CLINICAL SCIENCE
New-Onset Diabetes and Antiretroviral Treatments in HIV-Infected Adults in Thailand Prakit Riyaten, BSc,*† Nicolas Salvadori, MSc,†‡ Patrinee Traisathit, PhD,* Nicole Ngo-Giang-Huong, PhD,†‡§ Tim R. Cressey, PhD,†‡§ Prattana Leenasirimakul, MD,k Malee Techapornroong, MD,¶ Chureeratana Bowonwatanuwong, MD,# Pacharee Kantipong, MD,** Ampaipith Nilmanat, MD, BPA,†† Naruepon Yutthakasemsunt, MD,‡‡ Apichat Chutanunta, MD,§§ Suchart Thongpaen, MD,kk Virat Klinbuayaem, MD,¶¶ Luc Decker, PhD,†‡ Sophie Le Cœur, PhD,†‡## Marc Lallemant, MD, MSc,†‡§ Jacqueline Capeau, MD, PhD,*** Jean-Yves Mary, PhD,††† and Gonzague Jourdain, MD, PhD†‡§
Background: Use of several antiretrovirals (ARVs) has been shown to be associated with a higher risk of diabetes in HIV-infected adults. We estimated the incidence of new-onset diabetes and assessed the association between individual ARVs and ARV combinations, and diabetes in a large cohort in Thailand.
Methods: We selected all HIV-1–infected, nondiabetic, antiretroviral-naive adults enrolled in the Program for HIV Prevention and Treatment cohort (NCT00433030) between January 2000 and December 2011. Diabetes was defined as confirmed fasting plasma
Received for publication December 23, 2014; accepted February 18, 2015. From the *Department of Statistics, Faculty of Science, Chiang Mai University, Chiang Mai, Thailand; †Institut de recherche pour le développement (IRD) UMI 174-PHPT, Chiang Mai, Thailand; ‡Department of Medical Technology, Faculty of Associated Medical Sciences, Chiang Mai University, Chiang Mai, Thailand; §Department of Immunology and Infectious Diseases, Harvard School of Public Health, Boston, MA; kNakornping Hospital, Chiang Mai, Thailand; ¶Prapokklao Hospital, Chantaburi, Thailand; #Chonburi Hospital, Chonburi, Thailand; **Chiangrai Prachanukroh Hospital, Chiang Rai, Thailand; ††Hat Yai Hospital, Songkla, Thailand; ‡‡Nong Khai Hospital, Nong Khai, Thailand; §§Samutsakhon Hospital, Samutsakhon, Thailand; kkMahasarakham Hospital, Mahasarakham, Thailand; ¶¶Sanpatong Hospital, Chiang Mai, Thailand; ##Institut National d’Etudes Démographiques, Paris, France; ***Sorbonne Universités, INSERM UPMC UMR_S938, APHP, Univ Paris 06, Paris, France; and †††INSERM U1153, Team ECSTRA, Université Paris Diderot—Paris 7, Hôpital Saint-Louis, Paris, France. Supported by the Global Fund to Fight AIDS, Malaria and Tuberculosis (PR-A-N-008); Oxfam Great Britain, Thailand (THAA51); Ministry of Public Health, Thailand; and Institut de recherche pour le développement (IRD), France. P.R. received support scholarships from IRD UMI 174-PHPT; the Graduate School, Faculty of Science and the Center of Excellence for Innovation in Analytical Science and Technology (I-ANALY-S-T), Chiang Mai University, Thailand. The results of this study were partially presented as a poster at the 20th International AIDS Conference, July 20–25, 2014, Melbourne, Australia (MOPE182). The authors have no conflicts of interest to disclose. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.jaids.com). Correspondence to: Gonzague Jourdain, MD, PhD, 187/10, Changklan Road, Changklan, Muang, Chiang Mai 50100, Thailand (e-mail: [email protected]). Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
glucose $126 mg/dL or random plasma glucose $200 mg/dL. Incidence was the number of cases divided by the total number of person-years of follow-up. Association between ARVs and ARV combinations, and new-onset diabetes was assessed using Cox proportional hazards models.
Results: Overall, 1594 HIV-infected patients (76% female) were included. Median age at antiretroviral therapy initiation was 32.5 years. The incidence rate of diabetes was 5.0 per 1000 person-years of followup (95% confidence interval: 3.8 to 6.6) (53 cases). In analyses adjusted for potential confounders, exposure to stavudine + didanosine [adjusted hazard ratio (aHR) = 3.9; P = 0.001] and cumulative exposure $1 year to zidovudine (aHR = 2.3 vs. no exposure; P = 0.009) were associated with a higher risk of diabetes. Conversely, cumulative exposure $1 year to tenofovir (aHR = 0.4 vs. no exposure; P = 0.02) and emtricitabine (aHR = 0.4 vs. no exposure; P = 0.03) were associated with a lower risk. Conclusions: The incidence of diabetes in this predominantly female, young, lean population was relatively low. Although stavudine and didanosine have now been phased out in most antiretroviral therapy programs, our analysis suggests a higher risk of diabetes with zidovudine, frequently prescribed today in resource-limited settings. Key Words: HIV, diabetes, incidence, NRTIs, tenofovir, zidovudine (J Acquir Immune Defic Syndr 2015;69:453–459)
INTRODUCTION Antiretroviral therapy (ART) provides clear health and quality of life benefits to people living with HIV, but longterm adverse metabolic effects are of concern. Some antiretrovirals (ARVs), especially several protease inhibitors (PIs) and thymidine analogue nucleoside reverse transcriptase inhibitors (NRTIs), have been associated with the development of insulin resistance, drug-induced metabolic syndrome and diabetes.1–6 However, previous reports have not directly compared the risk of diabetes between NRTI-containing regimens. Moreover, the incidence of diabetes has not been evaluated in southeast Asian HIV-infected populations. HIV and diabetes are 2 chronic diseases with severe public health burdens in Thailand. In 2013, the estimated HIV
J Acquir Immune Defic Syndr Volume 69, Number 4, August 1, 2015
www.jaids.com |
453
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Acquir Immune Defic Syndr Volume 69, Number 4, August 1, 2015
Riyaten et al
prevalence was 1.1% in adults (age, 15–49 years),7 and the prevalence of diabetes in the adult population was 6.4%.8 Data on type 2 diabetes in HIV-infected patients in Thailand population are lacking. Our objective was to estimate the incidence of new-onset diabetes and assess the association between exposure and duration of exposure to ARVs and ARV combinations (ARVcs), and diabetes in a large HIV-infected adult cohort in Thailand.
METHODS Study Population The Program for HIV Prevention and Treatment (PHPT) cohort is an ongoing, observational, prospective, multicenter study initiated in 1999, which has enrolled more than 2400 HIVinfected patients from 50 public hospitals throughout Thailand (NCT00433030). HIV-1 infected adults (age, $18 years) who initiated ART between January 1, 2000, and December 31, 2011, nondiabetic according to chart review, and with at least 2 glucose measurements after ART initiation were included. ARTexperienced patients referred to PHPT sites and enrolled in the cohort were excluded. However, patients who received ARVs at PHPT sites solely for the prevention of mother-to-child transmission of HIV were not excluded.
Definition of Diabetes Following the American Diabetes Association guidelines, diabetes was defined as either a confirmed fasting plasma glucose $126 mg/dL or a confirmed random plasma glucose $200 mg/dL.9 The date and time of the last meal were systematically recorded at the time of blood draw. Blood glucose was considered as fasting if the last meal was more than 7 hours before the blood draw. Glycosylated hemoglobin was not part of the data collected in this cohort.
Baseline and Follow-up Data Baseline characteristics were retrieved from the closest assessment between 1 year before and 2 weeks after ART initiation: gender, date of birth, weight, height, Centers for Disease Control and Prevention (CDC) HIV classification,10 hepatitis B surface antigen and hepatitis C antibody, CD4 cell count, plasma HIV-1 RNA levels, glucose, triglycerides, and total cholesterol. Every 6 months, weight, CD4 cell count, HIV RNA load, glucose, triglycerides, and total cholesterol were measured as part of standard of care in this cohort. Clinical data collected on case report forms and copies of laboratory slips from the sites were forwarded to the PHPT Data Center for double data entry and routine data management.
