Secondary Logo

Journal Logo


Serum suppression of tumorigenicity 2 level is an independent predictor of all-cause mortality in HIV-infected patients

Thiébaut, Rodolphea,b,*; Hue, Sophiec,d,*; Le Marec, Fabiena; Lelièvre, Jean-Danielc,d; Dupon, Michelb; Foucat, Emilec,d; Lazaro, Estibalizb; Dabis, Françoisa,b; Duffau, Pierreb; Wittkop, Lindaa,b; Surenaud, Mathieuc; Pellegrin, Isabellea,b; Lacabaratz, Christinec; Bonnet, Fabricea,b; Lévy, Yvesc,d for the ANRS CO3 Aquitaine Cohort

Author Information
doi: 10.1097/QAD.0000000000001628



The AIDS-related morbidity and mortality have substantially decreased in people living with HIV since the availability of potent antiretroviral therapy (ART) [1]. Now that HIV viral load can be durably controlled and maintained at undetectable levels at least in the plasma, immune activation, immunoscenescence and systemic inflammation are contributing mechanisms leading to comorbidities in HIV-infected patients [2,3]. Hence, biomarkers such as IL-6, high-sensitivity C-reactive protein (hsCRP), D-dimer and soluble CD14 (sCD14) have been recognized as independent predictors of disease progression and mortality [4–10]. More specifically, sCD14, a marker of monocyte activation [11], has been reported to be associated with HIV disease progression in both HIV types 1 and 2, independently of T-cell activation, coagulation and inflammatory biomarkers [7,9,12,13]. IL-6, a marker of systemic inflammation, has been revealed as one of the first biomarker associated with all cause mortality [4], although its independent effect in HIV-infected patients has been mitigated [8].

Suppression of tumorigenicity 2 (ST2) is a member of the IL-1 receptor family that plays a major role in immune and inflammatory responses. Alternative promoter activation and splicing produce both a membrane-bound protein (ST2L) and a soluble ST2 (sST2). sST2 is present constitutively in human serum, in which it acts as a decoy receptor by binding free IL-33. Much of the variation in sST2 production has been shown to be driven by genetic factors [14].

Clinically, the ST2/IL-33 signaling pathway participates in the pathophysiology of a number of inflammatory and immune diseases related to Th2 activation, including asthma, ulcerative colitis, and inflammatory arthritis [15]. Animal models of cardiovascular disease demonstrated that IL-33 induction following vascular and cardiac stress was correlated with improved outcomes [16–18]. Furthermore, IL-33 reduces atherosclerosis [17], limits pressure overload-induced cardiac hypertrophy [16], and improves cardiac function and cardiomyocyte survival after myocardial infarction [19]. To summarize, sST2 acts as a decoy receptor and, by sequestering IL-33, sST2 antagonizes the cardioprotective effects of ST2L/IL-33 interaction. As a biomarker, circulating sST2 concentrations have been linked to worse prognosis in patients with heart failure [20,21] and also predict mortality and incident cardiovascular events in individuals without existing cardiovascular disease [22,23].

HIV-infected patients present an increased risk of cardiovascular mortality. Although the role of some antiretroviral drugs as well as traditional risk factors has been proven [24], a substantial part of the risk is still not explained leading to a restricted performance of predictive models based on traditional cardiovascular risk factors [25] and leaving the room for the discovery of new risk factors. The role of sST2 has been poorly explored in HIV-infected individuals so far. Higher levels of serum sST2 have been reported in HIV-1-infected patients as compared with HIV-negative controls [26,27] with the highest levels in those at early stage of HIV infection [27]. Recently, in a recent American cohort study of 332 HIV-infected patients with high rates of cardiovascular diseases, sST2 was independently associated with both cardiovascular dysfunction [relative risk (RR) = 1.36] and all-cause mortality (RR = 2.04). It was the strongest predictor among several biomarkers [28]. The prognostic role of sST2 needs now to be confirmed and explored in larger and more diverse HIV populations with proper adjustment for important clinical risk factors such as hemoglobinemia or predictive scores [2]. In the current study, we have estimated and validated the predictive capacity of sST2 on all-cause mortality taking into account sCD14 and IL-6 as well as internationally recognized prognostic scores [8,29] in a large cohort of 1800 HIV-1-infected patients in care and followed in France.


