Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Antivir Ther. Author manuscript; available in PMC 2009 August 4.
Published in final edited form as:
Antivir Ther. 2009; 14(4): 543–549.
PMCID: PMC2720522

The Effects of HIV-1 Viral Suppression and Non-Viral Factors on Quantitative Proteinuria in the HAART Era

Samir K. Gupta,1 Marlene Smurzynski,2 Nora Franceschini,3 Ronald J. Bosch,2 Lynda A. Szczech,4 and Robert C. Kalayjian5, for the AIDS Clinical Trials Group Longitudinal Linked Randomized Trials (ALLRT) Study Team



Proteinuria is associated with progressive renal disease and overall mortality in HIV-infected patients. However, the prevalence and correlates of quantitative proteinuria in the HAART era are unknown.


Spot urine protein to creatinine (P/Cr) ratios, an accepted measure of quantitative daily proteinuria, were measured annually since 2002 in participants of the AIDS Clinical Trials Group Longitudinal Linked Randomized Trials (ALLRT) cohort. We used linear regression models with general estimating equations to identify factors associated with the abnormal P/Cr thresholds of ≥0.2 and ≥1.0.


2857 participants, most of whom were receiving antiretroviral therapy, were analyzed. 16% and 3% had P/Cr levels ≥0.2 and ≥1.0, respectively, at first measurement. P/Cr levels did not change during a median follow-up of 3 (IQR 2, 4) years. Factors associated with P/Cr ≥0.2 at any measurement included greater age, lower glomerular filtration rate, female sex, antiretroviral therapy prior to entry into parent randomized trial, HIV-1 RNA level ≥ 400copies/ml, lower CD4 cell count, and history of hypertension, diabetes, or hepatitis C co-infection (all P<0.04). Black race and higher non-HDL-C levels were associated with P/Cr levels ≥1.0 but not with P/Cr levels ≥0.2. Hepatitis B co-infection and current use of adefovir, indinavir, and tenofovir were not associated with either P/Cr threshold.


Both HIV and non-HIV-related factors are associated with abnormal levels of proteinuria and identify those who are at greater risk of worse clinical outcomes. Several of these factors are differentially associated with lower and higher proteinuria thresholds.

Keywords: HIV, Proteinuria, Antiretroviral Therapy, Nephropathy


The presence of proteinuria predicts progression of chronic kidney disease, new AIDS-defining illness, and increased mortality in HIV-infected populations [1, 2]. Determining predictors of proteinuria in HIV-infected individuals may identify those at greater risk of adverse outcomes.

HIV-1 viral suppression with highly active antiretroviral therapy (HAART) reduces proteinuria [3] and improves reduced glomerular filtration rates [4]. Paradoxically, HAART is also associated with the development of diabetes [5], hypertension [6], and drug-related nephrotoxicity [7]. Most previous studies of the longitudinal effects of HAART on proteinuria are limited by their small sample size, short duration of follow-up, and use of dipstick measurements [8], which does not allow analysis of multiple competing factors that may influence proteinuria. Therefore, we performed an analysis of a large cohort of study participants enrolled into the AIDS Clinical Trials Group (ACTG) Longitudinal Linked Randomized Trials (ALLRT) cohort in which quantitative proteinuria has been measured annually since 2002.


Study participants

ALLRT is a prospective, longitudinal, cohort study comprising participants enrolled in ACTG parent clinical trials in which they receive randomized ART regimens or strategies and are followed longitudinally to evaluate long-term outcomes of ART [9]. Included in this analysis were HIV-infected subjects of age 13 years or older who enrolled into the ALLRT study and had at least one urine protein to creatinine (P/Cr) measurement during follow-up. Participants in trials that were ongoing or incorporated immune-based therapies were excluded in the current analyses. See the Appendix for a listing of included clinical trials for these analyses. All subjects provided written, informed consent to participate in the ALLRT cohort study, and this study was approved by Institutional Review Boards of each participating site.

