Clinical Science: Concise Communication
Viral load as an independent risk factor for opportunistic infections in HIV-infected adults and adolescents
Kaplan, Jonathan E.; Hanson, Debra L.; Jones, Jeffrey L.; Dworkin, Mark S.*; and the Adult and Adolescent Spectrum of HIV Disease Project Investigators
From the Division of HIV/AIDS Prevention–Surveillance and Epidemiology, National Center for HIV, STD, and TB Prevention, Centers for Disease Control and Prevention, Atlanta, Georgia. *See Appendix.
Note: This material was presented in part at the 39th Interscience Conference on Antimicrobial Agents and Chemotherapy. San Francisco, September 1999 [abstract 124].
Requests for reprints to: J. E. Kaplan, Division of HIV/AIDS Prevention, Mailstop D-21, Centers for Disease Control and Prevention, Atlanta, GA 30333, USA.
Received: 25 October 2000;
revised: 18 May 2001; accepted: 24 May 2001.
Objective: We investigated whether HIV plasma RNA (viral load; VL) predicts risk for opportunistic infections (OI) in HIV-infected persons, independent of CD4 lymphocyte count and other factors that might affect disease outcome.
Methods: Among persons who had initiated antiretroviral therapy (ART), we studied the risk for OI following a VL measurement in the Centers for Disease Control and Prevention Adult and Adolescent Spectrum of HIV Disease (ASD) Project, a medical record review study of HIV-infected persons in 11 US cities. Analysis was limited to persons who had initiated ART and who had VL data, primarily from the period 1996–1999. Persons were considered at risk for OI for 1 to 6 months after a given VL; risk for OI was assessed using a Poisson multiple regression model controlling for CD4 lymphocyte count, ART, and other variables potentially associated with development of OI: history of AIDS OI, age, sex, race, HIV risk category, OI prophylaxis, and calendar year.
Results: Although decreasing CD4 count was the strongest predictor of risk for OI [relative risk (RR), 13.3 for persons with CD4 lymphocyte count < 50 × 106/l compared with persons with CD4 lymphocyte count ≥ 500 × 106/l], increasing VL was independently associated with increased risk [RR, 1.6, 1.9, 2.7, and 3.5 for VL of 7000–19 999, 20 000–54 999, 55 000–149 999, and ≥ 150 000 copies/ml (by reverse transcription–PCR), respectively, compared with VL < 400]. Similar results were obtained when the risk period was reduced to 5, 4, 3, and 2 months after VL measurement.
Conclusions: VL is an independent risk factor for OI and should be considered in special situations, such as in decisions to discontinue primary or secondary OI prophylaxis after CD4 lymphocyte counts have increased in response to ART.
Since the beginning of the AIDS epidemic, many studies have addressed risk factors for development of AIDS-associated opportunistic infections (OI); decreasing CD4 lymphocyte count, an indicator of increasing immunodeficiency, has consistently been shown to be the strongest predictor of risk for OI [1–3]. Since the advent of highly active antiretroviral therapy (HAART), HIV plasma RNA, or viral load (VL), which is the best predictor of long-term clinical outcome and is used to monitor the effectiveness of antiretroviral therapy (ART), is now used routinely in clinical management [4–6]. Whether measurement of VL might also be useful in predicting risk of more immediate outcomes, such as development of OI, has received less attention [7–11]. We assessed the utility of VL as an independent risk factor for OI by analyzing data from the Centers for Disease Control and Prevention's (CDC) Adult and Adolescent Spectrum of HIV Disease (ASD) Project, a medical record review surveillance project that has been conducted in 11 US cities since 1990.
The methods for ASD have been described previously . Briefly, since 1990, HIV-infected persons 13 years of age or older have been identified and enrolled in ASD upon initiation of care at selected facilities in Atlanta, Dallas, Denver, Detroit, Los Angeles, New York, San Antonio, Seattle, Houston, New Orleans, and Bayamon, Puerto Rico, regardless of the stage of HIV infection. Patients’ medical records for the year before enrolment are reviewed for information on demographic characteristics, mode of HIV exposure, any history of AIDS-defining OI, and other clinical and laboratory findings. During 6 month follow-up intervals, the records are reviewed for information about AIDS-defining conditions, other illnesses, prescription medications, and laboratory tests. From 1990 through December 1999, more than 50 000 persons were enrolled in ASD.