Study population

All patients eligible for this study are enrolled in the ANRS CO3 Aquitaine Cohort, a prospective hospital-based cohort of HIV-1-infected patients under routine clinical management, in South-Western France [30]. The study protocol is approved by the local ethic committee and all patients included in the ANRS CO3 Aquitaine Cohort sign an informed consent. Patients were eligible for the current study if they had at least one plasma sample available in the cohort biobank. The date of the first available plasma sample was considered as the baseline and sST2 and sCD14 were measured on this sample. The best model to predict mortality was built on a first dataset and a validation study to measure the prediction error was performed on a new dataset available later on (from August 2015).

Causes of death

Information contained in the cohort questionnaire and based on the physician's report was used to determine the underlying cause of death according to the International Classification of Diseases, Tenth Revision rules: The underlying cause of death is the disease or injury, which initiates the train of morbid events leading to death. The algorithm of determination was adapted to specific concerns in HIV infection as in previous local studies and collaborations [31] and allowed categorization of deaths as follows: AIDS-related causes according to the 1993 US Centers for Disease Control clinical classification, deaths related to infection with hepatitis C virus (HCV) or hepatitis B virus (HBV) including hepatocellular carcinoma, other neoplasia and other causes not related to AIDS or HCV/HBV, and adverse effects of ART. The latter was considered the underlying cause of death only when this was the explicit conclusion of the physician. AIDS-defining causes were grouped in one underlying cause of death, and the frequency of AIDS-defining diseases was described secondarily. Where the standardized questionnaire was missing, the abstracted quarterly notifications were used to establish the underlying cause of death, if possible.


Samples were immediately centrifuged for storage at −80 °C. sST2, sCD14 and IL-6 were measured at distance using Luminex multiplex bead-based technology (R&D Systems, Minneapolis, Minneapolis, USA) and a Bio-Plex 200 instrument (BioRad, Hercules, California, USA) according to the manufacturers’ protocols on plasma diluted to 1/50. The assay has a lower detection limit of 0.05 ng/ml for sST2, 0.2 ng/ml for sCD14 and 0.26 pg/ml for IL-6 and higher limits of 49 ng/ml for sST2, 1300 ng/ml for sCD14 and 900 pg/ml for IL-6. For comparison, the limit of sST2 detection for the Presage assay (Critical Diagnostics, San Diego, California, USA) is 1.3 up to 200 ng/ml. The currently commercially available methods for measurement of sST2 (i.e. the Presage assay, the R&D assay and the MBL assay; Woburn, Massachusetts, USA) are not standardized.

The reliability of the measurements performed in our study is demonstrated by Supplementary Fig. 1, Coefficients of variation (SD of the mean) were used to assess precision. The mean of standard interplate percentages of coefficients of variations for sST2, sCD14 and IL-6 were less than 8%. A home-made internal control showed a good reproducibility of the assay. Mean coefficients of variations was less than 12% for sST2, less than 14% for sCD14 and less than 8% for IL-6. A good correlation between ELISA and Luminex technology was also observed. Laboratory technicians who performed the assays were blinded to all clinical data.

Statistical analyses

A multivariable Cox proportional hazard model was constructed using a backward selection strategy in which variables that were significant at 0.20 levels in unadjusted analyses were included in the multivariable analyses. Hence, the model was initially adjusted for sST2 concentration, age, sex, tobacco consumption, centers for disease control and prevention clinical classification, HIV transmission risk group, duration since first reported HIV seropositivity, BMI, hemoglobin, cholesterol, triglycerides, HDL, CD4+ nadir, last CD4+ cell count, time to CD4+ cell count more than 500 cells/μl before the first ART, undetectable viral load (≤50 copies/ml), time with viral load more than 500 copies/ml, ART treatment, time since the first ART prescription, HBV infection, HCV infection, estimated glomerular filtration rate (eGFR) less than 60 ml/min using the Modification of Diet in Renal Disease formula, dyslipidemia, diabetes, lipid-lowering drug treatment use, history of cardiovascular event, hypertension and cancer. The log-linearity of the effect of sST2 (and other continuous markers) was checked in comparison with fractional polynomials [32] and categorical variables.

Validation was performed in a study sample that became available after the learning sample and that included 246 patients among whom 15 died. It included the calculation of the c-index statistic which is the probability that a patient who died has a higher predicted probability of dying than a patient who survives [33]. The net reclassification index (NRI), which quantifies the significance of the improvement of the prediction using a new marker for event and nonevent individuals, was also calculated for a descriptive purpose [34] but was not consider as the main criteria because of its limitations [35]. Robustness analyses were performed with fluorescence data of ST2 and sCD14 instead of concentration and led to very similar results (not shown). Proportional hazard assumption was checked graphically as well as by producing Schoenfeld residuals (Supplementary Fig. 2, All analyses were performed with SAS 9.3 (SAS Institute, Cary, North Carolina, USA).