Measurements and definitions

For the purposes of these analyses, ‘baseline’ refers to values obtained at entry into the parent ACTG clinical trial. Urine samples collected annually in the ALLRT cohort since 2002 are used to measure protein and creatinine and were measured at local laboratories. Other laboratories measured along with these urine studies included serum creatinine, plasma HIV-1 RNA level, CD4 cell count, and lipid fractions. Fasting conditions at the time of blood draws were not uniform for all visits in ALLRT. Therefore, we used non-high density lipoprotein (HDL)-cholesterol (total cholesterol minus HDL-C) to estimate the proatherogenic lipid fraction as this measure is not affected by recent food intake. HIV-1 RNA level and CD4 cell count at baseline, demographics [age, sex, race (black vs. not black)], and clinical diagnoses were also analyzed. Patient-level data on individual glucose and blood pressure levels were not available for every study participant at each timepoint, so we examined diabetes and hypertension as either a reported history or a diagnosis anytime during the course of the study. Hepatitis B was defined by a positive hepatitis B surface antigen or core antibody, and hepatitis C was defined by either a listed diagnosis of hepatitis C or a positive hepatitis C antibody test at any time during the study. Quantitative proteinuria was defined as the ratio of spot urine protein to creatinine (P/Cr), which is an accepted surrogate measure of a 24 hour urine collection for proteinuria [10]. For example, a P/Cr measurement of 0.8 would imply that a 24-hour urine collection would contain 0.8 grams of protein. Glomerular filtration rate (GFR) was estimated using the abbreviated Modification of Diet in Renal Disease (MDRD) equation [11], in which values greater than 200ml/min/1.73m2 were truncated to 200 to reflect physiologic conditions.

We also analyzed data on the current use of adefovir, indinavir, and tenofovir as these antiretrovirals have previously been found to be associated with renal disease [12]. The dosing of adefovir was limited to that used for HIV treatment. Current use was defined as having received these antiretrovirals for at least 21 days within the 16 weeks prior to measurement of P/Cr.

Statistical analyses

We determined the rate of change in P/Cr since first available urine protein measurement. The slope of P/Cr and its 95% confidence intervals were assessed in mixed-effects linear models that assumed a random intercept and a random slope for each subject and a first-order autoregressive covariance structure. Because most participants in these analyses enrolled into their initial clinical trials prior to 2002 and because annual urine studies were not mandated prior to that time, we could not assess the change in quantitative proteinuria with initiation of HAART.

Generalized estimating equations were used to assess potential correlates of clinically significant proteinuria at two different levels, namely ≥0.2 or ≥1.0, at each annual P/Cr measurement. It is generally accepted that a P/Cr of 0.2 (or 0.2g of urine protein per day) is abnormal. Two different levels of quantitative proteinuria were chosen to explore the possibility that there may be differential predictors of lower level and higher level proteinuria. Correlates of higher levels of proteinuria, such as ≥3.0, were not examined due to the small number of participants with such levels (N=73).These models assumed a binomial distribution for the outcome variable and a first-order autoregressive covariance structure. Also included as possible baseline exploratory variables were estimated GFR, HIV-1 RNA level, CD4 cell count, history of prior antiretroviral therapy, hepatitis B, hepatitis C, and a fixed variable indicating history or on-study diagnosis of hypertension or diabetes. Time-updated, simultaneous variables included estimated GFR, CD4 cell count, non-HDL cholesterol, viral suppression to <400copies HIV-1 RNA/mL, and current use of adefovir, indinavir, or tenofovir. Variables significantly associated with one P/Cr threshold in initial analyses were included in the models for both P/Cr thresholds.

In addition to investigating the correlates of proteinuria at these two thresholds, we also investigated the correlates of proteinuria as a continuous variable using mixed effects linear models. These models assumed a first-order autoregressive covariance structure with random intercepts and random slopes and examined non-linear associations using cubed-root transformations (which provided the best approximation of a normal distribution for the observed data compared to log-transformations or squared-root transformations) of the outcome variable. In this continuous model, we defined HIV-1 RNA suppression as both a categorical variable at <400copies/mL and as a continuous variable as log10copies/mL while keeping all other variable definitions the same as described above. Model selection used backward elimination with significance defined as P < 0.05. All statistical analyses used SAS Version 9.0 (SAS Institute Inc., Cary, NC).


Characteristics of the study cohort

A total of 3290 subjects from ALLRT were potentially eligible. Of these, 433 were excluded due to lack of at least one urine P/Cr measurement while participating in ALLRT, leaving 2857 subjects available for these analyses. Table 1 shows the characteristics of this cohort both at baseline and the time of first P/Cr measurement. Of note, only 84 subjects had their first P/Cr measurement obtained within the first 16 weeks of initial randomized treatment.

