Skip Navigation

Low Rates of Antiretroviral Therapy among HIV-Infected Patients with Chronic Kidney Disease

  1. Andy I. Choi1,2,4,
  2. Rudolph A. Rodriguez6,
  3. Peter Bacchetti3,
  4. Paul A. Volberding2,4,
  5. Diane Havlir1,2,
  6. Daniel Bertenthal5,
  7. Alan Bostrom3, and
  8. Ann M. O'Hare6
  1. 1Department of Medicine, San Francisco General Hospital, Departments of Veterans Affairs Research Enhancement Award Program, San Francisco Veterans Affairs Medical Center, San Francisco, California
  2. 2Medicine and Veterans Affairs Research Enhancement Award Program, San Francisco Veterans Affairs Medical Center, San Francisco, California
  3. 3Epidemiology and Biostatistics, University of California, San Francisco, and Veterans Affairs Research Enhancement Award Program, San Francisco Veterans Affairs Medical Center, San Francisco, California
  4. 4Department of Medicine and Veterans Affairs Research Enhancement Award Program, San Francisco Veterans Affairs Medical Center, San Francisco, California
  5. 5Veterans Affairs Research Enhancement Award Program, San Francisco Veterans Affairs Medical Center, San Francisco, California
  6. 6Department of Medicine, Veterans Affairs Puget Sound Healthcare System and University of Washington, Seattle
  1. Reprints or correspondence: Dr. Andy I. Choi, San Francisco General Hospital, Box 1341, Renal Center, Bldg. 100, Rm. 350, 1001 Potrero Ave., San Francisco, CA 94110 (andy.choi{at}ucsf.edu).
 
Next Section

Abstract

Background. It is unknown whether chronic kidney disease (CKD) influences receipt of highly active antiretroviral therapy (HAART) among patients with the human immunodeficiency virus (HIV) and whether prescription practices contribute to excess mortality.

Methods. We conducted a retrospective observational study involving HIV-infected patients with established indications for HAART and an outpatient serum creatinine level measured in the Veterans Affairs health care system. Patients were followed up for the outcomes of HAART exposure (percentage of follow-up time treated with HAART), inadequate dose adjustment of renally eliminated antiretroviral medications, and time to death.

Results. A total of 1041 patients (8.5%) had CKD, defined as an estimated glomerular filtration rate (eGFR) <60 mL/min/1.73 m2. Compared with patients with an eGFR ⩾60 mL/min/1.73 m2, HAART exposure was 14% less (95% confidence interval [CI], 2%–24% less), 24% less (95% CI, 4% more to 45% less), 64% less (95% CI, 38%–79% less), and 49% less (95% CI, 32%–61% less) in patients who had an eGFR of 30–59 mL/min/1.73 m2, 15–29 mL/min/1.73 m2, and <15 mL/min/1.73 m2 (and were not receiving dialysis) and in patients receiving long-term dialysis, respectively. At study entry, 15.4% of patients with CKD received HAART unadjusted for the level of renal function. The adjusted hazard ratio for death was 1.36 (95% CI, 1.08–1.72) for patients with an eGFR of 30–59 mL/min/1.73 m2, 2.17 (95% CI, 1.43–3.27) for patients with an eGFR of 15–29 mL/min/1.73 m2, 5.97 (95% CI, 3.18–11.19) for patients with an eGFR <15 mL/min/1.73 m2, and 1.92 (95% CI, 1.30–2.82) for dialysis-dependent patients. Underexposure and inadequate dose adjustment of HAART were associated with 22.5%–35.5% of the excess mortality found among patients with different levels of CKD.

Conclusions. Underexposure and inadequate dose adjustment of HAART may contribute to excess mortality among HIV-infected patients with CKD.

Although many complications related to HIV infection have become less common in the era of HAART, the number of patients with end-stage renal disease caused by HIV-associated nephropathy doubled from 1995 through 2000, and the prevalence of non–dialysis-dependent chronic kidney disease (CKD), marked by proteinuria, is as high as 30% among HIV-infected persons [1, 2]. CKD is also a strong predictor of mortality among individuals with HIV infection. From 1995 through 1999, the proportion of deaths attributed to renal disease increased from 6.3% to 9.1% among HIV-infected persons nationally [3].