Study population and study sample

Among the 2113 patients of the ANRS CO3 Aquitaine Cohort with a least one sample available in the biobank, 1414 were included in the ‘learning study’ and 386 in the ‘validation study’. Baseline characteristics did not differ between patients included in the study with those excluded because no blood sample was available (data not shown). Dates of inclusion in the current study were distributed between 2006 and 2011. At study baseline, among the 1414 patients included in the main study, 1154 (81.6%) were treated with HAART [56% with two nucleoside reverse transcriptase inhibitors (NRTIs) and one boosted protease inhibitor, 24% with two NRTIs and one non-NRTI (NNRTI), 20% with other ART regimens]. The median duration of exposure to ART was 8 years [interquartile range (IQR): 3.6; 10.9]. The median baseline CD4+ cell count was 742 cells/μl (IQR: 545; 970). Other baseline characteristics, compared between patients alive and deceased, are presented in Table 1. During a median follow-up of 7.2 years (IQR: 6.0; 7.9), 93 deaths occurred leading to an incidence rate of 9.9 per 1000 patient-years [95% confidence interval (CI) 7.9–11.9]. The distribution of the underlying causes of death (Supplementary Table 1, was comparable with what had been reported in the French national survey of mortality in HIV-infected patients (that account for more than half of deaths in HIV-infected patients in France) [36], although non-AIDS-defining nonhepatitis-related malignancies appeared over-represented in the current study population (31 vs. 22%), followed by liver diseases (16 vs. 11%), cardiovascular diseases (10%) and AIDS-related diseases (7 vs. 25%).

Table 1:
Study population characteristics according to the clinical status at the end of the follow-up.

The median baseline Veterans Aging Cohort Study (VACS) clinical score [8] was 11 (IQR: 5; 19) in patients alive and 22 (IQR: 12; 41) in those deceased during the follow-up (P < 0.0001).

Soluble suppression of tumorigenicity 2, soluble CD14 and IL-6 distributions

The median sST2 concentration in baseline serum samples was 22.9 ng/ml (IQR: 17.7; 30.3). The concentration of sST2 was not associated with the baseline plasma HIV RNA level (Spearman correlation coefficient c = −0.04, P = 0.11). The baseline concentration of sST2 was higher in patients who died during follow-up (median 30.8 ng/ml, IQR: 21.5; 42.1) compared with those who stayed alive (median 22.6 ng/ml, IQR: 17.5; 29.6) (P < 10−4). The survival probabilities at 3 years (IQR) were significantly different according to the sST2 quartiles (Fig. 1): 0.6% (0.5; 0.6) for the first quartile, 1.4% (1.4; 1.5) for the second one, 2.9% (2.8; 3.0) for the third one and 7.2% (7.1; 7.4) for the fourth one (P < 0.10−4). Patients with baseline sST2 concentrations above the median (22.9 ng/ml) had a 2.76-fold (95% CI 1.75–4.34) increased risk of death compared with those below the median.

Fig. 1:
Survival probability according to soluble suppression of tumorigenicity 2 quartiles in the ANRS CO3 Aquitaine Cohort (n = 1414).

The median baseline sCD14 concentration was 1.57 μg/ml (IQR: 1.29; 1.93), being also higher in patients who died later on (1.78, IQR: 1.43; 2.36) compared with those who stayed alive (1.55, IQR: 1.28; 1.91; P = 10−4) (Fig. 2a). The survival probabilities at 3 years (IQR) varied also significantly according to sCD14 quartiles: 4.9% (4.8; 5.0), 3.7 (3.6; 3.8), 2.6 (2.5; 2.7) and 0.9 (0.8; 0.9), respectively (P = 0.0148).

Fig. 2:
(a) Measured baseline concentrations of soluble suppression of tumorigenicity 2 (a1), soluble CD14 (a2) and IL-6 (a3) according to the clinical status at the end of the study (deceased n = 93 or alive n = 1320). (b) Correlations between baseline soluble suppression of tumorigenicity 2 and soluble CD14 and IL-6 (Spearman correlation, P value for correlation c = 0).

