Download how to define treatment success and failure

PROCEEDINGS
HOW TO DEFINE TREATMENT SUCCESS AND FAILURE*
—
Jens Lundgren, MD
ABSTRACT
Highly active antiretroviral therapy (HAART)
has dramatically improved the prognosis for HIVinfected patients in the last 5 years, as demonstrated by several studies, including the EuroSIDA
study—in which patients’ CD4 cell counts of <20
cells/mL improved to 20 to 49 cells/mL and the
projected 25% mortality rate over the following
year was reduced to <10, and a collaborative
effort among 8 European cohorts, in which
researchers concluded that treatment with HAART
had increased and maintained patients’ CD4 cell
counts and conferred clinical protection despite
the discontinuation of a secondary prophylaxis.
To date, virologic failure has been managed
in several ways, including switch therapy, continued therapy, interruption of HAART, and using
interleukin-2 to boost CD4 counts. Trials examining the possibility of drug recycling are currently
underway, as are trials examining the rate at
which failing regimens increase plasma viral
load, accumulation of resistant mutations, and
immunologic failure.
Success or failure of treatment can be measured and defined in several ways; each method
of definition should be explored and compared.
Controlled trials must offer critical analyses of the
efficacy and treatment-limiting toxicities of all
treatments they examine.
(Advanced Studies in Medicine 2001;1(12):483-485)
*This article is based in part on a presentation given
by Dr Lundgren at the 1st International AIDS Society
Conference on HIV Pathogenesis and Treatment.
Correspondence to Jens Lundgren, MD, Coordinating
Centre, EuroSIDA/Copenhagen HIV Programme (CHIP),
Department of Infectious Diseases, Hvidovre Hospital,
University of Copenhagen, 2650 Hvidovre, Denmark.
Advanced Studies in Medicine
■
ighly active antiretroviral therapy
(HAART) strives to suppress viral replication, prevents resistance, preserves
and improves immune function, and
prevents the development of clinical
disease. It has dramatically improved the prognosis of
HIV-infected patients in the last 5 years: in 1994 and
1995, before these drugs became available in Europe,
approximately 22% of patients were expected to die
within the next year, and more than 30% were expected to develop a new AIDS-defining event. With the
introduction of triple combination therapy, the incidence of death and new AIDS-defining events has
dropped and continues to decline (Figure 1). Clearly,
HAART is successful where it is available, as demonstrated by several studies.
As illustrated by the EuroSIDA study, among
patients with a CD4 cell count of <20 cells/mL before
initiating HAART, as many as 25% of patients will die
within the following year.1 However, even with small
HAART-induced increments in the CD4 count, there
are dramatic decreases in the death rate, ie, an increase
in the CD4 count to between 20 and 49 cells/mL
reduces the risk of death within the following year to
<10% (Figure 2).
In a collaborative effort among 8 European
cohorts, we studied more than 300 patients who had
previously been diagnosed with Pneumocystis carinii
pneumonia. Patients initially had clinical signs and
symptoms of severe immune deficiency that permitted
the development of opportunistic infections. Patients
with a history of P carinii pneumonia usually require
secondary prophylaxis to prevent relapse. Between
October 1996 and January 2000, the patients in this
cohort study discontinued secondary prophylaxis
while being treated with at least 3 anti-HIV drugs after
having at least 1 peripheral-blood CD4 cell count of
H
483
PROCEEDINGS
MANAGING VIROLOGIC FAILURE
To date, virologic failure has been
managed in several ways, including
switch therapy, continued therapy,
interruption of HAART, and using
interleukin-2 to boost CD4 counts.
Each method has its strengths and
weaknesses. Switch therapy offers
renewed control, but may result in the
defeat of a newly introduced class of
drugs. Continued therapy may prevent
toxicity and preserve the activity of new
drugs, but bears the risk of additional
resistance mutations and class resistance. Interruption therapy, on the
other hand, is not costly, nor will it
increase toxicity, but can cause the CD4
count to plummet and can increase the
risk of clinical disease. Interleukin-2 can
be effective regardless of the plasma viral
load, but the clinical protection of this
method is yet unproved.
Further controlled trials should
increase our understanding of the
effects and safety of treatment interruptions. Trials examining the possibility of
drug recycling are currently under way,
as are trials examining the rate at which
failing regimens increase a patient’s plasma viral load, accumulation of resistant
mutations, and immunologic failure.
We do know, however, that HAART has
been successful: it keeps patients alive
and increases CD4 cell counts. Even in
patients who lack complete control of
viral replication and are therefore at
484
Figure 1. Clinical Effect of HAART: Incidence of AIDS and Death
1994-2000
Incidence (per 100 PYFU)
>200 cells/mL. CD4 cell count levels
were subsequently maintained at >200
cells/mL for a median of 11 months.
After a median follow-up period of 13
months, CD4 cell counts remained at
levels ≥ 200 cells/mL. Researchers concluded that treatment with HAART
had increased patients’ CD4 cell counts
and clearly conferred clinical protection, despite elimination of disease-specific prophylaxis.2
9/
94
9
>9
3/
3/
9
3/
3
9/
3/
98 9/98
99
97 /97
/9
95 /96 /96
95
9
-9
3/
9/
9/
3/
-3
9/
3/
9/
3/
9
9
9
9
9
9
9
/
9
/9
8
95
8
9
7
7
9
6
6
5
Calendar period
EuroSIDA
Reprinted with permission from The EuroSIDA study group, 2001, Coordinating Centre,
Copenhagen, Denmark.
Figure 2. Death Rate According to Latest CD4 Count in People
Treated with HAART, with Initial CD4 Count <20 Cells/mL
Reprinted with permission from The EuroSIDA study group, 2001, Coordinating Centre,
Copenhagen, Denmark.
Vol. 1, No. 12
■
December 2001
PROCEEDINGS
increased risk of drug resistance, the plasma viral load
still diminishes. Plasma viral load, however, remains a
weak predictor of the risk of clinical disease progression;
another strong predictor may be the current level of
hemoglobin.
Most controlled trials analyze a regimen’s ability to
suppress plasma viral load and the tolerability of the
regimen, since discontinuations are counted as virologic failures. Despite the important data often provided by trials that assess the efficacy and toxicity of
drugs, trial design, small patient cohorts, and short
duration of some studies may severely limit these data.
An efficacy-only endpoint requires continued followup in patients who discontinue treatment. Treatmentlimiting toxicity should be evaluated separately.
Extended follow-up until the completion of the trial,
regardless of whether the patient discontinues the
study medication, should be encouraged to better
determine the presence of late-onset adverse events
and resistance. Moreover, studies that focus solely on
effect, such as suppression of plasma viral load that is
Advanced Studies in Medicine
■
independent of discontinuations, may prove useful.
The risk of virologic failure associated with numbers
needed to treat vs the numbers needed to treat to harm
should also be examined.
Several definitions of treatment success and failure
exist; however, each should be explored and compared. Controlled trials must offer critical analyses of
the efficacy and treatment-limiting toxicities of all
treatments they examine.
REFERENCES
1. Miller V, Mocroft A, Reiss P, et al. Relations among CD4
lymphocyte count nadir, antiretroviral therapy, and HIV-1
disease progression: results from the EuroSIDA study. Ann
Intern Med. 1999;130:570-577.
2. Ledergerber B, Mocroft A, Reiss P, et al. Discontinuation of
secondary prophylaxis against Pneumocystic carinii pneumonia in patients with HIV infection who have a response
to antiretroviral therapy. Eight European Study Groups.
N Engl J Med. 2001;344:168-174.
485