The reasons for increased mortality among HIV-infected persons with CKD are unclear; however, in patients receiving long-term dialysis, under-prescription of HAART and inconsistencies in dosing of renally excreted antiretroviral medications have been suggested as possible contributors to excess mortality [4–7]. To date, very little is known about HAART use among HIV-infected patients with CKD and the influence it may have on mortality in this population. We examined dispensed antiretroviral medications in a national sample of HIV-infected veterans to define the impact of CKD on HAART exposure and dosing and to determine whether these factors may contribute to increased mortality.

Previous SectionNext Section

Patients and Methods

Data sources. The study cohort was assembled by linking data sources from the Department of Veterans Affairs (VA) health care system with several non-VA databases. The Immunology Case Registry is a national VA administrative database designed to monitor health care use among all HIV-infected VA patients [8]. We used the Immunology Case Registry to identify HIV-infected patients and to obtain antiretroviral medication pharmacy fill information, CD4+ cell counts, HIV RNA levels, inpatient admission data, and demographic and comorbidity information. The VA National Patient Care Database, Fee Basis files, and inpatient and outpatient Medicare claims were also used to supplement demographic and comorbidity information from the Immunology Case Registry [9]. Outpatient serum glucose and creatinine measurements were obtained from the VA Decision Support System Laboratory File [9]. Patients receiving long-term dialysis were identified using the United States Renal Data System [10], a comprehensive national end-stage renal disease registry. Vital status was ascertained by the VA Beneficiary Identification and Records Locator System [11–13].

Patients. We identified 12,315 veterans with (1) a confirmed diagnosis of HIV, (2) established indications for HAART, and (3) at least 1 outpatient serum creatinine level measured from 1 October 2001 through 30 September 2002 [14]. Patients entered the study at the time of their first creatinine measurement during the period 1 October 2001–30 September 2002. Indications for HAART were based on International AIDS Society-USA Panel Guidelines Year 2000 as a history of AIDS or laboratory tests within the previous 2-year period that revealed an CD4+ T lymphocyte count <350 cells/µL, an HIV RNA level >30,000 copies/mL, or a CD4+ cell count of 350–500 cells/µL with a HIV RNA level of 5000–30,000 copies/mL [14, 15].

Outcomes. We defined HAART exposure as the percentage of follow-up time during which a patient received HAART. The observation period began at the time of cohort entry and ended at the time of death or the end of follow-up (17 June 2005). HAART was defined as a regimen containing: (1) 2 nucleoside (or nucleotide) reverse-transcriptase inhibitors (NRTIs) with a protease inhibitor or a nonnucleoside reverse-transcriptase inhibitor; (2) the combination of 1 NRTI, 1 protease inhibitor, and 1 nonnucleoside reverse-transcriptase inhibitor; (3) ritonavir and saquinavir given in combination with 1 NRTI; or (4) abacavir and 3 NRTIs in the absence of a protease inhibitor or nonnucleoside reverse-transcriptase inhibitor.

We characterized dosing for all renally eliminated antiretroviral medications (zidovudine, lamivudine, stavudine, didanosine, zalcitabine, emtricitabine, and tenofovir) at the time of cohort entry. This cross-sectional analysis was based on antiretroviral prescriptions filled at a VA pharmacy that were active at the time of creatinine measurement. A medication was characterized as inadequately dose adjusted if the dispensed daily dose was greater than the maximum recommended dose for the patient's level of renal function according to Infectious Disease Society of America guidelines [2]. We examined the time from study entry to death through 17 June 2005.

Primary predictor. The primary predictor variable for all analyses was stage of CKD determined by the estimated glomerular filtration rate (eGFR), which was calculated using the abbreviated Modification of Diet in Renal Disease formula based on age, sex, race, and serum creatinine level [16, 17]. Level of renal function was classified at the time of cohort entry as follows: eGFR ⩾60 mL/min/1.73 m2 (normal or mildly reduced), 30–59 mL/min/1.73 m2 (moderately reduced), 15–29 mL/min/1.73 m2 and not receiving dialysis (severely reduced), <15 mL/min/1.73 m2 (renal failure), and long-term dialysis dependence. To assess for misclassification of level of renal function as a result of outpatient acute renal failure or irregularities in creatinine measurement, we conducted a subgroup analysis among patients with an additional creatinine measurement obtained at least 3 months after cohort entry, in which renal function was classified using the lower of the 2 creatinine levels.

