Prevalence and Predictors of Thyroid Dysfunction in Patients with HIV Infection and Acquired Immunodeficiency Syndrome: An Indian Perspective

Background. Predictors of thyroid dysfunction in HIV are not well determined. This study aimed to determine the prevalence and predictors of thyroid dysfunction in HIV infected Indians. Methods. Consecutive HIV patients, 18–70 years of age, without any severe comorbid state, having at least 1-year follow-up at the antiretroviral therapy clinic, underwent clinical assessment and hormone assays. Results. From initially screened 527 patients, 359 patients (61.44 ± 39.42 months' disease duration), having good immune function [CD4 count >200 cell/mm3: 90.25%; highly active antiretroviral therapy (HAART): 88.58%], were analyzed. Subclinical hypothyroidism (ScH) was the commonest thyroid dysfunction (14.76%) followed by sick euthyroid syndrome (SES) (5.29%) and isolated low TSH (3.1%). Anti-TPO antibody (TPOAb) was positive in 3.90%. Baseline CD4 count had inverse correlation with TPOAb after adjusting for age and body mass index. Stepwise linear regression revealed baseline CD4 count, TPOAb, and tuberculosis to be best predictors of ScH after adjusting for age, weight, duration of HIV, and history of opportunistic fungal and viral infections. Conclusion. Burden of thyroid dysfunction in chronic HIV infection with stable immune function is lower compared to pre-HAART era. Thyroid dysfunction is primarily of nonautoimmune origin, predominantly ScH. Severe immunodeficiency at disease onset, TPOAb positivity, and tuberculosis were best predictors of ScH.


Introduction
Dramatic improvements in survival following institution of highly active antiretroviral therapy (HAART) have given rise to increased occurrence of endocrinopathies in HIV infected patients in the last two decades, which are associated with significant morbidity. Thyroid dysfunction is among the commonest endocrinopathies in HIV. Undiagnosed thyroid dysfunction, even subclinical hypothyroidism, is associated with significant morbidity and poor quality of life [1][2][3][4]. Subtle thyroid dysfunction is common, believed to occur in as many as 35% of all HIV infected individuals [5][6][7]. In contrast overt thyroid dysfunction is less common, believed to involve 1-2% of all patients [6,7]. Prevalence of overt primary hypothyroidism in the general population and HIV infected individuals from different studies across the globe has been reported to be 0.3% and 0-2.6%, respectively [6][7][8][9]. Similarly the prevalence of subclinical hypothyroidism

