The occurrence of Simpson's paradox if site-level effect was ignored in the TREAT Asia HIV Observational Database

Kerr, Stephen J.; Chandrasekaran, Ezhilarasi; Huelgas, Aizobelle; Taecharoenkul, Sineenart; Teeraananchai, Sirinya; Wu, Gang; Ly, Penh Sun; Khol, V.; Zhang, F.J.; Zhao, Hong; Han, Nannan; Lee, M.P.; Li, P.C.K.; Lam, Wilson; Chan, Ya-Hsuan; Kumarasamy, N.; Saghayam, Suneeta; Ezhilarasi, C; Pujari, Sanjay; Joshi, Kedar; Gaikwad, Sunil; Chitalikar, Abhishek; Merati, Tuti Parwati; Wirawan, Dewa Nyoman; Yuliana, Fina; Yunihastuti, Evy; Imran, Darma; Widhani, Alvina; Oka, Shiro; Tanuma, Junko; Nishijima, Takeshi; Choi, Jun Yong; Na, S.; Kim, J.M.; Sim, Benedict Lh; Gani, Yasmin; David, R Jacobs; Kamarulzaman, Adeeba; Omar, S.F. Syed; Ponnampalavanar, Sasheela; Azwa, Iskandar; Mustafa, Mahiran; Nordin, Nina; Ditangco, Rossana; Uy, E; Bantique, R.; Wong, Wing‐Wai; Ku, Wen-Yuan; Wu, Po Chang; Ng, Oon Tek; Lim, Poh Lian; Lee, L.S.; Martínez-Vega, Ruth Aralí; Phanuphak, Praphan; Ruxrungtham, Kiat; Avihingsanon, Anchalee; Chusut, P.; Kiertiburanakul, Sasisopin; Sungkanuparph, Somnuek; Chumla, L.; Sanmeema, N.; Chaiwarith, Romanee; Sirisanthana, Thira; Kotarathititum, Wilai; Praparattanapan, Jutarat; Kantipong, Pacharee; Kambua, P.; Ratanasuwan, Winai; Sriondee, R.; Nguyen, Kinh Van; Bui, Virna; Nguyen, Dung Thi; Pham, Thuy Thanh; Cuong, DD; Ha, Hye-Sook; Sohn, Annette H.; Durier, Nicolas; Petersen, Boondarika; Cooper, David A.; Law, Matthew; Jiamsakul, Awachana; Boettiger, David

doi:10.1016/j.jclinepi.2016.01.030

Cited by 4 publications

(3 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We believe this is appropriate as sites without TB cases, even in the same country, could have quite different patient populations, thus potentially biasing risk factor models. 14,20 The power of analyses, which depends on the number of TB cases, is largely unaffected by this approach. A sensitivity analysis including participants at sites with no TB cases from the same countries as the study sites was performed (data not shown), and a lower TB incidence rate of 2.5 per 100 person-years was observed.…”

Section: Discussionmentioning

confidence: 99%

Factors Associated With and Characteristic of HIV/Tuberculosis Co-Infection: A Retrospective Analysis of SECOND-LINE Clinical Trial Participants

Henry

Jiamsakul

Law

et al. 2021

JAIDS Journal of Acquired Immune Deficiency Syndromes

Self Cite

View full text Add to dashboard Cite

Background: Tuberculosis (TB) is a common infection in people living with HIV. However, the risk factors for HIV/TB co-infection in second-line HIV therapy are poorly understood. We aimed to determine the incidence and risk factors for TB co-infection in SECOND-LINE, an international randomized clinical trial of second-line HIV therapy.Methods: We did a cohort analysis of TB cases in SECOND-LINE. TB cases included any clinical or laboratory-confirmed diagnoses and/or commencement of treatment for TB after randomization. Baseline factors associated with TB were analyzed using Cox regression stratified by site.Results: TB cases occurred at sites in Argentina, India, Malaysia, Nigeria, South Africa, and Thailand, in a cohort of 355 of the 541 SECOND-LINE participants. Overall, 20 cases of TB occurred, an incidence rate of 3.4 per 100 person-years (95% CI: 2.1 to 5.1). Increased TB risk was associated with a low CD4 + -cell count (#200 cells/mL), high viral load (.200 copies/mL), low platelet count (,150 •10 9 /L), and low total serum cholesterol (#4.5 mmol/L) at baseline. An increased risk of death was associated with TB, adjusted for CD4, platelets, and cholesterol. A low CD4 + -cell count was significantly associated with incident TB, mortality, other AIDS diagnoses, and virologic failure.Discussion: The risk of TB remains elevated in PLHIV in the setting of second-line HIV therapy in TB endemic regions. TB was associated with a greater risk of death. Finding that low CD4 + T-cell count was significantly associated with poor outcomes in this population supports the value of CD4 + monitoring in HIV clinical management.

