Abstract
A case-cohort analysis of human immunodeficiency virus (HIV)–infected individuals receiving antiretroviral therapy (ART) was performed within a multicountry randomized trial (PEARLS) to assess the prevalence of persistently elevated C-reactive protein (CRP) levels, based on serial measurements of CRP levels, and their association with HIV clinical failure. A persistently elevated CRP level in plasma (defined as ≥ 5 mg/L at both baseline and 24 weeks after ART initiation) was observed in 50 of 205 individuals (24%). A persistently elevated CRP level but not an elevated CRP level only at a single time point was independently associated with increased clinical failure, compared with a persistently low CRP level, despite achievement of virologic suppression. Serial monitoring of CRP levels could identify individuals who are at highest risk of HIV progression and may benefit from future adjunct antiinflammatory therapies.
Keywords: HIV, inflammation, CRP, persistent inflammation, antiretroviral therapy
Antiretroviral therapy (ART)–naive and ART-experienced human immunodeficiency virus (HIV)–infected individuals (including those who achieve virologic suppression) with higher inflammation, as measured by levels of inflammatory markers at a single time point, are at an increased risk of disease progression, including higher rates of morbidity and mortality [1–3]. A commonly used marker of inflammation, C-reactive protein (CRP), is of interest to the HIV field as a potential prognostic marker because it can be measured at relatively low cost and in many resource-limited settings, using a point-of-care assay [4]. In HIV, some but not all studies have observed an association of a single elevated CRP level with disease progression [3, 5]. However, both acute inflammation (which might subsequently resolve) and unresolved persistent inflammation can result in an elevated CRP level. Using a case-cohort sample nested within a multinational randomized trial (PEARLS) of ART efficacy in ART-naive HIV-infected individuals [6], we conducted longitudinal serial monitoring of CRP levels to assess the prevalence of and risk factors for a persistently elevated CRP level, defined as elevated CRP level before ART initiation and 24 weeks after ART initiation, and the association of CRP level with clinical failure beyond 24 weeks of ART.
METHODS
Study Population
A case-cohort study was nested within PEARLS, an AIDS Clinical Trial Group (ACTG) randomized clinical trial (ACTG A5175; clinical trials registration NCT00084136) of ART efficacy conducted between May 2005 and May 2010 [6]. In PEARLS, 1571 ART-naive HIV-infected adults (defined as individuals aged ≥18 years) with CD4+ T cell counts <300 cells/mm3 were recruited from nine countries. Exclusion criteria and random assignment of different treatment regimens in the parent study are described in detailed elsewhere [6]. Following ART initiation, participants were closely followed through week 96 after ART initiation.
A case-cohort design was chosen as it allows for the estimation of prevalence and the relationship between the factor of interest and multiple outcomes [7]. Our full case-cohort study included 255 individuals (58 cases and 197 controls), of whom 205 were from the randomly selected subcohort sample (Supplementary Figure 1). Cases had clinical failure, defined as an incident World Health Organization (WHO) stage 3 and 4 event or death 24–96 weeks after ART initiation. As a secondary analysis, we limited analyses to 222 individuals (42 cases and 180 controls; also referred to as the virologically suppressed case-cohort) who achieved virologic suppression (defined as a plasma HIV RNA level of <400 copies/mL) at week 24 (Supplementary Figure 1). Another secondary analysis included the comparison of individuals who had elevated CRP levels at weeks 0, 24, and 48 relative to individuals with low CRP levels at all 3 times.
Ethics Statement
Department of Health and Human Services guidelines on human experimentation were followed. Ethics committees and institutional review boards from Johns Hopkins University and participating site institutions approved this study.
Data Collection and Laboratory Analysis
Standardization across all sites and rigorous criteria were used to obtain clinical history from patients at baseline, 2 weeks after ART initiation, 4 weeks after initiation, and every 4 weeks thereafter, through 24 weeks after initiation, and every 8 weeks thereafter [6]. Laboratories for which external quality assurance was achieved measured CD4+ T-cell counts, serum hemoglobin levels, and serum albumin levels as previously described [6]. Using archived plasma that had been collected at baseline and 24 weeks after ART initiation, we measured plasma CRP levels at Johns Hopkins University, using an enzyme-linked immunosorbent assay kit (R&D Systems, Minneapolis, Minnesota).