Covariates. Multivariate analyses for both HAART exposure and mortality were adjusted for age, race and/or ethnicity, sex, and the presence of the following diagnosed conditions: hepatitis C virus infection, diabetes (defined as a diagnosis of diabetes or random serum glucose level >200 mg/dL during the 3 months before cohort entry), coronary artery disease, congestive heart failure, hypertension, atrial fibrillation, depression, drug abuse, dementia, alcohol abuse, AIDS, peripheral arterial disease, chronic obstructive pulmonary disease, and cerebrovascular disease. Comorbidities were designated using International Classification of Diseases, Ninth Revision, codes and Current Procedural Terminology codes [18–20]. Mortality estimates were additionally adjusted for CD4+ cell count and HIV RNA levels within the 6 months prior to and closest to cohort entry, inadequate dose adjustment of HAART, and the percentage of time that a patient was exposed to HAART.

Statistical analyses. Baseline characteristics were compared by level of eGFR using the χ2 test for categorical variables and analysis of variance for continuous variables. We examined the association between level of eGFR and HAART exposure using negative binomial regression. Although this method is usually used for count data, it can be applied to nonnegative numeric outcomes, and this parametric form provided the best match to the distribution of our outcome, which was very skewed, was nonnegative, and had many zeros. In addition to the aforementioned covariates, we also adjusted for the number of hospitalizations that occurred during follow-up to account for periods when a patient would not be expected to fill outpatient prescriptions. Overall P values for CKD categories were obtained using the Wald test. To determine whether patients who live longer have greater opportunity to obtain HAART (i.e., survivor treatment selection bias [21]), we repeated analyses among patients who were alive at the end of follow-up. Because indications for HAART evolved over the period of observation, we also conducted a sensitivity analysis using only patients with an established diagnosis of AIDS at baseline. The association between level of eGFR and inadequate dose adjustment of HAART was tested using the χ2 test.

We measured the association between level of eGFR and time to death using Cox proportional hazards models. These analyses were restricted to 7524 patients for whom complete CD4+ cell count and HIV RNA data were available. We examined the incremental effect of HAART exposure and inadequate dose adjustment, respectively, on the hazard ratio for death after controlling for all covariates listed above. The effect of HAART exposure was analyzed using an intention to treat approach [21]. To assess the mortality risk attributable to underexposure and inadequate dose adjustment of HAART among patients with CKD, we calculated the proportion of treatment effect explained by these covariates [22]. Bootstrap confidence intervals for the proportion of treatment effect explained were obtained using the bias-corrected percentile method. To determine the impact of missing data on results, mortality was reanalyzed among the group of patients excluded because of missing information. The proportional hazards assumption was checked using the Schoenfeld test and by comparing estimated log (-log [survivor function]) versus time survivor curves.

All regression models were checked using bootstrap methods and by obtaining robust standard error estimates. Analyses were conducted using Stata, version 9.0 (Stata). This study was approved by the Committee on Human Research at the University of California, San Francisco, and the VA San Francisco.

Previous SectionNext Section

Results

There were 1041 patients (8.5%) with CKD (eGFR <60 mL/min/1.73 m2) at the time of cohort entry (table 1). A total of 727 patients (5.9%) had an eGFR of 30–59 mL/min/1.73 m2, 109 (0.3%) had an eGFR of 15–29 mL/min/1.73 m2, 41 (0.3%) had an eGFR of <15 mL/min/1.73 m2, and 164 (1.3%) were dialysis dependent. A total of 6270 patients (51%) were black, and 5414 (44%) were white. At the time of entry into the study, 1729 patients (14.0%) were receiving HAART. Lower levels of HAART exposure at baseline were found among patients with CKD, with the lowest levels found among patients with an eGFR <15 mL/min/1.73 m2 (4.9%). Compared with patients with normal renal function, patients with CKD were more likely to be older; to be black; to have diagnosed diabetes, cardiovascular complications, dementia, hepatitis C, and AIDS; and to have lower CD4+ cell counts.

