Prevalence of hyperthyroidism and hypothyroidism in liver transplant recipients and associated risk factors

The prevalence of hyperthyroidism and hypothyroidism and associated risk factors are unknown in liver transplant recipients. We aimed to determine the prevalence of hyperthyroidism and hypothyroidism and associated risk factors in liver transplant recipients and to compare it with controls from the general population. As part of the Danish Comorbidity in Liver Transplant Recipients (DACOLT) Study, all Danish liver transplant recipients over the age of 20 were invited for measurements of concentrations of thyrotropin and thyroid hormones. The prevalence of hyperthyroidism and hypothyroidism was compared to age- and sex-matched controls from the Copenhagen General Population Study. Using logistic regression adjusted for age, sex, smoking, and body-mass index, we investigated potential risk factors. We recruited 489 liver transplant recipients and 1808 controls. Among liver transplant recipients, 14 (2.9%) had hyperthyroidism compared with 21 (1.2%) of controls (adjusted odds ratio [aOR] 2.24, 95% confidence interval [CI] 1.05–4.75, P = 0.04), while 42 (5.7%) had hypothyroidism compared with 139 (7.7%) of controls (aOR 0.68, 95% CI 0.43–1.08, P = 0.10). Female sex, and autoimmune hepatitis and primary sclerosing cholangitis as causes of transplantation were associated with hyperthyroidism after adjustments. Age, female sex, and autoimmune liver diseases as cause of transplantation were associated with hypothyroidism after adjustments. DACOLT is registered in ClinicalTrials.gov (NCT04777032).

liver transplantation with conflicting findings 8,9 .Thus, it is unknown if hyperthyroidism or hypothyroidism are comorbidities of particular concern in liver transplant recipients.
The aim of this study was to determine the prevalence of hyperthyroidism and hypothyroidism in liver transplant recipients and compare it to age-and sex-matched controls.In addition, we aimed to determine risk factors associated with hyperthyroidism and hypothyroidism in liver transplant recipients.We hypothesise that there is a higher prevalence of hyperthyroidism and hypothyroidism among liver transplant recipients than in controls and that transplantation-related variables, including causes of transplantation and immunosuppressants, are associated with hyperthyroidism and hypothyroidism.

Overview
The Danish Comorbidity in Liver Transplant Recipients (DACOLT) Study (ClinicalTrials.gov:NCT04777032) is a nation-wide, prospective, longitudinal cohort study that aims to determine the burden of comorbidities in liver transplant recipients.All Danish liver transplant recipients aged 20 or more have been invited, and > 85% have chosen to participate.The protocol has been published elsewhere 10 .The Copenhagen General Population Study (CGPS) is a prospective, longitudinal cohort study that recruits adults over the age of 20 from the general population in the Copenhagen area in Denmark 11 .For this cross-sectional study, we recruited liver transplant recipients from DACOLT and general population controls from CGPS who had available measurements of blood concentrations of thyrotropin (TSH), total thyroxine (tT4), total triiodothyronine (tT3) and free thyroxine (fT4).Controls were frequency matched on sex and age with a maximal time interval between examinations of five years.

Ethics declaration
The study was conducted in accordance with the Declaration of Helsinki.The DACOLT and CGPS studies have been approved by the Ethics Committee of the Capital Region of Denmark (approvals H-20052199 and H-KF-01-144/01).Oral and written informed consent was obtained from all subjects.All liver transplantations in Denmark are performed at the Copenhagen University Hospital -Rigshospitalet with prior consent from donors and recipients, and organs are not procured from prisoners.

Data acquisition
At inclusion, all participants in DACOLT and CGPS completed a questionnaire on smoking, alcohol intake, medical history and use of thyroid medications and antidiabetics, and they underwent a physical examination, including measurement of weight and height, performed by trained personnel.
Venous blood samples were collected between 2021 and 2023 from liver transplant recipients and between 2016 and 2017 from controls, and concentrations of TSH, tT4, tT3, and fT4 were measured at the laboratory of the hospital at which the sample was collected.
For liver transplant recipients, patient records were reviewed for the cause of liver transplantation, time since first liver transplantation, re-transplantation, and history of type 1 diabetes mellitus (T1DM).History of inflammatory bowel disease, coeliac disease, rheumatoid arthritis, psoriasis, multiple sclerosis, and/or any connective tissue disease was retrieved from the questionnaire.Data on use of immunosuppressants in liver transplant recipients at the time of inclusion were collected from the Shared Medication Record (FMK), which includes all prescribed out-patient medications in Denmark.Results of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) tests were retrieved from the Danish Microbiology Database that contains complete data from all departments of clinical microbiology in Denmark since 2010, including results of all SARS-CoV-2 polymerase chain reaction (PCR) and antigen tests from public test centres 12 .

