proteinase-3 (PR3) or MPO-ANCA. The modification of diet in renal disease study equation for estimating glomerular filtration rate (Modification of Diet in Renal Disease (MDRD 2006) was used at AAV diagnosis.34
Information on organ involvement and AAV disease activity that had been determined by the Birmingham Vasculitis Activity Score at 0 and 12 months was collected along with information regarding if and when a patient developed end-stage kidney disease (ESKD).
In patients with TD, additional clinical and laboratory data were collected when available to further confirm the TD diagnosis.
Statistical analysis
Statistical analysis was performed using IBM SPSS Statistics for Mac V.26.0–28.0 and R V.4.2.1 (R Foundation for Statistical Computing, Vienna, Austria). The differences in the frequency of categorical variables between groups were quantified using the χ2 test or Fisher’s exact test. For continuous and normally distributed variables, data are presented as means and SD, and Student’s t-test was used for comparison of groups. Data of continuous and non-normally distributed variables are presented as median and IQR, and the Mann-Whitney U test was used for comparison. A p value of <0.05 was considered significant in all analyses. Kaplan-Meier analysis was employed to study differences in survival and ESKD-free survival of patients with and without TD. The variables included in the univariable and multivariable Cox-regression model to explore predictors of TD were age at AAV diagnosis, sex, PR3-ANCA and serum creatinine at diagnosis. To examine the influence of pre-existing TD at the time of diagnosis of AAV on both patient and kidney survival, we conducted a Cox-regression analysis. We adjusted the outcome for age at AAV diagnosis and MPO-ANCA positivity. Furthermore, we investigated the interaction between these two covariates to illustrate any significant interaction effect. We also investigated the relationship between incident TD and mortality in patients with AAV, using a time-dependent Cox-regression model. Exposure was defined as diagnosis of TD post-AAV diagnosis (time-dependent variable) with death the outcome of interest.
Incidence rates of TD were calculated based on the number of patients with AAV with thyroid diagnosis as numerator and the sum of person-years (PY) of follow-up as the denominator. The PY of follow-up were calculated from date of AAV diagnosis to the date of diagnosis of TD, death or the conclusion of the study (1 January 2019, southern Sweden cohort; 1 October 2019, Cambridge cohort). Years of survival were calculated from date of AAV diagnosis to death or conclusion of study.
Results
A total of 644 individuals were included (Lund=325, Cambridge=319) of which 328 (51%) were female. Subjects were classified as having GPA (n=319, 49.5%), MPA (n=300, 46.5%) or EGPA (n=25, 4%) (table 1). Serology results were available for 584 patients (300 PR3-ANCA+ (47%), 284 MPO-ANCA+ (44%)).
Characteristics of 644 patients with AAV with and without pre-existing TD
Thyroid disease prior to diagnosis of AAV
One hundred subjects (15.5%) had received a diagnosis of TD before AAV diagnosis. Female sex and MPO-ANCA-associated disease were more common than in those without pre-existing TD (table 1). Patients with prior TD had less pronounced systemic inflammation, including lower C reactive protein (p=0.01), fewer general symptoms (p=0.007) and lower haemoglobin levels (p=0.03) compared with those without prior TD (table 1).
Date of thyroid diagnosis were available for 22 of these 100 patients. The median time from TD diagnosis to diagnosis of AAV was 5.5 years (IQR 2.9–18.1). Only 2 of the 100 patients were diagnosed with TD in the year preceding AAV diagnosis (10 and 11 months). The initial thyroid diagnoses were hypothyroidism (n=71) followed by hyperthyroidism (n=16), goitre (n=7) and thyroiditis (n=6). In several patients, more than one TD was recorded during the follow-up period, that is, in patients developing hypothyroidism after treatment for thyrotoxicosis.
Five patients received antithyroid drugs before being diagnosed with AAV. One patient had taken PTU a decade prior to AAV diagnosis. Another received PTU for Graves thyrotoxicosis 3 months before being diagnosed with AAV and a third was treated with MMI and PTU 12–18 months prior to AAV diagnosis. Two additional patients had taken MMI 4 and 12 years before AAV diagnosis.
