Possible effects of HIV on thyroid function in the modern combination antiretroviral therapy
(cART) era are largely unknown. We aimed to investigate the prevalence
and associated risk factors of thyroid dysfunction in well treated
people living with HIV (PLWH) and matched uninfected controls and to examine whether HIV is independently associated with thyroid dysfunction.
stimulating hormone (TSH), free thyroxine, total thyroxine and free
triiodothyronine were measured in 826 PLWH from the Copenhagen
co-morbidity in HIV
infection (COCOMO) Study and in 2,503 matched uninfected controls, and
medical treatment for hypo- or hyperthyroidism was recorded. Multinomial
logistic regression adjusting for known risk factor was used to examine
the association between HIV and thyroid dysfunction and multivariate linear regression to study the association between HIV and serum TSH concentrations.
PLWH were generally well-treated with 95% having undetectable viral
replication. Among PLWH and controls 31 (3.8%) and 114 (4.6%) had
hypothyroidism, and 7 (0.8%) and 21 (0.8%) had hyperthyroidism,
respectively. In adjusted analyses, we found no significant associations
between HIV and hypothyroidism OR 0.8 [0.6, 1.3] p = 0.40 or between HIV and hyperthyroidism OR 1.1 [0.5, 2.5] p = 0.91. Furthermore, serum TSH concentration was unrelated to HIV status (p = 0.6).
was no difference in prevalence of hyperthyroidism or hypothyroidism in
well treated PLWH compared to uninfected controls. HIV status was not associated with thyroid dysfunction or serum TSH concentration.
The only statistical significant associations were between female sex and both hypothyroidism (OR 1.9 [1.3, 2.8], p<0.01) and hyperthyroidism (OR 3.6 [1.6, 7.7], p<0.01) and between age and hypothyroidism (OR 1.2 [1.03-1.40], p=0.02), for each additional 10 years (estimates from model 1). The same models were used for multivariable linear regression to test the association between HIV and serum TSH concentration. None of the models showed any association between HIV and TSH. Neither current CD4 count (OR 1.0 [0.98, 1.01] p=0.81 and 1.0 [0.99, 1.04] p=0.28, hypothyroidism and hyperthyroidism respectively), CD4 nadir (OR 1.0 [0.97, 1.03] p= 0.76, 1.0 [0.94, 1.04] p=0.99, hypothyroidism and hyperthyroidism respectively) nor duration of HIV(OR 1.0 [0.96, 1.06] p= 0.73, 0.9 [0.82, 1.06] p=0.27, hypothyroidism and hyperthyroidism respectively) were found to be associated with thyroid dysfunction.
Prior to introduction of combination antiretroviral therapy (cART), abnormal thyroid function tests in people living with HIV (PLWH) were common and often due to the impact of non thyroidal illness. Less frequently, thyroid dysfunction was caused by opportunistic infections and neoplasms while development of autoimmune thyroid disease was rare. Initiation of cART results in suppression of viral replication and improvement of immune function. Since most thyroid diseases in iodine sufficient areas are autoimmune, improvement of immune function may, in theory, precipitate autoimmune thyroid disease[4,5]. Thus, immune reconstitution syndrome following cART has been linked to Graves’ disease. Previous studies have reported both higher than expected prevalence of especially subclinical thyroid disease[7–13] and no apparent increased prevalence of thyroid diseases in PLWH[14,15].
Common for these studies is the lack of uninfected controls. Only one study conducted in the cART era has compared prevalence of thyroid dysfunction in PLWH and uninfected controls. However, in this study only 2/3 of PLWH were on cART and the majority of both PLWH and controls were co-infected with hepatitis C virus. In that cohort of 362 PLWH and 300 controls no association between thyroid dysfunction and HIV was observed.
Consequently, it remains unknown whether HIV is an independent predictor of thyroid dysfunction in well-treated PLWH in the modern cART era. The aims of the study were to investigate the prevalence of thyroid dysfunction in PLWH and matched uninfected controls and to examine whether HIV is independently associated with thyroid dysfunction. We hypothesized that thyroid dysfunction would be more prevalent in PLWH compared to uninfected controls.
This study is the first to evaluate thyroid dysfunction in a large cohort of well-treated PLWH and uninfected controls in the modern cART era. We found no association between HIV and thyroid dysfunction. Furthermore, HIV-related variables were not associated with hypothyroidism or hyperthyroidism. Thus, no evidence was found to support increased risk of thyroid dysfunction in well-treated PLWH.
PLWH in our cohort were well-treated. Wiener et al. did not identify an association between HIV and thyroid dysfunction and their findings were comparable to ours despite their inclusion of a population that was less likely to be on cART (62%) and more likely to be coinfected with hepatitis C virus (70%) . Furthermore, the prevalence of thyroid dysfunction in PLWH in our study was in the same range as that reported in previous studies of PLWH that did not include controls[7–15]. The prevalence of hyper- and hypothyroidism among uninfected participants in our study was slightly lower than what was found in another study examining thyroid dysfunction in the general population of Copenhagen. This study found a prevalence of 1.8% and 5.8% for biochemical hyper- and hypothyroidism, respectively. However, 80% of participants in that study were female compared to less than 20% in our study which may explain the difference in prevalence. The HIV-related variables CD4 T cell count, CD4 T cell nadir and duration of HIV were not related to the prevalence of hypo- and hyperthyroidism. This is in accordance with Madge et al but in contrast to Beltran et al who found an association between low CD4 T cell count and hypothyroidism in a study including 350 PLWH. Strengths of our study include a large cohort of PLWH and well-characterized and well-matched uninfected controls. All participants are from the same geographical area and it is reasonable to assume a similar iodine intake. However, this region is characterized by mild iodine deficiency and the prevalence of thyroid dysfunction found in this study may not apply to regions with more severe iodine deficiency. Furthermore, the laboratory defined reference ranges were in accordance with the American Association of Clinical Endocrinologists and American
Thyroid Association cosponsored guidelines and European Thyroid association guidelines. A limitation is the necessity to rely on self-reported information about use of medication. Furthermore, T4 and T3 were only measured when TSH was abnormal. Thus, we were not able to investigate the prevalence of low T3 and T4 with normal TSH. In subclinical thyroid disease a repeated thyroid function test is generally recommended. Consequently, overdiagnosis of subclinical disease of both PLWH and controls may be expected.
In conclusion, prevalence of hyperthyroidism and hypothyroidism was similar in well-treated PLWH and uninfected controls. Furthermore, no association between HIV and thyroid dysfunction was found. Finally, HIV status was not associated with serum TSH concentrations. Thus, in well-treated PLWH there was no evidence of increased risk of thyroid dysfunction.