The median baseline IL-6 concentration was 0.63 pg/ml (IQR: 0.39; 1.11), being also higher in patients who died later on (1.29, IQR: 0.67; 2.91) compared with those who stayed alive (0.61, IQR: 0.38; 1.04, P = 10−4) (Fig. 2a). The survival probabilities at 3 years (IQR) varied also significantly according to IL-6 quartiles: 1.1% (1.0; 1.1), 1.8 (1.7; 1.9), 2.9 (2.8; 3.0) and 6.2 (6.1; 6.4), respectively (P < 0.10−4).

sST2 and sCD14 were correlated (Spearman correlation coefficient c = 0.36, P < 10−4, Fig. 2b) as well as sST2 and IL-6 (c = 0.22, P < 10−4, Fig. 2b). The three markers were weakly negatively correlated with baseline CD4+ cell count (Fig. 2b).

Predictive models

In multivariable analysis (Table 2), sST2 was strongly associated with survival with an increase of the risk of death of 20% for a concentration 10 ng/ml higher (hazard ratio = 1.25, P < 10−4). The other factors associated with poorer survival were men (hazard ratio = 2.27, P = 0.02), tobacco consumption (hazard ratio = 2.13 for ex-smokers and 2.85 for current smokers, P = 0.01), anemia (hazard ratio = 1.30 for 1 g/100 ml lower, P = 0.0002), duration of viral load more than 500 copies/ml (hazard ratio = 1.11 for 1 year less, P = 0.004), low glomerular filtration (hazard ratio = 3.16 if eGFR < 60 ml/min, P = 0.005), history of diabetes (hazard ratio = 2.17, P = 0.005) or cancer (hazard ratio = 2.30, P = 0.004) and hepatitis B virus infection (hazard ratio = 1.88, P = 0.053). When adjusted for the same prognostic factors, the effect of VACS score and IL-6 was independently associated with the risk of death, whereas the effect of sCD14 did not remain significantly associated with the risk of death (hazard ratio = 1.02 for 1 μg/ml, P = 0.17). In a model adjusted for all prognostic factors (Supplementary Table 2,, the effect of sST2 was still highly significant (hazard ratio = 1.21 for 10.0 ng/ml higher, P < 0.0001) as well as the IL-6 and the VACS score that were independently and positively associated with mortality (hazard ratio = 1.77 for 1 pg/ml higher, P = 0.041 and hazard ratio = 1.04 for 1 point higher, P < 0.0001, respectively). It is also worth noting that the effect of ST2 was independent of HBV infection, although ST2 was increased in those infected with HBV, HCV and coinfected compared with uninfected: 24.9 ng/ml (IQR: 18.2; 32.6), 25.2 (IQR: 18.7; 34.3) and 25.0 (IQR: 18.6; 32.7), respectively, vs. uninfected 22.2 ng/ml (IQR: 17.4; 29.1).

Table 2:
Factors associated with the risk of death in the ANRS CO3 Aquitaine Cohort (n = 1414 and 1301 for the univariable and multivariable models).

The linearity of the relationship between sST2 and mortality was verified (Fig. 3), meaning that, in adjusted analyses, there was no obvious threshold to define the normal range of sST2 values. Instead, the lower was sST2 at baseline, the better the clinical prognosis. Compared with those with less than 17.7 ng/ml (first quartile), patients with more than 30.3 ng/ml (fourth quartile) had a three-fold higher probability of dying during follow-up (hazard ratio = 3.31, 95% CI 1.53–7.15, P < 10−3).

Fig. 3:
Increase of the risk of death according to soluble suppression of tumorigenicity 2 values.The linear assumption of the increase of the risk is compatible with the most flexible formulation (fractional polynomial) that serves as reference. There was no obvious threshold in the increase of the risk of death according to soluble suppression of tumorigenicity 2 concentration.

Other predictors of survival and validation study

With regards to the predictive capacity at 3 years, the addition of sST2 in the model (resulting in the model presented Table 2) improved significantly the capacity to discriminate the patients who died to the others in term of c-index (0.700 with sST2 vs. 0.643 without sST2). With use of 3-year risk categories of 0–5, 5–10 and at least 10%, risk prediction was more accurate in models that included sST2. Among the 35 patients who died, three (8.6%) were reclassified into more accurate risk categories, whereas two (5.7%) were reclassified in a less accurate category (Supplementary Table 3, Among 1266 patients still alive at 3 years, 49 (3.9%) were reclassified into more accurate risk categories, whereas 19 (1.5%) were reclassified in a less accurate category. The NRI at 3 years was therefore improved (5.2%).

In the validation cohort (n = 386 including 31 deaths), the c-index was still higher with a model including sST2 than a model including VACS score and IL-6 (0.811 vs. 0.804), confirming the improvement of the prediction of the mortality with this biomarker when measured at baseline.


