Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
J Acquir Immune Defic Syndr. Author manuscript; available in PMC 2014 April 1.
Published in final edited form as:
PMCID: PMC3692572

Tenofovir use and urinary biomarkers among HIV-infected women in the Women's Interagency HIV Study (WIHS)



Tenofovir has been associated with renal tubular injury. Biomarkers that signal early tubular dysfunction are needed because creatinine rise lags behind tenofovir-associated kidney dysfunction. We examined several urinary biomarkers to determine if rises accompanying tenofovir initiation preceded creatinine changes.


Three urinary biomarkers of tubular impairment- neutrophil gelatinase-associated lipocalin (NGAL), N-acetyl- β -D-glucosaminidase (NAG), and β-2-microglobulin (β2MG)-were measured across three time points (one pre-tenofovir visit and two post tenofovir visits) in one hundred and thirty two HIV-positive women from the Women's Interagency HIV Study (WIHS). Women initiating HAART containing tenofovir were propensity score matched to women initiating HAART without tenofovir and women not on HAART.


There were no differences between groups for NGAL or NAG but β2MG was 19 times more likely to be elevated among tenofovir users at the 2nd post tenofovir visit compared to non-TDF users at the pre-tenofovir visit (p<0.01). History of proteinuria was associated with elevated NGAL (p <0.01). Factors associated with elevated NAG were GFR<60 ml/min, history of proteinuria, hepatitis C (p<0.01 for all) and diabetes mellitus (p=0.05). Factors associated with increased odds of elevated β2MG were HIV RNA>100,000 copies/ml, hepatitis C, boosted protease inhibitor (PI) use, and GFR<60 ml/min (p≤0.01 for all).


β2MG levels are elevated in women on tenofovir indicating probable early renal dysfunction. Biomarker elevation is additionally associated with baseline chronic kidney disease, uncontrolled viremia, and boosted PI use. Future studies are needed to explore urinary biomarker thresholds in identifying treated HIV-infected individuals at risk for renal dysfunction.

Keywords: Tenofovir, urinary biomarkers, HIV infected women


Persons with HIV are at high risk for the onset and progression of chronic kidney disease (CKD). In addition to HIV related kidney disease (HIV-associated nephropathy or HIVAN), the population of people living with HIV/AIDS is aging and developing comorbidities that predispose them to CKD such as diabetes mellitus, hypertension, and cardiovascular disease. Renal toxicity from antiretroviral and other medications also contributes to kidney dysfunction in HIV-infected persons. One such antiretroviral drug, tenofovir undergoes renal clearance by a combination of glomerular filtration and active proximal tubular secretion. It has also been implicated as a source of impaired kidney function in HIV-infected individuals [1-3].

While serum creatinine level is a well-accepted marker of kidney function, its disadvantages include its contemporaneous relationship with cellular damage. In addition, it is affected by a number of other factors including muscle mass, protein intake, proteinuria, race and age [4-5]. Therefore identifying biomarkers that reflect renal injury in real time and prior to the rise in creatinine could impact the delivery of care to patients with HIV. Given the multiple processes (direct HIV effects, antiretroviral effects and increased prevalence of CKD risk factors) that may affect kidney function in persons with HIV, markers that could predict kidney function decline would be of paramount interest to clinicians involved in HIV care.

Over the past decade, novel urine biomarkers specific for tubular injury have been found, which provide an earlier indicator of impairment. Urine neutrophil gelatinase-associated lipocalin (NGAL) is a member of the lipocalin family of proteins that is secreted into the urine by the thick ascending limb of Henle as well as the collecting ducts of the kidney. While normally expressed in low levels, increasing concentrations in urine are seen in the presence of epithelial injury and inflammation [6]. N-acetyl-β-D-glucosaminidase (NAG) is a proximal tubule lysosomal enzyme whose presence in the urine suggests proximal tubular damage [7]. β-2-microglobulin (β2MG) is a low molecular weight protein, found in all nucleated cells, freely filtered by glomeruli and catabolized by the proximal tubules [8].

These three biomarkers have been associated with acute kidney injury (AKI formerly acute renal failure) in people without HIV [9-10] and have been found to be elevated in patients with conditions associated with CKD such as diabetes [11]. However it is unclear whether these biomarkers could be clinically useful to monitor nephrotoxicity from antiretroviral therapy. This study sought to explore the relationships between tenofovirinitation and changesin levels of urinary biomarkers using longitudinal data from a stable outpatient populationof women with HIV.