Thyroid disease after diagnosis of AAV
Five-hundred forty-four patients with AAV had no identified TD at time of AAV diagnosis. Among these, 29 new cases (5.3%) of TD were diagnosed during follow-up: hypothyroidism in 17 (59%), hyperthyroidism in 8 (28%), goitre in 3 (10%) and thyroiditis in 1 (3%). There were no differences in demographics, clinical, laboratory characteristics including ANCA serotype or outcome of those who were diagnosed with TD post-AAV diagnosis versus those without TD (table 2). The median follow-up time from diagnosis of AAV to that of TD was 3.2 years (IQR 0.9–7.2). The onset of TD was within 6 months of AAV diagnosis in six patients (four hypothyroidism, two thyrotoxicosis). Twenty of the 29 (69%) diagnoses with TD were within 5 years of AAV diagnosis.
Characteristics of patients with new-onset TD and those without TD
The incidence rate of TD in AAV
During the 4522 PY follow-up, 29 patients with AAV were diagnosed with TD (table 3). The incidence rate of TD per 100 000 PY was 641 (95% CI 424 to 1238). In southern Sweden, the incidence rate was 831 (95% CI 401 to 1170), and in Cambridge, UK, it was 501 (95% CI 229 to 773) (p=0.1). The incidence rate for hypothyroidism was 376/100 000 PY and for hyperthyroidism 177/100 000 PY (table 3). The incidence rate was similar with respect to sex: 653 (95% CI 312 to 995) in females and 632 (95% CI 311 to 951) in males. The TD incidence rate in the 5 years postdiagnosis of AAV was 925/100 000 PY. Patients with PR3-ANCA-associated disease were estimated to have a slightly higher incidence rate of TD compared with those with MPO-ANCA-associated disease (p=0.4). However, in an analysis comparing TD incidence rate in the 5 years following AAV diagnosis, no significant difference was found (data not shown).
Incidence rate of TD per 100 000 person-years in individuals with AAV
The prevalence of thyroid disease in AAV
The southern Sweden cohort comprised 164 individuals with AAV at the end of the study on 1 January 2019. Thirty (23 female, 7 male) participants had a diagnosis of TD, resulting in a prevalence of 18%. The prevalence in females was 28.4% and in males 8.4%.
In the UK cohort, as of the study’s end on 1 October 2019, 68 out of 263 surviving individuals (26%) had TD. Due to the hospital-based nature of this cohort, the background population cannot be determined, hence we were unable to estimate the point prevalence.
Predictors of thyroid disease in AAV
The Cox-regression analysis revealed an increased risk for TD in patients with PR3-ANCA, however, the increase was not significant. In the multivariable analysis, no predictors of TD were identified (table 4).
Predictors of thyroid disease in patients with AAV
Outcome
Patients with thyroid disease prior to AAV
The median duration of follow-up from AAV diagnosis until death or end of study was 5 years (IQR 2–10) for patients with prior TD vs 9 years (IQR 4–13) for those with no prior TD (p<0.001). During the follow-up period, 92 out of 644 patients (14%) developed ESKD. Of these, 8 patients (8%) had prior TD, while 84 patients (15%) did not have prior TD (p=0.06). ESKD-free survival of patients with pre-existing TD at 1, 5 and 10 years was 96%, 92% and 92% vs 90%, 87% and 82% (log rank p=0.04) in those without TD, respectively (figure 1A).
(A) End-stage kidney disease-free survival of patients with and without pre-existing thyroid disease (TD). (B) Survival of patients with antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis with and without pre-existing TD. (C) End-stage kidney disease-free survival in subjects diagnosed with TD after AAV versus without TD.
During the follow-up period, 217 patients died (34%), with 23 (23%) having prior TD and 194 (36%) having no prior TD (p=0.02). In patients with TD prior to AAV diagnosis, the 1-year, 5-year and 10-year survival rates were 94%, 83% and 75% compared with 91%, 76% and 67% in those without prior TD, respectively (log rank p=0.08) (figure 1B).