Table 1
View larger version:
Table 1

Demographic and clinical characteristics of the study cohort at baseline, by renal function category.

HAART exposure. Overall, there were 16,405 total patient-years of antiretroviral treatment accounted for in the analysis. The median percentage of time treated with HAART was 4.8% (interquartile range [IQR], 0%–15.3%) among patients with normal renal function (eGFR ⩾60 mL/min/1.73 m2), 4.0% (IQR, 0%–13.3%) for patients with an eGFR of 30–59 mL/min/1.73 m2, 0% (IQR, 0%–10.1%) for patients with an eGFR of 15–29 mL/min/1.73 m2, 0% (IQR, 0%–3.8%) for patients with an eGFR <15 mL/min/1.73 m2, and 0% (IQR, 0%–7.3%) for dialysis-dependent patients. There was a strong relationship between level of renal function and HAART exposure that persisted after adjustment for patient characteristics (table 2). Notably, those with an eGFR <15 mL/min/1.73 m2 received only 0.36 (95% CI, 0.21–0.62) times as much HAART as those with an eGFR ⩾60 mL/min/1.73 m2. These estimates did not change substantially when the analysis was restricted to only those who survived the entire period of observation or to those with an established diagnosis of AIDS at baseline.

Table 2
View larger version:
Table 2

Relative proportion of time receiving HAART associated with chronic kidney disease.

HAART dosing. A total of 10,315 prescriptions for renally excreted NRTIs were dispensed by a VA pharmacy at the time of cohort entry. Dosing could be ascertained for 6768 of these prescriptions. For 227 (3.4%) of these medications, the prescribed dose exceeded the maximum recommended dose for the patient's level of renal function. Many of these dosing errors occurred among the same patients, with 160 individuals with CKD (15.4%) receiving at least 1 NRTI prescribed at an incorrect dose. Incorrect dosing occurred in 26 (0.2%) of the patients with an eGFR ⩾60 mL/min/1.73 m2, 105 (14.4%) of the patients with an eGFR of 30–59 mL/min/1.73 m2, 24 (20.2%) of the patients with an eGFR of 15–29 mL/min/1.73 m2, 7 (17.1%) of the patients with an eGFR <15 mL/min/1.73 m2, and 24 (14.6%) of the patients receiving long-term dialysis (P < .001 for those with an eGFR ⩾60 mL/min/1.73 m2 vs. all others). Underdosing of HAART was found in only 3 patients.

Mortality. Patients were followed up for a total of 53,503 person-years, with a median observation time of 3.4 years (IQR, 3.1–3.6 years). There were 308 deaths (29.6%) among patients with CKD (eGFR <60 mL/min/1.73 m2) and 1354 deaths (12.0%) in patients with an eGFR ⩾60 mL/min/1.73 m2. In adjusted analyses, we observed a clear relationship between level of renal function and hazard for death (table 3). Inadequate dose adjustment for HAART at baseline was associated with an increased risk of death (hazard ratio, 1.77; 95% CI, 1.27–2.46). The hazard ratios for death associated with CKD categories were attenuated when adjusted for clinical and demographic variables, underexposure, and inadequate dose adjustment of HAART (table 3). Point estimates for the proportion of CKD effect explained by underexposure and inadequate dose adjustment of HAART ranged from 22.2% to 35.5%, depending on the category of CKD. Among patients with missing CD4+ cell counts or HIV RNA levels, the proportion of treatment effect explained was similar but with wide confidence intervals: 9.8% (95% CI, -9.4% to 56.3%) for an eGFR of 30–59 mL/min/1.73 m2, 32.2% (95% CI, 9.7%–67.0%) for an eGFR of 15–29 mL/min/1.73 m2, 20.5% (95% CI, -626.5% to 96.3% ) for an eGFR <15 mL/min/1.73 m2, and 14.8% (95% CI -40.5% to 83.0%) for patients receiving dialysis. Results were similar for all outcomes when reanalyzed using CKD categories defined by the lower of 2 creatinine levels measured 3 months later.

Table 3
View larger version:
Table 3

Adjusted hazard ratios for mortality associated with chronic kidney disease (CKD), by renal function category.