censored at the date of death or last visit in case of voluntary withdrawal from the study or loss to follow-up, defined as an absence of visit or contact for at least 9 months. The overall and stratified incidence rates of diabetes were calculated as the number of new-onset diabetes divided by the total number of person-years of follow-up (PYFU). Confidence intervals (CIs) of the incidence rates were based on the Poisson distribution. The following potential baseline confounders in the association between exposure to ARVs and ARVcs, and new-onset diabetes were assessed using univariate and multivariate Cox proportional hazards models: gender, age, triglycerides, total cholesterol, HIV RNA load, hepatitis B surface antigen, hepatitis C antibody, CDC stage, calendar year of enrollment, time-updated body mass index (BMI) and time-updated CD4 cell count. All continuous variables were dichotomized according to the median values except BMI (categories: ,25 and $25 kg/m2) and CD4 count (,200 and $200 cells/mm3). Variables with a P value ,0.25 in univariate analyses were included in a multivariate model. A backward stepwise regression analysis was then performed to identify significant potential confounders in the multivariate analysis. The association between ARVs and ARVcs, and newonset diabetes was assessed using unadjusted and adjusted Cox proportional hazards models. Adjusted models included all significant potential confounders in the multivariate analysis. ARVcs were entered in the models as dichotomous variables (ever exposed or never exposed) and ARVs as timeupdated variables categorized in 3 groups based on the cumulative duration of exposure (no exposure, less than 1 year, or more than 1 year). The ARVs and ARVcs selected for the analysis were those received by at least 5% of the patients. Overall P values were based on partial likelihood ratio tests using Breslow method for ties, and specific P values on Wald tests. We tested the proportional hazards assumption of each model based on Schoenfeld residuals.11 Missing values at baseline were imputed using multivariate imputation by chained equations, based on the predictive mean matching approach for continuous variables and logistic regression for binary variables.12
Subanalysis
Study population characteristics were presented as medians and interquartile ranges (IQRs) for continuous variables and as counts and percentages for categorical variables. The follow-up period was defined as the period between baseline and the date of first diagnosis of diabetes, date of death, or of last clinic visit, whichever occurred first. Observations were
Because a significant proportion of patients were exposed sequentially to various antiretroviral drugs, we conducted a subanalysis restricted to those patients who received, exclusively and for a longer time than in the main analysis (at least 2 years), tenofovir disoproxil fumarate (TDF), zidovudine (ZDV), stavudine (d4T), or didanosine + stavudine (ddI + d4T) as part of their first-line regimen. The association between these 4 NRTI-containing regimens and new onset diabetes was assessed and compared using a Cox proportional hazards model adjusted for all variables significantly different across the 4 groups (tested using Fisher exact tests). All reported P values are 2 sided and P values ,0.05 were considered statistically significant. All analyses were performed using Stata software version 10.1 (Stata Corp, College Station, TX).
454
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Statistical Analysis
| www.jaids.com
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Acquir Immune Defic Syndr Volume 69, Number 4, August 1, 2015
Diabetes and Antiretrovirals in Thailand
TABLE 1. Baseline Characteristics of the Study Population According to the Occurrence or Not of Diabetes During Follow-up and Incidence Rates Diabetes Characteristics Overall Gender Female Male Age (years) ,30 $30 BMI (kg/m2) ,25 $25 Hepatitis B surface antigen (n = 1538) Positive Negative Hepatitis C antibody (n = 1521) Positive Negative Triglycerides (mg/dL) (n = 1023) ,140 $140 Total cholesterol (mg/dL) (n = 1026) ,160 $160 HIV RNA load (log10 copies/mL) (n = 1485) ,5 $5 CD4 count (cells/mm3) (n = 1567) ,200 $200 CDC stage A B or C Calendar year of enrollment 2000–2005 2006–2011
Incidence of Diabetes
No Diabetes, n (%)
n (%)
PYFU
IR
95% CI
1541 (97)
53 (3)
10,507
5.0
3.8 to 6.6
1181 (78) 360 (22)
37 (70) 16 (30)
8113 2394
4.6 6.7
3.2 to 6.3 3.8 to 10.9
569 (37) 972 (63)
9 (17) 44 (83)
3843 6664
2.3 6.6
1.1 to 4.4 4.8 to 8.9
1347 (87) 194 (13)
37 (70) 16 (30)
9345 1162
4.0 13.8
2.8 to 5.5 7.9 to 22.4
76 (5) 1411 (95)
1 (1) 50 (99)
500 9597
2.0 5.2
0.1 to 11.1 3.9 to 6.9
57 (4) 1413 (96)
3 (6) 48 (94)
397 9573
7.6 5.0
1.6 to 22.1 4.0 to 6.6
541 (54) 453 (46)
10 (34) 19 (66)
3303 2531
3.0 7.5
1.5 to 5.6 4.5 to 11.7
519 (52) 477 (48)
15 (50) 15 (50)
3199 2656
4.7 5.6
2.6 to 7.7 3.2 to 9.3
878 (61) 563 (39)
26 (59) 18 (41)
5823 3792
4.5 4.7
2.9 to 6.5 2.8 to 7.5
1082 (71) 434 (29)
44 (86) 7 (14)
7777 2579
5.7 2.7
4.1 to 7.6 1.1 to 5.6
946 (61) 595 (39)
28 (53) 25 (47)
6353 4154
4.4 6.0
2.9 to 6.4 3.9 to 8.9
822 (53) 719 (47)
33 (62) 20 (38)
6656 3851
5.0 5.2
3.4 to 7.0 3.2 to 8.0
Comparison tests are listed in Table 2 IR, incidence rate per 1000 person-years of follow-up.
Ethical Considerations All participants provided written informed consent before participation in this study. The PHPT cohort study protocol and amendments received ethical clearance from the Ethics Committees of the Thai Ministry of Public Health and the Faculty of Associated Medical Sciences, Chiang Mai University, Thailand.
RESULTS
The remaining 1594 patients were included in this analysis. There were no patients with diabetes at entry in the cohort.
Baseline Characteristics Of the 1594 participants, 1218 (76%) were female. At baseline, the median age was 32.5 years (IQR: 28.2–37.7), BMI was 20.5 kg/m2 (IQR: 18.6–23.0), HIV RNA load was 4.8 log10 copies per milliliter (IQR: 4.2–5.2), and CD4 count was 137 cells per cubic millimeter (IQR: 67–211).
Study Population A total of 2092 HIV-infected adults on ART were enrolled between January 2000 and December 2011. Of these patients, 203 received ART before enrollment and 295 had no or only 1 glucose measurement during the follow-up period. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Follow-up The median duration of follow-up was 6.9 years (IQR: 4.9–8.2). During follow-up, 91% of the patients were exposed www.jaids.com |
455
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Acquir Immune Defic Syndr Volume 69, Number 4, August 1, 2015
Riyaten et al
to lamivudine (3TC), 62% to zidovudine, 58% to TDF, 48% to emtricitabine, 42% to stavudine, 14% to didanosine, 58% to efavirenz (EFV), 50% to nevirapine and 33% to ritonavir (see Table, Supplemental Digital Content 1, http://links.lww.com/QAI/A673, which describes the main ARVs and ARVcs administered at ART initiation and during follow-up). A total of 65 (4%) patients died, 158 (10%) were lost to follow-up, and 246 (15%) voluntarily withdrew from the study, most often because they moved to another province. On treatment, 1287 of 1475 patients (87%) had a viral load ,400 copies per milliliter at 1 year after ART initiation, 1240 of 1318 (94%) at 3 years, and 1117 of 1162 (96%) at 5 years.
P = 0.009) were associated with a higher risk of diabetes. Conversely, exposure to TDF + 3TC + EFV (aHR = 0.1; 95% CI: 0.05 to 0.5; P = 0.001) and cumulative exposure $1 year to TDF (aHR = 0.4 vs. no exposure; 95% CI: 0.2 to 0.9; P = 0.02) and emtricitabine (aHR = 0.4 vs. no exposure; 95% CI: 0.2 to 0.9; P = 0.03) were associated with a lower risk of diabetes (Table 3; see Table in Supplemental Digital Content 1, http://links.lww.com/QAI/A673, which describes the main ARVs and ARVcs administered at ART initiation and during follow-up).
Subanalysis
Incidence of Diabetes A total of 53 patients were diagnosed with diabetes over 10,507 PYFU, yielding an overall incidence rate of 5.0 per 1000 PYFU (95% CI: 3.8 to 6.6) (Table 1). Incidence rates were 6.7 per 1000 PYFU in men (95% CI: 3.8 to 10.9) and 4.6 per 1000 PYFU in women (95% CI: 3.2 to 6.3).
Factors Associated With New-Onset Diabetes The results of the univariate and multivariate analyses to identify non–drug-related risk factors of diabetes are presented in Table 2. In the multivariate analysis, older age, higher triglycerides, and time-updated BMI were independently associated with the risk of diabetes. However, HIVrelated characteristics (CDC stage, HIV RNA load, and time-updated CD4 cell count) were not associated with the risk of diabetes. The analysis assessing the role of ARVs and ARVcs on the risk of diabetes showed that after adjusting for age, triglycerides, and time-updated BMI, exposure to ddI + d4T [adjusted hazard ratio (aHR) = 3.9; 95% CI: 1.7 to 9.0; P = 0.001] and ZDV + 3TC + EFV (aHR = 2.2; 95% CI: 1.2 to 3.8; P = 0.007) and cumulative exposure $1 year to zidovudine (aHR = 2.3 vs. no exposure; 95% CI: 1.2 to 4.2;
A subanalysis conducted to assess and compare the association between 4 NRTI-containing regimens and new-onset diabetes showed that a total of 520 HIVinfected patients received, exclusively and for at least 2 years, TDF, zidovudine, stavudine, or ddI + d4T as part of their first-line regimen. At baseline, 63% were female, and the median age was 34.1 years (IQR: 29.5–40.1). A total of 329 (63%) patients received a first-line ART containing TDF, 28% zidovudine, and 7% stavudine (usually in addition to lamivudine or emtricitabine), and 2% received ddI + d4T. Gender, age, hepatitis B surface antigen, and hepatitis C antibody were significantly different across the 4 groups and were thus included in the adjusted model (see Table, Supplemental Digital Content 2, http://links.lww.com/QAI/A673, which presents the baseline characteristics of the subanalysis). In the adjusted analysis, patients exposed to a first-line regimen containing zidovudine (aHR = 6.6; 95% CI: 1.6 to 27.5; P = 0.01) or ddI + d4T (aHR = 73.2; 95% CI: 11.5 to 465.2; P , 0.001) were at a significantly higher risk of diabetes than those exposed to a first-line regimen containing TDF (see Table, Supplemental Digital Content 3, http://links.lww.com/QAI/A673, which evaluates the association between NRTI-containing first-line regimens and the occurrence of diabetes during follow-up).