Patients with pre-existing TD demonstrated better renal survival compared with those without prior TD (HR 0.47, 95% CI 0.22 to 1.01, p=0.05). The presence of TD before the diagnosis of AAV significantly correlated with improved renal survival after adjusting for age at AAV diagnosis and MPO-positivity (HR 0.39, 95% CI 0.18 to 0.86, p=0.02). Additionally, patients with pre-existing TD exhibited a trend towards better patient survival than those without prior TD (HR 0.66, 95% CI 0.40 to 1.07, p=0.09), a trend that persisted even after adjusting for age at AAV diagnosis and MPO-positivity (HR 0.64, 95% CI 0.39 to 1.06, p=0.08).
Patients diagnosed with thyroid disease after AAV diagnosis
Five patients (17%) of those diagnosed with TD after AAV developed ESKD vs 79 (15%) patients without TD, p=0.8. ESKD-free survival at 1, 5 and 10 years, respectively, in patients diagnosed with TD after AAV was 96%, 89% and 79% vs 89%, 86% and 83% (log rank p=0.9) in those without TD (figure 1C). During follow-up period, eight patients (28%) with new-onset TD died vs 186 (36%) without TD, p=0.4.
The time-dependent Cox-regression analysis showed TD onset after diagnosis of AAV to be associated with increased risk of death with an HR of 2.37 (95% CI 0.73 to 7.63). The association was not significant.
Discussion
In this study involving two large cohorts of patients with AAV in Sweden and the UK, 20% of individuals with AAV were diagnosed with TD at some time during their lifetime. Five per cent were diagnosed with TD after AAV onset, with the majority occurring in the 5 years following AAV diagnosis and 18% of individuals living with AAV were found to have TD. The observed incidence of TD in AAV was higher than has been reported in the general population.35–39 Our study revealed a higher survival and renal survival rate in patients with TD prior to the diagnosis of AAV compared with those without pre-existing TD.
We found differences in patient demographics and laboratory characteristics with respect to the temporal relationship of TD and AAV diagnoses. Patients with TD prior to the diagnosis of AAV were more likely to be female and MPO-ANCA+, as reported in previous studies.15–18 Our findings also reflect common clinical and laboratory findings of lower inflammation and fewer general symptoms in patients with MPO-ANCA.40 Data of individuals with established AAV who are later diagnosed with TD are scarce. We found no differences in clinical, laboratory or demographic features, including ANCA type and sex, between patients with AAV who developed TD during follow-up and those who did not. There were also no differences in outcome of patients with AAV who developed TD after the diagnosis of AAV and those with no TD.
The differences in clinical and laboratory characteristics of patients with TD onset before versus after diagnosis of AAV may be related to the inflammatory activity of vasculitis and to the treatment used. TD was most often diagnosed within 5 years of AAV diagnosis. This could stem from disease activity or from stress of being diagnosed with AAV.41 Survey bias19 must also be considered, as persons with chronic inflammatory disease such as AAV are more likely to be screened by treating physicians when presenting with non-specific symptoms.9 It should also be remembered that thyroid-stimulating hormone (TSH), triiodothyronine and thyroxine can be disrupted in times of serious illness, and the condition may be identified as non-thyroidal.42 We reviewed cases 12 months after AAV diagnosis to reduce the risk of including such cases. TD is one of the most common morbidities in the general population. It varies depending on iodine availability and geographic area and is far more common in females.10 35 37–39 43 These factors may also contribute to regional differences in patient characteristics and TD incidence.
Previous studies have demonstrated a high rate of TD in patients with other systemic disease such as rheumatoid arthritis44 and systemic lupus erythematosus.45 The mechanisms underlying the association between TD and inflammatory rheumatic disease are not well known. Plausible explanations have been explored, for example, genetic susceptibility and polymorphisms in genes such as CTLA4 and PTNP22, which affect regulation of T-cell activation, play a role in several autoimmune diseases including autoimmune TD.46 A higher percentage of antithyroid antibodies are seen in patients with systemic disease, and the amino acid pattern in MPO and thyroid peroxidase is similar, although it does not seem to have clinical relevance.47 In AAV, there is an interplay of thyroxine, ANCA and macrophage migration inhibitory factor that is not fully understood.48 Only 3% of our subjects with AAV and TD had a history of exposure to PTU prior to AAV, precluding a role of drug-induced vasculitis in the cohort. This is comparable to what has been reported in previous studies.15–17
There are previous reports of elevated TSH and lower T3 concentrations in chronic kidney disease,49 the focus of this study has been on the clinical diagnosis of TD. Paradoxically, the individuals with prior TD have less risk of ESKD in our study.
