|
There are three randomized trials which evaluate the
ICD as a treatment for patients with previous documented sustained VT or VF. Two of these,
The Cardiac Arrest Study Hamburg (CASH)2 and the
Canadian Implantable Defibrillator Study (CIDS)3 will
be reported next year. The Antiarrhythmic vs Implantable Defibrillator (AVID) Study4 has recently been reported. There were 1016 patients
randomized to receive either an ICD or drug therapy, which was specified as either
amiodarone or sotalol. Forty-five percent of enroled patients had VF and the rest had
hemodynamically unstable VT. Drug therapy was randomly allocated in patients eligible for
either drug. Only 13 of 509 patients randomized to drug therapy actually were discharged
from hospital on sotalol; the rest received amiodarone. The mean dose of amiodarone at one
year was 331 mg/day and 87% of patients remained on amiodarone at one year. There was a
marked imbalance in beta blocker use between ICD and amiodarone patients with 45% of ICD
patients receiving this therapy compared to 13% of drug therapy patients.
In the AVID study there was a reduction in mortality with the ICD. Over a mean
follow-up of 18 months, crude death rates were 15.8 ± 3.2% for the ICD versus 24.0 + 3.7%
for drugs (p < 0.02). The relative risk reductions at one, two and three years were 39
± 20% 27 ± 21% and 31 ± 21% (± 95% confidence intervals). Adjustment for imbalances in
baseline features and concomitant therapy had little effect on the main result. There was
no sub-group in whom there was a significantly greater or lesser effect of the ICD. The
average unadjusted lifetime extension conferred by the ICD at 3 years was 3.2 months.
This study provides reasonable evidence that the ICD prolongs life compared to drug
therapy. The fact that the vast majority of drug treatment patients actually received
amiodarone and stayed on the drug during follow-up is a major strenght of the study
because amiodarone is the only drug, other than beta blockers, for which there is evidence
from randomized controlled trials of a benefical effect on arrhythmic death and overall
mortality. However the size of the relative risk reduction is moderate and the
prolongation of life is only modest, only just over 3 months. The 95% confidence interval
about the risk reduction is quite broad and the results are consistent with a relative
reduction as small as only about 10% at 2 years which would translate into a prolongation
in life of only a few weeks. Considering the substantial costs of therapy initiation 7
(26) with the ICD and the higher rate of re-hospitalization with the ICD compared to
amiodarone (60% vs 56% p = 0.04), it is clear that the cost per year of life saved is very
high.
The results of CIDS and CASH will be complimentary to AVID, providing much longer
average duration of follow-up and more than doubling the number of events. These trials
will very likely be statistically consistent with AVID but may give different point
estimates of the treatment effect. CIDS and CASH will be important because if they support
the findings of AVID, the confidence interval will tighten, giving greater security in the
conclusions of AVID. There is some concern, however that AVID has over-estimated the
benefit of the ICD. Trials, like AVID, which terminate prematurely due to observation of a
benefit have a tendency, in general, to over-estimate the benefit. There was also an
imbalance in AVID in the use of beta blocker therapy, which favoured the ICD treatment
limb. A meta-analysis of the three trials will provide the most accurate assessment of the
true benefit of the ICD. Meta-analysis will also be valuable because, by means of pooling
data, it may also be able to identify sub-groups of patients with substantially greater or
lesser benefit from the ICD. This would be important because it would allow resources to
be directed most appropriately.
|
