Docetaxel with prednisone or prednisolone for the treatment of ...

Progression-free survival
It is not possible to perform a meta-analysis as the
definitions of progression-free survival vary widely
between the three trials.
Response rate
It is not possible to perform a meta-analysis as the
definitions of response rate vary widely between
the three trials.
Health-related quality of life
It is not possible to perform a meta-analysis as no
data on QoL were reported for the trial by Berry
and colleagues and the definitions and
instruments used to measure HRQoL vary widely
between the other two trials. In addition, only 66%
of patients in the CALGB 9182 trial were assessed
at baseline and at least one follow-up, hence this
analysis is not true ITT. However, in the two
studies that measured HRQoL, the mitoxantrone
groups had statistically significant improvements
compared with the corticosteroid groups. Due to
the limited follow-up for this outcome, these
benefits should not be overstated.
Pain
It is not possible to perform a meta-analysis as no
data on pain were reported for the trial by Berry
and colleagues and the definitions and
instruments used to measure pain vary between
the two remaining trials. In addition, only 66% of
patients in the CALGB 9182 trial were assessed at
baseline and at least one follow-up, hence this
analysis is not true ITT. However, in the two
studies that measured pain, the mitoxantrone
groups had statistically significant improvements
compared with the corticosteroid groups. Due to
the limited follow-up for this outcome, these
benefits should not be overstated.
PSA decline
It would be possible to obtain a pooled estimate
for the RR of PSA decline, as all three trials have
reported information on the proportion of
patients who experienced a PSA reduction of at
least 50% from baseline. However, PSA decline is
not a very informative outcome in itself. As we
have managed to obtain a pooled estimate for
overall survival, it is unnecessary to obtain a
pooled estimate for PSA decline.
Adverse effects of treatment
All three trials that assessed the effectiveness of
mitoxantrone plus a corticosteroid versus a
corticosteroid measured the adverse effects of
treatment using the NCI CTC. However, various
adverse effects of treatment were reported for each
© Queen’s Printer and Controller of HMSO 2007. All rights reserved.
Health Technology Assessment 2007; Vol. 11: No. 2
trial, limiting the opportunities to obtain pooled
estimates for any single adverse effect of
treatment. Also, given the nature of adverse effects
being specific to the interventions received,
obtaining pooled estimates for the adverse effects
of mitoxantrone plus a corticosteroid versus a
corticosteroid has limited use in further indirect
comparisons.
Comparison of all treatments versus
mitoxantrone plus corticosteroids
This chapter has presented the median length of
follow-up, the median survival and HR for overall
survival for each identified trial. Each HR has
been presented using mitoxantrone plus a
corticosteroid as the common comparator. Only
the results for overall survival have been
presented, because the definitions and
measurements of the other outcomes varied across
the trials and therefore it is impossible to make
any comparisons between trials for any other
outcome, as discussed previously. However, in the
two studies comparing mitoxantrone plus a
corticosteroid with a corticosteroid alone that
measured HRQoL and pain responses (CCI-
NOV22 and CALGB 9182), the mitoxantrone
groups had statistically significant improvements
compared with the corticosteroid groups. In
addition, in the trial comparing mitoxantrone plus
prednisone with docetaxel plus prednisone (TAX
327), 3-weekly docetaxel plus prednisone resulted
in statistically significant improvements in terms of
QoL and pain compared with mitoxantrone plus
prednisone. Therefore, it is not expected that the
addition of these outcomes would change the
conclusions based on the findings of this analysis.
The results are presented in Table 17.
From the data presented in Table 17, it can be seen
that only two treatments are statistically superior
compared with mitoxantrone plus prednisone in
terms of overall survival: 3-weekly docetaxel plus
prednisone [HR = 0.76 (95% CI: 0.62 to 0.94)]
and docetaxel plus estramustine [HR = 0.80 (95%
CI: 0.67 to 0.97)]. All other chemotherapy
regimens, except mitoxantrone plus prednisone
plus clodronate, show higher survival rates in
comparison with mitoxantrone plus prednisone.
However, the difference is not statistically
significant. Mitoxantrone plus prednisone plus
clodronate and also corticosteroids alone show
lower survival rates in comparison with
mitoxantrone plus prednisone but, again, the
difference is not statistically significant.
From these data, it could be assumed that
docetaxel plus prednisone is statistically superior
37