Study Design and Measurements

Women's Interagency HIV Study (WIHS)

The rationale and methods of the WIHS have been previously described [12-13]. Briefly, the WIHS is a multicenter, prospective cohort study of the natural history of HIV-infection among women, conducted in Chicago, Los Angeles, New York City, the San Francisco Bay Area, and Washington DC. From October 1994 through November 1995, 2,623 (2,054 HIV+ and 569 HIV-) study participants were enrolled. From October 2001 through September 2002, an additional 1,143 (737 HIV+ and 406 HIV-) participants were enrolled for a total of 3,766 (2,791 HIV+ and 975 HIV-) women. Participants were recruited from HIV primary care clinics, hospital-based outpatient infectious diseases clinics, research programs, community outreach sites, women's support groups, drug rehabilitation programs, HIV testing sites, and referrals from previously enrolled participants. HIV-infection was determined with ELISA and confirmed via Western Blot. A standardized interview-based survey was used at enrollment to collect demographic data and prior medical, sexual, and drug use history. Women are evaluated semi-annually to obtain weight, CD4 lymphocyte count, HIV RNA level, albumin, creatinine, and other testing. Urine is collected annually and stored in the central repository. Proteinuria was qualitatively assessed during the first seven study visits. Women were described as having proteinuria if at least two urine analyses demonstrated the qualitative presence of ≥+1 protein. Race, hepatitis C antibody status, and history of proteinuria were determined prior to study baseline. All other factors were time-varying and were determined at each visit biomarkers were measured. The study was approved by the institutional review boards (IRBs) at WIHS Institutions and informed consent was obtained from every participant.

Study Population

Among WIHS participants, we included in the current study HIV-infected women initiating a highly active antiretroviral therapy regimen (HAART) containing tenofovir who had available stored urine specimens and serum creatinine measurements. For each included participant, three visits were selected for urine testing: the visit one year prior to the initial report of tenofovir (study baseline), the initial visit that tenofovir was reported (matching visit, 2nd measure) and one year after the initial report of tenofovir (3rd measure). Participant visits contributed by HIV-infected women using HAART without tenofovir (non-TDF HAART users) were matched to the visit that tenofovir was initially reported for each TDF HAART user via propensity score matching with a tolerance of 0.02 [14]. Propensity scores were estimated from a logistic multivariate regression model containing the following predictors: glomerular filtration rate (GFR) calculated using the CKD-EPI equation [15], CD4 cell count (CD4), calendar year, HAART duration, HAART interruptions, body mass index, HIV RNA level, and use of a ritonavir-boosted Protease Inhibitor (PI). The study design and matching structure are shown in Figure 1. A total of 45 person-visits contributed by 45 individual non-TDF HAART users were matched to the visit of the first report of tenofovir for 45 individual TDF HAART users.

Figure 1
The matched design of the study is represented by the three arrows, which highlight the three groups of HIV-infected women (women on HAART initiating tenofovir and two comparison groups) and the three time points at which the urinary biomarkers were measured. ...

As an additional comparison group, participant visits contributed by HIV-infected women who had not used HAART (but who may have used HAART in the past) within the past 3-9 months (non-HAART users) were matched to the visit that tenofovir was initially reported for each TDF HAART user via propensity score matching as described above. For this matched analysis, the propensity scores were estimated from a logistic multivariate model containing the following predictors: race, history of proteinuria, GFR, CD4, calendar year, body mass index, HIV RNA, hepatitis C (HCV) antibody status, and smoking history. A total of 45 person-visits contributed by 42 individual non-HAART users were matched to the visit of first reported tenofovir use for the same 45 individual TDF HAART users described above.


Stored urine specimens were measured for NGAL, NAG, and β2MG for each woman at each of the three time points. Urine NGAL level was assessed with the ARCHITECT® assay (Abbott Diagnostics, Abbott Park, IL) that specifically detected human NGAL. Urine NAG activity was assessed using a NAG kit (Roche Diagnostics, Indiana) and β2MG was assayed using enzyme immunoassay kit from ALPCO Ltd., New Hampshire; the intra-assay and interassay variation coefficients were less than 5%. Serum creatinine was measured locally at each WIHS site.