TABLE 2. Association Between Baseline and Follow-up Characteristics, and the Occurrence of Diabetes During Follow-up Univariate Analysis Characteristics At ART initiation Male gender Age $30 yr Positive hepatitis B surface antigen Positive hepatitis C antibody Triglycerides $140 mg/dL Total cholesterol $160 mg/dL HIV RNA load $5 log10 copies/mL CDC stage B or C Calendar year of enrollment $2006 Time-updated variables CD4 count $200 cells/mm3 BMI $25 kg/m2
Diabetes
Total
HR (95% CI)
16 44 2 3 32 27 22 25 20
376 1016 79 64 709 766 613 620 739
1.5 2.9 0.8 1.4 2.1 1.2 1.2 1.4 1.0
— —
— —
Multivariate Analysis P
aHR (95% CI)
P
2.7) 6.1) 3.3) 4.5) 3.7) 2.1) 2.0) 2.3) 1.9)
0.20 0.001 0.73 0.59 0.006 0.49 0.63 0.27 0.87
— 2.7 (1.3 to 5.7) — — 1.8 (1.0 to 3.1) — — — —
— 0.007 — — 0.04 — — — —
1.1 (0.4 to 3.2) 3.1 (1.7 to 5.4)
0.80 ,0.001
— 2.7 (1.5 to 4.7)
— 0.001
(0.8 (1.4 (0.2 (0.4 (1.2 (0.7 (0.7 (0.8 (0.6
to to to to to to to to to
HR, hazard ratio. Bolded P values correspond to P,0.05 (for both Tables 2 and 3).
456
| www.jaids.com
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Acquir Immune Defic Syndr Volume 69, Number 4, August 1, 2015
Diabetes and Antiretrovirals in Thailand
TABLE 3. Association Between ARVs and ARVcs, and the Occurrence of Diabetes Unadjusted Analysis ARVs and ARVcs ARVs (time-updated exposure to) TDF ,1 yr vs. never exposed $1 yr vs. never exposed ZDV ,1 yr vs. never exposed $1 yr vs. never exposed Stavudine ,1 yr vs. never exposed $1 yr vs. never exposed Didanosine ,1 yr vs. never exposed $1 yr vs. never exposed Emtricitabine ,1 yr vs. never exposed $1 yr vs. never exposed Lamivudine ,1 yr vs. never exposed $1 yr vs. never exposed Efavirenz ,1 yr vs. never exposed $1 yr vs. never exposed Nevirapine ,1 yr vs. never exposed $1 yr vs. never exposed Ritonavir ,1 yr vs. never exposed $1 yr vs. never exposed ARVcs (ever exposed to) ddI + d4T d4T + 3TC + EFV d4T + 3TC + NVP TDF + 3TC + EFV TDF + FTC + EFV TDF + FTC + NVP TDF + FTC + IDV/r ZDV + 3TC + EFV ZDV + 3TC + NVP ZDV + 3TC + IDV/r
0.14 0.43 0.11 0.05 0.14 0.02 0.89 0.72 0.80 0.96 0.87 0.81 0.19 0.45 0.14 0.43 0.66 0.21 0.66 0.89 0.36 0.89 0.65 0.90 0.72 0.43 0.68
1.4 (0.6 to 3.5) 0.6 (0.3 to 1.2) 1.9 (0.8 to 4.3) 2.1 (1.1 to 3.9) 0.8 (0.2 to 3.3) 1.1 (0.6 to 2.0) 0.8 (0.1 to 6.1) 0.9 (0.3 to 2.8) 1.5 (0.6 to 3.9) 0.6 (0.3 to 1.2) 1.2 (0.4 to 3.6) 1.5 (0.8 to 3.0) 1.1 (0.4 to 3.3) 1.3 (0.7 to 2.4) 0.7 (0.2 to 3.1) 1.0 (0.5 to 1.7) 1.5 (0.5 to 4.3) 1.2 (0.6 to 2.3) 2.3 0.9 1.0 0.2 0.7 0.6 2.2 2.2 0.8 0.8
(1.0 (0.3 (0.5 (0.1 (0.4 (0.2 (0.9 (1.3 (0.4 (0.4
to to to to to to to to to to
Adjusted Analysis* P
HR (95% CI)
5.0) 2.5) 1.7) 0.6) 1.2) 1.7) 5.2) 3.8) 1.5) 1.8)
0.06 0.90 0.93 ,0.001 0.20 0.33 0.10 0.008 0.45 0.54
aHR (95% CI)
P
1.7 (0.6 to 4.9) 1.3 (0.6 to 2.6)
0.04 0.59 0.02 0.03 0.13 0.009 0.60 0.69 0.39 0.98 0.90 0.90 0.06 0.62 0.03 0.17 0.49 0.06 0.90 0.97 0.65 0.85 0.56 0.99 0.50 0.30 0.49
3.9 1.0 1.1 0.1 0.6 0.6 2.2 2.2 0.8 1.0
0.001 0.99 0.67 0.001 0.08 0.26 0.09 0.007 0.54 0.90
1.3 (0.5 to 3.2) 0.4 (0.2 to 0.9) 1.9 (0.8 to 4.4) 2.3 (1.2 to 4.2) 0.7 (0.2 to 3.1) 1.3 (0.7 to 2.4) 0.9 (0.1 to 6.4) 0.9 (0.3 to 3.0) 1.3 (0.5 to 3.5) 0.4 (0.2 to 0.9) 1.5 (0.5 to 4.4) 1.9 (1.0 to 3.8) 1.0 (0.3 to 3.1) 1.1 (0.6 to 2.0) 0.6 (0.2 to 2.9) 1.0 (0.6 to 1.8)
(1.7 to 9.0) (0.4 to 2.8) (0.6 to 2.1) (0.05 to 0.5) (0.3 to 1.1) (0.2 to 1.6) (0.9 to 5.6) (1.2 to 3.8) (0.5 to 1.6) (0.4 to 2.1)
HR, hazard ratio; NVP, nevirapine; FTC, emtricitabine; IDV/r, ritonavir-boosted indinavir. *Adjusted for age, triglycerides, and time-updated BMI.
DISCUSSION In this large cohort of HIV-infected patients in Thailand, we found a 5.0 per 1000 PYFU incidence rate of diabetes. This rate is similar to those reported from the Data Collection on Adverse events of Anti-HIV Drugs study (5.7 per 1000 PYFU)13 and the Swiss HIV Cohort Study (4.4 per 1000 PYFU)14 but much lower than those reported from the US Multicenter AIDS Cohort Study (47 per 1000 PYFU),15 the US Women’s Interagency HIV Study (17 per 1000 PYFU in the oldest study and at least 25 per 1000 PYFU, depending on the ART, in the most recent study),16,17 a cohort in Italy Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
(20.6 per 1000 PYFU),18 the French APROCO-COPILOTE cohort (14.1 per 1000 PYFU),19 and a case–control study in Taiwan (13.1 per 1000 PYFU).20 To our knowledge, this is the first study to report the incidence of new-onset diabetes in a large cohort of HIV-infected patients on ART in southeast Asia. This cohort has a large proportion of women because it was primarily enrolling women who previously participated in clinical trials for the HIV prevention of mother-to-child transmission. The high proportion of women (76%), the relatively young age of the patients (median age of 37.2 years in men and 32.4 years in women), and the low BMI values www.jaids.com |
457
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Riyaten et al
J Acquir Immune Defic Syndr Volume 69, Number 4, August 1, 2015
may explain the lower incidence rate observed in our cohort. Other differences between the studies may also contribute to different incidence rates, including genetic background, risk factors usually associated with mode of HIV acquisition, stage of HIV disease, type of ART received, drug adherence, nutritional and behavioral factors, and definition of diabetes. Our incidence rate in HIV-infected patients can be compared with that in the general population in Thailand. Two cohort studies have reported diabetes incidence rates in Bangkok, one in professional and office workers in 43 establishments, the other one in employees of a large university hospital, all aged 35–60 years.21,22 These rates were 11.4 and 7.8 per 1000 PYFU, respectively, compared with 7.8 in our cohort in the same age category (data not shown). Of note, the mean age and BMI in these studies were higher than those in our cohort. The contribution of the age and BMI to the risk of diabetes is well known to be major and was confirmed in our study. Taking into account the role of age and BMI, our data are compatible with other reports showing a higher risk in the HIV population. In univariate analyses, older age, higher triglycerides, and time-updated BMI were associated with a higher risk of diabetes. These risk factors are well known in both HIVinfected and HIV-uninfected populations, and our findings are consistent with other studies.13,14,19,23,24 In multivariate analyses adjusting for age, triglycerides, and time-updated BMI, cumulative exposure $1 year to zidovudine and exposure to ddI + d4T or ZDV + 3TC + EFV were associated with a higher risk of diabetes. The strong association between some nucleosides and the risk of diabetes is consistent with previous studies and might be explained by the role of these drugs in the development of insulin resistance.13,14,17–20 The subanalysis conducted in patients exposed, exclusively and for at least 2 years, to 1 of the 4 NRTI-containing regimens showed also that zidovudine and ddI + d4T were significantly associated with a higher risk of diabetes. We did not observe an association between PI use and risk of diabetes, but the number of patients on PIs was low. The association between PIs and the risk of diabetes have yielded conflicting results: most studies found an increased risk associated with some PIs such as indinavir,14–16,18–20 others reported no association,17,23 and the D:A:D study reported a decreased risk with ritonavir.13 A challenge in this cohort study is that the analysis of the association between drugs and the risk of diabetes cannot compare exposed vs. truly unexposed patients because all patients were on ART. Therefore, the results of the analysis are only valid for patients on different combinations, and we cannot determine if patients on TDF are at a higher risk of diabetes than those who would not have received nucleos(t) ides. This remains an issue to be investigated, considering that more and more patients will start TDF-containing regimens at any level of CD4 following recent WHO recommendations.25 Data from cohorts in Western countries, where NRTI-sparing regimens are increasingly used, may provide information on this. This study has other limitations. Our study population is predominantly composed of relatively young women and
thus may not be representative of the adult population living with HIV in Thailand and southeast Asia. It is unlikely that other potential confounders such as HIV-related inflammation or nutritional factors that were not studied in this cohort guided the choice of the treatments. Also, some data at baseline were not measured and were imputed for the Cox proportional hazards analyses. Finally, the number of patients with new-onset diabetes was too small to use time-updated cumulative variables to assess the association between ARVcs and the risk of diabetes, as this was done for ARVs. In summary, the incidence rate of diabetes in this lean and predominantly young, female population was low as compared with the rates found in other cohorts and to the general adult population in Thailand. The use of zidovudine, as well as the combinations ddI + d4T and ZDV + 3TC + EFV, was strongly associated with an increased risk of diabetes. Our results provide further support for the current WHO guidelines recommending phasing out stavudine and didanosine, starting TDF-containing first-line regimens, and preserving zidovudine for the second line.