Previous SectionNext Section

Comment

In this large, national cohort of HIV-infected veterans with established indications for antiretroviral therapy, patients with CKD were less likely to receive HAART than were patients with normal renal function. Among those receiving antiretroviral therapy at baseline, inadequate NRTI dose adjustment for level of renal function was common among patients with CKD. Collectively, underexposure and inadequate dose adjustment of HAART were associated with 22.2%–35.5% of excess mortality risk among patients with CKD, depending on the level of renal function.

Our results are consistent with those of earlier cross-sectional studies reporting considerable variability in NRTI dosing and low rates of HAART use among patients receiving long-term dialysis [4–7, 23]. To our knowledge, this study is the first to demonstrate that patients with HIV infection and non–dialysis-dependent CKD are less likely to receive HAART when indications exist, compared with their counterparts with normal renal function. There are several possible reasons that HIV-infected patients with CKD were less likely to receive HAART. Many antiretroviral medications are renally cleared, providing a greater opportunity for alterations in adverse effect profiles and complex drug interactions [24]. The relatively high frequency of inadequate NRTI dose adjustment found among patients with CKD raises the question of whether such practices contribute to the development of drug toxicity, leading to regimen changes, intermittent adherence, and HAART intolerance [25]. Competing medication priorities and comorbid illnesses among patients with CKD may also limit HAART exposure. In this cohort, patients with renal disease were more likely to have AIDS, diabetes, and cardiovascular complications. Therefore, concurrent medical issues may take precedence in these patients leading to higher rates of underexposure to HAART or nonadherence related to increased pill burden.

Our study has important implications for clinical practice. Recent Infectious Diseases Society of America guidelines identify HIV-related renal disease as an indication for HAART [2]. Because HIV-related kidney diseases account for ∼60% of cases in a renal biopsy series of HIV-infected persons [26, 27], our findings demonstrate the need to investigate low rates of HAART exposure in patients with CKD [28]. Furthermore, we observed an attenuation of mortality risk after adjustment for underexposure and inadequate dose adjustment of HAART. This raises the question of whether interventions to improve HAART prescribing practices among patients with CKD could substantially improve survival.

This study has the following limitations. We relied on VA pharmacy records to characterize HAART exposure; therefore, patients receiving medications outside of the VA system would be misclassified as not receiving HAART in this study. We believe that the effect of missing pharmacy data was unlikely to have influenced study results [29], because veterans lack financial motivation to obtain medications elsewhere [30], VA health care coverage is regularly used only for prescription benefits [31], and misclassification is unlikely to occur differentially by level of renal function. Overall rates of HAART exposure were low, but they were consistent with previous reports when the same patient-selection criteria and definition for HAART were applied [32, 33]. Our results are striking, considering that all patients received care at a VA facility, an equal access health care system in which the effect of a patient's ability to pay for medications or access to a prescribing physician should be minimized.

Although we adjusted for a large number of sociodemographic and clinical characteristics, our results may reflect residual confounding attributable to factors, such as homelessness, psychiatric comorbidity, and substance abuse, that may not be completely captured by diagnostic codes [34–38]. Nonetheless, the magnitude of the adjusted association between CKD with exposure to HAART makes this less likely, and previous literature does not clearly support the association of these factors with renal function.

The Modification of Diet in Renal Disease equation has not been independently validated as a measure of glomerular filtration rate among HIV-infected persons. Nevertheless, the prognostic significance of eGFR among patients with HIV infection is relevant to clinical practice, because estimation of glomerular filtration rate using the Modification of Diet in Renal Disease equation has been widely incorporated into routine clinical care and is specifically recommended by Infectious Diseases Society of America Guidelines for assessment of renal function in the HIV-infected population [2, 39].

The assessment of NRTI dosing was limited to a fixed point in time and may not reflect dynamic dosing changes made in response to changes in renal function. Clinicians may have had reasons for over- or under-dosing NRTIs that we were unable to discern from our data sources, such as accommodating interactions with other medications. It is more likely, however that many cases of unadjusted dosing were missed, because most drug interactions require dose reductions [40]. Results were also confirmed when subsequent creatinine measurements 3 months later were used in sensitivity analyses.

Unfortunately, information on drug-related symptoms, provider expertise, site characteristics, or cause of death was not available to us. Therefore, we were unable to characterize the role of these factors in mediating prescription practices and mortality in the setting of CKD. Further investigation is needed to determine if deaths were related to CKD itself, cardiovascular complications, AIDS, or adverse medication events.