In this study exploring the performance of a new prognostic marker, we took the opportunity of a large cohort enrolling patients since the diagnosis of HIV, with long-term follow-up under routine clinical management in a multicenter setting in south-western France [30] and a bio-bank available. We showed that sST2 is a significant predictor of all-cause mortality. This predictive capacity was confirmed on a validation sample. Our study confirmed also the predictive value of death of plasma levels of sCD14 and IL-6 as previously reported [7,29]. We adjusted for the VACS index, a well known validated predictor of all-cause mortality when measured at any single time point from 1 to 5 years after ART initiation in several HIV cohorts [2,8]. When adjusting on the VACS index, the effect of sST2 and IL-6, but not of sCD14, remained significantly associated with the risk of death with a fair discriminative capacity for ST2 compared with previous studies [8,28]. Moreover, we showed the risk of death increases gradually with baseline sST2 serum levels, reaching a 2.3-fold increase in patients with highest values. Because of the continuous increase of the risk, it was impossible to define threshold values for the risk and therefore normal range within the current study. We took the advantage of the high number of clinical parameters measured and collected during the follow-up of the Aquitaine cohort, and were thus able to perform large-scale multivariable analyses and to confront the predictive value of sST2 to several other factors influencing the survival of HIV-infected patients. Although we were not able to adjust for every biomarker studied in the literature, the observation that the effect of sST2 was independent of the VACS index is highly informative as this clinical score has been strongly associated with biomarkers of inflammation and captures the effect of a large number of them (hsCRP, D-dimer, cystatin C, IL-6 and TNF-alpha) [3,8,37]. In a previously published study, the addition of D-dimer and sCD14, but not IL-6, improves the predictive accuracy of the VACS index for mortality [8]. The only one article published on the association between ST2 and mortality in HIV-infected patients shows that IL-6 was no longer associated with all-cause mortality after adjustment for ST2 [28]. In the current study, IL-6 stayed independently associated with the mortality after adjustment for the VACS score and sST2 although borderline significant. The improvement of risk stratification by the addition of sST2 to the model was modest but consistent and it was in the same range of previously reported capacity of the other biomarkers [8]. Hence, our data extended a report showing an association of sST2 with all-cause mortality in HIV-infected patients [28], in a larger and different population including additional clinical and biological markers.

In addition to any clinical implication of such results, our study suggests the importance of the ST2/IL-33 pathway in the clinical progression of this infection and also opens new hypotheses in understanding the physiopathology of HIV. Even if we do not provide the full link about the physiopathological role of sST2 in all-cause mortality in our cohort, recent reports provide more insights about the systemic and ubiquitous importance of the ST2/IL-33 pathway beyond its role in the pathophysiology of cardiovascular disease. A link between this circuit and homeostasis of cells including T regulator cells in adipose tissue [38,39], lung, central nervous system [40], obesity, insulin resistance [41] has indeed been recently reported. Thus, we may speculate that by interfering with IL-33, sST2 may more generally impair tissue repair and downregulate the inflammatory or activation processes, thus worsening chronic activation in tissues and leading ultimately to severe non-AIDS events. Beyond its prognostic value in predicting clinical outcomes, our data also identified sST2 as a possible new player in the physiopathology of systemic organ dysfunctions leading to various causes of death. This hypothesis is reinforced by the fact that sST2 has been reported to be a prognostic biomarker for the development of a systemic and highly inflammatory disease such as graft-vs.-host disease (GVHD) including mortality in these patients [42]. Blocking sST2 with a neutralizing antibody was shown to decrease the production of proinflammatory cytokines and to increase the frequency of anti-inflammatory molecules and cells, improving GVHD severity and mortality [43]. Overall, these results validate sST2/IL-33 as an emerging key pathway that could be involved in systemic disease including HIV infection.

Our results have potential clinical and research consequences. First, this marker could be proposed for the risk stratification of HIV-infected patients enrolled in clinical research protocols testing new strategies aiming to improve clinical outcomes and/or to prevent the risk of death. Second, in addition to the control of HIV replication with ART, biological interventions will be necessary for the long-term control of infection at individual and population levels, notwithstanding the search for a cure. With this respect, our results reveal sST2 as an emerging targetable biomarker for HIV disease [44]. This underlines the need for future studies to understand the interplay between HIV and sST2/IL-33 circuit and its role in organ dysfunctions.


Special thanks to Matthias Bruyand and Charlotte Lewden for their contribution to the general organization of the cohort-team during these years. The ANRS CO3 Aquitaine Cohort is funded in part by the French National AIDS and viral hepatitis Research Agency (ANRS) through its Coordination Action no. 7 (AC7) with additional support from the National Institute for Medical Research (INSERM U1219), the Bordeaux School of Public Health (ISPED) and the Regional Coordination for HIV (COREVIH Aquitaine) of the Bordeaux University Hospital.