458
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
| www.jaids.com
ACKNOWLEDGMENTS The authors thank all the patients who participated in the PHPT cohort study. They are also grateful to Nirattiya Jaisieng, Suriyan Tanasri, Sanuphong Chailoet, Kanchaya Yoddee and Kanchana Than-in-at, who performed data management, Rukchanok Peongjakta, who reviewed the study population, and Linda Harrison, Nontiya Homkham, Patumrat Sripan and Rapeepan Suaysod, who provided statistical advice. They also thank the site principal investigators: Nakornping: Prattana Leenasirimakul; Prapokklao: Malee Techapornroong and Sirinee Wipavakul; Chonburi: Chureeratana Bowonwatanuwong; Chiangrai Prachanukroh: Pacharee Kantipong; Rayong: Sukit Banchongkit and Sriprapar Ariyadej; Mae Chan: Surachai Piyaworawong and Sudanee Buranabanjasatean; Samutsakhon: Apichat Chutanunta; Phayao Provincial: Guttiga Halue; Bhumibol Adulyadej: Sinart Prommas; Hat Yai: Ampaipith Nilmanat; Sanpatong: Virat Klinbuayaem; Lamphun: Nuananong Luekamlung; Mahasarakham: Chalongchai Thundee and Suchart Thongpaen; Doi Saket: Preecha Sirichithaporn; Buddhachinaraj: Somboon Tansuphasawasdikul; Nong Khai: Naruepon Yutthakasemsunt and Sukunya Krabkraikaew; Maharat Nakhon Ratchasima: Rittha Lertkoonalak and Yongyuth Jongjirawisan; Bhuddasothorn: Pakorn Wittayapraparat; Ratchaburi: Wanna Sariyacheva and Pensiriwan Sang-a-gad; Lampang: Panita Pathipvanich; Mae on: Nopporn Pattanapornpun; Chiang Kham: Yuwadee Buranawanitchakorn and Chaiwat Putiyanun; San Sai: Worawut Cowatcharagul; Mae Sai: Sura Kunkongkapan, Rattakarn Paramee, and Sirisak Nanta; Regional Health Promotion Centre 6, Khon Kaen: Narong Winiyakul and Kraisorn Vivatpatanakul; Samutprakarn: Naree Eiamsirikit; Nakhonpathom: Rapat Pittayanon; Phan: Sivaporn Jungpichanvanich and Sookchai Theansavettrakul; Pranangklao: Sripetcharat Mekviwattanawong, Paiboon Lucksanapisitkul,
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Acquir Immune Defic Syndr Volume 69, Number 4, August 1, 2015
and Surachai Pipatnakulchai; Khon Kaen: Janyaporn Ratanakosol; Nopparat Rajathanee: Jeerapat Wongchinsri; Somdej Prapinklao: Pasri Maharom, Nithi Ponganant, and Ratchanee Wirojsakulchai; Kalasin: Pramual Thaingamsilp; Queen Sirikit: Nonglug Doungtong and Nopporn Prasongmanee; Kranuan Crown Prince: Amporn Rattanaparinya and Ruangyot Thongdej; Phaholpolpayuhasaena: Patinun Chirawatthanaphan; Roi-et: Boonyong Jeerasuwannakul; Sankhampang: Narongdej Pipattanawong; Banglamung: Prateep Kanjanavikai; Chiang Dao: Airada Saipanya; and Vachira Phuket: Somnuk Chirayus. Finally, they thank Sukon Prasitwattanaseree and Kate Grudpan from Chiang Mai University. REFERENCES 1. Dube MP, Johnson DL, Currier JS, et al. Protease inhibitor-associated hyperglycaemia. Lancet. 1997;350:713–714. 2. Visnegarwala F, Krause KL, Musher DM. Severe diabetes associated with protease inhibitor therapy. Ann Intern Med. 1997;127:947. 3. Carr A, Samaras K, Burton S, et al. A syndrome of peripheral lipodystrophy, hyperlipidaemia and insulin resistance in patients receiving HIV protease inhibitors. AIDS. 1998;12:F51–F58. 4. Carr A, Samaras K, Thorisdottir A, et al. Diagnosis, prediction, and natural course of HIV-1 protease-inhibitor-associated lipodystrophy, hyperlipidaemia, and diabetes mellitus: a cohort study. Lancet. 1999; 353:2093–2099. 5. Hadigan C. Diabetes, insulin resistance, and HIV. Curr Infect Dis Rep. 2006;8:69–75. 6. Brown TT, Glesby MJ. Management of the metabolic effects of HIV and HIV drugs. Nat Rev Endocrinol. 2011;8:11–21. 7. UNAIDS. Thailand: HIV and AIDS estimates. 2013. Available at: http:// www.unaids.org/en/regionscountries/countries/thailand/. Accessed September 25, 2014. 8. Guariguata L, Whiting DR, Hambleton I, et al. Global estimates of diabetes prevalence for 2013 and projections for 2035. Diabetes Res Clin Pract. 2014;103:137–149. 9. American Diabetes Association. Standards of medical care in diabetes. Diabetes Care. 2012;35(suppl 1):S11–S63. 10. Centers for Disease Control and Prevention. 1993 revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR Recomm Rep. 2012;41:1–19.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Diabetes and Antiretrovirals in Thailand
11. Grambsch PM, Therneau TM. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika. 1994;69:239–241. 12. Little RJA, Rubin DB. Statistical Analysis With Missing Data. 2nd ed. New York, NY: John Wiley & Sons; 2002. 13. De Wit S, Sabin CA, Weber R, et al. Incidence and risk factors for new-onset diabetes in HIV-infected patients: the data collection on adverse events of anti-HIV drugs (D:A:D) study. Diabetes Care. 2008; 31:1224–1229. 14. Ledergerber B, Furrer H, Rickenbach M, et al. Factors associated with the incidence of type 2 diabetes mellitus in HIV-infected participants in the Swiss HIV Cohort Study. Clin Infect Dis. 2007;45:111–119. 15. Brown TT, Cole SR, Li X, et al. Antiretroviral therapy and the prevalence and incidence of diabetes mellitus in the multicenter AIDS cohort study. Arch Intern Med. 2005;165:1179–1184. 16. Justman JE, Benning L, Danoff A, et al. Protease inhibitor use and the incidence of diabetes mellitus in a large cohort of HIV-infected women. J Acquir Immune Defic Syndr. 2003;32:298–302. 17. Tien PC, Schneider MF, Cole SR, et al. Antiretroviral therapy exposure and incidence of diabetes mellitus in the Women’s Interagency HIV Study. AIDS. 2007;21:1739–1745. 18. Brambilla AM, Novati R, Calori G, et al. Stavudine or indinavircontaining regimens are associated with an increased risk of diabetes mellitus in HIV-infected individuals. AIDS. 2003;17:1993–1995. 19. Capeau J, Bouteloup V, Katlama C, et al. Ten-year diabetes incidence in 1046 HIV-infected patients started on a combination antiretroviral treatment. AIDS. 2012;26:303–314. 20. Lo YC, Chen MY, Sheng WH, et al. Risk factors for incident diabetes mellitus among HIV-infected patients receiving combination antiretroviral therapy in Taiwan: a case-control study. HIV Med. 2009;10:302–309. 21. Jiamjarasrangsi W, Aekplakorn W. Incidence and predictors of type 2 diabetes among professional and office workers in Bangkok, Thailand. J Med Assoc Thai. 2005;88:1896–1904. 22. Jiamjarasrangsi W, Lohsoonthorn V, Lertmaharit S, et al. Incidence and predictors of abnormal fasting plasma glucose among the university hospital employees in Thailand. Diabetes Res Clin Pract. 2008;79:343–349. 23. Butt AA, McGinnis K, Rodriguez-Barradas MC, et al. HIV infection and the risk of diabetes mellitus. AIDS. 2009;23:1227–1234. 24. Hughes CA, Cashin RP, Eurich DT, et al. Risk factors for new-onset diabetes mellitus in patients receiving protease inhibitor therapy. Can J Infect Dis Med Microbiol. 2005;16:230–232. 25. WHO. Consolidated Guidelines on the Use of Antiretroviral Drugs for Treating and Preventing HIV Infection: Recommendations for a Public Health Approach. Geneva, MD: WHO; 2013.
www.jaids.com |
459
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
New-Onset Diabetes and Antiretroviral Treatments in HIV-Infected Adults in Thailand Prakit Riyaten, BSc,*† Nicolas Salvadori, MSc,†‡ Patrinee Traisathit, PhD,* Nicole Ngo-Giang-Huong, PhD,†‡§ Tim R. Cressey, PhD,†‡§ Prattana Leenasirimakul, MD,k Malee Techapornroong, MD,¶ Chureeratana Bowonwatanuwong, MD,# Pacharee Kantipong, MD,** Ampaipith Nilmanat, MD, BPA,†† Naruepon Yutthakasemsunt, MD,‡‡ Apichat Chutanunta, MD,§§ Suchart Thongpaen, MD,kk Virat Klinbuayaem, MD,¶¶ Luc Decker, PhD,†‡ Sophie Le Cœur, PhD,†‡## Marc Lallemant, MD, MSc,†‡§ Jacqueline Capeau, MD, PhD,*** Jean-Yves Mary, PhD,††† and Gonzague Jourdain, MD, PhD†‡§
Background: Use of several antiretrovirals (ARVs) has been shown to be associated with a higher risk of diabetes in HIV-infected adults. We estimated the incidence of new-onset diabetes and assessed the association between individual ARVs and ARV combinations, and diabetes in a large cohort in Thailand.
