icon_folder.gif   Conference Reports for NATAP  
  ICAAC 41st Interscience Conference on Antimicrobial Agents and Chemotherapy
Chicago, Illinois, December 16-19
Back grey_arrow_rt.gif
The Prevalence of Antiretroviral Drug Resistance in the US
Reported by Jules Levin
  Doug Richman, from the University of San Diego at California and a resistance researcher of note, reported on the prevalence of resistance to HIV drugs in a large group of patients participating in a study. Additional contributing authors to this study were the Virologic Company, the RAND corporation, Sam Bozzette, S Morton, S Chien, N Hellman T Wrin, and K Dawson.
This report was presented by Richman in a Late Breaker oral session Tues December 18 at the ICAAC meeting in Chicago. He said the true prevalence of HIV drug resistance in the US has been difficult to ascertain from studies of selected populations of patients. So, the objective of this study is to estimate the prevalence of antiretroviral drug resistance in a highly defined group of study patients who receive care in the US.
Richman described the study patients and study methods: the HIV Cost and Service Study (HCSUS) is a longitudinal study representative of all US HIV-infected adults who received medical care in early 1996 and survived to early 1999; 1906 blood samples obtained by Quest Diagnostics were looked at and represent 208,000 US patients (weighted n).
36% of patients (n=698) had <500 copies/ml (weighted n=76,000). 63% (n=1200) had >500 copies/ml (weighted n=132,000). Drug resistance testing was performed on 1080 samples (weighted n=117,000). Richman said that results were weighted to represent the 132,000 (63%) of the 209,000 US adult HIV patients under care in early 1996 who survived until 1999 and had >500 copies/ml viral load.
Demographics of the patients (weighted n=117,000): age- 55% 35-49, 37% 18-34, 7% >50. 75% male. Risk groups- 25% IVDU, 46% MSM, 21% heterosexual, 8% other. Race: 44% non-Hispanic white, 38% non-Hispanic black, 14% Hispanic, 3% other. Median viral about 30,000. Education: 31% HS grad., 27% some HS, 42% college. 40% of patients were from the South, 26% from the Northeast, 9% from the Midwest, and 25% from the West.
Editorial note: my impression is that 40% of patients from the South may not represent a true picture of the demographics of HIV in the US.
Practice size of doctor: 0-10 3%, 11-100 19%, 101-500 60%, >500 18%. Health insurance: none 16%, Medicaid 32%, Private 25%, Medicare +/- other insurance 27%.
Drug resistance testing was performed by the Virologic Phenosense assay. Resistance was defined by the presence of reduced drug susceptibility compared to a drug-sensitive reference virus. The cut-off for sensitivity by the patient sample virus to the drug was:
  • DdI and d4T: >1.7 fold