show abstract

Section: Discussionmentioning

confidence: 99%

Factors Associated With and Characteristic of HIV/Tuberculosis Co-Infection: A Retrospective Analysis of SECOND-LINE Clinical Trial Participants

Henry

Jiamsakul

Law

et al. 2021

JAIDS Journal of Acquired Immune Deficiency Syndromes

Self Cite

View full text Add to dashboard Cite

show abstract

“…All analyses were stratified by site to account for the potential disparities in health care systems and differences in clinical practices across the participating sites 16 . A two‐sided level of 5% statistical significance was used for all inferences.…”

Section: Methodsmentioning

confidence: 99%

“…All analyses were stratified by site to account for the potential disparities in health care systems and differences in clinical practices across the participating sites. 16 A twosided level of 5% statistical significance was used for all inferences. The analyses were performed using sas software version 9.4 (SAS Institute, Cary, NC) and stata software version 14.2 (STATA Corp.).…”

Section: Metabolic Syndrome and Allcause Mortalitymentioning

confidence: 99%

Weight changes, metabolic syndrome and all‐cause mortality among Asian adults living with HIV

et al. 2021

Self Cite

View full text Add to dashboard Cite

Objectives: We investigated weight changes following antiretroviral therapy (ART) initiation, the development of metabolic syndrome (MetS) and its association with all-cause mortality among Asian adults living with HIV.Methods: Participants enrolled in a regional Asian HIV-infected cohort with weight and height measurements at ART initiation were eligible for inclusion in the analysis. Factors associated with weight changes and incident MetS (according to the International Diabetic Federation (IDF) definition) were analysed using linear mixed models and Cox regression, respectively. Competing-risk regression models were used to investigate the association of MetS with all-cause mortality.Results: Among 4931 people living with HIV (PLWH), 66% were male. At ART initiation, the median age was 34 [interquartile range (IQR) 29-41] years, and the median (IQR) weight and body mass index (BMI) were 55 (48-63) kg and 20.5 (18.4-22.9) kg/m 2 , respectively. At 1, 2 and 3 years of ART, overall mean (± standard deviation) weight gain was 2.2 (±5.3), 3.0 (±6.2) and 3.7 (±6.5) kg, respectively. Participants with baseline CD4 count ≤ 200 cells/µL [weight difference (diff) = 2.2 kg; 95% confidence interval (CI) 1.9-2.5 kg] and baseline HIV RNA ≥ 100 000 HIV-1 RNA copies/mL (diff = 0.6 kg; 95% CI 0.2-1.0 kg), and those starting with integrase strand transfer inhibitor (INSTI)-based ART (diff = 2.1 kg; 95% CI 0.7-3.5 kg vs. nonnucleoside reverse transcriptase inhibitors) had greater weight gain. After exclusion of those with abnormal baseline levels of MetS components, 295/3503 had incident MetS [1.18 (95% CI 1.05-1.32)/100 person-years (PY)]. The mortality rate was 0.7 (95% CI 0.6-0.8)/100 PY. MetS was not significantly associated with all-cause mortality in the adjusted model (P = 0.236).Conclusions: Weight gain after ART initiation was significantly higher among those initiating ART with lower CD4 count, higher HIV RNA and an INSTI-based regimen after controlling for baseline BMI. Greater efforts to identify and manage MetS among PLWH are needed.

show abstract

Improving Prognostic Web Calculators: Violation of Preferential Risk Independence

AlemiFarrokh

LevyCari

CitronBruce

et al. 2016

Journal of Palliative Medicine

View full text Add to dashboard Cite

show abstract

The occurrence of Simpson's paradox if site-level effect was ignored in the TREAT Asia HIV Observational Database

Cited by 4 publications

References 28 publications

Factors Associated With and Characteristic of HIV/Tuberculosis Co-Infection: A Retrospective Analysis of SECOND-LINE Clinical Trial Participants

Factors Associated With and Characteristic of HIV/Tuberculosis Co-Infection: A Retrospective Analysis of SECOND-LINE Clinical Trial Participants

Weight changes, metabolic syndrome and all‐cause mortality among Asian adults living with HIV

Improving Prognostic Web Calculators: Violation of Preferential Risk Independence

Contact Info

Product

Resources

About