Methods: We selected all HIV-1–infected, nondiabetic, antiretroviral-naive adults enrolled in the Program for HIV Prevention and Treatment cohort (NCT00433030) between January 2000 and December 2011. Diabetes was defined as confirmed fasting plasma
Received for publication December 23, 2014; accepted February 18, 2015. From the *Department of Statistics, Faculty of Science, Chiang Mai University, Chiang Mai, Thailand; †Institut de recherche pour le développement (IRD) UMI 174-PHPT, Chiang Mai, Thailand; ‡Department of Medical Technology, Faculty of Associated Medical Sciences, Chiang Mai University, Chiang Mai, Thailand; §Department of Immunology and Infectious Diseases, Harvard School of Public Health, Boston, MA; kNakornping Hospital, Chiang Mai, Thailand; ¶Prapokklao Hospital, Chantaburi, Thailand; #Chonburi Hospital, Chonburi, Thailand; **Chiangrai Prachanukroh Hospital, Chiang Rai, Thailand; ††Hat Yai Hospital, Songkla, Thailand; ‡‡Nong Khai Hospital, Nong Khai, Thailand; §§Samutsakhon Hospital, Samutsakhon, Thailand; kkMahasarakham Hospital, Mahasarakham, Thailand; ¶¶Sanpatong Hospital, Chiang Mai, Thailand; ##Institut National d’Etudes Démographiques, Paris, France; ***Sorbonne Universités, INSERM UPMC UMR_S938, APHP, Univ Paris 06, Paris, France; and †††INSERM U1153, Team ECSTRA, Université Paris Diderot—Paris 7, Hôpital Saint-Louis, Paris, France. Supported by the Global Fund to Fight AIDS, Malaria and Tuberculosis (PR-A-N-008); Oxfam Great Britain, Thailand (THAA51); Ministry of Public Health, Thailand; and Institut de recherche pour le développement (IRD), France. P.R. received support scholarships from IRD UMI 174-PHPT; the Graduate School, Faculty of Science and the Center of Excellence for Innovation in Analytical Science and Technology (I-ANALY-S-T), Chiang Mai University, Thailand. The results of this study were partially presented as a poster at the 20th International AIDS Conference, July 20–25, 2014, Melbourne, Australia (MOPE182). The authors have no conflicts of interest to disclose. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.jaids.com). Correspondence to: Gonzague Jourdain, MD, PhD, 187/10, Changklan Road, Changklan, Muang, Chiang Mai 50100, Thailand (e-mail: [email protected]). Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
glucose $126 mg/dL or random plasma glucose $200 mg/dL. Incidence was the number of cases divided by the total number of person-years of follow-up. Association between ARVs and ARV combinations, and new-onset diabetes was assessed using Cox proportional hazards models.
Results: Overall, 1594 HIV-infected patients (76% female) were included. Median age at antiretroviral therapy initiation was 32.5 years. The incidence rate of diabetes was 5.0 per 1000 person-years of followup (95% confidence interval: 3.8 to 6.6) (53 cases). In analyses adjusted for potential confounders, exposure to stavudine + didanosine [adjusted hazard ratio (aHR) = 3.9; P = 0.001] and cumulative exposure $1 year to zidovudine (aHR = 2.3 vs. no exposure; P = 0.009) were associated with a higher risk of diabetes. Conversely, cumulative exposure $1 year to tenofovir (aHR = 0.4 vs. no exposure; P = 0.02) and emtricitabine (aHR = 0.4 vs. no exposure; P = 0.03) were associated with a lower risk. Conclusions: The incidence of diabetes in this predominantly female, young, lean population was relatively low. Although stavudine and didanosine have now been phased out in most antiretroviral therapy programs, our analysis suggests a higher risk of diabetes with zidovudine, frequently prescribed today in resource-limited settings. Key Words: HIV, diabetes, incidence, NRTIs, tenofovir, zidovudine (J Acquir Immune Defic Syndr 2015;69:453–459)
INTRODUCTION Antiretroviral therapy (ART) provides clear health and quality of life benefits to people living with HIV, but longterm adverse metabolic effects are of concern. Some antiretrovirals (ARVs), especially several protease inhibitors (PIs) and thymidine analogue nucleoside reverse transcriptase inhibitors (NRTIs), have been associated with the development of insulin resistance, drug-induced metabolic syndrome and diabetes.1–6 However, previous reports have not directly compared the risk of diabetes between NRTI-containing regimens. Moreover, the incidence of diabetes has not been evaluated in southeast Asian HIV-infected populations. HIV and diabetes are 2 chronic diseases with severe public health burdens in Thailand. In 2013, the estimated HIV
J Acquir Immune Defic Syndr Volume 69, Number 4, August 1, 2015
www.jaids.com |
453
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Acquir Immune Defic Syndr Volume 69, Number 4, August 1, 2015
Riyaten et al
prevalence was 1.1% in adults (age, 15–49 years),7 and the prevalence of diabetes in the adult population was 6.4%.8 Data on type 2 diabetes in HIV-infected patients in Thailand population are lacking. Our objective was to estimate the incidence of new-onset diabetes and assess the association between exposure and duration of exposure to ARVs and ARV combinations (ARVcs), and diabetes in a large HIV-infected adult cohort in Thailand.
METHODS Study Population The Program for HIV Prevention and Treatment (PHPT) cohort is an ongoing, observational, prospective, multicenter study initiated in 1999, which has enrolled more than 2400 HIVinfected patients from 50 public hospitals throughout Thailand (NCT00433030). HIV-1 infected adults (age, $18 years) who initiated ART between January 1, 2000, and December 31, 2011, nondiabetic according to chart review, and with at least 2 glucose measurements after ART initiation were included. ARTexperienced patients referred to PHPT sites and enrolled in the cohort were excluded. However, patients who received ARVs at PHPT sites solely for the prevention of mother-to-child transmission of HIV were not excluded.
Definition of Diabetes Following the American Diabetes Association guidelines, diabetes was defined as either a confirmed fasting plasma glucose $126 mg/dL or a confirmed random plasma glucose $200 mg/dL.9 The date and time of the last meal were systematically recorded at the time of blood draw. Blood glucose was considered as fasting if the last meal was more than 7 hours before the blood draw. Glycosylated hemoglobin was not part of the data collected in this cohort.
Baseline and Follow-up Data Baseline characteristics were retrieved from the closest assessment between 1 year before and 2 weeks after ART initiation: gender, date of birth, weight, height, Centers for Disease Control and Prevention (CDC) HIV classification,10 hepatitis B surface antigen and hepatitis C antibody, CD4 cell count, plasma HIV-1 RNA levels, glucose, triglycerides, and total cholesterol. Every 6 months, weight, CD4 cell count, HIV RNA load, glucose, triglycerides, and total cholesterol were measured as part of standard of care in this cohort. Clinical data collected on case report forms and copies of laboratory slips from the sites were forwarded to the PHPT Data Center for double data entry and routine data management.
