icon-    folder.gif   Conference Reports for NATAP  
  18th CROI
Conference on Retroviruses
and Opportunistic Infections
Boston, MA
February 27 - March 2, 2011
Back grey_arrow_rt.gif
CD4 Decline Before ART Faster in More Recent Years in ATHENA Cohort: "may have implications for planning on when to start cART."
  18th Conference on Retroviruses and Opportunistic Infections, February 27-March 2, 2011, Boston
Mark Mascolini
Before subtype B-infected gay men in the ATHENA cohort start antiretroviral therapy (ART), downward CD4 slopes have become significantly steeper in more recent years, according to two of three models used to calculate slopes [1].
Earlier the ATHENA investigators reported higher viral loads and lower CD4 counts at viral set point (9 to 27 months after seroconversion) in subtype B-infected cohort members during recent years than 10 years ago [2]. To look more closely at initial CD4 count and CD4 slope before ART, the researchers focused on gay men infected with subtype B HIV-1 and born in the Western Europe or North America. Study participants had laboratory evidence of recent infection or a maximum of 1 year between the last negative and first positive HIV-1 antibody test. Everyone had one or more CD4 counts and viral loads available from 9 to 27 months after HIV seroconversion and before they started ART.
Gauging CD4 decline before ART is not straightforward, the investigators noted, because people with faster declines start ART sooner. To account for this dropout trend, they devised three models with different assumptions about dropout pattern: (1) a random effects model, (2) a selection model, and (3) a pattern-mixture model. All models included patient age and year of seroconversion. The random effects model assumes that dropout depends only on observed CD4 count and will be biased if dropout depends on unobserved CD4 cell counts after dropout. The selection and pattern-mixture analyses model dropout and longitudinal CD4 count simultaneously. The pattern-mixture analysis models an interaction effect between seroconversion periods and dropout. "Dropout" meant starting ART, getting an AIDS diagnosis, having a CD4 count under 100, or dying.
The analysis included 607 men, 111 infected from 1984 to 1995, 139 infected from 1996 to 2002, and 357 infected from 2003 to 2007. Median age at seroconversion was 35.2 in 1984-1995, 34.6 in 1996-2002, and 37.9 in 2003-2007. Median first CD4 count between 9 and 27 months after seroconversion fell steadily over the three study periods:
1984-1995: 580 (interquartile range [IQR] 450 to 850)
1996-2002: 550 (IQR 450 to 720)
2003-2007: 510 (IQR 390-650)
Among men who seroconverted from 1984 through 1995, the three models gave similar estimates for CD4 decline (slope) because dropout was infrequent. In 1996-2002 and 2003-2007, when dropout became more frequent, the pattern-mixture and selection models estimated a steeper CD4 decline than did the random effects model. The pattern-mixture model and selection model both calculated steeper CD4 slopes in the second two periods, though the slopes were similar in 1996-2002 and 2003-2007:
Pattern-mixture model estimated median CD4 decline at 9 to 48 months (overall P = 0.02)
1984-1995: -263
1996-2002: -320
2003-2007: -315
Selection model estimated median CD4 decline at 9 to 48 months (overall P = 0.02)
1984-1995: -272
1996-2002: -327
2003-2007: -320
The ATHENA team concluded that standard random effects models underestimate the CD4 decline slope before ART begins. The two models that take dropout into account suggest that CD4 declines before ART grew steeper with the arrival of triple therapy in 1996. The investigators did not speculate on reasons for this steeper CD4 fall in recent years, but they suggested their findings "may have implications for planning on when to start cART."
In their earlier publication, the ATHENA team did offer a hypothesis to explain their findings [2]. As Luuk Gras noted in an e-mail message to NATAP, "we speculate that the higher viral load and lower CD4 cell count at viral setpoint could be due to [viral] selection. We speculate that in the cART era more transmissions occur during the primary infection phase characterized by very high viral load levels than in the pre-cART era. If setpoint viral load is (in part) hereditary, this could explain the higher viral load, lower CD4 cell count, and steeper CD4 declines after seroconversion in more recent years."

1. Gras L, Geskus R, van Sighem A, et al. Faster CD4 cell count decline before the start of ART in patients with HIV-1 seroconversion in more recent calendar years. 18th Conference on Retroviruses and Opportunistic Infections. February 27-March 2, 2011. Boston. Abstract 1082.
2. Gras L, Jurriaans S, Bakker M, van Sighem A, et al. Viral load levels measured at set-point have risen over the last decade of the HIV epidemic in the Netherlands. PLoS One. 2009;4(10):e7365. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2753776/?tool=pubmed