Statistical analysis

All biomarker values were log transformed and the unadjusted geometric means were examined graphically. For multivariable analyses, NGAL and NAG were treated as log normally distributed while β2MG was treated as a dichotomous outcome (> 0.5 versus ≤ 0.5 μg/mL) as a result of a large percentage of undetectable values (limit of detection =0.06 μg/ml).This dichotomized outcome for β2MG is consistent with previous reports [16], but there is no established cutoff value for either NGAL or NAG in previous studies. The transformed biomarker values were fit using generalized linear models, with generalized estimating equations used to adjust standard errors to account for the repeated measures over time and matched exposure groups [17]. Time trends over the three biomarker visits were modeled piecewise linearly with a spline at the matching visit (2nd measure). Models were simplified to linear if the estimated change from the 2nd to 3rd measure was non-significant, indicating no departure from linearity over time across the three visits. Final models using the non- tenofovir HAART group as the comparator were adjusted for age as continuous, race (white, black and Latina), proteinuria and diabetes history and WIHS baseline HCV antibody status as binary indicators, eGFR<60 ml/min|1.73 m2, hypertension (systolic>140 mmHg or diastolic>90 mmHg), current smoking status, categorized body mass index (BMI: <18.5, 18.5-25, 25-30, >30 kg/m2), CD4<200 cells/mm3, HIV RNA>100,000 copies/ml, years since HAART initiation, proportion of visits no HAART use was reported since initiation, and boosted PI use. For the additional analysis using a non-HAART comparison group, final models used the same covariates with the exception of the HAART use-related variables. All covariates except race, history of proteinuria, and baseline hepatitis C antibody status were time-varying.

Missing data were imputed according to whether the factor was fixed (proteinuria history was missing for 8 women and diabetes history for 4 women) or time-dependent (9 BMIs, 7 systolic and diastolic blood pressures, 6 eGFR measures, 5 CD4 cell counts, and 3 HIV RNA measures). For fixed covariates, the proportion of cases among individuals with complete data was used to randomly assign case status to individuals with missing data. For time-dependent covariates, missing values were interpolated from neighboring values in time using the average between the last non-missing value prior to and the first non-missing value after the missing value.


Urinary biomarkers were measured at three time points for three groups of women based on subsequent exposure to antiretroviral therapy (Figure 1). This comprised 135 person-visits from 132 individual women. At baseline, the median age of the women initiating HAART with tenofovir was 42 years and the racial distribution was 27% white/other, 49% black and 24% Latina (Table 1). Eighteen percent of women in the group had a history of proteinuria at baseline and the average eGFR was 98ml/min. Only 4 women reported use of stavudine or atazanavir – both of which have been linked to renal tubular dysfunction - in any of the study visits [18-19]. Following propensity score matching all factors were similar between the tenofovir HAART users and the non-tenofovir HAART users except for the proportion of women with unknown history of proteinuria (0% vs. 18%, p<0.01), median HIV RNA (3.5 vs. 1.9, p=0.01), and categorized HIV RNA (p=0.01).

Table 1
Comparison of characteristics at study baseline

The geometric means and 95% confidence limits (95% CIs) for NGAL did not change over time and were not different among exposure groups in the unadjusted analysis (Appendix Figure, top row). While NAG levels increased over time, there were no notable differences among exposure groups (Appendix Figure, middle row). In contrast, β2MG levels increased over time among women initiating HAART with tenofovir (Appendix Figure, bottom row). The proportion of women with elevated β2MG (>0.5 μg/ml) increased from 7% to 40% among tenofovir HAART users (p < 0.01), from 11% to 16% among non-tenofovir HAART users (p 0.29), and 13% to 18% among non-HAART users, from the 1st to the 3rd measurement. We calculated Spearman correlation coefficients for the log-transformed values of the 3 urinary biomarkers. The correlation between NGAL and NAG was 0.47 (p<0.0001). Restricted to pairs with positive β2MG, correlation between β2MG and NGAL was 0.29 (p<0.0001) and between β2MG and NAG was 0.52 (p<0.01).