censored at the date of death or last visit in case of voluntary withdrawal from the study or loss to follow-up, defined as an absence of visit or contact for at least 9 months. The overall and stratified incidence rates of diabetes were calculated as the number of new-onset diabetes divided by the total number of person-years of follow-up (PYFU). Confidence intervals (CIs) of the incidence rates were based on the Poisson distribution. The following potential baseline confounders in the association between exposure to ARVs and ARVcs, and new-onset diabetes were assessed using univariate and multivariate Cox proportional hazards models: gender, age, triglycerides, total cholesterol, HIV RNA load, hepatitis B surface antigen, hepatitis C antibody, CDC stage, calendar year of enrollment, time-updated body mass index (BMI) and time-updated CD4 cell count. All continuous variables were dichotomized according to the median values except BMI (categories: ,25 and $25 kg/m2) and CD4 count (,200 and $200 cells/mm3). Variables with a P value ,0.25 in univariate analyses were included in a multivariate model. A backward stepwise regression analysis was then performed to identify significant potential confounders in the multivariate analysis. The association between ARVs and ARVcs, and newonset diabetes was assessed using unadjusted and adjusted Cox proportional hazards models. Adjusted models included all significant potential confounders in the multivariate analysis. ARVcs were entered in the models as dichotomous variables (ever exposed or never exposed) and ARVs as timeupdated variables categorized in 3 groups based on the cumulative duration of exposure (no exposure, less than 1 year, or more than 1 year). The ARVs and ARVcs selected for the analysis were those received by at least 5% of the patients. Overall P values were based on partial likelihood ratio tests using Breslow method for ties, and specific P values on Wald tests. We tested the proportional hazards assumption of each model based on Schoenfeld residuals.11 Missing values at baseline were imputed using multivariate imputation by chained equations, based on the predictive mean matching approach for continuous variables and logistic regression for binary variables.12
Subanalysis
Study population characteristics were presented as medians and interquartile ranges (IQRs) for continuous variables and as counts and percentages for categorical variables. The follow-up period was defined as the period between baseline and the date of first diagnosis of diabetes, date of death, or of last clinic visit, whichever occurred first. Observations were
Because a significant proportion of patients were exposed sequentially to various antiretroviral drugs, we conducted a subanalysis restricted to those patients who received, exclusively and for a longer time than in the main analysis (at least 2 years), tenofovir disoproxil fumarate (TDF), zidovudine (ZDV), stavudine (d4T), or didanosine + stavudine (ddI + d4T) as part of their first-line regimen. The association between these 4 NRTI-containing regimens and new onset diabetes was assessed and compared using a Cox proportional hazards model adjusted for all variables significantly different across the 4 groups (tested using Fisher exact tests). All reported P values are 2 sided and P values ,0.05 were considered statistically significant. All analyses were performed using Stata software version 10.1 (Stata Corp, College Station, TX).
454
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Statistical Analysis
| www.jaids.com
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Acquir Immune Defic Syndr Volume 69, Number 4, August 1, 2015
Diabetes and Antiretrovirals in Thailand
TABLE 1. Baseline Characteristics of the Study Population According to the Occurrence or Not of Diabetes During Follow-up and Incidence Rates Diabetes Characteristics Overall Gender Female Male Age (years) ,30 $30 BMI (kg/m2) ,25 $25 Hepatitis B surface antigen (n = 1538) Positive Negative Hepatitis C antibody (n = 1521) Positive Negative Triglycerides (mg/dL) (n = 1023) ,140 $140 Total cholesterol (mg/dL) (n = 1026) ,160 $160 HIV RNA load (log10 copies/mL) (n = 1485) ,5 $5 CD4 count (cells/mm3) (n = 1567) ,200 $200 CDC stage A B or C Calendar year of enrollment 2000–2005 2006–2011
Incidence of Diabetes
No Diabetes, n (%)
n (%)
PYFU
IR
95% CI
1541 (97)
53 (3)
10,507
5.0
3.8 to 6.6
1181 (78) 360 (22)
37 (70) 16 (30)
8113 2394
4.6 6.7
3.2 to 6.3 3.8 to 10.9
569 (37) 972 (63)
9 (17) 44 (83)
3843 6664
2.3 6.6
1.1 to 4.4 4.8 to 8.9
1347 (87) 194 (13)
37 (70) 16 (30)
9345 1162
4.0 13.8
2.8 to 5.5 7.9 to 22.4
76 (5) 1411 (95)
1 (1) 50 (99)
500 9597
2.0 5.2
0.1 to 11.1 3.9 to 6.9
57 (4) 1413 (96)
3 (6) 48 (94)
397 9573
7.6 5.0
1.6 to 22.1 4.0 to 6.6
541 (54) 453 (46)
10 (34) 19 (66)
3303 2531
3.0 7.5
1.5 to 5.6 4.5 to 11.7
519 (52) 477 (48)
15 (50) 15 (50)
3199 2656
4.7 5.6
2.6 to 7.7 3.2 to 9.3
878 (61) 563 (39)
26 (59) 18 (41)
5823 3792
4.5 4.7
2.9 to 6.5 2.8 to 7.5
1082 (71) 434 (29)
44 (86) 7 (14)
7777 2579
5.7 2.7
4.1 to 7.6 1.1 to 5.6
946 (61) 595 (39)
28 (53) 25 (47)
6353 4154
4.4 6.0
2.9 to 6.4 3.9 to 8.9
822 (53) 719 (47)
33 (62) 20 (38)
6656 3851
5.0 5.2
3.4 to 7.0 3.2 to 8.0
Comparison tests are listed in Table 2 IR, incidence rate per 1000 person-years of follow-up.
Ethical Considerations All participants provided written informed consent before participation in this study. The PHPT cohort study protocol and amendments received ethical clearance from the Ethics Committees of the Thai Ministry of Public Health and the Faculty of Associated Medical Sciences, Chiang Mai University, Thailand.
RESULTS
The remaining 1594 patients were included in this analysis. There were no patients with diabetes at entry in the cohort.
Baseline Characteristics Of the 1594 participants, 1218 (76%) were female. At baseline, the median age was 32.5 years (IQR: 28.2–37.7), BMI was 20.5 kg/m2 (IQR: 18.6–23.0), HIV RNA load was 4.8 log10 copies per milliliter (IQR: 4.2–5.2), and CD4 count was 137 cells per cubic millimeter (IQR: 67–211).
Study Population A total of 2092 HIV-infected adults on ART were enrolled between January 2000 and December 2011. Of these patients, 203 received ART before enrollment and 295 had no or only 1 glucose measurement during the follow-up period. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Follow-up The median duration of follow-up was 6.9 years (IQR: 4.9–8.2). During follow-up, 91% of the patients were exposed www.jaids.com |
455
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Acquir Immune Defic Syndr Volume 69, Number 4, August 1, 2015
Riyaten et al
to lamivudine (3TC), 62% to zidovudine, 58% to TDF, 48% to emtricitabine, 42% to stavudine, 14% to didanosine, 58% to efavirenz (EFV), 50% to nevirapine and 33% to ritonavir (see Table, Supplemental Digital Content 1, http://links.lww.com/QAI/A673, which describes the main ARVs and ARVcs administered at ART initiation and during follow-up). A total of 65 (4%) patients died, 158 (10%) were lost to follow-up, and 246 (15%) voluntarily withdrew from the study, most often because they moved to another province. On treatment, 1287 of 1475 patients (87%) had a viral load ,400 copies per milliliter at 1 year after ART initiation, 1240 of 1318 (94%) at 3 years, and 1117 of 1162 (96%) at 5 years.
P = 0.009) were associated with a higher risk of diabetes. Conversely, exposure to TDF + 3TC + EFV (aHR = 0.1; 95% CI: 0.05 to 0.5; P = 0.001) and cumulative exposure $1 year to TDF (aHR = 0.4 vs. no exposure; 95% CI: 0.2 to 0.9; P = 0.02) and emtricitabine (aHR = 0.4 vs. no exposure; 95% CI: 0.2 to 0.9; P = 0.03) were associated with a lower risk of diabetes (Table 3; see Table in Supplemental Digital Content 1, http://links.lww.com/QAI/A673, which describes the main ARVs and ARVcs administered at ART initiation and during follow-up).
Subanalysis
Incidence of Diabetes A total of 53 patients were diagnosed with diabetes over 10,507 PYFU, yielding an overall incidence rate of 5.0 per 1000 PYFU (95% CI: 3.8 to 6.6) (Table 1). Incidence rates were 6.7 per 1000 PYFU in men (95% CI: 3.8 to 10.9) and 4.6 per 1000 PYFU in women (95% CI: 3.2 to 6.3).
Factors Associated With New-Onset Diabetes The results of the univariate and multivariate analyses to identify non–drug-related risk factors of diabetes are presented in Table 2. In the multivariate analysis, older age, higher triglycerides, and time-updated BMI were independently associated with the risk of diabetes. However, HIVrelated characteristics (CDC stage, HIV RNA load, and time-updated CD4 cell count) were not associated with the risk of diabetes. The analysis assessing the role of ARVs and ARVcs on the risk of diabetes showed that after adjusting for age, triglycerides, and time-updated BMI, exposure to ddI + d4T [adjusted hazard ratio (aHR) = 3.9; 95% CI: 1.7 to 9.0; P = 0.001] and ZDV + 3TC + EFV (aHR = 2.2; 95% CI: 1.2 to 3.8; P = 0.007) and cumulative exposure $1 year to zidovudine (aHR = 2.3 vs. no exposure; 95% CI: 1.2 to 4.2;
A subanalysis conducted to assess and compare the association between 4 NRTI-containing regimens and new-onset diabetes showed that a total of 520 HIVinfected patients received, exclusively and for at least 2 years, TDF, zidovudine, stavudine, or ddI + d4T as part of their first-line regimen. At baseline, 63% were female, and the median age was 34.1 years (IQR: 29.5–40.1). A total of 329 (63%) patients received a first-line ART containing TDF, 28% zidovudine, and 7% stavudine (usually in addition to lamivudine or emtricitabine), and 2% received ddI + d4T. Gender, age, hepatitis B surface antigen, and hepatitis C antibody were significantly different across the 4 groups and were thus included in the adjusted model (see Table, Supplemental Digital Content 2, http://links.lww.com/QAI/A673, which presents the baseline characteristics of the subanalysis). In the adjusted analysis, patients exposed to a first-line regimen containing zidovudine (aHR = 6.6; 95% CI: 1.6 to 27.5; P = 0.01) or ddI + d4T (aHR = 73.2; 95% CI: 11.5 to 465.2; P , 0.001) were at a significantly higher risk of diabetes than those exposed to a first-line regimen containing TDF (see Table, Supplemental Digital Content 3, http://links.lww.com/QAI/A673, which evaluates the association between NRTI-containing first-line regimens and the occurrence of diabetes during follow-up).