Conflicts of interest

There are no conflicts of interest.

Organization of the ANRS CO3 Aquitaine Cohort Study Group

Aquitaine Cohort (France) Coordination: F. Bonnet, F. Dabis

Scientific committee: F. Bonnet, S. Bouchet, D. Breilh, G. Chêne, F. Dabis, M. Dupon, H. Fleury, V. Gaborieau, D. Lacoste, D. Malvy, P. Mercié, P. Morlat, D. Neau, I. Pellegrin, JL. Pellegrin, S. Reigadas, S. Tchamgoué

Epidemiology and Methodology: G. Chêne, F. Dabis, C. Fagard, S. Lawson-Ayayi, R. Thiébaut, L. Wittkop.

Infectious Diseases and Internal Medicine: K. André, F. Bonnet, N. Bernard, L. Caunègre, C. Cazanave, I. Chossat, C. Courtault, FA. Dauchy, S. De Witte, D. Dondia, M. Dupon, P. Duffau, H. Dutronc, S. Farbos, I. Faure, M. Foissac, V. Gaborieau, Y. Gerard, C. Greib, M. Hessamfar-Joseph, Y. Imbert, D. Lacoste, P. Lataste, E. Lazaro, D. Malvy, J. Marie, M. Mechain, P. Mercié, E. Monlun, P. Morlat, D. Neau, A. Ochoa, JL. Pellegrin, M. Pillot-Debelleix, T. Pistone, I. Raymond, MC. Receveur, P. Rispal, L. Sorin, S. Tchamgoué, C. Valette, MA. Vandenhende, MO. Vareil, JF. Viallard, H. Wille, G. Wirth.

Immunology: JF. Moreau, I. Pellegrin.

Virology: H. Fleury, ME. Lafon, S. Reigadas, P. Trimoulet.

Pharmacology: S. Bouchet, D. Breilh, F. Haramburu, G. Miremont-Salamé

Data collection, Project Management and Statistical Analyses: MJ. Blaizeau, I. Crespel, M. Decoin, S. Delveaux, F. Diarra, C. D’Ivernois, C. Hanappier, D. Lacoste, S. Lawson-Ayayi, O. Leleux, F. Le Marec, E. Lenaud, J. Mourali, E. Pernot, A. Pougetoux, B. Uwamaliya-Nziyumvira, R. Nawabzad.

IT department and eCRF development: V. Conte, G. Palmer, V. Sapparrart, D. Touchard.