Definitions
A persistently elevated CRP level was defined as a CRP level of ≥ 5 mg/L at both baseline and 24 weeks after ART initiation in the primary analysis and at baseline, 24 weeks after initiation, and 48 weeks after initiation in a secondary analysis [8]. WHO cutoffs for sex-specific hemoglobin level (males, <13.0 g/dL; nonpregnant women, <12.0 g/dL) and albumin level (<3.5 g/dL) were used to define anemia [9] and hypoalbuminemia [10].
Statistical Analysis
Prevalence was estimated using the random subcohort. Differences in prevalence of persistently elevated CRP levels, stratified by categorical covariates, were assessed in the full case-cohort by weighted chi-squared tests with inverse probability weights, using Stata, version 13), with similar results observed from a Fisher exact test of the random subcohort. Univariable and multivariable Cox proportional hazards models, using S-plus Tibco Spotfire, version 7.2), were used to determine the association of persistent inflammation with clinical failure. The Cox models were stratified by country and treatment, and we accounted for this stratification by robust standard errors and Barlow weighting [7].
RESULTS
Prevalence and Correlates of Persistently Elevated CRP Level
Using the randomly sampled subcohort of 205 individuals, 91 (44%) had persistently low CRP levels, 33 (17%) had low CRP levels at baseline but elevated CRP levels at 24 weeks, 31 (15%) had elevated CRP levels at baseline but low CRP levels at 24 weeks, and 50 (24%) had persistently elevated CRP levels at both baseline and week 24 after ART initiation. Prevalence was similar when this analysis was limited to individuals who had virologic suppression 24 weeks after ART. Median CRP values at baseline and week 24 for the group with persistently elevated CRP levels were similar to those of the other groups who had elevated CRP levels at either baseline or week 24 (Supplementary Figure 2A and 2B), suggesting that this group was different not because it had higher levels of CRP but because it had persistently elevated CRP levels. Among several covariates tested, anemia (P = .0004) and hypoalbuminemia (P = .0001) were the only risk factors significantly associated with persistently elevated CRP levels (Table 1).
Table 1.
Characteristics of the Study Population, by Inflammation Status, as Measured by C-Reactive Protein (CRP) Level
| Characteristic | Overall (n = 255) |
Persistently Elevated CRP Level,a Subjects, No. (%) (n = 70) |
No Persistently Elevated CRP Level, Subjects, No. (%) (n = 185) |
P Valueb |
|---|---|---|---|---|
| Sex | ||||
| Male | 131 (53) | 37 (29) | 94 (71) | .95 |
| Female | 124 (47) | 33 (29) | 91 (71) | |
| Age, y | ||||
| <30 | 60 (24) | 15 (26) | 45 (74) | .79 |
| 30–40 | 114 (43) | 33 (31) | 81 (69) | |
| >40 | 81 (33) | 22 (28) | 59 (72) | |
| Country | ||||
| Brazil | 35 (19) | 9 (26) | 26 (74) | .09 |
| Haiti | 29 (7) | 3 (10) | 26 (90) | |
| India | 4 (2) | 1 (25) | 3 (75) | |
| Malawi | 41 (20) | 15 (37) | 26 (63) | |
| Peru | 30 (9) | 11 (37) | 19 (63) | |
| South Africa | 29 (14) | 12 (41) | 17 (59) | |
| Thailand | 26 (6) | 3 (12) | 23 (88) | |
| US | 31 (15) | 5 (16) | 26 (84) | |
| Zimbabwe | 30 (8) | 11 (37) | 19 (63) | |
| Race | ||||
| White | 18 (9) | 4 (23) | 14 (77) | .50 |
| Black | 147 (58) | 44 (32) | 103 (68) | |
| Hispanic | 58 (24) | 18 (30) | 40 (70) | |