Methods
Antiretroviral therapy (ART) clinic at our institute has been functional since 2004, established by the National AIDS Control Organization (NACO), India, and the World Health Organization (WHO). The clinic provides for all the necessary investigations, medications (including highly active antiretroviral therapy [HAART]), counseling, and education to all patients with HIV infection. Consecutive ambulatory patients, 18-70 years of age, with serologically documented HIV infection, in stable clinical condition without any acute, severe illness, attending the ART clinic of our hospital were considered. Severely ill patients with multiple comorbid states, which would warrant hospital admission, were excluded. Patients with vitamin-D and/or calcium supplementation in the last 6 months were excluded. Patient records were reviewed and patients having clinical data of at least 1 year of follow-up were further evaluated. Patients whose CD4 cell count records at diagnosis and at least one followup (6-12 months after diagnosis) available were included in the study. The study protocol was explained to the considered patients and only those who gave informed written consent were included. The institutional ethics committee approved the study protocol. The period of the study was from August 2014 till September 2015.
Data were collected from the patients and their records regarding the duration of diagnosis of HIV infection and details of HAART. Data was also collected regarding past or current evidence of infections including opportunistic infections (bacterial, viral, and fungal). History of clinical features suggestive of hypothyroidism or hyperthyroidism was taken. All patients underwent detailed clinical assessment, including anthropometry. The patients were called the subsequent day in fasting state for blood sampling. Blood samples of 5 mL each were collected in plain and EDTA Vacutainer (Becton Dickinson). Serum was separated from blood collected in plain Vacutainer and processed immediately for routine biochemical analysis and one aliquot of serum was stored at −20 degrees' Celsius for specific immunological (hormonal) assays. EDTA sample was processed for hematological analysis.
The reference range of fT3, fT4, and TSH in our laboratory is 2-4.4 pg/mL, 0.6-2.2 ng/dL, and 0.5-5 mIU/L, respectively. Anti-TPO antibody titer <50 IU/mL is considered to be negative or normal. In accordance with previous literature, primary overt hypothyroidism was defined as patients having elevated TSH along with low fT4 [5,6]. Subclinical hypothyroidism was defined as patients having elevated TSH along with normal thyroid hormone levels. Subclinical hyperthyroidism was defined as patients having suppressed TSH levels with normal thyroid hormone levels [5,6]. Overt hyperthyroidism was defined as patients having suppressed TSH levels along with elevated fT4 and/or fT3. Sick euthyroid syndrome, a physiological adaptive phenomenon to conserve energy during periods of extreme stress and infection, was defined as patients having isolated low fT3 or low fT4 with low fT3 levels, along with low or normal TSH levels [10]. Euthyroidism was defined as clinically asymptomatic patients having normal fT3, fT4, and TSH levels. It has been reported that 1.3-6.8% HIV infected patients have isolated low TSH with normal thyroid hormone levels [11]. Hence, patients with normal fT3 and fT4 with low TSH were grouped separately. A few studies have also noted isolated low fT4 (up to 6.8% patients), a pattern distinct from sick euthyroid syndrome in patients with HIV infection, which is believed to be due to an alternation in hypothalamic-pituitary-thyroid axis [6]. Hence these patients were also grouped separately in our study. Serum 25-hydroxy-vitamin-D (25OHD) levels ≥30 ng/mL were defined as vitamin-D sufficiency, 20-30 ng/mL as vitamin-D insufficiency, and <20 ng/mL as vitamin-D deficiency [12].
Immune reconstitution inflammatory syndrome (IRIS) in HIV infected patients is characterized by clinical deterioration in a patient secondary reestablishment of immunity following initiation of HAART [13]. It is usually observed in patients with low baseline CD4 count, which increases rapidly following initiation of HAART. HAART has been linked to increased occurrence of autoimmunity and autoimmune disorders [13,14]. IRIS has been defined as an increased CD4 count above 200 cells/mm 3 in patients who previously had CD4 counts lower than 100-200 cells/mm 3 [15,16]. Hence patients in our study with baseline CD4 counts less than 200 cells/mm 3 , which increased to >200 cells/mm 3 at the first follow-up following initiation of HAART, were defined to have IRIS. Parameters like age, duration of HIV infection, occurrence of opportunistic infections, baseline and posttreatment CD4 counts, IRIS, antiretroviral drugs used, vitamin-D status, and autoimmunity (anti-TPO antibody) were evaluated for their role in predicting the occurrence of thyroid dysfunction.
Subtle thyroid dysfunction is common in HIV infected individuals, believed to involve as many as 35% of all individuals [5][6][7]. Hence for keeping a power of 80% and type-I error at 5%, it has been calculated that we need to recruit at least 246 patients in our study for accurate assessment of thyroid dysfunction.

Statistical Analysis
Normality of the distribution of variables was assessed using the Kolmogorov-Smirnov test. Independent -test and Wilcoxon rank sum test were done for normally distributed and skewed variables, respectively. Chi-square tests were used for categorical variables. Pearson's or Spearman's correlation coefficient was calculated for normally distributed and skewed variables, respectively. Multiple linear regression analyses were done to determine variables that independently influenced the occurrence of thyroid dysfunction after adjusting for factors in different models. A value <0.05 was considered statistically significant. SPSS version 20 was used for analyses.