1. Smith CJ, Ryom L, Weber R, Morlat P, Pradier C, Reiss P, et al. Trends in underlying causes of death in people with HIV from 1999 to 2011 (D:A:D): a multicohort collaboration. Lancet 2014; 384:241–248.
2. Justice AC, Modur SP, Tate JP, Althoff KN, Jacobson LP, Gebo KA, et al. Predictive accuracy of the Veterans Aging Cohort Study index for mortality with HIV infection: a North American cross cohort analysis. J Acquir Immune Defic Syndr 2013; 62:149–163.
3. Duffau P, Wittkop L, Lazaro E, le Marec F, Cognet C, Blanco P, et al. Association of immune-activation and senescence markers with non-AIDS-defining comorbidities in HIV-suppressed patients. AIDS 2015; 29:2099–2108.
4. Kuller LH, Tracy R, Belloso W, De Wit S, Drummond F, Lane HC, et al. Inflammatory and coagulation biomarkers and mortality in patients with HIV infection. PLoS Med 2008; 5:e203.
5. Rodger AJ, Fox Z, Lundgren JD, Kuller LH, Boesecke C, Gey D, et al. Activation and coagulation biomarkers are independent predictors of the development of opportunistic disease in patients with HIV infection. J Infect Dis 2009; 200:973–983.
6. Neuhaus J, Jacobs DR Jr, Baker JV, Calmy A, Duprez D, La Rosa A, et al. Markers of inflammation, coagulation, and renal function are elevated in adults with HIV infection. J Infect Dis 2010; 201:1788–1795.
7. Sandler NG, Wand H, Roque A, Law M, Nason MC, Nixon DE, et al. Plasma levels of soluble CD14 independently predict mortality in HIV infection. J Infect Dis 2011; 203:780–790.
8. Justice AC, Freiberg MS, Tracy R, Kuller L, Tate JP, Goetz MB, et al. Does an index composed of clinical data reflect effects of inflammation, coagulation, and monocyte activation on mortality among those aging with HIV?. Clin Infect Dis 2012; 54:984–994.
9. Thiébaut R, Charpentier C, Damond F, Taieb A, Antoine R, Capeau J, et al. Association of soluble CD14 and inflammatory biomarkers with HIV-2 disease progression. Clin Infect Dis 2012; 55:1417–1425.
10. So-Armah KA, Tate JP, Chang C-CH, Butt AA, Gerschenson M, Gibert CL, et al. Do biomarkers of inflammation, monocyte activation and altered coagulation explain excess mortality between HIV infected and uninfected people?. J Acquir Immune Defic Syndr 2016; 72:206–213.
11. Shive CL, Jiang W, Anthony DD, Lederman MM. Soluble CD14 is a nonspecific marker of monocyte activation. AIDS 2015; 29:1263–1265.
12. Hunt PW, Sinclair E, Rodriguez B, Shive C, Clagett B, Funderburg N, et al. Gut epithelial barrier dysfunction and innate immune activation predict mortality in treated HIV infection. J Infect Dis 2014; 210:1228–1238.
13. Tenorio AR, Zheng Y, Bosch RJ, Krishnan S, Rodriguez B, Hunt PW, et al. Soluble markers of inflammation and coagulation but not T-cell activation predict non-AIDS-defining morbid events during suppressive antiretroviral treatment. J Infect Dis 2014; 210:1248–1259.
14. Ho J, Chen W, Chen M. Common genetic variation at the IL1RL1 locus regulates IL-33/ST2 signaling. J Clin Invest 2013; 123:4208–4218.
15. Molofsky AB, Savage AK, Locksley RM. Interleukin-33 in tissue homeostasis, injury, and inflammation. Immunity 2015; 42:1005–1019.
16. Sanada S, Hakuno D, Higgins LJ, Schreiter ER, McKenzie ANJ, Lee RT. IL-33 and ST2 comprise a critical biomechanically induced and cardioprotective signaling system. J Clin Invest 2007; 117:1538–1549.
17. Miller AM, Xu D, Asquith DL, Denby L, Li Y, Sattar N, et al. IL-33 reduces the development of atherosclerosis. J Exp Med 2008; 205:339–346.
18. Sánchez-Más J, Lax A, Asensio-López MDC, Fernandez-Del Palacio MJ, Caballero L, Santarelli G, et al. Modulation of IL-33/ST2 system in postinfarction heart failure: correlation with cardiac remodelling markers. Eur J Clin Invest 2014; 44:643–651.
19. Seki K, Sanada S, Kudinova AY, Steinhauser ML, Handa V, Gannon J, et al. Interleukin-33 prevents apoptosis and improves survival after experimental myocardial infarction through ST2 signaling. Circ Heart Fail 2009; 2:684–691.
20. Ky B, French B, McCloskey K, Rame JE, McIntosh E, Shahi P, et al. High-sensitivity ST2 for prediction of adverse outcomes in chronic heart failure. Circ Heart Fail 2011; 4:180–187.
21. deFilippi C, Daniels LB, Bayes-Genis A. Structural heart disease and ST2: cross-sectional and longitudinal associations with echocardiography. Am J Cardiol 2015; 115:59B–63B.
22. Chen LQ, De Lemos JA, Das SR, Ayers CR, Rohatgi A. Soluble ST2 is associated with all-cause and cardiovascular mortality in a population-based cohort: the Dallas heart study. Clin Chem 2013; 59:536–546.
23. Lei Z, Mo Z, Zhu J, Pang X, Zheng X, Wu Z, et al. Soluble ST2 plasma concentrations predict mortality in HBV-related acute-on-chronic liver failure. Mediators Inflamm 2015; 2015:535938.