In adjusted models, no differences were seen for NGAL between groups in the main comparison (tenofovir HAART versus non-tenofovir HAART) at baseline or over time. However, NAG rose among tenofovir HAART users by 20% per year on average (p<0.05) but did not differ between groups over time (figure 2). NGAL and NAG were both elevated among women with renal parameters suggestive of CKD at baseline (Table 2). Values of both were between 69% and 62% higher, respectively, among women with a history of proteinuria as compared to women without a history of proteinuria (p<0.01 for both). Similarly, women with eGFR less than 60 ml/min at study baseline had approximately a 60% greater value of NGAL (p0.10) and 80% greater value of NAG (p0.01). With respect to medical history, two other factors were marginally associated with higher NGAL and NAG levels. These include years since initiation of HAART for NGAL (p0.04), and a history of diabetes mellitus (p0.05) and hepatitis C (p<0.01) for NAG. Race was also highly associated with NGAL and NAG level. Latina women had NGAL and NAG levels two times as high as white women (p0.01 for both).

Figure 2
Estimated changes over time in biomarker level or odds of an increased level from the multivariate generalized linear models adjusted for confounding factors. All odds ratios are relative to baseline for the control group, non-TDF HAART on the left and ...
Table 2
Factors associated with elevated NGAL and NAG in multivariable analysis

In the secondary comparison, levels and change over time for NGAL and NAG among HAART users initiating tenofovir were compared to a group not taking HAART. NGAL and NAG levels were found to be similar and there was no significant change over time. However, NAG increased among tenofovir users by 21% per year on average compared to baseline (p <0.05) (consistent with results from the main analysis) and also rose among non-tenofovir HAART users by 33% per year on average compared to baseline (p<0.05) (compared to non-HAART group). African-American women as well as Latina women had higher NAG levels compared to whites (p0.01). Levels of NGAL and NAG were not significantly different among women with a history of proteinuria in this analysis, but were more than twice as high among women with estimated GFR<60ml/min (p0.01).


In the adjusted model of β2MG, tenofovir HAART users had 19 times the odds of having an elevated β2MG at the 3rd time point compared to 2.8 for non-tenofovir HAART users at the 3rd time point (p < 0.01). Furthermore, CKD risk factors were associated with an elevated β2MG. Women with GFR<60 ml/min (OR 5.67 p 0.01) and HCV (OR 9.3, p <0.01) were more likely to have elevated β2MG while smokers (OR0.27, p0.04) and Black women (OR0.2, p0.02) were less likely to have elevated β2MG. However, controlling for eGFR, proteinuria was not associated with likelihood of elevated β2MG (p0.22). With respect to HIV specific parameters, women with HIV RNA >100,000 copies/ml (OR 23.89, p<0.01) and boosted PI use (OR 8.97, p <0.01) were more likely to have elevated β2MG.

In the secondary comparison, a different pattern of predictors of elevatedβ2MG was noted. The estimated odds of having an elevated β2MG level at the 3rd time point was 3.9 times higher among tenofovir initiators compared to 1.1 times higher for non-HAART users at the 3rd time point (p0.05). The CKD risk factors of an estimated GFR<60 and HCV were still significant factors. CD4 ≤ 200 cells/mm3was also a significant predictor of elevated β2MG (OR 9.0, p<0.05) and obesity was associated with a decreased odds of elevated β2MG (OR 0.22, p<0.05)


In this cross-sectional analysis of a cohort of HIV infected women (WIHS), we evaluated the changes in three significant urinary biomarkers over time in a group of women who initiate a tenofovir-containing antiretroviral regimen compared to women not initiating tenofovir or not on HAART. At study baseline, there were no differences between groups for NGAL, NAG and β2MG. The increase over time for NGAL and NAG was not significant among TDF users compared to the non-TDF users. β2MG was more likely to be elevated among TDF users at the 3rd time point compared to non-TDF users, indicating that this marker may be an important indicator of TDF related kidney dysfunction. Even though increases over time for NGAL and NAG was not significant among TDF users compared to the non-TDF users , our results suggest that elevations in NGAL and NAG among TDF users are more generally related to comorbid diseases commonly associated with chronic kidney disease (CKD). In contrast, the factors specific to HIV infection such as antiretroviral regimen (HAART with tenofovir and regimen including a boosted PI), lower CD4 and higher HIV RNA level were associated with greater urinary β2MG levels.