Table 1
Characteristics of the study cohort (N=2857) at baseline and at first proteinuria (P/Cr) measurement.

The median (IQR) time from baseline enrollment to the first P/Cr measurement was 80 (16, 208) weeks, thus reflecting that most participants had been enrolled in a randomized trial for at least a year before their first quantitative proteinuria measurement was performed as part of the ALLRT protocol This is reflected in the improvement in CD4 cell counts and virologic suppression between the baseline values and those obtained at first P/Cr measurement (Table 1). The median (IQR) years of follow-up was 3 (2, 4). The numbers of subjects who had P/Cr measured at follow-up years 1, 2, 3, and 4 were 2232, 1913, 1549, and 987, respectively. There were no significant differences between times from baseline to first P/Cr measurement or between follow-up times in ALLRT for those with and without P/Cr ≥0.2 at first measurement (data not shown), thereby suggesting the absence of bias in retention rates between these groups.

The maximum percentages of subjects who received at least 21 days of adefovir, indinavir, and tenofovir at any P/Cr measurement were 0.2%, 2.1%, and 8.9%, respectively.

Changes in quantitative proteinuria over time

The median (IQR) P/Cr level at first measurement was 0.09 (0.06, 0.16). The number of subjects with P/Cr >0.2 and >1.0 at first measurement was 534 (16%) and 94 (3%), respectively. In the overall cohort, the mean (95% CI) slope change in P/Cr over time was 0.006 (−0.006, 0.019)/year, which was not significantly different than 0 (P=0.32). However, changes in slope did differ based on initial P/Cr categorization. In subjects with an initial P/Cr <0.2, the change (95% CI) in mean P/Cr over time was 0.015 (0.009, 0.021), which was significantly different than 0 (P<0.001). Subjects with P/Cr ≥1.0 at first measurement also had a significant (P=0.04) change in mean P/Cr levels over time of −0.270 (0.02, −0.52)/year. In subjects with an initial P/Cr ≥0.2 and <1.0, the change in P/Cr over time was 0.014 (−0.015, 0.04)/year; this result was not significantly different than 0 (P=0.35).

Correlates of proteinuria

Table 2 shows the final, separate multivariable models for factors associated with having a P/Cr level ≥0.2 or ≥1.0 at any measurement time during the ALLRT study. Diabetes, hepatitis C co-infection, lower current GFR, and lower current CD4 cell count were all associated with both levels of proteinuria (all P≤0.01). Interestingly, older age, female sex, receipt of antiretroviral therapy prior to baseline, and lack of viral suppression to less than 400 copies/ml were associated with P/Cr ≥0.2 (all P≤0.04) but were not associated with P/Cr ≥1.0. There was a trend (P=0.08) towards an association between current use of tenofovir with P/Cr ≥0.2 but not with P/Cr ≥1.0. Of note, diagnosis of hypertension was significantly associated with P/Cr ≥0.2 but not ≥1.0; since the odds ratios were the same for each threshold, the lack of significant for the higher cutoff was likely due to lack of power. Conversely, black race (P=0.02) and higher non-HDL-cholesterol (P=0.03) were each associated with P/Cr ≥1.0 but not with P/Cr ≥0.2. Hepatitis B co-infection, current use of indinavir, and current use of adefovir were not associated with either level of P/Cr.

Table 2
Factors associated with spot urine protein to creatinine ratios (P/Cr) ≥0.2 and ≥1.0 [odds ratios (95% CI)] and P/Cr as a continuous variable.

P/Cr was also modeled as a continuous variable (not as a categorical variable with specific thresholds) across the entire range of proteinuria results. In this model (Table 2) older age, lower current GFR, female sex, hypertension, diabetes, hepatitis C co-infection, receipt of antiretroviral therapy prior to baseline, , lower current CD4 cell count, and longer duration of follow-up were all significantly associated with higher levels of P/Cr (all P<0.05). Black race, current use of indinavir, current use of adefovir, current non-HDL-cholesterol, and hepatitis B co-infection, however, were not associated with higher levels of P/Cr. There was a trend towards an association between current tenofovir use and higher P/Cr (P=0.06) in this continuous model. HIV-1 RNA less than 400copies/mL was not associated with P/Cr as a continuous variable (P=0.08). However, when HIV-1 RNA was defined as a continuous function (log10copies/mL), with all other independent variables identical, higher viral load then became significantly correlated with higher P/Cr (P=0.002). When age was considered as a time-updated variable, there were no changes in the results of this continuous model except that duration of follow-up no longer became a significant correlate of P/Cr.


