What CD4 Cell Count Levels are Associated With a Reduced Risk of Cancer? 'When To Begin HAART'
Reported by Jules Levin
5th IAS Capetown July 19-22 2009
N. Crum-Cianfl one*1,2, K. Huppler Hullsiek1,3, A. Ganesan1, V. Marconi1, A. Weintrob1, R.V. Barthel1, B. Agan1, and the Infectious Disease Clinical Research Program 1. Infectious Disease Clinical Research Program, Bethesda, USA, 2. Naval Medical Center San Diego, San Diego, USA, and 3. University of Minnesota, Minneapolis, USA
From Jules: Regarding the question of 'when to begin HAART', risk for non-AIDS cancers are 2-fold higher (100% higher) when CD4s are below 700 and 3.7 fold higher when CD4 is 200-349 in the pre-HAART era in this study and also 2-fold higher for AIDS-defining risk when CD4s are <500 in the pre-HAART era (see tables 5a and 5d). This suggests that in the HAART era currently starting HAART at 700 CD4s reduces non-AIDS cancer risk and starting at 500 CD4 count reduces AIDS cancer risk. In other words, these data suggest to start HAART at 700 CD4 to optimally reduce cancer risk. This is something I said 10 years ago and I'm sure others raised this concern also when discussing this in the context of when to begin HAART: that risk for cancers are reduced if you keep the CD4 count higher and start HAART before the CD4 declines below the normal level which is I think generally considered to be 500-700 at the low end of the spectrum. When taking a treatment interruption and thereby allowing CD4 to decline and viral load to increase for a period of time it appears you can increase risk for cancers. A study at IAS I emailed yesterday found cumulative exposure to detectable viral load and independently cumulative exposure to CD4 <350 increased NHL risk, and there have been similar studies reported recently, at CROI 2008 & 2009. Another question is if cd4 is 1000 does that reduce cancer risk further and perhaps that depends on the quality of the cd4s.
While the risk of ADCs was highest at CD4 counts < 200 cells/mm3, increased risk was also present at counts <500. In the HAART era CD4 count of 350-499 in HAART era increased risk 3.6 fold compared to CD4 >500 and in pre-HAART era risk was 1.9 fold higher. For CD4 count 200-349 risk was 7.6 fold higher compared to >500 CD4s while the risk was 3.4 fold higher. For <200 CD4s the risk for ADCs in HAART era was 29.8 fold higher compared to >500 cd4s and the risk was 20.2 fold higher in pre-HAART era.
There was an association between NADCs and CD4 counts < 700 cells/mm3 during follow-up. This association was mainly seen in the pre-HAART era and for skin cancers.
During the HAART era, there were no associations between non-skin NADCs and CD4 strata.
Study limitations include the small numbers of nonskin NADCs and the variety of cancers in this category.
Further studies examining large numbers of individual NADCs are needed to further examine this relationship.
Maintaining robust CD4 counts is useful in reducing AIDS-defining cancers among HIV patients, but did not have a significant impact on NADCs in the HAART era.
AIDS-defining cancers: KS, non-Hodgkin's lymphoma, invasive cervical cancer
Non-AIDS defining cancers: skin cancers, anal carcinoma, Hodgkin's disease, prostate cancer and other types.
Study Population Characteristics:
· At Time of HIV Diagnosis:
- Demographics and HIV characteristics are shown in Table 1
Table 1: Baseline Study Characteristics at the
Time of HIV Diagnosis
At Time of Cancer Event or Censoring:
- Characteristics of those with and without cancer
are shown in Table 2
Table 2: Study Characteristics at Cancer
Event or Censoring
CD4 Cell Counts among HIV-Infected Persons with and
without a Cancer Event
The CD4 count varied among those who had cancer compared to those without a cancer event (Table 3).
Persons who developed cancer, particularly an ADC, had a lower CD4 nadir before the cancer and a lower CD4 count proximal to the time of diagnosis compared to those without cancer
The median CD4 counts proximal to ADC and NADC events were 53 and 454 cells/mm3, respectively.
Table 3: Description of Median (IQR) CD4 Counts
by Cancer Status
Incidence Rates of Cancer by Proximal CD4 Cell Count
The incidence rate of cancers varied by the proximal CD4 cell count (Tables 4a and 4b).
Overall, the highest rate of ADCs was 44.7 per 1000 PYs at CD4 counts of <200 cells/mm3. As the CD4 count increased, the incidence of ADCs progressively
decreased with the lowest rate of 1.3 per 1000 PYs with a CD4 count of >700 cells/mm3.
The incidence of NADCs was similar at CD4 counts of <200 cells/mm3 (5.1 per 1000 PYs) and 500-700 cells/mm3 (5.6 per 1000 PYs). The rates fell to 3.6 per 1000 PYs at CD4 counts >700 cells/mm3.
The rates of ADCs and NADCs also varied by HAART era (Table 4a & 4b). Compared to the pre-HAART era, the rate of ADCs were lower in each CD4 strata during the HAART era. On the other hand, rates of NADCs were slightly higher in the HAART vs. pre-HAART eras.
The rates of non-skin NADCs were relatively constant over all the CD4 cell strata, whereas skin cancers occurred at lower rates at CD4 counts>700 cells/mm3.
Table 4a: Incidence Rates of AIDS-Defining Cancers
Per 1000 Person Years by Proximal CD4 Cell Count:
Overall, Pre-HAART*, HAART** Eras
Table 4b: Incidence Rates of Non-AIDS-Defining
Cancers Per 1000 Person Years by Proximal CD4 Cell
Count: Overall, Pre-HAART*, HAART** Eras
Time to Event Analyses
Compared to CD4 counts > 500 cells/mm3, the risk of an ADC was elevated at CD4 counts <200 (HR 21.5; 95 CI% 13.9-33.3), 200-349 (HR 3.7; 95% CI 2.2-6.1) and 350-499 (HR 2.1; 95% CI 1.2-3.6) (Table 5a).
Table 5a: Multivariate Time to Event Analyses* for
Factors Associated with AIDS-Defining Cancer
The risk of an NADC event was not significantly increased for CD4 counts < 500 (HR 1.2; 95% CI 0.9-1.6, p=0.32) compared to > 500.
However, compared to those with CD4 count > 700 cells/mm3, the risk of NADCs were increased at CD4 counts < 500 (HR 1.7; 95% CI 1.1-2.7) and 500-699 (HR 1.9; 95% CI 1.1-3.1) (Table 5b).
The reduction in the risk of NADCs at CD4 cell counts >700 was predominantly seen in the pre-HAART era.
Table 5b: Multivariate Time to Event Analyses* for
Factors Associated with Non-AIDS-Defining Cancer
*adjusted for age at HIV diagnosis, race, and gender; adjusted and stratified by year of HIV diagnosis
HR, hazards ratio; CI, confidence interval
The associations between CD4 counts and two categories of NADCs (skin and non-skin) are shown in Tables 5c and 5d.
Non-skin NADCs were not significantly related to CD4 counts, whereas HIV-infected persons with CD4 counts >700 had a lower risk of skin cancers.
The findings regarding NADCs and CD4 counts were not significantly changed when HAART was adjusted for in the models.
Table 5c: Multivariate Time to Event Analyses* for
Factors Associated with Skin Cancer
Table 5d: Multivariate Time to Event Analyses*
for Factors Associated with Non-Skin