24. Friis Moller N, Reiss P, Sabin CA, Weber R, Monforte AD, El Sadr W, et al. Class of antiretroviral drugs and the risk of myocardial infarction. N Engl J Med 2007; 356:1723–1735.
25. Friis-Møller N, Thiébaut R, Reiss P, Weber R, Monforte AD, De Wit S, et al. Predicting the risk of cardiovascular disease in HIV-infected patients: the data collection on adverse effects of anti-HIV drugs study. Eur J Cardiovasc Prev Rehabil 2010; 17:491–501.
26. Miyagaki T, Lin S, Chan LS. High levels of soluble ST2 and low levels of IL-33 in sera of patients with HIV infection. J Invest Dermatol 2011; 131:794–796.
27. Mehraj V, Jenabian M-A, Ponte R, Lebouché B, Costiniuk C, Thomas R, et al. The plasma levels of soluble ST2 as a marker of gut mucosal damage in early HIV infection. AIDS 2016; 30:1617–1627.
28. Secemsky EA, Scherzer R, Nitta E, Wu AHB, Lange DC, Deeks SG, et al. Novel biomarkers of cardiac stress, cardiovascular dysfunction, and outcomes in HIV-infected individuals. JACC Hear Fail 2015; 3:591–599.
29. Hsu DC, Ma YF, Hur S, Li D, Rupert A, Scherzer R, et al. Plasma IL-6 levels are independently associated with atherosclerosis and mortality in HIV-infected individuals on suppressive antiretroviral therapy. AIDS 2016; 30:2065–2074.
30. Thiébaut R, Morlat P, Jacqmin-Gadda H, Neau D, Mercié P, Dabis F, et al. Clinical progression of HIV-1 infection according to the viral response during the first year of antiretroviral treatment. AIDS 2000; 14:971–978.
31. Lewden C, Salmon D, Morlat P, Bevilacqua S, Jougla E, Bonnet F, et al. Causes of death among human immunodeficiency virus (HIV)-infected adults in the era of potent antiretroviral therapy: emerging role of hepatitis and cancers, persistent role of AIDS. Int J Epidemiol 2005; 34:121–130.
32. Binder H, Sauerbrei W, Royston P. Comparison between splines and fractional polynomials for multivariable model building with continuous covariates: a simulation study with continuous response. Stat Med 2013; 32:2262–2277.
33. Pencina MJ, D’Agostino RB. OverallC as a measure of discrimination in survival analysis: model specific population value and confidence interval estimation. Stat Med 2004; 23:2109–2123.
34. Pencina MJ, D’Agostino RB, Vasan RS. Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond. Stat Med 2008; 27:157–172.
35. Hilden J, Gerds TA. A note on the evaluation of novel biomarkers: do not rely on integrated discrimination improvement and net reclassification index. Stat Med 2014; 33:3405–3414.
36. Morlat P, Roussillon C, Henard S, Salmon D, Bonnet F, Cacoub P, et al. Causes of death among HIV-infected patients in France in 2010 (national survey). AIDS 2014; 28:1181–1191.
37. Mooney S, Tracy R, Osler T, Grace C. Elevated biomarkers of inflammation and coagulation in patients with HIV are associated with higher framingham and VACS risk index scores. PLoS One 2015; 10:e0144312.
38. Bapat SP, Myoung Suh J, Fang S, Liu S, Zhang Y, Cheng A, et al. Depletion of fat-resident Treg cells prevents age-associated insulin resistance. Nature 2015; 528:137–141.
39. Vasanthakumar A, Moro K, Xin A, Liao Y, Gloury R, Kawamoto S, et al. The transcriptional regulators IRF4, BATF and IL-33 orchestrate development and maintenance of adipose tissue-resident regulatory T cells. Nat Immunol 2015; 16:276–285.
40. Pomeshchik Y, Kidin I, Korhonen P, Savchenko E, Jaronen M, Lehtonen S, et al. Interleukin-33 treatment reduces secondary injury and improves functional recovery after contusion spinal cord injury. Brain Behav Immun 2015; 44:68–81.
41. Brestoff JR, Kim BS, Saenz SA, Stine RR, Monticelli LA, Sonnenberg GF, et al. Group 2 innate lymphoid cells promote beiging of white adipose tissue and limit obesity. Nature 2015; 519:242–246.
42. Vander Lugt MT, Braun TM, Hanash S, Ritz J, Ho VT, Antin JH, et al. ST2 as a marker for risk of therapy-resistant graft-versus-host disease and death. N Engl J Med 2013; 369:529–539.
43. Reichenbach DK, Schwarze V, Matta BM, Tkachev V, Lieberknecht E, Liu Q, et al. The IL-33/ST2 axis augments effector T-cell responses during acute GVHD. Blood 2015; 125:3183–3192.
44. Zhang J, Ramadan AM, Griesenauer B, Li W, Turner MJ, Liu C, et al. ST2 blockade reduces sST2-producing T cells while maintaining protective mST2-expressing T cells during graft-versus-host disease. Sci Transl Med 2015; 7:308ra160.

† See composition which is listed in the Acknowledgements section.


biomarkers; HIV infection; mortality; suppression of tumorigenicity 2

Supplemental Digital Content

Copyright © 2017 Wolters Kluwer Health, Inc.