To our knowledge, this is the largest cohort study to date evaluating correlates of quantitative proteinuria over time in HIV-infected persons. The first major finding was that there was no overall change in P/Cr levels over time in this cohort. Although the persistent prevalences of nearly 16% and 3%, respectively, for P/Cr ≥0.2 and ≥1.0 are still quite high compared to the general population [13], most subjects had little proteinuria at first measurement in this cohort. This likely precluded observing any further reductions in P/Cr in the overall cohort. It is possible that any further proteinuria reductions that may have been achieved in the overall cohort by increasing the frequency of viral suppression during follow-up assessments may have been offset by worsening control of chronic co-morbidities, such as diabetes, hypertension, and hepatitis C co-infection. However, we could not examine patient-level data on serum glucose, blood pressure, or hepatitis C viral load levels, which precludes our ability to determine if improvement in these factors would have resulted in lower proteinuria levels. We did observe a significant reduction in P/Cr of nearly 0.3 per year in the subset of participants who specifically had a P/Cr ≥1.0 at initial measurement. On the other hand, we also observed an increase in P/Cr of 0.015 per year in those whose initial measurement was <0.2. These results may have resulted from changes in diabetes and hypertension control, changes in CD4 cell count or HIV or HCV viral loads, or development of drug nephrotoxicity. We cannot exclude the possibility that these results were simply due to chance, i.e. a ‘regression to the mean’ effect, and thus should be interpreted with caution.

It was not surprising to find that previously reported risk factors, such as older age, lower CD4 cell count, lower GFR, higher HIV-1 log10RNA level, diabetes, hypertension, and hepatitis C co-infection were associated with higher P/Cr in our continuous model as these variables have previously been associated with HIV-related proteinuria in other studies [1, 8, 1416]. The association between previous antiretroviral use and higher P/Cr may reflect longer durations of HIV infection and perhaps longer durations of uncontrolled viremia prior to entry into an ACTG trial. The higher risk for proteinuria in women corroborates a recent finding that women may be at higher risk for incident chronic kidney disease, regardless of racial category [17]. However, it should be noted that we did not use sex-specific cutoffs for P/Cr for our threshold analyses. We chose not to use such thresholds for P/Cr for simplicity and as such cutoffs have not yet been commonly used in general HIV care.

We also found that tenofovir was marginally associated with P/Cr ≥0.2 and with P/Cr as a continuous variable but was not associated at all with P/Cr ≥1.0. These results are similar to those found in a previous case-control study of tenofovir vs. non-tenofovir containing regimens in which tenofovir use was associated only with low-levels of abnormal proteinuria, which may represent renal tubular defects rather than glomerular protein loss [18]. The lack of truly significant associations between current tenofovir use and P/Cr in these analyses may have been due to limited power as tenofovir use was not used frequently in the parent clinical trials included in these analyses.

Another important new finding from this study was that there may be unique predictors for higher degrees of proteinuria. Given that HIV-associated nephropathy is almost exclusively found in patients of African descent, likely as a result of a recently identified genetic predisposition based around polymorphisms of the myosin heavy chain 9 (MYH9) gene [19], it would be reasonable to assume that black race would be associated with proteinuria. However, we found that black race, along with higher levels of non-HDL-C, was only associated with P/Cr greater than 1.0 and not P/Cr in the continuous model. On the other hand, viral suppression and older age were not related to this higher level of proteinuria. These results suggest that a race-specific predisposition to proteinuria may outweigh the effects of viral suppression in developing higher degrees of proteinuria. Recent epidemiologic studies suggest that HIV-infected blacks are at greater risk for developing end stage renal disease compared to HIV-infected patients of other races and ethnicities even though the prevalence of earlier stage chronic kidney disease are similar between black and white patients [17]. The finding here that black subjects are more likely to have greater amounts of proteinuria corroborate these findings and indicate a greater likelihood of having accelerated progression of kidney disease.

We also observed the well-known association between pro-atherogenic lipid fractions and higher levels of proteinuria as a marker of the nephrotic syndrome in HIV-infected patients [20]. Further studies are required to determine if lipid-lowering therapies will reduce proteinuria and subsequently delay renal disease progression in HIV-infected patients.