Results
Five hundred and twenty-seven consecutive patients with HIV infection were screened at the ART clinic, of which 370 patients who fulfilled all criteria were considered for inclusion into the study. Reasons for exclusion from the study included severe associated illness ( = 23), multiple comorbid states like chronic liver disease, chronic kidney disease ( = 11), previous steroid use ( = 25), and history of calcium or vitamin-D supplementation in last 6 months ( = 41). Fortysix patients were excluded because of less than 1-year followup and 11 were excluded due to incomplete records. Of the 370 considered patients, 11 refused to consent to the study. Hence 359 patients (225 males and 134 females) who fulfilled all criteria and gave informed written consent underwent clinical assessment, hormonal evaluation, and analysis. Prevalence of vitamin-D deficiency (<20 ng/mL) and insufficiency (<30 ng/mL) among the study cohort was 55.71% (200/359) and 89.69% (322/359), respectively. Severe vitamin-D deficiency (<10 ng/mL) was observed in 9.19% (33/359) patients. At the time of diagnosis of HIV infection, 60.20% (216/359), 32.60% (117/359), and 7.20% (26/359) patients had CD4 count <200 cell/mm 3 , 200-500 cell/mm 3 , and >500 cell/mm 3 , respectively. The mean duration of HIV infection was 61.44 ± 39.42 months. Three hundred and nineteen (88.86%) patients were on HAART at the time of inclusion into the study. At the time of hormonal analysis, 9.75% (35/359), 58.50% (210/359), and 31.75% (114/359) patients had CD4 count <200 cell/mm 3 , 200-500 cell/mm 3 , and >500 cell/mm 3 , respectively. One hundred and forty-five patients (40.39%) had history of tuberculosis. None of the patients in this study had active tuberculosis at the time of recruitment. Six patients were on isoniazid and rifampicin at the time of recruitment as a part of maintenance phase of antitubercular therapy.
Patients with history of IRIS were older ( = 0.049), were more likely to be males ( = 0.007), had lower BMI ( = 0.002), higher history of tuberculosis ( = 0.002), and higher use of protease inhibitors ( < 0.001), and had significantly lower baseline ( < 0.001) and current CD4 cell counts ( = 0.005) ( Table 2). Serum fT3 was significantly higher in patients with history of IRIS ( = 0.036) ( Table 2). The occurrence of different types of thyroid dysfunction was comparable in patients with history of IRIS as compared to those without (Table 2).
An inverse correlation was observed between baseline CD4 count ( = 0.031) and anti-TPO antibody titers, which persisted even after adjusting for age and body mass index ( = 0.032) ( Table 3). Similarly an inverse correlation was observed in CD4 count at present with TSH levels, both at baseline ( = 0.043) and after adjusting for age and body mass index ( = 0.049) ( Table 3). Stepwise linear regression analysis revealed that anti-TPO antibody titers and CD4 cell count at the time of initial diagnosis of HIV infection were the 2 best predictors of occurrence of subclinical hypothyroidism, at baseline (Model-1), after adjusting for age and duration of HIV infection (Model-2), and after adjusting for variables in Model-2 plus weight and history of opportunistic fungal and viral infections (Model-3) ( Table 4). Increased anti-TPO antibody titers and lower baseline CD4 count were independent predictors of increased occurrence of subclinical hypothyroidism. Previous history of tuberculosis tended to be a good predictor of subclinical hypothyroidism later in life both at baseline ( = 0.084) and after adjusting for variables in Model-2 ( = 0.087) and Model-3 ( = 0.065) ( Table 4).

Discussion
The occurrence of sick euthyroid syndrome among HIV infected patients is highly variable ranging from 1.3% to 11.6% in different studies [11,[16][17][18][19]. Stable, ambulatory, asymptomatic patients, with a large majority being on HAART (88.86%) with stable immune function (as evidenced by only 9.75% patients having CD4 count <200 cell/mm 3 at the time of recruitment into this study and hormonal assessment) may explain the low occurrence of sick euthyroid syndrome in our study cohort. Heterogeneity in the disease profile of the patients evaluated in different studies (duration of infection, severity of immunodeficiency, in-patient versus out-patient, associated comorbidities, and functional status) may explain this variation. A pilot study from central India reported a high prevalence of subclinical (30%) and overt hypothyroidism (10.66%) ; all discreet variables have been expressed as absolute numbers (percentage); Wilcoxon rank sum test was done for analysis; normally distributed continuous variables were analyzed using unpaired -test; normality checked using Kolmogorov-Smirnov test; < 0.05 considered statistically significant; # value calculated using Chi-square test; * viral infections include hepatitis-B, hepatitis-C; HAART: highly active antiretroviral therapy; NRTI: nucleoside reverse transcriptase inhibitors; NNRTI: nonnucleoside reverse transcriptase inhibitor; PI: protease inhibitors; zidovudine (AZT), lamivudine (3TC), stavudine (d4T), and/or tenofovir (TDF) were the NRTIs received by the patients; nevirapine (NVP) or efavirinez (EFV) was NNRTIs received by the patients; atazanavir (ATV) or ritonavir (RTV) was the PI received by the patients; 25OHD: 25-hydroxy-vitamin-D; IRIS: immune reconstitution inflammatory syndrome; ALP: alkaline phosphate.
in a cohort of 150 HAART naïve newly diagnosed HIV infected patients [17]. Subclinical hypothyroidism was the commonest type of thyroid dysfunction observed in our study cohort documented in 14.76% patients. The prevalence is comparable to previous reports from other countries [6][7][8][9]. However this is in contrast to a previous report from our institute where a very high occurrence of subclinical hypothyroidism (53%) was documented in patients newly diagnosed with HIV infection [20]. It is important to highlight that the patients evaluated in that study were newly diagnosed with HIV infection, had advanced immunodeficiency (mean CD4 count: 147.1 ± 84 cell/mm 3 , 70.1% had CD4 count <200 cell/mm 3 ), and were HAART naïve [20]. This is consistent with data reported from pre-HAART era and from patients newly diagnosed with HIV infection not on HAART, where the occurrence of thyroid dysfunction has been reported to be higher ranging from 10 to 40% [17,21]. This is in contrast to our patients having mean disease duration of 5 years, with a large majority being on HAART (88.6%), being clinically stable, asymptomatic, and ambulatory with good immune function (90.25% having CD4 count >200 cell/mm 3 ; 31.75% having CD4 count >500 cell/mm 3 ).

