Baseline inflammation and coagulation markers and changes over four weeks during a treatment interruption are strongly linked to HIV viraemia and risk of mortality

Simon Collins, HIV i-Base

Lewis Kuller presented further analysis from the SMART study. In summary, this international study randomised around 5,500 patients to CD4-guided treatment interruptions or continuous treatment, was stopped early because of excess death in the intermittent treatment group: with 55 vs 30 deaths over the first 16 months, most not related to opportunistic infections.

Baseline markers of inflammation and clotting as they related to total mortality, the changes of these variables, and how they related to outcome.

Two matched controls for each death prior to January 2006 when the study was stopped, were matched on a variety of variables. Logistic regression analysis was used to estimate odds ratios (OR) for mortality with adjustment for cardiovascular, HIV, co infection risk factors and other demographic factors. Baseline characteristics relating to deaths included lower CD4 counts, and a higher proportion of current smoker, diabetes, treated blood pressure and CVD, in the cases compared to the controls, all of which were adjusted for.

Markers analysed included inflammatory markers: serum amyloid A and serum amyloid P, C-reactive protein (CRP) – acute phase proteins (pentraxins), and IL-6 (a major inflammatory marker and stimulant of CRP and stimulant of release of tissue factor from smooth muscle cells in the endothelium).

The markers of coagulation that were measured were D-dimer (a measure of the breakdown of fibrogenesis that is a powerful risk factor for CVD) and prothrombin fragments 1+2 (F/1.2).

All markers have been strongly associated with cardiovascular risk in HIV-negative patients.

Kuller first reported that many of these markers were elevated at baseline and that this was significantly associated with all cause mortality. Adjusted OR (4th vs 1st quartiles) showed an almost 12-fold increased risk associated with elevated IL-6 and a 26-fold increased risk associated with increased D-dimer (both p<0.0001). CRP was associated with a 3-fold risk (p=0.03).

IL-6 and D-dimer, but not the other markers, also increased over 4 weeks in patients not on treatment, compared to those on continuous treatment (p<0.0001). Changes in both markers were strongly correlated with increases in viral load. Change from baseline to week four levels in IL-6 (OR 5.3, 95%CI 1.6-17.1; p=0.006) and D-dimer (OR 5.0. 95%CI 1.3 -18.9; p=0.02) were also strongly correlated with risk of death.

Kuller suggested that activation of coagulation and inflammation markers could impact multiple organs: the association with all cause mortality was stronger that to CVD.

This may occur as a consequence of a challenge to the endothelium. Viraemia affects vascular endothelium which increases production of tissue factor transcription, which activate the extrinsic clotting pathways. Elevated d-dimer and IL-6 clearly identified patients at highest risk of death, at a level unseen with any other marker in HIV-negative predictive biomarker studies (which report 1-2-fold increased risk. He concluded by suggesting that this opened the questions of whether this may be a useful marker to monitor in patient management.

These findings were initially presented to the IAS conference last July. [2]


  1. Kuller L et al. Elevated Levels of Interleukin-6 and D-dimer Are Associated with an Increased Risk of Death in Patients with HIV. Oral abstract 139.
    This oral presentation is available to view online from the conference website.
  2. See HTB August/September 2007

Links to other websites are current at date of posting but not maintained.