TABLE 2. Association Between Baseline and Follow-up Characteristics, and the Occurrence of Diabetes During Follow-up Univariate Analysis Characteristics At ART initiation Male gender Age $30 yr Positive hepatitis B surface antigen Positive hepatitis C antibody Triglycerides $140 mg/dL Total cholesterol $160 mg/dL HIV RNA load $5 log10 copies/mL CDC stage B or C Calendar year of enrollment $2006 Time-updated variables CD4 count $200 cells/mm3 BMI $25 kg/m2
Diabetes
Total
HR (95% CI)
16 44 2 3 32 27 22 25 20
376 1016 79 64 709 766 613 620 739
1.5 2.9 0.8 1.4 2.1 1.2 1.2 1.4 1.0
— —
— —
Multivariate Analysis P
aHR (95% CI)
P
2.7) 6.1) 3.3) 4.5) 3.7) 2.1) 2.0) 2.3) 1.9)
0.20 0.001 0.73 0.59 0.006 0.49 0.63 0.27 0.87
— 2.7 (1.3 to 5.7) — — 1.8 (1.0 to 3.1) — — — —
— 0.007 — — 0.04 — — — —
1.1 (0.4 to 3.2) 3.1 (1.7 to 5.4)
0.80 ,0.001
— 2.7 (1.5 to 4.7)
— 0.001
(0.8 (1.4 (0.2 (0.4 (1.2 (0.7 (0.7 (0.8 (0.6
to to to to to to to to to
HR, hazard ratio. Bolded P values correspond to P,0.05 (for both Tables 2 and 3).
456
| www.jaids.com
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Acquir Immune Defic Syndr Volume 69, Number 4, August 1, 2015
Diabetes and Antiretrovirals in Thailand
TABLE 3. Association Between ARVs and ARVcs, and the Occurrence of Diabetes Unadjusted Analysis ARVs and ARVcs ARVs (time-updated exposure to) TDF ,1 yr vs. never exposed $1 yr vs. never exposed ZDV ,1 yr vs. never exposed $1 yr vs. never exposed Stavudine ,1 yr vs. never exposed $1 yr vs. never exposed Didanosine ,1 yr vs. never exposed $1 yr vs. never exposed Emtricitabine ,1 yr vs. never exposed $1 yr vs. never exposed Lamivudine ,1 yr vs. never exposed $1 yr vs. never exposed Efavirenz ,1 yr vs. never exposed $1 yr vs. never exposed Nevirapine ,1 yr vs. never exposed $1 yr vs. never exposed Ritonavir ,1 yr vs. never exposed $1 yr vs. never exposed ARVcs (ever exposed to) ddI + d4T d4T + 3TC + EFV d4T + 3TC + NVP TDF + 3TC + EFV TDF + FTC + EFV TDF + FTC + NVP TDF + FTC + IDV/r ZDV + 3TC + EFV ZDV + 3TC + NVP ZDV + 3TC + IDV/r
0.14 0.43 0.11 0.05 0.14 0.02 0.89 0.72 0.80 0.96 0.87 0.81 0.19 0.45 0.14 0.43 0.66 0.21 0.66 0.89 0.36 0.89 0.65 0.90 0.72 0.43 0.68
1.4 (0.6 to 3.5) 0.6 (0.3 to 1.2) 1.9 (0.8 to 4.3) 2.1 (1.1 to 3.9) 0.8 (0.2 to 3.3) 1.1 (0.6 to 2.0) 0.8 (0.1 to 6.1) 0.9 (0.3 to 2.8) 1.5 (0.6 to 3.9) 0.6 (0.3 to 1.2) 1.2 (0.4 to 3.6) 1.5 (0.8 to 3.0) 1.1 (0.4 to 3.3) 1.3 (0.7 to 2.4) 0.7 (0.2 to 3.1) 1.0 (0.5 to 1.7) 1.5 (0.5 to 4.3) 1.2 (0.6 to 2.3) 2.3 0.9 1.0 0.2 0.7 0.6 2.2 2.2 0.8 0.8
(1.0 (0.3 (0.5 (0.1 (0.4 (0.2 (0.9 (1.3 (0.4 (0.4
to to to to to to to to to to
Adjusted Analysis* P
HR (95% CI)
5.0) 2.5) 1.7) 0.6) 1.2) 1.7) 5.2) 3.8) 1.5) 1.8)
0.06 0.90 0.93 ,0.001 0.20 0.33 0.10 0.008 0.45 0.54
aHR (95% CI)
P
1.7 (0.6 to 4.9) 1.3 (0.6 to 2.6)
0.04 0.59 0.02 0.03 0.13 0.009 0.60 0.69 0.39 0.98 0.90 0.90 0.06 0.62 0.03 0.17 0.49 0.06 0.90 0.97 0.65 0.85 0.56 0.99 0.50 0.30 0.49
3.9 1.0 1.1 0.1 0.6 0.6 2.2 2.2 0.8 1.0
0.001 0.99 0.67 0.001 0.08 0.26 0.09 0.007 0.54 0.90
1.3 (0.5 to 3.2) 0.4 (0.2 to 0.9) 1.9 (0.8 to 4.4) 2.3 (1.2 to 4.2) 0.7 (0.2 to 3.1) 1.3 (0.7 to 2.4) 0.9 (0.1 to 6.4) 0.9 (0.3 to 3.0) 1.3 (0.5 to 3.5) 0.4 (0.2 to 0.9) 1.5 (0.5 to 4.4) 1.9 (1.0 to 3.8) 1.0 (0.3 to 3.1) 1.1 (0.6 to 2.0) 0.6 (0.2 to 2.9) 1.0 (0.6 to 1.8)
(1.7 to 9.0) (0.4 to 2.8) (0.6 to 2.1) (0.05 to 0.5) (0.3 to 1.1) (0.2 to 1.6) (0.9 to 5.6) (1.2 to 3.8) (0.5 to 1.6) (0.4 to 2.1)
HR, hazard ratio; NVP, nevirapine; FTC, emtricitabine; IDV/r, ritonavir-boosted indinavir. *Adjusted for age, triglycerides, and time-updated BMI.
DISCUSSION In this large cohort of HIV-infected patients in Thailand, we found a 5.0 per 1000 PYFU incidence rate of diabetes. This rate is similar to those reported from the Data Collection on Adverse events of Anti-HIV Drugs study (5.7 per 1000 PYFU)13 and the Swiss HIV Cohort Study (4.4 per 1000 PYFU)14 but much lower than those reported from the US Multicenter AIDS Cohort Study (47 per 1000 PYFU),15 the US Women’s Interagency HIV Study (17 per 1000 PYFU in the oldest study and at least 25 per 1000 PYFU, depending on the ART, in the most recent study),16,17 a cohort in Italy Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
(20.6 per 1000 PYFU),18 the French APROCO-COPILOTE cohort (14.1 per 1000 PYFU),19 and a case–control study in Taiwan (13.1 per 1000 PYFU).20 To our knowledge, this is the first study to report the incidence of new-onset diabetes in a large cohort of HIV-infected patients on ART in southeast Asia. This cohort has a large proportion of women because it was primarily enrolling women who previously participated in clinical trials for the HIV prevention of mother-to-child transmission. The high proportion of women (76%), the relatively young age of the patients (median age of 37.2 years in men and 32.4 years in women), and the low BMI values www.jaids.com |
457
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Riyaten et al
J Acquir Immune Defic Syndr Volume 69, Number 4, August 1, 2015
may explain the lower incidence rate observed in our cohort. Other differences between the studies may also contribute to different incidence rates, including genetic background, risk factors usually associated with mode of HIV acquisition, stage of HIV disease, type of ART received, drug adherence, nutritional and behavioral factors, and definition of diabetes. Our incidence rate in HIV-infected patients can be compared with that in the general population in Thailand. Two cohort studies have reported diabetes incidence rates in Bangkok, one in professional and office workers in 43 establishments, the other one in employees of a large university hospital, all aged 35–60 years.21,22 These rates were 11.4 and 7.8 per 1000 PYFU, respectively, compared with 7.8 in our cohort in the same age category (data not shown). Of note, the mean age and BMI in these studies were higher than those in our cohort. The contribution of the age and BMI to the risk of diabetes is well known to be major and was confirmed in our study. Taking into account the role of age and BMI, our data are compatible with other reports showing a higher risk in the HIV population. In univariate analyses, older age, higher triglycerides, and time-updated BMI were associated with a higher risk of diabetes. These risk factors are well known in both HIVinfected and HIV-uninfected populations, and our findings are consistent with other studies.13,14,19,23,24 In multivariate analyses adjusting for age, triglycerides, and time-updated BMI, cumulative exposure $1 year to zidovudine and exposure to ddI + d4T or ZDV + 3TC + EFV were associated with a higher risk of diabetes. The strong association between some nucleosides and the risk of diabetes is consistent with previous studies and might be explained by the role of these drugs in the development of insulin resistance.13,14,17–20 The subanalysis conducted in patients exposed, exclusively and for at least 2 years, to 1 of the 4 NRTI-containing regimens showed also that zidovudine and ddI + d4T were significantly associated with a higher risk of diabetes. We did not observe an association between PI use and risk of diabetes, but the number of patients on PIs was low. The association between PIs and the risk of diabetes have yielded conflicting results: most studies found an increased risk associated with some PIs such as indinavir,14–16,18–20 others reported no association,17,23 and the D:A:D study reported a decreased risk with ritonavir.13 A challenge in this cohort study is that the analysis of the association between drugs and the risk of diabetes cannot compare exposed vs. truly unexposed patients because all patients were on ART. Therefore, the results of the analysis are only valid for patients on different combinations, and we cannot determine if patients on TDF are at a higher risk of diabetes than those who would not have received nucleos(t) ides. This remains an issue to be investigated, considering that more and more patients will start TDF-containing regimens at any level of CD4 following recent WHO recommendations.25 Data from cohorts in Western countries, where NRTI-sparing regimens are increasingly used, may provide information on this. This study has other limitations. Our study population is predominantly composed of relatively young women and
thus may not be representative of the adult population living with HIV in Thailand and southeast Asia. It is unlikely that other potential confounders such as HIV-related inflammation or nutritional factors that were not studied in this cohort guided the choice of the treatments. Also, some data at baseline were not measured and were imputed for the Cox proportional hazards analyses. Finally, the number of patients with new-onset diabetes was too small to use time-updated cumulative variables to assess the association between ARVcs and the risk of diabetes, as this was done for ARVs. In summary, the incidence rate of diabetes in this lean and predominantly young, female population was low as compared with the rates found in other cohorts and to the general adult population in Thailand. The use of zidovudine, as well as the combinations ddI + d4T and ZDV + 3TC + EFV, was strongly associated with an increased risk of diabetes. Our results provide further support for the current WHO guidelines recommending phasing out stavudine and didanosine, starting TDF-containing first-line regimens, and preserving zidovudine for the second line.