Previous SectionNext Section

Conclusion

In this large national cohort of HIV-infected veterans with established indications for HAART, patients with CKD were less likely to receive HAART than were patients with healthy renal function. Furthermore, when patients with CKD did receive antiretroviral medications, renally excreted NRTIs were frequently prescribed at a dose that was inadequately adjusted for level of renal function. Underexposure and inadequate dose adjustment of HAART was associated with 22.2%–35.5% of excess mortality observed among patients with CKD. These findings identify HAART prescription practices as a target for quality improvement among HIV-infected patients with CKD.

Previous SectionNext Section

Acknowledgments

Financial support. National Institutes of Health (T32-DK07219, K24-AI51982, P30-AI27763, and K23-AG028980-01), the Veterans Affairs Health Services Research and Development Career Development Award, the National Kidney Foundation Fellowship Grant, Veterans Affairs San Francisco Research Enhancement Award Program, and the Veterans Affairs Public Health Strategic Health Care Group.

Potential conflicts of interest. P.A.V. has been a consultant for Pfizer, GlaxoSmithKline, Bristol-Myers Squibb, Gilead, Boehringer Ingelheim, Merck, and Schering-Plough; has received honoraria from Gilead and Bristol-Myers Squibb; and owns stock in Immune Response. All other authors: no conflicts.

  • Received April 15, 2007.
  • Revision received August 14, 2007.
  • Accepted November 6, 2007.
Previous Section
 

