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

1
The Effects of HIV-1 Viral Suppression and
Non-Viral Factors on Clinically Significant
Proteinuria in the HAART Era Samir Gupta, Nora
Franceschini, Lynda Szczech, Marlene Smurzynski,
Robert Kalayjian, and the ACTG ALLRT Study Team
974
Contact Information SK Gupta, MD, MS Division of
Infectious Diseases, Indiana University School of
Medicine Wishard Hospital 1001 W. 10th
Street West Building 430 Indianapolis, IN, 46202,
USA sgupta1_at_iupui.edu
ABSTRACT (Revised) Background Proteinuria is a
marker for progressive HIV-related chronic kidney
disease. However, predictors of proteinuria in
the HAART era are unknown. Methods Urine
protein to creatinine ratios (P/Cr) were measured
annually since 2002 in the ALLRT, a longitudinal
observational cohort study of subjects
prospectively randomized to receive HAART
regimens or treatment strategies in select ACTG
trials. Predictors of higher P/Cr were assessed
using multivariable linear mixed effects
models. Results Initial P/Cr was measured at
least once in 2,806 ALLRT participants (83 male,
28 black, 8 reported history of DM, 21
reported history of HTN, 67 were ARV-naïve at
baseline, 11 were HCV AB, and 5 were HBV AG).
Baseline characteristics included median (IQR)
estimated GFR 98.4 (85.6, 115.6) mL/min/1.73m2,
age 39 (33, 46) yrs, CD4 count 233 (98, 385)
cells/µL, and log10HIV-1 RNA 4.64 (4.01, 5.27)
c/mL. Median (IQR) time from baseline to initial
P/Cr was 1.9 (0.9, 4.5) years, and median (IQR)
number of P/Cr measurements was 3 (2, 4). Median
(IQR) initial P/Cr was 0.094 (0.063, 0.155) 17
had P/Cr gt0.2. Change in P/Cr depended on initial
values (interaction P0.005) with an improvement
(Plt0.0001) in those with initial P/Cr gt0.2 and an
increase (Plt0.0001) in those with initial P/Cr
0.2. Predictors of P/Cr gt0.2 included older age,
lower current CD4 count, lower current eGFR,
higher current log10HIV-1 RNA level, reported
history of diabetes, HCV AB, and ever use of
indinavir. Conclusions Changes in quantitative
proteinuria depended on initial levels in this
HIV-infected cohort. Current CD4, viral load, and
GFR (but not baseline levels) significantly
predicted higher P/Cr. Several non-HIV factors
also exist to identify those more likely to have
proteinuria and who should be treated
aggressively to prevent progressive kidney
disease.

• BACKGROUND
• Proteinuria predicts progressive kidney disease,
  new AIDS-defining illness, and overall mortality
  in the HIV-infected population
• Reduction in proteinuria with viral suppression
  may be offset by antiretroviral toxicity or
  co-morbidities (diabetes, hypertension, viral
  hepatitis)
• Quantitative proteinuria has higher specificity
  and lower detection limits compared to dipstick
  measurements
• The determinants of quantitative proteinuria
  over time in the HAART era are not well-defined

• DISCUSSSION
• Changes in quantitative proteinuria
  significantly depended on initial measurement of
  P/Cr
• Significant predictors of higher P/Cr included
  current CD4, viral load, GFR, hepatitis C,
  history of diabetes, and previous use of
  indinavir
• Black race, female sex, history of
  antihypertensive drug use, and current tenofovir
  use were marginally associated with higher P/Cr
• Baseline GFR, CD4, and viral load did not affect
  current P/Cr levels
• Data limited by not having baseline proteinuria
  levels
• Improvement in proteinuria will depend on
  controlling both HIV and non-HIV factors

• CHANGES IN PROTEINURIA OVER TIME
• Proteinuria in overall cohort did not change
• Significant interaction between level of
  initially measured P/Cr and change over time (P
  for interaction 0.005)
• For initial P/Cr gt 0.2, P/Cr declined by 0.03
  (SE 0.01)g/g per year (Plt0.0001)
• For initial P/Cr 0.2, P/Cr increased by 0.009
  (SE 0.001)g/g per year (Plt0.0001)
• Proportions of those with P/Cr gt 0.2 at each
  time point remained nearly 17 at each time point

• PREDICTORS OF HIGHER PROTEINURIA
• Variable ß estimate (SE) P
• Age per year 0.009 (0.0003) 0.04
• Black race (vs. non-Black) 0.01 (0.008) 0.09
• Female sex 0.02 (0.009) 0.09
• Current CD4
• per cell/µL -0.00002 (0.00001) 0.05
• Current GFR per
• mL/min/1.732 -0.0005 (0.0002) 0.01
• Current log10viral load 0.008 (0.003) 0.02
• History of diabetes 0.04 (0.01) 0.01
• History of anti-HTN meds 0.02 (0.01) 0.06
• HCV AB 0.03 (0.01) 0.03
• Current tenofovir use 0.03 (0.02) 0.06
• Ever used indinavir 0.03 (0.01) 0.001
• Year of follow-up 0.002 (0.001) 0.2
• Baseline CD4, baseline viral load, baseline GFR,
  history of ARV use prior to parent entry, HBV
  AG, ever use of tenofovir, current use of
  indinavir, ever/current use of adefovir, and
  non-HDL-C were not entered into the final model
  due to lack of univariable association

• METHODS
• Quantitative spot urine protein to creatinine
  ratios (P/Cr) were measured annually since 2002
  in subjects enrolled into the ALLRT study, an
  observational cohort of participants of various
  parent ACTG randomized trials of treatment
  strategies
• Subjects with at least one P/Cr (N2806) were
  included
• Baseline defined as time of parent study
  initiation
• Multivariable mixed effects linear models
  (significant skewing of data required using
  square root-transformation for P/Cr ) were
  constructed to assess predictors of higher
  proteinuria

ACKNOWLEDGEMENTS This work was supported in part
by NIH/NIAID grants to the AIDS Clinical Trials
Group (AI68636) and the SDAC/Harvard School of
Public Health (AI38855).