In agreement with Kermani et al,18 we found no significant differences in renal involvement with respect to presence of TD, although differences have been reported.15–17 We observed significantly better renal survival and a tendency towards greater overall survival in patients with TD at AAV diagnosis. Even after adjustment for covariates such as MPO-ANCA and age at diagnosis, TD remained associated with better outcome. Our data provide no explanation for these findings. The only difference observed between those with and without TD at the time of AAV diagnosis was lower systemic inflammation in the TD group. Our study did not identify predictors of TD onset after AAV diagnosis. However, the small number of events and potential missing data may have impacted the results.
We observed the highest TD incidence rate in the 5 years following AAV diagnosis. Similar patterns of comorbidity onset early in the course of AAV therapy have been demonstrated in venous thromboembolic diseases,7 stroke5 and severe infections.4 19 Although direct comparison is not possible, the TD incidence rate among our patients with AAV was twice that reported in the general population in the two countries of this study and throughout Europe.43 Incidence of hypothyroidism in the UK is 243–297/100 000 PY (male 60–101, female 350–498/100 000),35 36 39 and incidence of hyperthyroidism in southern Sweden is 25.8–27.6/100 000 PY (male 10.1–11, female 40.6–44.1/100 000).37 38 In our previous comorbidity study, we calculated an incidence rate ratio of thyroid diagnosis of 2.1 in individuals from southern Sweden with AAV to that of the reference population.50 The 18% prevalence of TD in patients with AAV (female, 28.4%; male, 8.4%) is higher than reported in the population of the corresponding geographic area in Sweden (all 11.5%; female 20.3%; male 2.4%).28 Earlier retrospective studies have shown comparable prevalence of TD in vasculitis ranging from 10% to 21%.13 15–17
Our study limitations include incomplete retrospective data on AAV disease severity and treatment, lack of ethnicity data and limited systemic histopathological classification due to ethical constraints. Furthermore, limited data on thyroid function, auto-antibodies and overlapping TD could affect inclusion of specific TD. The differences between the two cohorts, one population-based and one hospital-based, must be considered. The low number of events in incidence estimates impact conclusions with respect to differences in incidence and to predictors of TD in AAV.
The study also has several strengths. The data originated from large well-characterised cohorts of patients with AAV, and diagnoses and classifications were confirmed by case record review. For TD case identification, we used comprehensive search methods in both cohorts to identify a substantial majority, if not all, cases.
Individuals with AAV suffer a high rate of TD, especially during the 5 years following AAV diagnosis. This important comorbidity should be taken into consideration in patients with AAV who suffer non-specific manifestations such as fatigue that could be attributed to TD as well as to AAV. Longitudinal prospective studies are needed to understand the temporal relationship between AAV and TD and may increase our understanding of the co-occurrence of these diseases and shed light on a potential common causative pathway.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information. Raw data are protected by confidentiality laws in Sweden and cannot be shared. All data relevant to the study are included in the article. Please contact the corresponding author.
Ethics statements
Patient consent for publication
Ethics approval
For the Swedish cohort, the research was approved by the Swedish Ethics Review Authority (2010-517). For the UK cohort, in accordance with the UK National Health Service (NHS) Research Ethics Committee guidelines, approval was not required, because the study consisted of retrospective data and all treatment decisions were made prior to our analysis. The study was conducted in compliance with criteria of the Declaration of Helsinki.
References
Footnotes
Contributors All authors were involved in drafting the article or revising it for intellectual content, and all authors approved the final version to be published. Study conception and design: AW, RS, MS and AJM. Data acquisition: AW, RS and AJM. Analysis and interpretation of data: AW, RS, MS, DJ and AJM. Manuscript guarantor: AJM.
Funding This study was supported by grants from the Swedish Research Council (Vetenskapsrådet: 2019-01655), Faculty of Medicine, Lund University (ALF-medel), The Swedish Rheumatism Association, the Swedish Medical Society, Alfred Österlund’s Foundation, King Gustaf V's 80-year foundation Anna-Greta Crafoords foundation for rheumatology research (to AJM) Lund University (ST-ALF) (AW).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.