Advanced immunosuppression (CD4 <50 cells/mm3) has been associated with a decline in creatinine clearance in a prior observational study [20]. However, another study did not find an association between CD4<200 cells/mm3 and higher β2MG excretion in HIV-infected patients on TDF compared to other nucleoside reverse transcriptase inhibitors [21]. Therefore, our study uniquely demonstrates that immunosuppression affects β2MG levels which has not previously been described. Similarly, uncontrolled HIV viremia has been linked to renal dysfunction in a large randomized trial [22]. Our findings of the association between uncontrolled HIV viremia with urinary biomarker elevation supports the findings from a recent cohort study which noted an association between detectable HIV RNA (> 400 copies/ml) and elevated cystatin C, another biomarker of renal function [23]. This suggests that the development of these biomarkers as clinical guides will need to take into account the medical history of the patient as well as the effectiveness of their HIV directed care.

Our findings regarding β2MG confirm prior work demonstrating higher levels of urinary β2MG among persons on tenofovir containing antiretroviral regimen and the impact of concurrent PI use. In the ASSERT study, subjects receiving TDF/FTC fixed dose combination (FDC) had a 72% greater increase at week 24 and a 133% greater increase at week 48 in urinary β2MG than subjects taking ABC/3TC FDC [24]. In addition, a greater decline in kidney function has been noted in persons receiving tenofovir with a boosted PI compared to tenofovir without a PI [25]. Furthermore, an observational cohort of patients receiving tenofovir demonstrated that subgroups receiving tenofovir with ritonavir-boosted lopinavir had greater urinary β2MG levels than those receiving tenofovir without a PI [16].

In addition, our findings confirm prior reports that metabolic factors may play a role in modulating biomarker levels. Low body weight (defined by <60 kg) has been associated with renal function decline in persons on TDF compared to abacavir containing regimen [26] and greater urinary β2MG levels [16]. Likewise, our study found that obesity was protective against elevated β2MG in the TDF group compared to the non-HAART group. Furthermore, we found an association between current smoking and reduced β2MG levels in the TDF group compared to non-TDF HAART users contradicting prior reports of smoking and elevated β2MG in an environmental cohort [27].

Prior studies have been limited in their ability to examine the relationships between CKD risk factors and these biomarkers due to the exclusion of patients with proteinuria or preexisting history of renal disease, our results reveal the importance of preexisting kidney disease in the interpretation of urinary biomarker elevations in patients on tenofovir. These results thus add to the literature on the clinical utility of these markers for screening for sub-clinical renal toxicity in HIV-infected patients.

The interpretation of the results of this study may be limited to women because of gender differences in these biomarker levels. In an HIV negative population with type 1 diabetes, urinary NGAL levels were significantly higher in females compared to males [28], which may reflect an estrogen-mediated difference in protein expression in renal or urinary tract tissues [29]. Furthermore, gender and race differences have been noted in NAG levels. In a cross-sectional study of the early natural history of cardiovascular disease, Black women had the higher levels of urinary NAG compared to Black men but White men and women had similar levels [30]. Studies assessing β2MG levels in HIV-infected persons have mostly been in men and have not assessed for gender differences [16, 24, and 31].

Other limitations to our analysis should be noted. Our conclusion regarding the impact of GFR <60 on biomarkers is limited by the small sample of women with GFR<60. Further the level of urinary biomarkers is influenced by urine concentration. Urine creatinine measurements, typically used to standardize for urine concentration, were unavailable in this study. However, the impact of not accounting for differences in urine concentration would likely be to attenuate effects by introducing additional variability into the analysis and thus our results may be conservative.

In summary, this analysis examined the value of the urinary biomarkers NGAL, NAG and β2MG as indicators of kidney injury resulting from tenofovir use. Our results demonstrate the potential utility of urinary β2MG levels in predicting patients at risk for loss of kidney function due to tenofovir use. Further studies are needed to determine when it is appropriate to use this urinary biomarker as well as how frequently to monitor. If validated, this biomarker may have clinical utility in identifying higher risk individuals allowing appropriate diagnostic and therapeutic interventions to be delivered earlier with a potential positive impact on kidney function.