Limitations to these analyses should be noted. First, we could not evaluate in this cohort the possibility that proteinuria associated with previously untreated HIV infection would be reduced by antiretroviral therapy as has been shown in other smaller studies [3, 8], since most individuals were already on treatment at first P/Cr measurement. However, as the ALLRT study continues and includes participants of additional, large, randomized clinical trials, we will be able to repeat these analyses to include larger numbers of untreated participants with P/Cr measured pre-treatment and longitudinally after treatment. In addition, the subjects in these analyses were required to be medically stable with relatively preserved renal function, generally with serum creatinine values no greater than 1.5 times the upper limit of normal at the participant’s local laboratory, to enter the randomized trials. Thus, the cohort did not include large numbers of subjects with more pronounced renal dysfunction and possibly greater proteinuria levels. Furthermore, the study population had relatively low rates of co-morbidities, including hepatitis B and C co-infection, as compared to those reported in other outpatient populations [21], thereby further limiting external generalizability. In addition, these findings may not be generalizable to HIV-infected persons who could not enter randomized trials. We also acknowledge that the use of local lab results instead of performing all tests at a centralized laboratory likely increased the variability of the proteinuria levels and may have reduced our ability to find significant correlates of proteinuria in these analyses.

Offsetting these limitations is the unique feature of the ALLRT study cohort, which is the inclusion of participants enrolled in trials in which their treatment regimens were randomized. This allowed us to investigate more confidently the effects of HIV treatment strategies without introduction of potentially unmeasured or unknown confounders. The measurement of quantitative proteinuria levels instead of urine dipstick levels also allowed us to investigate more precisely changes in proteinuria over time and predictors at various thresholds.

In summary, mild proteinuria remains highly prevalent in HIV-infected persons in the HAART era and is associated with multiple factors, both HIV-related and HIV-unrelated. As such, our findings may directly impact clinical management. In keeping with the Infectious Diseases Society of America Guidelines [22], the results of this study suggest that monitoring for proteinuria is justified in patients who are older, are black, do not have viral suppression, have lower CD4 cell counts, who have either diabetes, hypertension, or hepatitis C co-infection, and possibly those who are receiving tenofovir. Monitoring may also be justified for women and those with higher non-HDL-C levels. Further research is required to determine if strategies that affect the potentially modifiable risk factors identified in this study will result in reduction of proteinuria and consequently improve clinical outcomes in the HIV-infected population.


This work was supported in part by National Institute of Allergy and Infectious Diseases grants to the AIDS Clinical Trials Group (AI68636, AI38858), SDAC/Harvard School of Public Health (AI68634, AI38855), Indiana University School of Medicine (AI25859), MetroHealth Medical Center (AI25879), Duke University Medical Center (AI069484). S.K.G. was supported by NIH/NHLBI K23 HL073682. N.F. is supported by AHA0675001N, NIH/NIDDK RO1 DK068336, and R01 HL089651. We thank all the ALLRT team, participating ACTG sites, and especially the ALLRT participants.


Presented in part: Poster #974, 15th Conference on Retroviruses and Opportunistic Infections, 2008, Boston, MA, USA.

Conflicts of Interest

S.K.G. has received consulting fees, advisory fees, and honoraria from Gilead Sciences, Inc and grant support from Tibotec Therapeutics. All other authors: no conflicts.


