Ribavirin: A meta-analysis

There is an acute respiratory infections review group of the
Cochrane Collaboration - a group devoted to systematically reviewing
the evidence from randomized controlled trials (RCT's) using
methodologically accepted approaches. Two reviews have been
submitted so far. One is of bronchodilators and steroids for
bronchiolitis.

The senior author is Elaine Wang, MD at the Hospital for Sick
Children in Toronto who is an epidemiologist and specialist in
pediatric infectious diseases. The results of her meta-analysis
should be available in a couple of months.

Elaine and I also submitted a systematic review of Ribavirin for RSV
to the Cochrane Database and our manuscript was also recently
accepted for publication in Archives of Pediatrics and Adolescent
Medicine. For what it is worth, here is a summary of our
conclusions. You will need to refer to the published manuscript for
further details.

Ribavirin has an absolute MINIMUM COST (not charges) of $1460 per day
(usual course of 3 to 7 days). (Yes, the actual costs are probably
much higher but nobody can argue about the way I calculated this
minimum).

Therefore, improvements in clinically important outcomes such as
mortality reduction, a decrease in respiratory deterioratiion, or a
decrease in hospital stay are needed to make this drug
cost-effective. Therefore, it is probably not cost-effective to use
ribavirin in populations already at low risk of death or respiratory
deterioration.

There are 3 trials of high risk patients (2 ventilated and one
non-ventilated) which analyze mortality or have strict criteria for
defining respiratory deterioration. If you pool the data from these
3 trials using meta-analysis (overall death rate about 14%) you will
see a strong trend for a decrease in mortality and a decrease in
respiratory deterioration (all 3 trials show a trend in the direction
of decrease for ribavirin for these 2 outcomes and the trend when the
3 are pooled is quite strong and almost reaches statistical
significance). Failure to reach statistical significance is not
surprising because the overall sample sizes even with 3 trials pooled
(99 patients for mortality and 116 patients for respiratory
deterioration) are not sufficient to rule out a 20% mortality
decrease (or even a 50%) decrease with 80% power. Therefore, a
larger study of at least 250 patients per treatment arm is needed
using only high risk patients
(premature, cardiac, already on vent, immunosuppressed, chronic lung
disease) who have a mortality rate or a probability of respiratory
deterioration of 10% or more (respiratory deterioration would need
to be explicitly defined by increasing PEEP and FiO2 needs on the
ventilator as it was in the previous vent trials as an outcome). A
trial of this size would rule out a 50% mortality reduction if no
significant difference was found. To rule out a 20% decrease in
mortality with the drug with 80% power you need over 1000 patients.
If the actual mortality rate in your population is lower than 10%,
you would also need many more patients.

Conclusion: Ribavirin shows a trend towards a decrease in mortality
and a trend towards a decrease in respiratory deterioration in high
risk patients but the sample is insufficient to rule in or out a
positive benefit with certainty.

What is needed would be a multicenter study to obtain an adequate
number of patients. ICN pharmaceuticals was not interested in
funding such a study when approached by Canadian investigators.
Other trials of interventions in patients with respiratory failure
may well suffer from the same sample size issues as ribavirin
studies.

Using smaller samples when studying ribavirin or other interventions
in this group will lead to the problem of comparing only intermediate
outcomes such as length of time on the ventilator, ICU stay,
oxygenation etc and these are flawed by the amount of variability
inherent in our methods of management and their questionable overall
clinical importance. We are working on developing a computerized
protocol for ventilatory management of pediatric patients with acute
hypoxemic respiratory failure at the Pulmonary Informatics Research
Group, LDS Hospital, University of Utah which would help overcome
some of this variability (possibly systematic bias) that could be
masking true outcome effects.

Last point: The nebulized water control used in the strongly
positive Smith study - I reviewed the entire literature on this and
Dr. McBride was correct when he stated that the amounts of water
needed to incite bronchospasm in patients with known reactive airways
are huge in comparison the amount going through the circuit with a
SPAG drying chamber. The water control in the Smith study probably
did not reach the patient and the extreme length of time on the vent
for the control group in that study was probably more a result of
what is seen in tiny trials like that (28 patients total) - spurious
results.

I admit that this review does not add a huge amount to the comments
of many experts. However, it may be wise to not take a strong
approach for or against ribavirin and to think of it as a drug that
needs further study and a drug which might hold promise for patients
at high risk of mortality or severe respiratory failure.


Adrienne G. Randolph, MD
Pulmonary Informatics Research Division
LDS Hospital
(current email address: randolph_a@a1.tch.harvard.edu)

Back to File Cabinet Index