| Asian | 31 (9) | 4 (14) | 27 (86) | |
| Body mass indexc | ||||
| <18.5 | 19 (8) | 5 (30) | 14 (70) | .95 |
| 18–25 | 170 (64) | 48 (29) | 122 (71) | |
| ≥25 | 66 (28) | 17 (27) | 49 (73) | |
| Prior tuberculosis diagnosis | ||||
| No | 222 (88) | 61 (29) | 161 (71) | .93 |
| Yes | 33 (12) | 9 (11) | 24 (72) | |
| Prevalent tuberculosis at baseline | ||||
| No | 245 (96) | 66 (28) | 179 (72) | .42 |
| Yes | 10 (4) | 4 (41) | 6 (59) | |
| Prior/current AIDS | ||||
| No | 235 (92) | 62 (28) | 173 (72) | .27 |
| Yes | 20 (8) | 8 (40) | 12 (60) | |
| Hepatitis B virus antigen | ||||
| Negative | 244 (97) | 67 (29) | 177 (71) | .67 |
| Positive | 11 (3) | 3 (23) | 8 (77) | |
| Treatment arm | ||||
| A | 92 (36) | 30 (34) | 62 (66) | .21 |
| B | 84 (33) | 18 (22) | 66 (78) | |
| C | 79 (31) | 22 (29) | 57 (71) | |
| CD4+ T-cell count, cells/mm3 | ||||
| <100 | 73 (27) | 26 (36) | 47 (64) | .18 |
| 100–200 | 88 (33) | 16 (22) | 72 (78) | |
| >200 | 94 (40) | 28 (30) | 66 (70) | |
| Log viral load, copies/mL | ||||
| <5 | 114 (44) | 30 (27) | 84 (73) | .71 |
| ≥5 | 141 (56) | 40 (30) | 101 (70) | |
| Albumin level, g/dLd | ||||
| <3.5 | 61 (27) | 29 (48) | 32 (52) | .0001 |
| ≥3.5 | 194 (73) | 41 (22) | 153 (78) | |
| Anemiae | ||||
| No | 119 (47) | 22 (17) | 97 (83) | .0004 |
| Yes | 134 (53) | 47 (39) | 87 (61) | |
Abbreviation: ART, antiretroviral therapy.
a Defined as an elevated CRP level (ie, ≥5 mg/L) at both week 0 and week 24 after ART initiation.
b By weighted χ2 tests, using inverse probability weights.
c Defined as the weight in kilograms divided by the height in meters squared.
d An albumin level of ≤3.5 g/dL was used to define hypoalbuminemia.
e Sex-specific hemoglobin level cutoffs (males, <13.0 g/dL; nonpregnant women, <12.0 g/dL) were used to define anemia.
Association of Persistently Elevated CRP Level With Clinical Failure
Of 58 cases, 36 (62%) had WHO stage III events (of which 16 had pulmonary tuberculosis), 17 (29%) had WHO stage IV events (of which 5 had extrapulmonary tuberculosis), and 5 (9%) died. Individuals who had an elevated CRP level at baseline only or at 24 weeks and had no clinical failure had a survival curve similar to that for individuals with a persistently low CRP level (Supplementary Figure 3A). However, individuals with a persistently elevated CRP level had a substantially lower survival rate (Supplementary Figure 3A).
Compared to individuals with persistently low CRP levels, individuals with persistently elevated CRP levels had an increased hazard of clinical failure (hazard ratio [HR], 7.04; 95% confidence interval [CI], 2.87–17.27), but not individuals with an elevated CRP level only at either baseline (HR, 1.72; 95% CI, .55–5.538) or at 24 weeks after ART initiation (HR, 1.06; 95% CI, .37–3.05) in univariable models (Table 2). After adjustment for sex, age, body mass index, baseline CD4+ T-cell count, viral load, and prior tuberculosis (model 1; adjusted HR, 8.97; 95% CI, 3.16–25.52) or after further adjustment for baseline albumin level, hemoglobin level, tuberculosis, and prior/baseline AIDS (model 2; adjusted HR, 4.45; 95% CI, 1.42–13.93), individuals with a persistently elevated CRP level remained at an increased hazard of clinical failure (Table 2). Adjustment for week 24 CD4+ T-cell count instead of baseline CD4+ T-cell count gave similar results (data not shown).
Table 2.