1. Gupta SK, Mamlin BW, Johnson CS, Dollins MD, Topf JM, Dube MP. Prevalence of proteinuria and the development of chronic kidney disease in HIV-infected patients. Clin Nephrol. 2004;61:1–6. [PubMed]
2. Szczech LA, Hoover DR, Feldman JG, Cohen MH, Gange SJ, Gooze L, et al. Association between renal disease and outcomes among HIV-infected women receiving or not receiving antiretroviral therapy. Clin Infect Dis. 2004;39:1199–1206. [PubMed]
3. Szczech LA, Edwards LJ, Sanders LL, van der Horst C, Bartlett JA, Heald AE, et al. Protease inhibitors are associated with a slowed progression of HIV-related renal diseases. Clin Nephrol. 2002;57:336–341. [PubMed]
4. Kalayjian RC, Franceschini N, Gupta SK, Szczech LA, Mupere E, Bosch RJ, et al. Suppression of HIV-1 replication by antiretroviral therapy improves renal function in persons with low CD4 cell counts and chronic kidney disease. AIDS. 2008;22:481–487. [PMC free article] [PubMed]
5. Brown TT, Cole SR, Li X, Kingsley LA, Palella FJ, Riddler SA, 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. [PubMed]
6. Crane HM, Van Rompaey SE, Kitahata MM. Antiretroviral medications associated with elevated blood pressure among patients receiving highly active antiretroviral therapy. AIDS. 2006;20:1019–1026. [PubMed]
7. Gupta SK. Tenofovir-associated Fanconi syndrome: review of the FDA adverse event reporting system. Aids Patient Care STDS. 2008;22:99–103. [PubMed]
8. Gupta SK, Parker RA, Robbins GK, Dube MP. The effects of highly active antiretroviral therapy on albuminuria in HIV-infected persons: results from a randomized trial. Nephrol Dial Transplant. 2005;20:2237–2242. [PMC free article] [PubMed]
9. Smurzynski M, Collier AC, Koletar S, Bosch RJ, Wu K, Bastow B, et al. AIDS Clinical Trials Group Longitudinal Linked Randomized Trials (ALLRT): Rationale, Design, and Baseline Characteristics. HIV Clin Trials. 2008;9:268–281. [PMC free article] [PubMed]
10. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Am J Kidney Dis. 2002;39:S1–S266. [PubMed]
11. Levey AS, Green T, Kusek JW, Beck GJ, Group MS. A simplified equation to predict glomerular filtration rate from serum creatinine (Abstract A0828) J Am Soc Nephrol. 2000;11
12. Mocroft A, Kirk O, Gatell J, Reiss P, Gargalianos P, Zilmer K, et al. Chronic renal failure among HIV-1-infected patients. AIDS. 2007;21:1119–1127. [PubMed]
13. Jones CA, Francis ME, Eberhardt MS, Chavers B, Coresh J, Engelgau M, et al. Microalbuminuria in the US population: third National Health and Nutrition Examination Survey. Am J Kidney Dis. 2002;39:445–459. [PubMed]
14. Szczech LA, Gange SJ, van der Horst C, Bartlett JA, Young M, Cohen MH, et al. Predictors of proteinuria and renal failure among women with HIV infection. Kidney Int. 2002;61:195–202. [PubMed]
15. Szczech LA, Grunfeld C, Scherzer R, Canchola JA, van der Horst C, Sidney S, et al. Microalbuminuria in HIV infection. AIDS. 2007;21:1003–1009. [PMC free article] [PubMed]
16. Jung O, Bickel M, Ditting T, Rickerts V, Welk T, Helm EB, et al. Hypertension in HIV-1-infected patients and its impact on renal and cardiovascular integrity. Nephrol Dial Transplant. 2004;19:2250–2258. [PubMed]
17. Lucas GM, Lau B, Atta MG, Fine DM, Keruly J, Moore RD. Chronic Kidney Disease Incidence, and Progression to End-Stage Renal Disease, in HIV-Infected Individuals: A Tale of Two Races. J Infect Dis. 2008 [PMC free article] [PubMed]
18. Mauss S, Berger F, Schmutz G. Antiretroviral therapy with tenofovir is associated with mild renal dysfunction. AIDS. 2005;19:93–95. [PubMed]
19. Kopp JB, Smith MW, Nelson GW, Johnson RC, Freedman BI, Bowden DW, et al. MYH9 is a major-effect risk gene for focal segmental glomerulosclerosis. Nat Genet. 2008;40:1175–1184. [PMC free article] [PubMed]
20. Groggel GC, Cheung AK, Ellis-Benigni K, Wilson DE. Treatment of nephrotic hyperlipoproteinemia with gemfibrozil. Kidney Int. 1989;36:266–271. [PubMed]
21. Wyatt CM, Winston JA, Malvestutto CD, Fishbein DA, Barash I, Cohen AJ, et al. Chronic kidney disease in HIV infection: an urban epidemic. AIDS. 2007;21:2101–2103. [PubMed]
22. Gupta SK, Eustace JA, Winston JA, Boydstun II, Ahuja TS, Rodriguez RA, et al. Guidelines for the management of chronic kidney disease in HIV-infected patients: recommendations of the HIV Medicine Association of the Infectious Diseases Society of America. Clin Infect Dis. 2005;40:1559–1585. [PubMed]