CD4 and VL measurements and prescription of ART
CD4 lymphocyte counts have been recorded since the beginning of the project; beginning in 1997, VL measurements were abstracted for patients followed up during 1995 and later. For both CD4 lymphocyte count and VL, the highest, lowest, and most recent values are recorded during each 6 month interval, with the dates of these tests. The exact VL measurement (or the lower limit of detectability if VL is undetectable), in copies/ml, is recorded, with the type of test used [reverse transcription (RT)–PCR or branched-chain DNA (bDNA)]. Prescription of antiretroviral drugs is recorded for each 6 month interval.
Risk factor analysis
Because the focus of this analysis was risk of OI as a function of VL, the analysis included all persons who had initiated ART, had at least one VL measurement recorded in 1995 or later, and had at least one month of follow time after a VL measurement. The endpoints in the analysis were diagnoses of new (not previously diagnosed) AIDS-defining OI, according to the 1993 CDC AIDS case definition . Risk for OI was assessed using a Poisson regression model. The model initially included factors that might influence development of OI, namely CD4 lymphocyte count, VL, ART, history of AIDS OI, age, sex, race, HIV risk mode, prophylaxis against Pneumocystis carinii pneumonia (PCP) and disseminated Mycobacterium avium complex (MAC) disease, and calendar year. The final model used to estimate risk associated with specific ranges of VL was reduced to include only factors significant in the preliminary model.
VL measurements were categorized according to the Department of Health and Human Services/Henry J. Kaiser Foundation guidelines for use of ART as follows (copies/ml): < 400 (RT–PCR) or < 500 (bDNA); 400–1499 (RT–PCR) or 500–999 (bDNA); 1500–6999 (RT–PCR) or 1000–2999 (bDNA); 7000–19 999 (RT–PCR) or 3000–9999 (bDNA); 20 000–54 999 (RT–PCR) or 10 000–29 999 (bDNA); 55 000–149 999 (RT–PCR) or 30 000–99 999 (bDNA); and ≥ 150 000 (RT–PCR) or ≥ 100 000 (bDNA) . Persons were considered at risk for newly diagnosed OI at a given VL for 1–6 months after that particular measurement, unless another measurement intervened, in which case the person was considered at risk for 1–6 months at that VL level. Follow up time was censored after 6 months if no intervening VL was available, at death, at loss to follow-up, or in December 1999, the cut-off date for the analysis.
Each person-month of observation at a given VL level was associated with the most recent CD4 lymphocyte count within the previous 6 months. CD4 lymphocyte counts were grouped as follows: 0–49 × 106/l, 50–199 × 106/l, 200–349 × 106/l, 350–499 × 106/l, or ≥ 500 × 106/l, or missing (if no CD4 measurements were available within the previous 6 months). Each person-month was also associated with prescription of an antiretroviral regimen, as recorded during the previous 6 month interval, as follows: HAART [two nucleoside reverse transcriptase inhibitors (NRTI) plus at least one protease inhibitor or non-nucleoside reverse transcriptase inhibitor], dual therapy (two NRTI), other ART (any other regimen, e.g., a single NRTI plus a protease inhibitor, NRTI monotherapy), or no ART. Each person-month was also associated with presence or absence of a history of an AIDS-defining OI.
A total of 10 885 persons with at least 1 month of follow time after initiation of ART and after a VL measurement were included in the analysis. Of these persons, 75% were male; 35% were White, 42% were African–American, and 22% were Hispanic; the median age (at observation) was 38 years. HIV risk modes were recorded as: men who have sex with men, 45%; injecting drug use, 16%; both men who have sex with men and injecting drug use, 7%; and heterosexual contact, 13%. A total of 47 079 VL measurements were recorded for these persons in 1995 or later (88% by RT–PCR, 13% by bDNA).
A total of 163 774 person-months were included in the analysis. CD4 lymphocyte count information was missing for 9.7% of the person-months. For those months of observation in which CD4 information was available, by our definition, the CD4 lymphocyte count was obtained within 6 months of the person-month of interest; it was obtained in the same month as the VL for 66% of the person-months of observation. Thirty-six percent of person-months were associated with a history of AIDS OI. Fifty-seven percent of person-months were associated with prescription of HAART during the previous 6 month interval, 23% with dual therapy, 12% with other ART, and 8% with no ART. A total of 1409 new diagnoses of OI were recorded during the person-months of observation.
The unadjusted incidences of OI, stratified by CD4 lymphocyte count and VL, are shown in Fig. 1. It is noteworthy that patients with the highest VL were at significant risk for OI (i.e., ≥ 3 cases/100 person-years) at all ranges of CD4 lymphocyte count (Fig. 1).
