ESAs for CancerRelated Anemia
TMR Journal Club
Shuen Tan
October 6, 2009
Tonelli M, Hemmelgarn B, Reiman T, et
al. Benefits and harms of
erythropoiesis-stimulating agents for
anemia related to cancer: a metaanalysis. CMAJ 180(11): E62-71,
2009.
Anemia in Cancer
Related to cancer
 Related to chemotherapy


Associated with:
 quality of life
  survival

ESAs: Benefit vs. Harm

Benefits



Improved QOL
Decreased
transfusions
? survival

Harm





Thromboembolism
HTN
Stimulate tumour
growth
(Cost)
? survival
Methods
Systematic review of ESAs for the
treatment of cancer-related anemia
 Published and unpublished RCTs

Epoetin or darbepoetin vs. control
 Adults age>18
 Cancer-related anemia
 >30 subjects in each group
 English, French, Spanish, or Mandarin

“Any” Outcomes
Mortality
 Cardiac events
 Hospital admission
 Quality of life
 Hypertension
 RBC transfusion
 Adverse events

Subgroups

American Society of Clinical Oncology
criteria
Baseline hemoglobin
 Chemotherapy (or not)
 Target hemoglobin

Results

52 trials met criteria
4 trials in perioperative patients
 30 trials in solid tumours
 10 trials in hematologic cancer
 11 trials included both
 Mean duration 12 (2-28) weeks

Cardiovascular events
MI, stroke, CHF, revascularization
 14 trials



RR 1.12 (0.83-1.5)
Hypertension
17 trials
 RR 1.41 (0.94-2.12)

Tumour response

2 trials

No difference for complete response
• RR 0.88 (0.69-1.12)

No difference for partial response
• RR 0.70 (0.44-1.11)
Serious Adverse Events

21 trials

Increased risk in ESA groups
• RR1.16 (1.08-1.25)

Thrombotic events
13 trials
 Increased risk in ESA groups

• RR 1.69 (1.27-2.24)
Sub-group analyses
No differences between any groups
and total study population
 Do the American Society of Clinical
Oncology guidelines permit
identification of patients most likely to
benefit?

Validity
1. Did the authors ask a focused clinical
question?
For the most part, Yes
 Well-defined patient group
 Broad but reasonable disease
category
 No specific outcomes stated


Included studies that “reported one or
more outcomes”
2. Were the criteria used to select articles
for inclusion appropriate?
Yes
 Study type well-defined
 Patients well-defined
 Therapy well-defined
 Outcomes not defined
 Languages reasonably dealt with


Only 25/2025 studies could not be
assessed because of translation
3. Is it unlikely that important, relevant
studies were missed?
Yes
 Very thorough search strategy



Multiple databases


ESAs extensively exploded
CancerLit?
Unpublished literature searched

Some papers included that were
published after search dates
3. Is it unlikely that important, relevant
studies were missed?
Very small percentage of citations
could not be retrieved (32/2025)
 Abstracts screened by 2 reviewers



All flagged articles retrieved
Full text assessed by 2 reviewers for
inclusion

Disagreements resolved with a third
reviewer
4. Was the validity of the included studies
appraised (study quality)?


Yes
Chalmers index





More detailed version of the Jadad score
Randomization, blinding, and handling of
withdrawals
Rated statistical analysis, presentation of
results, and source of funding as well
Not entirely clear if the final rating was
subjective
Does not appear to have been incorporated
into the meta-analysis
5. Were assessments of studies
reproducible (data abstraction)?
Yes?
 One reviewer abstracted data
 A second reviewer checked for
accuracy

6. Were the results similar from study to
study (homogeneity)?
Yes
 Random effects model
 Quantified heterogeneity with the I2
statistic (=0% for all calculations)


Calculates proportion of total variation
in the estimates of treatment effects
that is due to heterogeneity between
studies
Issues related to the included studies


Overall, systematic review was well done
Limitations to applying results to patients


Short follow-up time (median 12 weeks)
Low-moderate quality scores
• Many had unblinded treatment groups

Majority were privately funded
Relevance
Will the results change my practice?
Are the outcomes important to my
patients?
Populations of Interest
Solid organ vs. hematologic cancer
 Anemic (100 vs. 120) vs. very anemic
(<100)
 Long-term (>12 weeks) vs. short-term
(<12 weeks) vs. really short-term (2-3
doses)
 Chemo vs. no chemo

Shuen’s Thoughts -- Con

Increased reluctance to recommend use in
pre-op anemia




risk of thrombosis with surgical stress
response
Thoracic, bowel surgery not typically
associated with high blood loss
Lower targets, as few doses as possible
Anemia associated with survival, but does
treatment with ESA make a difference?
Shuen’s Thoughts -- Pro
Risk-benefit study of ESA vs. no ESA
but does not compare to other
treatments (e.g. transfusion)
 Improved quality of life
 ? Uncertain diagnosis
 Based on studies of low-moderate
quality

CancerCare’s Patients
Any changes in practice?
 Is transfusion a “better” treament for
anemia than Epo?
 Non-random oncologist #1



Following ASCO guidelines
Non-random oncologist #2
Rarely uses epo
 Transfusion more common

Patients

Can a reasonable patient choose?




 quality of life with  survival
 quality of life with  survival
? quality of life with transfusion and ?
Survival
Cause of increased mortality unclear




Thromboembolism?
Tumour progression?
Something else?
Dose or duration-dependent?
Conclusion
The use of ESAs in cancer patients is
associated with increased mortality
and serious adverse events, but
improved quality of life and decreased
transfusions
 The ASCO guidelines for the use of
ESAs do not appear to identify a
lower-risk, higher-benefit group

Conclusion

ESAs should be used cautiously in
any patient with a cancer diagnosis,
regardless of type of cancer,
chemotherapy, or degree of anemia

But are not contraindicated…
Questions?
Descargar

ESAs for Cancer-Related Anemia