Inflammatory and Coagulation Biomarkers and Mortality in Patients with HIV Infection

October 2008

Inflammatory and Coagulation Biomarkers and Mortality in Patients with HIV Infection
Kuller LH, et al. PLoS Medicine 2008;5:e203. [PubMed Abstract]

Background: The Strategies for Management of Anti-Retroviral Therapy (SMART) trial showed increased mortality for patients randomized to intermittent antiretroviral treatment compared to continuous viral suppression based on CD4 cell count. The question posed in this trial concerns the hypothesis that HIV viremia associated with drug discontinuation resulted in inflammatory changes that increased the risk of mortality.

Methods: The “Biomarker Sub-study” of SMART was done with 1,415 patients enrolled in the US using specimens shipped frozen to a central repository. The analysis included four inflammatory markers including: a high sensitivity C-reactive protein (hsCRP), IL-6, amyloid A, and amyloid P, and two coagulations markers, D-dimer and prothrombin fragment 1+2 (F 1.2). These biomarkers were analyzed blindly and correlated for the drug conservation (DC) and the viral suppression (VS) groups in the parent trial. There were two studies: the first was a nested case-controlled study to correlate biomarkers with mortality; the second study compared the DC and the VS groups for biomarker changes.

Results: There were three biomarkers that showed important associations with mortality: hsCRP, IL-6, and D-dimer. There were 55 deaths in the DC group and 30 in the VS group. The following Table (1) summarizes the baseline data for the 85 deaths in both groups compared to 170 controls that were matched by age, sex and date of randomization.

TABLE 1 Baseline Data for Biomarkers
	Deaths n=85**	Controls n=179**	P
hsCRP (ug/mL)*	4.3	2.1	0.005
IL-6 (pg/mL)*	3.8	.3	<0.0001
D-dimer (ug/mL)*	0.5	0.3	<0.0001

*Median levels.
**No significant difference between cases and controls for drug conservation versus viral suppression.

The following Table (2) provides data for these three biomarkers for the latest levels prior to death which, again, shows highly significant correlations between these biomarkers and death.

TABLE 2 Risk of Death Based on Latest Level of Biomarkers
Biomarker	OR	95% CI	P value
hsCRP	2.4	1.4-4.2	0.003
IL-6	2.0	1.2-3.1	0.006
D-dimer	2.2	1.1-4.1	0.02

The second study evaluated biomarkers for 249 participants in the DC and 250 in the VS participants at study entry and at one month after randomization to drug discontinuation (DC). The following Table (3) summarizes the change in biomarkers based on viral load for IL-6 and D-dimer:

TABLE 3 Change in log 10 IL-6 and D-dimer from Baseline to One Month Correlated with Viral Load: Change in Biomarkers
Viral load	No.	IL-6* (log 10 pg/mL)	D-dimer (log 10 ug/mL)**
<400	34	0.08	0.0
400-10,000	30	0.14	0.04
10,000-50,000	29	0.20	0.11
>50,000	39	0.33	0.28

*P=0.0003 for trend.
**P=0.0005 for trend.

These data show a highly significant correlation for the change at one month based on virologic rebound following drug discontinuation.

Conclusions: The authors conclude that IL-6 and D-dimer correlated strongly with mortality and that drug interruption (DC group) further increased the risk of death with further increases in IL-6 and D-dimer.

Comment: This is an important paper that was published in 2008 and was reviewed here because it has such an important impact on current concepts regarding HIV infection, treatment and cardiovascular disease risk. The above summary is a distillate of an enormous amount of information that includes several thousand analyses of biomarkers done in the context of the study. Most important is the third Table that shows the correlation between viral load and change in IL-6 and D-dimer. The strongest correlations were with baseline data at study entry, but there was also an increased risk of death in participants who had discontinuation of HAART that showed a very strong correlation between virologic rebound for IL-6 and D-dimer. These data fit with the current concept of HIV infection as a health risk independent of immune-compromise.