Association of Persistently Elevated C-Reactive Protein (CRP) Levels With Clinical Failure
| Subjects, CRP Level, Time After ART Initiation | Subjects, No. (%) | Hazard Ratio (95% CI) |
||
|---|---|---|---|---|
| Univariable Model | Multivariable Model 1a,b | Multivariable Model 2b,c | ||
| Overall | ||||
| Low at baseline | ||||
| Low at 24 wks | 107 (42) | Reference | Reference | Reference |
| High at 24 wks | 39 (15) | 1.72 (.55–5.38) | 1.43 (.34–6.05) | 2.17 (.43–10.90) |
| High at baseline | ||||
| Low at 24 wks | 39 (15) | 1.06 (.37–3.05) | 1.47 (.32–6.68) | 0.71 (.14–3.56) |
| High at 24 wks | 70 (28) | 7.04 (2.87–17.27) | 8.97 (3.16–25.52) | 4.45 (1.42–13.93) |
| With virological suppression | Multivariable Model 3a,d | Multivariable Model 4c,d | ||
| Low at baseline | ||||
| Low at 24 wks | 92 (42) | Reference | Reference | Reference |
| High at 24 wks | 36 (16) | 2.10 (.57–7.69) | 2.11 (.43–10.41) | 3.98 (.89–17.72) |
| High at baseline | ||||
| Low at 24 wks | 36 (16) | 1.24 (.38–4.06) | 1.36 (.27–6.96) | 1.06 (.16–7.13) |
| High at 24 wks | 58 (26) | 8.65 (3.12–23.95) | 9.49 (3.01–29.94) | 6.73 (1.86–24.36) |
An elevated CRP level (ie, ≥5 mg/L) at baseline and week 24 after ART initiation was used to categorize individuals into one of 4 groups: (1) low CRP level at baseline and week 24 after ART initiation, (2) low CRP level at baseline and elevated CRP level at week 24, (3) an elevated CRP level at baseline and a low CRP level at week 24, and (4) elevated CRP levels at baseline and week 24 (ie, persistently elevated CRP level). The association of these groups with clinical failure, relative to the group with low CRP levels at baseline and week 24, was assessed using univariable and multivariable Cox regression models.
Abbreviations: ART, antiretroviral therapy; BMI, body mass index; CI, confidence interval.
a Adjusted for sex, age, BMI, CD4+ T-cell count, viral load, and prior tuberculosis.
b Conducted on the final case cohort (n = 255) for primary analysis.
c Adjusted for sex, age, BMI, CD4+ T-cell count, viral load, and prior tuberculosis, hemoglobin level, albumin level, prior/current AIDS at baseline, and prevalent tuberculosis at baseline.
d Limited to subjects who achieved viral suppression at week 24 after ART initiation (n = 222).
When we limited the analysis to individuals who had virologic suppression 24 weeks after ART initiation, a similar association was observed in univariable models (HR, 8.65; 95% CI, 3.12–23.95) and multivariable models with adjustment for the same variables as in model 1 (multivariable model 3; adjusted HR, 9.49; 95% CI, 3.01–29.94) and model 2 (multivariable model 4; adjusted HR, 6.73; 95% CI, 1.86–24.36; Table 2).
This association of a persistently elevated CRP level with an increased hazard of clinical failure relative to that for individuals with a persistently low CRP level was further confirmed (Supplementary Figure 3B) in both the full cohort (adjusted HR, 15.92; 95% CI, 2.56–98.59), as well as in individuals who had virologic suppression 48 weeks after ART initiation (adjusted HR, 35.1; 95% CI, 2.91–423), even when the definition of persistently elevated CRP level was modified to include individuals with a high CRP level at weeks 0, 24, and 48 after ART initiation.
DISCUSSION
Using serial measurements of CRP levels before and 24 weeks after ART initiation, we showed that the prevalence of persistently elevated CRP levels was approximately 25% in our study of HIV-infected adults from predominantly low-income and middle-income settings and that anemia and hypoalbuminemia were risk factors for persistently elevated CRP levels. Importantly, individuals with persistently elevated CRP levels (despite achievement of virologic suppression), but not individuals with elevated CRP levels at only 1 time point, had an increased risk of clinical failure between week 24 and week 96 after ART initiation as compared to individuals with persistently low CRP levels. Our results suggest that serial CRP measurements within the first year of ART could be particularly useful to identify HIV-infected adults at highest risk of subsequent treatment failure and those who might benefit from approaches to reduce inflammation.