Definitions
Hyperthyroidism was defined as TSH concentration below reference range or self-reported use of medications for hyperthyroidism irrespective of TSH concentration.Hypothyroidism was defined as TSH concentration above reference range or self-reported use of medications for hypothyroidism irrespective of TSH concentration.Euthyroidism was defined as TSH concentration within reference range and no self-reported use of medications for either hyperthyroidism or hypothyroidism.
For participants who were not treated with thyroid medications, overt hyperthyroidism was defined as TSH concentration below reference range and fT4 and/or tT3 concentrations above reference ranges, while subclinical hyperthyroidism was defined as TSH concentration below reference range and fT4 and tT3 concentrations within reference ranges.Overt hypothyroidism was defined as TSH concentration above reference range and fT4 below reference range, while subclinical hypothyroidism was defined as TSH concentration above reference range and fT4 concentration within reference range.Biochemical euthyroidism was defined as TSH concentration within reference range.
Assays and reference ranges of the analysing laboratories are provided in the Supplementary Table S1.Ethnicity was based on the origin of the participants' grandparents as reported by the participant.T1DM was defined as use of insulin for a recorded T1DM diagnosis in the medical record.BMI categories were defined in accordance with WHO definitions 13 : Underweight (< 18.5 kg/m 2 ), normal weight (18.5-24.9kg/m 2 ), overweight (25.0-29.9kg/m 2 ), and obese (≥ 30.0 kg/m 2 ).Smoking status was defined as either current, former, or never smoker, while cumulative smoking was defined as pack-years (years smoking 20 cigarettes per day) in former or current smokers.Autoimmune liver disease was defined as autoimmune hepatitis (AIH), primary biliary cholangitis (PBC), primary sclerosing cholangitis (PSC) and/or other autoimmune liver disease as causes of liver transplantation.Non-thyroidal autoimmune disease was defined as either autoimmune liver disease, T1DM,

Statistical analyses
For univariable and multivariable analyses, we used univariable and multivariable logistic regressions to calculate the odds ratios (OR) and adjusted odds ratios (aOR) of each variable of interest, respectively.Participants with missing data were omitted in the regression models.Model assumptions were assessed graphically.The outcome was hyperthyroidism or hypothyroidism.The additional classification in overt and subclinical hyperthyroidism and hypothyroidism was used for descriptive statistics but not in the regression models.
An adjusted base model was defined a priori and included age, sex, smoking status, and BMI.Age, sex, smoking status, cumulative smoking, BMI, ethnicity, previous re-transplantation, cause of transplantation, time since first liver transplantation, use of calcineurin inhibitor at inclusion, use of antimetabolite at inclusion, use of azathioprine at inclusion, use of prednisolone at inclusion, and use of everolimus at inclusion were tested one at a time in liver transplant recipients before and after adjustment for the base model.Use of calcineurin inhibitor at inclusion was treated as a categorical variable: tacrolimus, ciclosporin, or none of these at inclusion.Use of antimetabolite at inclusion was also treated as a categorical variable: mycophenolic acid/mycophenolate mofetil (MPA/MMF), azathioprine, or none of these at inclusion.Afterwards, adjusted for the base model, liver transplantation status and interactions between liver transplantation status and age, sex, smoking status, cumulative smoking, BMI, and ethnicity were tested one at a time.
A priori, we planned three sensitivity analyses.First, excluding liver transplant recipients transplanted < 6 months prior to inclusion.Second, excluding liver transplant recipients with non-thyroidal autoimmune disease.Third, we tested if analysing laboratory or out-patient centre were associated with hyperthyroidism or hypothyroidism in liver transplant recipients.Post hoc, we performed two sensitivity analyses.First, we excluded all participants with history of COVID-19.Second, we additionally adjusted the base model for ethnicity.
Two-sided p-values of 0.05 or less were deemed statistically significant.For all analyses, we used R (version 1.2.5001) with the sandwich package (version 3.0-1) 14 .