References

  1. ↵
    United States Renal Data System 2001 annual data report. Bethesda, MD: National Institutes of Health and the National Institute of Diabetes and Digestive and Kidney Diseases; 2001.
  2. ↵
    1. Gupta SK,
    2. Eustace JA,
    3. Winston JA,
    4. 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;40:1559-85.
    FREE Full Text
  3. ↵
    1. Selik RM,
    2. Byers RH Jr,
    3. Dworkin MS
    . Trends in diseases reported on U.S. death certificates that mentioned HIV infection, 1987–1999. J Acquir Immune Defic Syndr 2002;29:378-87.
    MedlineWeb of Science
  4. ↵
    1. Rodriguez RA,
    2. Mendelson M,
    3. O'Hare AM,
    4. Hsu LC,
    5. Schoenfeld P
    . Determinants of survival among HIV-infected chronic dialysis patients. J Am Soc Nephrol 2003;14:1307-13.
    Abstract/FREE Full Text
    1. Szczech LA,
    2. Kalayjian R,
    3. Rodriguez R,
    4. Gupta S,
    5. Coladonato J,
    6. Winston J
    . The clinical characteristics and antiretroviral dosing patterns of HIV-infected patients receiving dialysis. Kidney Int 2003;63:2295-301.
    CrossRefMedlineWeb of Science
    1. Abbott KC,
    2. Trespalacios FC,
    3. Agodoa LY,
    4. Ahuja TS
    . HIVAN and medication use in chronic dialysis patients in the United States: analysis of the USRDS DMMS Wave 2 study. BMC Nephrol 2003;4:5-0000.
    CrossRefMedline
  7. ↵
    1. Izzedine H,
    2. Launay-Vacher V,
    3. Deray G
    . Inadequacy of antiretroviral drugs dosage adjustment in HIV patients receiving dialysis. Kidney Int 2003;64:2324-0000.
    MedlineWeb of Science
  8. ↵
    1. Backus L,
    2. Mole L,
    3. Chang S,
    4. Deyton L
    . The immunology case registry. J Clin Epidemiol 2001;54(Suppl 1):12-5.
    CrossRef
  9. ↵
    1. Maynard C,
    2. Capko MK
    . Data resources in the Department of Veterans Affairs. Diabetes Care 2004;27(Suppl 2):22-6.
    CrossRef
  10. ↵
    1. Collins AJ,
    2. Kasiske B,
    3. Herzog C,
    4. et al
    . Excerpts from the United States Renal Data System 2003 annual data report: atlas of end-stage renal disease in the United States. Am J Kidney Dis 2003;42(6 Suppl 5):5-7. S1-320.
  11. ↵
    1. Cowper DC,
    2. Kubal JD,
    3. Maynard C,
    4. Hynes DM
    . A primer and comparative review of major US mortality databases. Ann Epidemiol 2002;12:462-8.
    CrossRefMedlineWeb of Science
    1. Fisher SG,
    2. Weber L,
    3. Goldberg J,
    4. Davis F
    . Mortality ascertainment in the veteran population: alternatives to the National Death Index. Am J Epidemiol 1995;141:242-50.
    Abstract/FREE Full Text
  13. ↵
    1. Page WF,
    2. Braun MM,
    3. Caporaso NE
    . Ascertainment of mortality in the U.S. veteran population: World War II veteran twins. Mil Med 1995;160:351-5.
    MedlineWeb of Science
  14. ↵
    1. Centers for Diease Control Prevention
    . 1993 Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR Recomm Rep 1992;41(RR-17):1-19.
    Medline
  15. ↵
    1. Carpenter CC,
    2. Cooper DA,
    3. Fischl MA,
    4. et al
    . Antiretroviral therapy in adults: updated recommendations of the International AIDS Society-USA Panel. JAMA 2000;283:381-90.
    Abstract/FREE Full Text
  16. ↵
    1. Levey AS,
    2. Bosch JP,
    3. Lewis JB,
    4. Greene T,
    5. Rogers N,
    6. Roth D
    . A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 1999;130:461-70.
    Abstract/FREE Full Text
  17. ↵
    1. Levey AS,
    2. Coresh J,
    3. Balk E,
    4. Kausz AT,
    5. Levin A,
    6. Steffes MW,
    7. et al
    . National Kideny Foundation practice guidelines for chronic kidney disease: evaluation, classificiation, and stratification. Ann Intern Med 2003;139:137-47.
    Abstract/FREE Full Text
  18. ↵
    1. Borzecki AM,
    2. Wong AT,
    3. Hickey EC,
    4. Ash AS,
    5. Berlowitz DR
    . Identifying hypertension-related comorbidities from administrative data: what's the optimal approach? Am J Med Qual 2004;19:201-6.
    Abstract/FREE Full Text
  19. Description of ICD-9 comorbidity codes used by the Veterans Aging Cohort Study. 2007. Available at: http://www.vacohort.org/ICD9_Codes.html. Accessed 15 April 2007.
  20. ↵
    1. O'Hare AM,
    2. Bertenthal D,
    3. Shlipak MG,
    4. Sen S,
    5. Chren MM
    . Impact of renal insufficiency on mortality in advanced lower extremity peripheral arterial disease. J Am Soc Nephrol 2005;16:514-9.
    Abstract/FREE Full Text
  21. ↵
    1. Glesby MJ,
    2. Hoover DR
    . Survivor treatment selection bias in observational studies: examples from the AIDS literature. Ann Intern Med 1996;124:999-1005.