A preliminary multivariate model indicated that only CD4 lymphocyte count, VL, antiretroviral therapy, history of AIDS OI, and calendar year were significantly associated with incidence of OI; therefore, only these variables were included in the final model. Person-time, numbers of OI cases, and risk ratios (RR) estimated from the final Poisson regression analysis are shown in Table 1. CD4 lymphocyte count was the strongest risk factor for developing an OI; that is, the highest risk ratios observed in this analysis were associated with very low CD4 counts. However, increasing VL also conferred an independent risk for OI, with RR of 1.6, 1.9, 2.7, and 3.5 for VL of 7000–19 999, 20 000–54 999, 55 000–149 999, and ≥ 150 000 copies/ml (by RT–PCR), respectively (Table 1). History of AIDS OI [RR, 1.7; 95% confidence interval (CI), 1.5–1.9] and earlier calendar year (RR, 1.1/year from 1999 back through 1995; 95% CI, 1.1–1.2) also conferred independent risk for OI; HAART (RR, 0.7; 95% CI, 0.6–0.9) and dual ART (RR, 0.8; 95% CI, 0.7–1.0) were protective.
Although numbers of cases of individual OI were fewer, we also estimated the RR associated with CD4 lymphocyte count and VL for development of PCP, non-tuberculosis mycobacteriosis (MAC or M. kansasii in the ASD database), and Candida esophagitis, the three most frequently occurring OI. For each of these OI, risk was associated with decreasing CD4 lymphocyte count. Independent risk was also conferred by increasing VL as follows: for PCP RR, 1.9 (95% CI, 1.2–2.9), 1.3 (95% CI, 0.8–2.2), 2.6 (95% CI, 1.6–4.1), and 3.6 (95% CI, 2.2–5.9) for VL 7000–19 999, 20 000–54 999, 55 000–149 999, and ≥ 150 000 copies/ml (by RT–PCR), respectively; for non-tuberculosis mycobacteriosis RR, 0.6 (95% CI, 0.2–2.3), 1.3 (95% CI, 0.5–3.4), 3.2 (95% CI, 1.5–7.0), and 3.9 (95% CI, 1.8–8.4), respectively; and for Candida esophagitis RR, 3.4 (95% CI, 1.7–7.0), 3.3 (95% CI, 1.7–6.7), 4.9 (95% CI, 2.5–9.4), and 6.4 (95% CI, 3.4–12.2), respectively.
We also examined whether the risk for OI associated with VL levels was the same at all ranges of CD4 lymphocyte counts by testing two-way interaction effects between CD4 lymphocyte count and VL categories in the multiple regression model. We were specifically interested in risk estimates associated with VL ranges at CD4 lymphocyte counts ≥ 200 × 106/l, as persons with CD4 lymphocyte counts in this range would not qualify for OI prophylaxis according to current guidelines . The RR associated with developing a new OI in this group were 1.9 (95% CI, 1.2–2.9), 1.3 (95% CI, 0.8–2.2), 2.6 (95% CI, 1.6– 4.1), and 3.6 (95% CI, 2.2–5.9) for persons with VL 7000–19 999, 20 000–54 999, 55 000–149 999, and ≥ 150 000, respectively. Numbers of OI were too small to assess the risk specifically for PCP at CD4 lymphocyte count ≥ 200 × 106/l, or for non-tuberculosis mycobacteriosis at CD4 lymphocyte count ≥ 50 × 106/l.
Finally, as VL measurements can change rapidly in response to change in use of ART, we repeated the analysis of risk for new OI by defining different periods of risk for OI following a specific VL measurement: 2 months, 3 months, 4 months, and 5 months. Although person-months of risk at each VL level decreased and the confidence intervals widened, the point estimates of risk at the various VL levels were comparable to those in the analysis using 6 months of risk (data not shown).
Our findings confirm the observations of many others that CD4 lymphocyte count appears to be the strongest predictor of the development of OI in HIV-infected persons; they also confirm observations by others that this association continues to be valid in the era of HAART [8,15,16]. This observation is still seen after controlling for VL, data which has generally not been available to other investigators at the time of their analyses.
However, our findings also indicate that VL above 7000 copies/ml as measured by RT–PCR (3000 copies by bDNA) is an independent risk factor for OI. This conclusion has been suggested previously by investigators in the Multicenter AIDS Cohort and the Adult Clinical Trials Group [7,8]; in these studies, baseline VL was found to be an independent risk factor for PCP, MAC, and cytomegalovirus disease. Our study extends this observation using a time-dependent analysis in which patients could be at risk at different VL levels as these measurements were obtained during the course of their illness.