Table 3
Factors associated with elevated β2MG level in multivariable analysis

Supplementary Material


Data in this manuscript were collected by the Women's Interagency HIV Study (WIHS) Collaborative Study Group with centers (Principal Investigators) at New York City/Bronx Consortium (Kathryn Anastos); Brooklyn, NY (Howard Minkoff); Washington DC, Metropolitan Consortium (Mary Young); The Connie Wofsy Study Consortium of Northern California (Ruth Greenblatt); Los Angeles County/Southern California Consortium (Alexandra Levine); Chicago Consortium (Mardge Cohen); Data Coordinating Center (Stephen Gange).

The WIHS is funded by the National Institute of Allergy and Infectious Diseases (UO1-AI-35004, UO1-AI-31834, UO1-AI-34994, UO1-AI-34989, UO1-AI-34993, and UO1-AI-42590) and by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (UO1-HD-32632). The study is co- funded by the National Cancer Institute, the National Institute on Drug Abuse, and the National Institute on Deafness and Other Communication Disorders. Funding is also provided by the National Center for Research Resources (UCSF-CTSI Grant Number UL1 RR024131). The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health.

Reagents to perform ARCHITECT Urine NGAL testing were provided by Abbott.


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


1. Antoniou T, Raboud J, Chirhin S, et al. Incidence of and risk factors for tenofovir-induced nephrotoxicity: a retrospective cohort study. HIV Med. 2005;6:284–290. [PubMed]
2. Barrios A, Garcia-Benayas T, Gonzalez-Lahoz J, et al. Tenofovir-related nephrotoxicity in HIV-infected patients. AIDS. 2004;18:960–963. [PubMed]
3. Karras A, Lafaurie M, Furco A, et al. Tenofovir-related nephrotoxicity in human immunodeficiency virus-infected patients: three cases of renal failure, Fanconi syndrome, and nephrogenic diabetes insipidus. Clin Infect Dis. 2003;36:1070–1073. [PubMed]
4. Coresh J, Astor BC, Greene T, et al. Prevalence of chronic kidney disease and decreased kidney function in the adult US population: Third National Health and Nutrition Examination Survey. Am J Kidney Dis. 2003;41(1):1–12. [PubMed]
5. Lew SW, Bosch JP. Effect of diet on creatinine clearance and excretion in young and elderly healthy subjects and in patients with renal disease. Journal of American Society of Nephrology. 1991;2(4):856–865. [PubMed]
6. Mishra J, Ma Q, Prada A, et al. Identification of neutrophil gelatinase-associated lipocalin as a novel early urinary biomarker for ischemic renal injury. J Am Soc Nephrol. 2003;14(10):2534–2543. [PubMed]
7. Price RG. The role of NAG (N-acetyl-beta-D-glucosaminidase) in the diagnosis of kidney disease including the monitoring of nephrotoxicity. Clin Nephrol. 1992;38(Suppl 1):S14–S19. [PubMed]
8. Thielemans N, Lauwerys R, Bernard A. Competition between albumin and low-molecular-weight proteins f or renal tubular uptake in experimental nephropathies. Nephron. 1994;66(4):453–458. [PubMed]
9. Nickolas TL, O'Rouke MJ, Yang J, et al. Sensitivity and specificity of a single emergency department measurement of urinary neutrophil gelatinase-associated lipocalin for diagnosing acute kidney injury. Ann Intern Med. 2008;148(11):810–819. [PMC free article] [PubMed]
10. Doi K, Negishi K, Ishizu T, et al. Evaluation of new acute kidney injury biomarkers in a mixed intensive care unit. Crit Care Med. 2011;39(11):2464–2469. [PubMed]
11. Nauta FL, Boertien WE, Bakker SJ, et al. Glomerular and tubular damage markers are elevated in patients with diabetes. Diabetes Care. 2011;34(4):975–981. [PMC free article] [PubMed]
12. Barkan SE, Melnick SL, Preston-Martin S, et al. The Women's Interagency HIV Study. WIHS Collaborative Study Group. Epidemiology. 1998;9(2):117–125. [PubMed]