    Abstract/FREE Full Text
  22. ↵
    1. Freedman LS,
    2. Graubard BI,
    3. Schatzkin A
    . Statistical validation of intermediate endpoints for chronic diseases. Stat Med 1992;11:167-78.
    MedlineWeb of Science
  23. ↵
    1. Ahuja TS,
    2. Grady J,
    3. Khan S
    . Changing trends in the survival of dialysis patients with human immunodeficiency virus in the United States. J Am Soc Nephrol 2002;13:1889-93.
    Abstract/FREE Full Text
  24. ↵
    1. Blanchard JN,
    2. Wohlfeiler M,
    3. Canas A,
    4. King K,
    5. Lonergan JT
    . Pancreatitis with didanosine and tenofovir disoproxil fumarate [corrected]. Clin Infect Dis 2003;37:e57-62.
    Abstract/FREE Full Text
  25. ↵
    1. Servoss JC,
    2. Kitch DW,
    3. Andersen JW,
    4. Reisler RB,
    5. Chung RT,
    6. Robbins GK
    . Predictors of antiretroviral-related hepatotoxicity in the Adult AIDS Clinical Trial Group 1989 through 1999. J Acquir Immune Defic Syndr 2006;43:320-3.
    CrossRefMedlineWeb of Science
  26. ↵
    1. Ross MJ,
    2. Klotman PE
    . Recent progress in HIV-associated nephropathy. J Am Soc Nephrol 2002;13:2997-3004.
    FREE Full Text
  27. ↵
    1. Szczech LA,
    2. Gupta SK,
    3. Habash R,
    4. et al
    . The clinical epidemiology and course of the spectrum of renal diseases associated with HIV infection. Kidney Int 2004;66:1145-52.
    CrossRefMedlineWeb of Science
  28. ↵
    1. Schwartz EJ,
    2. Szczech LA,
    3. Ross MJ,
    4. Klotman ME,
    5. Winston JA,
    6. Klotman PE
    . Highly active antiretroviral therapy and the epidemic of HIV+ end-stage renal disease. J Am Soc Nephrol 2005;16:2412-20.
    Abstract/FREE Full Text
  29. ↵
    1. Justice AC,
    2. Dombrowski E,
    3. Conigliaro J,
    4. et al
    . Veterans Aging Cohort Study (VACS): overview and description. Med Care 2006;44(Suppl 2):13-24.
    CrossRef
  30. ↵
    Veterans' health care: use of VA services by Medicare-eligible veterans. Washington, DC: United States General Accounting Office; 1994.
  31. ↵
    1. Shen Y,
    2. Hendricks A,
    3. Zhang S,
    4. Kazis LE
    . VHA enrollees' health care coverage and use of care. Med Care Res Rev 2003;60:253-67.
    Abstract/FREE Full Text
  32. ↵
    1. Sulkowski MS,
    2. Moore RD,
    3. Mehta SH,
    4. Chaisson RE,
    5. Thomas DL
    . Hepatitis C and progression of HIV disease. JAMA 2002;288:199-206.
    Abstract/FREE Full Text
  33. ↵
    1. Kilbourne AM,
    2. Justice AC,
    3. Rollman BL,
    4. et al
    . Clinical importance of HIV and depressive symptoms among veterans with HIV infection. J Gen Intern Med 2002;17:512-20.
    CrossRefMedlineWeb of Science
  34. ↵
    1. Cook RL,
    2. Sereika SM,
    3. Hunt SC,
    4. Woodward WC,
    5. Erlen JA,
    6. Conigliaro J
    . Problem drinking and medication adherence among persons with HIV infection. J Gen Intern Med 2001;16:83-8.
    CrossRefMedlineWeb of Science
    1. Singh N,
    2. Squier C,
    3. Sivek C,
    4. Wagener M,
    5. Nguyen MH,
    6. Yu VL
    . Determinants of compliance with antiretroviral therapy in patients with human immunodeficiency virus: prospective assessment with implications for enhancing compliance. AIDS Care 1996;8:261-9.
    CrossRefMedlineWeb of Science
    1. Walsh JC,
    2. Dalton M,
    3. Gazzard BG
    . Adherence to combination antiretroviral therapy assessed by anonymous patient self-report. AIDS 1998;12:2361-3.
    MedlineWeb of Science
    1. Bangsberg D,
    2. Tulsky JP,
    3. Hecht FM,
    4. Moss AR
    . Protease inhibitors in the homeless. JAMA 1997;278:63-5.
    Abstract/FREE Full Text
  38. ↵
    1. Singh N,
    2. Berman SM,
    3. Swindells S,
    4. et al
    . Adherence of human immunodeficiency virus-infected patients to antiretroviral therapy. Clin Infect Dis 1999;29:824-30.
    MedlineWeb of Science
  39. ↵
    K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Am J Kidney Dis 2002;39(2 Suppl 1):1-266.
  40. ↵
    Guidelines for the use of antiretroviral agents in HIV-infected adults and adolescents 4 May 2006. Available at: http://www.aidsinfo.nih.gov/ContentFiles/AdultandAdolescentGL.pdf. Accessed 5 September 2006.
| Table of Contents

This Article

  1. Clin Infect Dis. (2007) 45 (12): 1633-1639. doi: 10.1086/523729
  1. AbstractFree
  2. » Full Text (HTML)Free
  3. Full Text (PDF)Free

Classifications

Share

  1. Email this article

Navigate This Article

Current Issue

  1. March 1, 2012 54 (5)
  1. Clinical Infectious Diseases
  1. Alert me to new issues

Most