  • Abacavir: >4.5 fold

  • All other drugs: >2.5 fold

    Editorial note: Using these cutoffs I question whether they in some instance may be too low and thus perhaps overestimating the numbers of patients with drug resistance and the severity of the resistance. In particular I question the 2.5 cut-off for all other drugs. I don't question the cutoffs for ddI, d4T and abacavir, but 2.5 may be too low for protease inhibitors. However, I do recognize that I think the development of drug resistance to HIV antiretrovirals is an increasing concern both for society and for individual patients.
    78% of patients with detectable HIV viral load had any drug resistance. If you assume that patients with <500 copies/ml had no drug resistance this would still represent 50% of the entire US HIV-infected population. But even with an undetectable viral load a patient can still have drug resistance, which is being suppressed by therapy.
    70% had resistance to a NRTI with 3TC being the most often drug people had resistance to. 42% had resistance to a PI. 31% had resistance to a NNRTI. 51% had resistance to 2 or more classes of drugs. And 14% had resistance to 3 classes of drugs. Richman reminded the audience that these numbers represent the time point at 1999 which was 3 years into the HAART era and it is now 3 years later. Again, these percentages assume there is no drug resistance when viral load is <500.
    Factors Associated with Drug Resistance
    The lowest historical CD4 count a patient had was associated with drug resistance. So if a patient had 0-49 CD4s at one point in their past they were 90% likely to have drug resistance. If CD4s were 50-199 they were 80% likely to have had drug resistance. When cd4s were 200-499 there was a 70% chance the patient would develop drug resistance. And if the lowest past CD4 count was >500 patient the likelihood of developing resistance was 50%. The current CD4 count was not associated with developing resistance. So no matter what your CD4 count was at the time of the blood sample it didn't appear to affect whether or not you had drug resistance. Higher baseline viral load was also associated with a risk for developing drug resistance. Baseline viral load and the lowest CD4 count in the past were statistically significant factors (p<0.05) in predicting the development of drug resistance.
    Risk Group
    MSM had a slightly higher risk for developing resistance than IVDUs, heterosexuals and others. Males had a higher risk for drug resistance than women. Whites had a slightly higher risk and individuals with more education were at slightly higher risk for developing resistance.
    Using a multivariate analysis the risk factors associated with developing drug resistance were low CD4s, male gender, higher viral load, geographic region with Midwest being the least risky (Richman said this was probably because there may be less access to treatment in the Midwest). Practice size of the doctor was also a factor. Patients who were seeing doctors who treated more HIV-infected patients were least likely to develop resistance. So the experience of the treating physician was a factor. Patients seeing doctors with more experience in treating HIV were less likely to develop drug resistance.
    In summary, Richman concluded
    • that 63% of adults in care who survived the first two (I think he meant 3 yrs.) years of the HAART era had viral load >500.

    • 78% of these patients had resistance to at least one drug and 50% had resistance to multiple classes of HIV antiretroviral drugs.

    • 87% of patients who were currently taking therapy had resistance while 41% who were not currently taking therapy had resistance (p=0.0001). Although he didn't mention it I assume this could be because resistance may not be detectable when a person is not taking the drug.

    • Having advanced HIV disease was associated with risk for developing resistance (84% of patients with CDC stage C).

    • A lower nadir CD4 count was associated with risk for developing resistance (50% when CD4s were >500 vs 80% when CD4s were <200, 90% when CD4s were <50.

    Editorial note: These last two factors suggest the a higher risk for developing resistance when the initiation of therapy is delayed too long. Certainly, waiting until a patient has CD4s below 200 and when a patient has an opportunistic infection or AIDS defining event appears associated with risk for developing resistance.
    • Some factors related to greater access to care are associated with a higher risk for developing resistance: whites, highly educated, privately insured, lower current viral load, and men having sex with men were risk factors identified in this study.

    • Seeing a doctor who is treating fewer HIV-infected patients is associated with a higher risk for developing drug resistance.

    In his final comments, Richman warned about the implications of this information for future treatment efficacy and transmission of drug resistance virus.
    Editorial note: We need more HIV drugs/treatments that will be effective against HIV viruses with resistance to the current drugs. Treatment may be less effective for patients who contract HIV viruses with drug resistance. A few preliminary studies from other researchers have already shown that the response to treatment may not be as good for individuals who had drug resistant virus transmitted to them. As well, this data shows the high and increasing likelihood that drug resistance is being transmitted. Unless improved and more effective prevention methods are successful, perhaps an increasing number of newly infected patients will have drug resistant virus. Prevention approaches have not been successful so far in avoiding this situation. Perhaps new approaches are needed. One suggestion which may help to a degree is to use resistance testing after a person is diagnosed with HIV. Although this method is certainly not foolproof because resistant virus may not always be detectable, this may help identify some patients with resistance. If you can find which drugs the patient may have resistance to, then you can try selecting a regimen that may have the most success. It is not now generally accepted that resistance testing should be used for newly diagnosed patients. But perhaps the DHHS Guidelines committee and doctors and researchers should reconsider and discuss this further. Resistance testing is expensive and increases the burden on our health system. Another important factor is that most ordinary HIV treating care providers do not understand how to use and interpret resistance testing. As well. there have also been questions raised about the relaibility of some of the testing and test interpretation by some commercial labs. We need better oversite of these problems. More intense education on resistance testing for doctors, care providers and patients is crucial. The improper interpretation and application of resistance testing in selecting a regimen can do more harm than good.