458
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
| www.jaids.com
ACKNOWLEDGMENTS The authors thank all the patients who participated in the PHPT cohort study. They are also grateful to Nirattiya Jaisieng, Suriyan Tanasri, Sanuphong Chailoet, Kanchaya Yoddee and Kanchana Than-in-at, who performed data management, Rukchanok Peongjakta, who reviewed the study population, and Linda Harrison, Nontiya Homkham, Patumrat Sripan and Rapeepan Suaysod, who provided statistical advice. They also thank the site principal investigators: Nakornping: Prattana Leenasirimakul; Prapokklao: Malee Techapornroong and Sirinee Wipavakul; Chonburi: Chureeratana Bowonwatanuwong; Chiangrai Prachanukroh: Pacharee Kantipong; Rayong: Sukit Banchongkit and Sriprapar Ariyadej; Mae Chan: Surachai Piyaworawong and Sudanee Buranabanjasatean; Samutsakhon: Apichat Chutanunta; Phayao Provincial: Guttiga Halue; Bhumibol Adulyadej: Sinart Prommas; Hat Yai: Ampaipith Nilmanat; Sanpatong: Virat Klinbuayaem; Lamphun: Nuananong Luekamlung; Mahasarakham: Chalongchai Thundee and Suchart Thongpaen; Doi Saket: Preecha Sirichithaporn; Buddhachinaraj: Somboon Tansuphasawasdikul; Nong Khai: Naruepon Yutthakasemsunt and Sukunya Krabkraikaew; Maharat Nakhon Ratchasima: Rittha Lertkoonalak and Yongyuth Jongjirawisan; Bhuddasothorn: Pakorn Wittayapraparat; Ratchaburi: Wanna Sariyacheva and Pensiriwan Sang-a-gad; Lampang: Panita Pathipvanich; Mae on: Nopporn Pattanapornpun; Chiang Kham: Yuwadee Buranawanitchakorn and Chaiwat Putiyanun; San Sai: Worawut Cowatcharagul; Mae Sai: Sura Kunkongkapan, Rattakarn Paramee, and Sirisak Nanta; Regional Health Promotion Centre 6, Khon Kaen: Narong Winiyakul and Kraisorn Vivatpatanakul; Samutprakarn: Naree Eiamsirikit; Nakhonpathom: Rapat Pittayanon; Phan: Sivaporn Jungpichanvanich and Sookchai Theansavettrakul; Pranangklao: Sripetcharat Mekviwattanawong, Paiboon Lucksanapisitkul,
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
J Acquir Immune Defic Syndr Volume 69, Number 4, August 1, 2015
and Surachai Pipatnakulchai; Khon Kaen: Janyaporn Ratanakosol; Nopparat Rajathanee: Jeerapat Wongchinsri; Somdej Prapinklao: Pasri Maharom, Nithi Ponganant, and Ratchanee Wirojsakulchai; Kalasin: Pramual Thaingamsilp; Queen Sirikit: Nonglug Doungtong and Nopporn Prasongmanee; Kranuan Crown Prince: Amporn Rattanaparinya and Ruangyot Thongdej; Phaholpolpayuhasaena: Patinun Chirawatthanaphan; Roi-et: Boonyong Jeerasuwannakul; Sankhampang: Narongdej Pipattanawong; Banglamung: Prateep Kanjanavikai; Chiang Dao: Airada Saipanya; and Vachira Phuket: Somnuk Chirayus. Finally, they thank Sukon Prasitwattanaseree and Kate Grudpan from Chiang Mai University. REFERENCES 1. Dube MP, Johnson DL, Currier JS, et al. Protease inhibitor-associated hyperglycaemia. Lancet. 1997;350:713–714. 2. Visnegarwala F, Krause KL, Musher DM. Severe diabetes associated with protease inhibitor therapy. Ann Intern Med. 1997;127:947. 3. Carr A, Samaras K, Burton S, et al. A syndrome of peripheral lipodystrophy, hyperlipidaemia and insulin resistance in patients receiving HIV protease inhibitors. AIDS. 1998;12:F51–F58. 4. Carr A, Samaras K, Thorisdottir A, et al. Diagnosis, prediction, and natural course of HIV-1 protease-inhibitor-associated lipodystrophy, hyperlipidaemia, and diabetes mellitus: a cohort study. Lancet. 1999; 353:2093–2099. 5. Hadigan C. Diabetes, insulin resistance, and HIV. Curr Infect Dis Rep. 2006;8:69–75. 6. Brown TT, Glesby MJ. Management of the metabolic effects of HIV and HIV drugs. Nat Rev Endocrinol. 2011;8:11–21. 7. UNAIDS. Thailand: HIV and AIDS estimates. 2013. Available at: http:// www.unaids.org/en/regionscountries/countries/thailand/. Accessed September 25, 2014. 8. Guariguata L, Whiting DR, Hambleton I, et al. Global estimates of diabetes prevalence for 2013 and projections for 2035. Diabetes Res Clin Pract. 2014;103:137–149. 9. American Diabetes Association. Standards of medical care in diabetes. Diabetes Care. 2012;35(suppl 1):S11–S63. 10. Centers for Disease Control and Prevention. 1993 revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR Recomm Rep. 2012;41:1–19.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Diabetes and Antiretrovirals in Thailand
11. Grambsch PM, Therneau TM. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika. 1994;69:239–241. 12. Little RJA, Rubin DB. Statistical Analysis With Missing Data. 2nd ed. New York, NY: John Wiley & Sons; 2002. 13. De Wit S, Sabin CA, Weber R, et al. Incidence and risk factors for new-onset diabetes in HIV-infected patients: the data collection on adverse events of anti-HIV drugs (D:A:D) study. Diabetes Care. 2008; 31:1224–1229. 14. Ledergerber B, Furrer H, Rickenbach M, et al. Factors associated with the incidence of type 2 diabetes mellitus in HIV-infected participants in the Swiss HIV Cohort Study. Clin Infect Dis. 2007;45:111–119. 15. Brown TT, Cole SR, Li X, et al. Antiretroviral therapy and the prevalence and incidence of diabetes mellitus in the multicenter AIDS cohort study. Arch Intern Med. 2005;165:1179–1184. 16. Justman JE, Benning L, Danoff A, et al. Protease inhibitor use and the incidence of diabetes mellitus in a large cohort of HIV-infected women. J Acquir Immune Defic Syndr. 2003;32:298–302. 17. Tien PC, Schneider MF, Cole SR, et al. Antiretroviral therapy exposure and incidence of diabetes mellitus in the Women’s Interagency HIV Study. AIDS. 2007;21:1739–1745. 18. Brambilla AM, Novati R, Calori G, et al. Stavudine or indinavircontaining regimens are associated with an increased risk of diabetes mellitus in HIV-infected individuals. AIDS. 2003;17:1993–1995. 19. Capeau J, Bouteloup V, Katlama C, et al. Ten-year diabetes incidence in 1046 HIV-infected patients started on a combination antiretroviral treatment. AIDS. 2012;26:303–314. 20. Lo YC, Chen MY, Sheng WH, et al. Risk factors for incident diabetes mellitus among HIV-infected patients receiving combination antiretroviral therapy in Taiwan: a case-control study. HIV Med. 2009;10:302–309. 21. Jiamjarasrangsi W, Aekplakorn W. Incidence and predictors of type 2 diabetes among professional and office workers in Bangkok, Thailand. J Med Assoc Thai. 2005;88:1896–1904. 22. Jiamjarasrangsi W, Lohsoonthorn V, Lertmaharit S, et al. Incidence and predictors of abnormal fasting plasma glucose among the university hospital employees in Thailand. Diabetes Res Clin Pract. 2008;79:343–349. 23. Butt AA, McGinnis K, Rodriguez-Barradas MC, et al. HIV infection and the risk of diabetes mellitus. AIDS. 2009;23:1227–1234. 24. Hughes CA, Cashin RP, Eurich DT, et al. Risk factors for new-onset diabetes mellitus in patients receiving protease inhibitor therapy. Can J Infect Dis Med Microbiol. 2005;16:230–232. 25. WHO. Consolidated Guidelines on the Use of Antiretroviral Drugs for Treating and Preventing HIV Infection: Recommendations for a Public Health Approach. Geneva, MD: WHO; 2013.
www.jaids.com |
459
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