An interesting observation in this analysis was that HAART was associated with an improved outcome in a multivariate model that controlled for VL, which is presumably the mechanism by which HAART is effective. Prescription of HAART may be a marker for other, unmeasured variables associated with better health outcome, including socioeconomic factors. The protective effect of HAART in the model was modest, with an upper 95% confidence limit approaching unity.
Our findings are limited by the observational nature of the ASD database, as well as by the fact that patients could be included only if VL data had been obtained. VL data were available for 11 464 of 19 329 (59.3%) persons observed after January 1995 and after initiation of ART; persons without VL information or without follow-up after the VL measurement were not included. Excluded persons were more likely to be African–American, to be injecting drug users, and to have low CD4 lymphocyte counts and a history of AIDS OI (data not shown); the findings regarding race and HIV risk mode reflect differences in access to health care that have been described previously [17,18]. Whether inclusion of these persons, had VL data been available, would have altered the associations between CD4 lymphocyte count, VL, and incidence of OI is unknown but seems unlikely. Race and HIV risk mode were not significant factors in our analysis, and CD4 lymphocyte count and history of AIDS OI were included in our multivariate model.
Another limitation is that the frequency of VL measurements in persons included in the analysis was variable. However, when we repeated the analyses using different periods of risk following each VL measurement, the results were comparable.
Finally, the VL obtained in this study were processed at different laboratories; this may be true even for samples from the same individual. Hence, the comparability of measurements cannot be assured. However, most VL measurements were made by RT–PCR; this test is approved by the Food and Drugs Administration as a basis for decisions regarding use of ART. In addition, categorizing VL values as we did would be expected to reduce the effect of the variability in these measurements on the outcome of the analysis.
The results of these analyses suggest that VL should be considered when making decisions regarding chemoprophylaxis in HIV-infected persons, that is, for example, in persons with high VL who do not otherwise meet criteria for prophylaxis. Similarly, a high VL might weigh against a decision to discontinue prophylaxis in persons whose CD4 lymphocyte counts have increased in response to HAART. Such use of VL information has already been suggested in current guidelines , although clinical trial data evaluating the use of VL information in making these decisions are lacking and will be difficult to obtain given the low incidence of OI in persons whose CD4 counts have increased to above prophylaxis thresholds [19–21].
The authors thank the Adult and Adolescent Spectrum of HIV Disease Investigators. The authors acknowledge the long-term commitment and efforts of former Los Angeles County ASD Principal Investigator, F. Sorvillo, and thank him for his contributions to this study over more than a decade. We thank S. Cox and C. Kiernan for secretarial and graphics assistance.
1. Phair J, Munoz A, Detels R, Kaslow R, Rinaldo C, Saah A. The risk ofPneumocystis cariniipneumonia among men infected with human immunodeficiency virus type 1. Multicenter AIDS Cohort Study Group.
N Engl J Med 1990, 32: 161–165.
2. Crowe SM, Carlin JB, Stewart KI, Lucas CR, Joy JF. Predictive value of CD4 lymphocyte numbers for the development of opportunistic infections and malignancies in HIV-infected persons. J Acquir Immune Defic Syndr 1991, 4: 770–776.
3. Moore RD, Chaisson RE. Natural history of opportunistic disease in an HIV-infected urban clinical cohort. Ann Intern Med 1996, 124: 633–642.
4. Mellors JW, Rinaldo CR Jr, Gupta P, White RM, Todd JA, Kingsley LA. Prognosis in HIV-1 infection predicted by the quantity of virus in plasma. Science 1996, 272: 1167–1170.
5. Department of Health and Human Services/Henry J. Kaiser Family Foundation Panel on Clinical Practices for the Treatment of HIV Infection. Guidelines for the use of antiretroviral agents in HIV-infected adults and adolescents.
February 2001. Available at: http://www.hivatis.org.
6. Carpenter CCJ, Cooper DA, Fischl MA. et al
. Antiretroviral therapy in adults: updated recommendations of the International AIDS Society–USA panel. JAMA 2000, 283: 381–390.
7. Lyles RH, Chu C, Mellors JW. et al
. Prognostic value of plasma HIV RNA in the natural history ofPneumocystis cariniipneumonia, cytomegalovirus andMycobacterium aviumcomplex. Multicenter AIDS Cohort Study.
AIDS 1999, 13: 341–349.
8. Williams P, Swindells S, Currier JS. Joint effects of HIV-1 RNA levels and CD4 lymphocyte cells on the risk of specific opportunistic infections. AIDS 1999, 13: 1035–1044.
9. Yerly S, Perneger TV, Hirschel B. et al
. A critical assessment of the prognostic value of HIV-1 RNA levels and CD4+ cell counts in HIV-infected patients. Arch Intern Med 1998, 158: 247–252.