Results
We included 489 liver transplant recipients and 1808 age-and sex-matched controls.Among liver transplant recipients and controls, 54% were male, and the median age was 55.3 (interquartile range [IQR] 45.8-63.8) in liver transplant recipients and 56.5 (IQR 47.7-64.8) in controls.While smoking habits were similar, fewer liver transplant recipients than controls had current intake of alcohol, normal weight, and Scandinavian ethnicity (Table 1).The most common cause of liver transplantation was autoimmune liver disease (43%), followed by alcoholic or cryptogenic cirrhosis (18%) (Table 2).
Before adjustments, use of prednisolone was not associated with hyperthyroidism in liver transplant recipients (OR 2.90 [0.90-9.37],P = 0.08) but became significant after adjustment for age, sex, smoking, and BMI (aOR 3.50 [1.03-11.9],P = 0.04), and after a further adjustment for time since first liver transplantation (aOR 3.45 [1.01-11.8],P = 0.048).Use of calcineurin inhibitor and use of antimetabolite were not associated with hyperthyroidism or hypothyroidism before or after adjustments.

Sensitivity analyses
Excluding liver transplant recipients who have been transplanted less than 6 months prior to inclusion did not change any results.
In a sensitivity analysis excluding liver transplant recipients with non-thyroidal autoimmune disease, the parameter estimate of liver transplantation status did not change for hyperthyroidism but became non-significant (OR 2.49 [0.88-7.07],P = 0.09).For hypothyroidism, the parameter estimate of liver transplantation status became borderline significant (OR 0.54 [0.28-1.03],P = 0.06).Furthermore, results did not differ from the primary analyses for sex, smoking status, cumulative smoking, BMI, and ethnicity.However, the association between hyperthyroidism and age did not reach significance (aOR 2.10 [0.83-5.26],P = 0.12).
In an analysis adjusted for age, sex, smoking, and BMI, analysing laboratory and out-patient centre were not associated with hyperthyroidism or hypothyroidism.