In our study, median CRP levels before ART were similar to those after ART. Our results suggest that looking at only median CRP values is not sufficient to identify the various risk groups where certain individuals change their inflammation status after ART while others do not. Whether anemia and hypoalbuminemia, risk factors identified in this study, lead to persistent inflammation or are a result of persistent inflammation remains to be determined. Other potential risk factors for persistent inflammation, including concurrent infections and low-level viremia not detected by our assay, warrant further study.
There is continued debate about the predictive role of a cross-sectional measure of the CRP level in HIV-infected adults because the level of CRP can be transiently elevated during a variety of acute illnesses and thus is nonspecific [3, 5]. In contrast to cross-sectional measures, serial measurements of CRP levels have been shown to have prognostic value in several conditions, including cardiovascular diseases [11]. However, this is the first study, to our knowledge, to use serial measurements of CRP levels among HIV-infected individuals to categorize individuals into different risk groups and identify a subset of individuals with persistently elevated CRP levels to be at much higher hazards of future treatment failure.
The biological mechanism for the relationship of a persistently elevated CRP level with HIV treatment failure could be explained through its involvement in general inflammation [12], where persistent inflammation is known to increase risk factors for HIV disease progression: oxidative stress, anemia, and micronutrient deficiency, among others [13–15]. Whether CRP level has any specific effects in addition to the general effects of inflammation remains to be determined.
An alternate explanation is that CRP levels are elevated because of underlying occult opportunistic infections, which could then precipitate treatment failure. However, this does not seem to be the case in our study, as individuals with persistently elevated CRP levels had a similar median time to clinical failure (52–55 weeks) and a similar number of events by 36 weeks after ART initiation as the other groups (we would have expected occult infections to present symptoms by this time).
Importantly, we show that individuals with an elevated CRP level at just 1 time point are not at an increased risk of treatment failure, and this might represent individuals with an acute infection at 1 time point that is resolved during the other time point. In contrast, individuals with a persistently elevated CRP level likely represent individuals with an advanced disease stage in which inflammation is not being controlled despite ART-induced viral suppression. As CRP values could decrease and increase again within 24 weeks for some individuals, our data suggests that, while every individual might not truly have had a persistently elevated CRP level, this is powerful approach to identify groups at highest risk who could benefit from antiinflammatory treatments (eg, statins) and more-intense monitoring.
Because this was a not a randomized controlled trial, we may have had unmeasured confounders and cannot prove causal relationships. However, our prospective design is a strength of our study. Another limitation is that we only measured the CRP level at 2 different time points and 24 weeks apart. However, this relationship was also confirmed when we measured the CRP level at 3 different times. Another limitation is the potential for survivor bias, as the exposure definitions only allowed for the inclusion of participants who did not have clinical failure before 24 weeks of treatment. However, it is unlikely that this would have affected the direction of the effect size, as an elevated baseline CRP level was present in a higher proportion of individuals who had clinical failure prior to 24 weeks after ART initiation (52%), compared with individuals who had clinical failure 24 weeks after ART initiation (40%), suggesting that our effect size is actually an underestimation. Nevertheless, our conclusions remain consistent for individuals who do not die or did not respond to treatment during the first 24 weeks of ART.
We conclude that a persistently elevated CRP level is associated with increased clinical failure, including among individuals with viral suppression, and that serial monitoring of CRP is a potentially useful approach for identifying individuals at highest risk for clinical failure. Future studies are needed to determine whether CRP and inflammation is causal and whether appropriate antiinflammatory interventions can reduce HIV disease progression.
Supplementary Data
Supplementary materials are available at http://jid.oxfordjournals.org. Consisting of data provided by the author to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the author, so questions or comments should be addressed to the author.
Notes
Acknowledgments. We thank the PEARLS study participants, for volunteering their time and efforts.
R. S. conceived the research question, conducted the data analysis, and wrote the primary version of the manuscript. W.-T. Y., N. G., R. C. B., and A. B. contributed to data interpretation and manuscript review. S. B., N. M., C. K., S. P., W. S., B. S., S. P., S. T., C. R., J. R. L., S. W. C., and P. S. contributed to data collection and manuscript review. R. D. S. contributed to study design, laboratory testing, and manuscript review. T. B. C. contributed to study design, data collection, oversight of study implementation, and manuscript review. A. G. obtained funding and contributed to study conception, design, and manuscript writing and review. All authors meet criteria for authorship as recommended by the International Committee of Medical Journal Editors and were fully responsible for all aspects of manuscript development.