10. Staszewski S, de Masi R, Hill AM, Dawson D. HIV-1 RNA, CD4 cell count and risk of progression to AIDS and death during treatment with HIV-1 reverse transcriptase inhibitors. AIDS 1998, 12: 1991–1997.
11. Powderly WG. The interaction of opportunistic infections and HIV replication. AIDS 1999, 13: 1603–1606.
12. Farizo KM, Buehler JW, Chamberland ME. et al
. Spectrum of disease in persons with human immunodeficiency virus infection in the United States. JAMA 1992, 267: 1798–1805.
13. Centers for Disease Control and Prevention. 1993 revised classification system for HIV infection and extended surveillance case definition for AIDS among adolescents and adults. MMWR 1992, 41 (RR–17): 4–19.
14. USPHS/IDSA Prevention of Opportunistic Infections Working Group. 1999 USPHS/IDSA guidelines for the prevention of opportunistic infections in persons infected with human immunodeficiency virus. MMWR 1999, 48 (RR–10): 1–66.
15. Forrest DM, Seminari E, Hogg RS. et al
. The incidence and spectrum of AIDS-defining illnesses in persons treated with antiretroviral drugs. Clin Infect Dis 1998, 27: 1379–1385.
16. Chene G, Binquet C, Moreau J-F. et al
. Changes in CD4+
cell count and the risk of opportunistic infection or death after highly active antiretroviral treatment. AIDS 1998, 12: 2313–2320.
17. Moore RD, Stanton D, Gopalan R, Chiasson RE. Racial differences in the use of drug therapy for HIV disease in an urban community. N Engl J Med 1994, 330: 763–768.
18. Shapiro MF, Morton SC, McCaffrey DF. et al
. Variations in the care of HIV-infected adults in the United States: results from the HIV Cost and Services Utilization Study. JAMA 1999, 281: 2305–2315.
19. Furrer H, Egger M, Opravil M. et al
. Discontinuation of primary prophylaxis against Pneumocystis carinii
pneumonia in HIV-1 infected adults treated with combination antiretroviral therapy. N Engl J Med 1999, 340: 1301–1306.
20. El-Sadr WM, Burman WJ, Grant LB. et al
. Discontinuation of prophylaxis forMycobacterium aviumcomplex disease in HIV-infected patients who have a response to antiretroviral therapy. Terry Beirn Community Programs for Clinical Research on AIDS.
N Engl J Med 2000, 342: 1085–1092.
21. Dworkin MS, Hanson DL, Kaplan JE, Jones JL, Ward JW. Risk for preventable opportunistic infections in persons with AIDS after antiretroviral therapy increases CD4+
T-lymphocyte counts above prophylaxis thresholds. J Infect Dis 2000, 182: 611–615.
Adult and Adolescent Spectrum of HIV Disease Investigators
M. Thompson, J. Gable, AIDS Research Consortium of Atlanta, Atlanta; S. Odem, S. Melville, Texas Department of Health, Austin; A. Davidson, D. L. Cohn, C. Rietmeijer, Denver Department of Health and Hospitals, Denver; L. L. Wotring, E. D. Mokotoff, Michigan Department of Community Health, Detroit; W. McNeely, K. Reynolds, Houston Department of Health and Human Services, Houston; P. Keiser, University of Texas Southwestern, Dallas; J. Turner, D. Masters, Los Angeles County Department of Health Services, Los Angeles; A. Morse and S. Broyles, Louisiana Office of Public Health, New Orleans; J. Sackoff, The City of New York Department of Health, New York City; J. Otero, R. Hunter, M. de los Angeles Gomez, University Central del Caribe, Bayamon; S. Miranda, Puerto Rico Department of Health, San Juan; S. Buskin, S. G. Hopkins, B. Sohlberg, Seattle-King County Department of Public Health, Seattle. Cited Here...
HIV; viral load; opportunistic infection; risk
© 2001 Lippincott Williams & Wilkins, Inc.
What does "Remember me" mean?
By checking this box, you'll stay logged in until you logout. You'll get easier access to your articles, collections,
media, and all your other content, even if you close your browser or shut down your
To protect your most sensitive data and activities (like changing your password),
we'll ask you to re-enter your password when you access these services.
What if I'm on a computer that I share with others?
If you're using a public computer or you share this computer with others, we recommend
that you uncheck the "Remember me" box.
Highlight selected keywords in the article text.
Data is temporarily unavailable. Please try again soon.
Readers Of this Article Also Read