Discussion
In this nation-wide cross-sectional study of all Danish liver transplant recipients, we found that liver transplant recipients have twice the odds of having hyperthyroidism compared to age-and sex matched controls from the general population independently of age, sex, smoking, and BMI, but similar odds of hypothyroidism.All untreated participants with hyperthyroidism had subclinical disease.This finding is clinically important, as in  the general population, even subclinical hyperthyroidism requires monitoring and is associated with higher risk of cardiovascular morbidity, fractures, and overall mortality in older adults 15 .
All but one of liver transplant recipients with untreated hypothyroidism had subclinical disease.With current guidelines, treatment is not necessarily indicated for subclinical hypothyroidism 16 , but hypothyroidism, including subclinical hypothyroidism, before liver transplantation for hepatocellular carcinoma is associated with poorer survival 17,18 .Among liver transplant recipients, we found a combined prevalence of hyperthyroidism and hypothyroidism of 8.7%, which differs from two Brazilian studies of highly selected participants transplanted mainly for cirrhosis after hepatitis C virus (HCV) infection or alcohol use.The first study included 30 liver transplant recipients and found no cases of thyroid dysfunction 6 months after transplantation 9 , while the other study included 46 liver transplant recipients with type 2 diabetes mellitus and found a prevalence of 13% 8 .Significantly smaller sample size, selection based on specific comorbidities, and different causes of transplantation in the previous studies may explain the differences.
In this study, we saw that increasing age and female sex were important risk factors associated with both hyperthyroidism and hypothyroidism, while other European ethnicity than Scandinavian was associated with hypothyroidism.One reference range may not sufficiently encompass physiological differences among euthyroid Table 4. Associations of potential demographic variables and lifestyle with hyperthyroidism in liver transplant recipients.aOR, odds ratio adjusted for age, sex, smoking, and BMI; BMI, body-mass index: underweight (< 18.5), normal (18.5-24.9),overweight (25.0-25.9),obese (≥ 30.0);CI, confidence interval; NA, not available due to low number of cases; OR, odds ratio.www.nature.com/scientificreports/persons of different ethnicities 19 , but the finding is in line with the Rotterdam Study of the general population that found that age, sex, and genetic predisposition were important risk factors associated with thyroid function 20 .In the Rotterdam Study, TSH concentrations decreased in smokers and increased with higher BMI.Although we could not compare TSH concentrations directly across laboratories due to different reference values at different study sites, we did not find an association between hyperthyroidism or hypothyroidism and either smoking or BMI, indicating that these factors are less prominent in hyperthyroidism or hypothyroidism in liver transplant recipients than age, sex, and ethnicity.For lifestyle variables, missing data was < 2.5% and is unlikely to affect the results.Morphologic alterations of the thyroid and non-thyroidal illness syndrome are common in severe liver disease 21,22 .Thyroid hormone levels may normalise after liver transplantation 23,24 , but the long-term thyroid function in liver transplant recipients is poorly described.We found evidence that being liver transplanted modifies the effect of age on hypothyroidism, and odds of having hypothyroidism are higher with increasing age among liver transplant recipients than in the general population.This has not previously been reported in liver transplant recipients, but the phenomenon of a chronic condition modifying the effect of age on age-related comorbidities has been reported among, e.g., people with HIV 25 .Overall, the interaction suggests that clinical suspicion of or screening for hyperthyroidism and hypothyroidism in ageing liver transplant recipients should be considered earlier than in the general population, as the long-term effect of subclinical disease is uncertain in liver transplant recipients.
We found that AIH and PSC as causes of transplantation were associated with hyperthyroidism, while PBC seemed to drive the association between hypothyroidism and autoimmune liver diseases, overall, as causes of transplantation.This is in line with previous studies of the general population where PBC and AIH were associated with hypothyroidism due to AITD 7,[26][27][28] .The association between autoimmune liver diseases and hyperthyroidism is less well explored, but abnormal liver function tests are common in both hyperthyroidism and hypothyroidism, and multiple autoimmune diseases, including autoimmune liver disease and AITD, often coexist 7,29 .
Immunosuppressants may impact on the prevalence and severity of AITD 29 .Calcineurin inhibitors decrease regulatory T cell activity, which has been proposed to promote AITD 30,31 .Standard immunosuppression after liver transplantation in our clinics includes a calcineurin inhibitor, an antimetabolite, and corticosteroids 32 .In this study, we found that use of prednisolone was associated with hyperthyroidism in the adjusted models.This effect is probably due to confounding by indication, as corticosteroids in our clinics are often discontinued a year after transplantation, but frequently continued in recipients with autoimmune disease.We did not find evidence to suggest that other immunosuppressants are associated with hyperthyroidism or hypothyroidism in liver transplant recipients.
As COVID-19 may induce changes in thyroid function, comparing liver transplant recipients included during the pandemic with controls included before the pandemic may potentially impact the results 33,34 .However, access to COVID-19 tests was free and unrestricted for the public in Denmark, and we had access to all test results.As most data were collected during 2021 and early 2022, a minority of liver transplant recipients had a history of COVID-19 35 .Therefore, we were able to exclude participants with history of COVID-19 in a sensitivity analysis, and this did not affect our findings.
Strengths of this study include the large cohort size of around 85% of all liver transplant recipients in Denmark.With matched controls from the general population, investigation of interactions between liver transplantation and possible risk factors are possible.Furthermore, clinical measurements were performed uniformly in liver transplant recipients and in controls.With nation-wide access to hospital records, we also had access to detailed and complete data on transplantation-related variables and immunosuppressants on all liver transplant recipients.
A limitation is the use of different laboratories.Controls are recruited from an area of previous mild iodine deficiency prior to introduction of mandatory iodine fortification in 2000, while some liver transplant recipients live in areas of previous moderate iodine deficiency.After iodine fortification, geographical differences in hyperthyroidism incidence have disappeared by 2006-2007 36 , although there was still a higher incidence of hypothyroidism in areas of mild iodine deficiency by 2008 37 .In this contemporary cohort, we saw no difference in prevalence across the transplantation centres, so geographical differences are unlikely to account for the findings.Furthermore, the ethnic distribution of liver transplant recipients and controls from the capital area differed, but findings were consistent after further adjustment for ethnicity.
As data on autoimmune diseases among controls were not available, we were not able to include non-thyroidal autoimmune diseases in the adjusted models.Furthermore, the number of cases with hyperthyroidism limits the power to investigate potential risk factors.With a cross-sectional design, we cannot determine if the found associations reflect causation, and future longitudinal studies are warranted, preferably including pre-and posttransplantation thyroid function tests to investigate the incidence of de novo hyperthyroidism and hypothyroidism after liver transplantation.The high proportion of autoimmune diseases as causes of transplantation is a limitation to generalisability of this study to areas with a high proportion of other causes, e.g., metabolic dysfunction-associated steatohepatitis or viral hepatitis.
In conclusion, among liver transplant recipients, the prevalence of hyperthyroidism was twice that in ageand sex-matched controls from the general population.Higher clinical suspicion of hyperthyroidism may be warranted in clinics that follow liver transplant recipients as they age.

Table 3 .
Thyroid function in liver transplant recipients and controls.LT, liver transplant.a For the biochemical classification, only participants not using thyroid medications were included.
Significant values are in[bold]

Table 5 .
Associations of transplantation-related variables with hyperthyroidism.AIH, autoimmune hepatitis; aOR, odds ratio adjusted for age, sex, smoking, and BMI; CI, confidence interval; LT, liver transplantation; NA, not available due to low number of cases; OR, odds ratio; PBC, primary biliary cholangitis; PSC, primary sclerosing cholangitis.Significant values are in[bold]