Disclaimer. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Allergy and Infectious Diseases (NIAID) or the National Institutes of Health (NIH). Funders had no role in study design, data collection, analysis, publication decision, or manuscript preparation.
Financial support. This work was supported by the NIAID (grants R01AI080417, UM1AI068636, UM1AI068634, UM1AI106701, and UM1AI069465), the National Institute of Mental Health, and the National Institute of Dental and Craniofacial Research, NIH. The parent trial, A5175, was also supported in part by Boehringer-Ingelheim, Bristol-Myers Squibb, Gilead Sciences, and GlaxoSmithKline.
Potential conflicts of interest. T. B. C. is an advisory board member for Gilead Sciences. A. G. has received grant funding from Gilead Foundation. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
References
- 1.Lau B, Sharrett AR, Kingsley LA et al. . C-reactive protein is a marker for human immunodeficiency virus disease progression. Arch Intern Med 2006; 166:64–70. [DOI] [PubMed] [Google Scholar]
- 2.Kuller LH, Tracy R, Belloso W et al. . Inflammatory and coagulation biomarkers and mortality in patients with HIV infection. PLoS Med 2008; 5:e203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ledwaba L, Tavel JA, Khabo P et al. . Pre-ART levels of inflammation and coagulation markers are strong predictors of death in a South African cohort with advanced HIV disease. PloS One 2012; 7:e24243. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Yoon C, Davis JL, Huang L et al. . Point-of-care C-reactive protein testing to facilitate implementation of isoniazid preventive therapy for people living with HIV. J Acquir Immune Defic Syndr 2014; 65:551–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.McDonald B, Moyo S, Gabaitiri L et al. . Persistently elevated serum interleukin-6 predicts mortality among adults receiving combination antiretroviral therapy in Botswana: results from a clinical trial. AIDS Res Hum Retroviruses 2013; 29:993–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Campbell TB, Smeaton LM, Kumarasamy N et al. . Efficacy and safety of three antiretroviral regimens for initial treatment of HIV-1: a randomized clinical trial in diverse multinational settings. PLoS Med 2012; 9:e1001290. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Barlow WE, Ichikawa L, Rosner D, Izumi S. Analysis of case-cohort designs. J Clin Epidemiol 1999; 52:1165–72. [DOI] [PubMed] [Google Scholar]
- 8.Sage EK, Noursadeghi M, Evans HE et al. . Prognostic value of C-reactive protein in HIV-infected patients with Pneumocystis jirovecii pneumonia. Int J STD AIDS 2010; 21:288–92. [DOI] [PubMed] [Google Scholar]
- 9.McLean E, Cogswell M, Egli I, Wojdyla D, de Benoist B. Worldwide prevalence of anaemia, WHO Vitamin and Mineral Nutrition Information System, 1993–2005. Public Health Nutr 2009; 12:444–54. [DOI] [PubMed] [Google Scholar]
- 10.Sudfeld CR, Isanaka S, Aboud S et al. . Association of serum albumin concentration with mortality, morbidity, CD4 T-cell reconstitution among Tanzanians initiating antiretroviral therapy. J Infect Dis 2013; 207:1370–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Gabay C, Kushner I. Acute-phase proteins and other systemic responses to inflammation. N Engl J Med 1999; 340:448–54. [DOI] [PubMed] [Google Scholar]
- 12.Pepys MB, Hirschfield GM. C-reactive protein: a critical update. J Clin Invest 2003; 111:1805–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Drain PK, Kupka R, Msamanga GI, Urassa W, Mugusi F, Fawzi WW. C-reactive protein independently predicts HIV-related outcomes among women and children in a resource-poor setting. AIDS 2007; 21:2067–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Rimaniol AC, Zylberberg H, Zavala F, Viard JP. Inflammatory cytokines and inhibitors in HIV infection: correlation between interleukin-1 receptor antagonist and weight loss. AIDS 1996; 10:1349–56. [DOI] [PubMed] [Google Scholar]
- 15.Ford ES, Liu S, Mannino DM, Giles WH, Smith SJ. C-reactive protein concentration and concentrations of blood vitamins, carotenoids, and selenium among United States adults. Eur J Clin Nutr 2003; 57:1157–63. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.