13. Bacon MC, Von Wyl W, Alden C, et al. The Women's Interagency HIV Study: an observational cohort brings clinical sciences to the bench. Clin Diagn Lab Immunol. 2005 Sep;12(9):1013–1019. [PMC free article] [PubMed]
14. Rosenbaum PR, Rubin D. Constructing a control group using multivariate matched sampling methods that incorporate the propensity score. Am Stat. 1985;39:33–38.
15. Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150(9):604–612. [PMC free article] [PubMed]
16. Gatanaga H, Tachikawa N, Nikuchi Y, et al. Urinary beta2-microglobulin as a possible sensitive marker for renal injury caused by tenofovir disoproxil fumarate. AIDS Res Hum Retroviruses. 2006;22(8):744–748. [PubMed]
17. Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics. 1986;42:121–130. [PubMed]
18. Nelson M, Azwa A, Sokwala A, et al. Fanconi syndrome and lactic acidosis associated with stavudine and lamivudine therapy. AIDS. 2008 Jul 11;22(11):1374–6. [PubMed]
19. Dauchy FA, Lawson-Ayayi S, de La Faille R, et al. Increased risk of abnormal proximal renal tubular function with HIV infection and antiretroviral therapy. Kidney Int. 2011 Aug;80(3):302–9. [PubMed]
20. Gallant JE, Parish MA, Keruly JC, et al. Changes in renal function associated with tenofovir disoproxil fumarate treatment, compared with nucleoside reverse-transcriptase inhibitor treatment. Clin Infect Dis. 2005 Apr 15;40(8):1194–8. [PubMed]
21. Kinai E, Hanabusa H. Progressive renal tubular dysfunction associated with long-term use of tenofovir DF. AIDS Res Hum Retroviruses. 2009 Apr;25(4):387–94. [PubMed]
22. El-Sadr WM, Lundgren JD, Neaton JD, et al. Strategies for Management of Antiretroviral Therapy (SMART) Study Group,CD4+ count-guided interruption ofantiretroviral treatment. N Engl J Med. 2006 Nov 30;355(22):2283–96. [PubMed]
23. Overton ET, Patel P, Mondy K, et al. Cystatin C and baseline renal function among HIV-infected persons in the SUN Study. AIDS Res Hum Retroviruses. 2012 Feb;28(2):148–55. [PubMed]
24. Post FA, Moyle GJ, Stellbrink HJ, et al. Randomized comparison of renal effects, efficacy, and safety with once-daily abacavir/lamivudine versus tenofovir/emtricitabine, administered with efavirenz, in antiretroviral-naive, HIV-1-infected adults: 48-week results from the ASSERT study. J Acquir Immune Defic Syndr. 2010;55(1):49–57. [PubMed]
25. Goicoechea M, Liu S, Best B, et al. Greater tenofovir-associated renal function decline with protease inhibitor-based versus nonnucleoside reverse-transcriptase inhibitor-based therapy. J Infect Dis. 2008;197(1):102–108. [PubMed]
26. Nishijima T, Gatanaga H, Komatsu H, et al. Renal function declines more in tenofovir-than abacavir-based antiretroviral therapyin low-body weight treatment-naïve patients with HIV infection. PLoS One. 2012;7(1):e29977. [PMC free article] [PubMed]
27. Chaumont A, De Winter F, Dumont X, et al. The threshold level of urinary cadmium associated with increased urinary excretion of retinol-binding protein and beta 2-microglobulin: a re-assessment in a large cohort of nickel-cadmium battery workers. Occup Environ Med. 2011 Apr;68(4):257–64. [PMC free article] [PubMed]
28. Thrailkill KM, Moreau CS, Cockrell GE, et al. Disease and gender-specific dysregulation of NGAL and MMP-9 in type 1 diabetes mellitus. Endocrine. 2010;37(2):336–343. [PMC free article] [PubMed]
29. Seth P, Porter D, Lahti-Domenici J, et al. Cellular and molecular targets of estrogen in normal human breast tissue. Cancer Res. 2002;62(16):4540–4544. [PubMed]
30. Gan SL, Chia SE, Chio LF, et al. Urinary beta 2- and alpha1-microglobulin levels in normal subjects in Singapore. Occup Med. 1994;44(4):195–200. [PubMed]
31. Agirbasli M, Radhakrishnamurthy B, Jiang X, et al. Urinary N-acetyl-beta-D-glucosaminidase changes in relation to age, sex, race, and diastolic and systolic blood pressure in a young adult biracial population.The Bogalusa Heart Study. Am J Hypertens. 1996;9(2):157–161. [PubMed]