6
Journal of Thyroid Research   Anti-TPO antibody titers, baseline CD4 counts, and history of tuberculosis were the best predictors of subclinical hypothyroidism in HIV infected patients in our study. Anti-TPO antibody positivity has previously been documented to be predictors of subclinical and overt hypothyroidism in normal and postpartum individuals [22,23]. However it needs to be emphasized that, in spite of being a good predictor of subclinical hypothyroidism in our study, the prevalence of elevated anti-TPO antibody titers in our cohort was only 3.9%, which is lower than that observed in the general population (3-15%) [24]. A previous study of 642 normal individuals from northern India revealed the prevalence of anti-TPO positivity to be 21% [25].
The inverse relation observed between current CD4 count and TSH even after adjusting for variables highlights the possible link between increased immunodeficiency and elevated TSH. Further, in the setting of low prevalence of elevated anti-TPO antibody titers, low baseline CD4 count being a strong independent predictor of subclinical hypothyroidism highlights the importance of early immunodeficiency and HIV infection per se having an important role in the genesis of subclinical hypothyroidism later in life. It is likely that patients with lower CD4 counts had higher viral load. However, HIV viral load was not evaluated in our study and is a limitation of this report. Previous history of tuberculosis being an important independent predictor of hypothyroidism in our study also highlights the importance of immunodeficiency in the pathogenesis of thyroid dysfunction (as tuberculosis is more common in the immunodeficient state). A high rate of hypothyroidism (54%) was reported from 69 HIV infected patients with multidrug resistant (MDR) tuberculosis from Mumbai, India [26]. Use of rifampicin, para-aminosalicylate (PAS), and ethionamide for treating tuberculosis has been linked to the increased occurrence of hypothyroidism in these patients [26].
The inverse relationship observed between baseline CD4 count and anti-TPO antibody titers raises the possibility of increased immunodeficiency early in the course of disease being linked to increased occurrence of autoimmunity later. Increased autoantibody expression has previously been documented in patients of HIV infection with lower CD4+ cell counts, believed to be the result of a direct effect of the virus on endothelium, hematopoietic cells, and different tissues leading to enhancement of the cytotoxic activity of immune cells and autoantigen expression [27,28]. Induction of immune restoration by HAART has been linked to thyroid dysfunction in some studies [29,30]. This was primarily observed with regard to occurrence of hyperthyroidism and Graves' disease [17,29,30]. Graves' disease has most commonly been reported 12-36 months after HAART initiation [17]. Subclinical hyperthyroidism was not observed in our study. Also the occurrence of overt hyperthyroidism was too low in our study (0.01%) to be able to evaluate the factors responsible for their occurrence. In our study, IRIS was not a predictor of thyroid dysfunction.
In our study, none of the antiretroviral agents were predictors of occurrence of subclinical hypothyroidism. Studies evaluating the link between HAART and thyroid dysfunction have given conflicting results, with some but not all studies observing a link [6,7]. A large cross-sectional study in 350 patients did not observe any link between HAART and occurrence of subclinical hypothyroidism [6]. Afhami et al. documented similar observations in a cross-sectional study of 85 HIV infected patients [31]. A higher than expected incidence of hypothyroidism (10.7 per 1000 treated patient years) but not hyperthyroidism (3.7 per 1000 treated patient years)