Bisphosphonates and GI cancers

Report
Exposure to bisphosphonates and risk
of non-gastrointestinal cancers:
nested case-control studies
Yana Vinogradova, Carol Coupland, Julia Hippisley-Cox
SAPC 2013 , Nottingham
Background
 Bone loss and bisphosphonates
 Laboratory linkages with reduced cancer risk
Anti-tumor properties, accumulation in different organs
 Treatment for osteoporosis and cancer risk
 Reduced risk of breast cancer
•
•
•
•
Chlebowski (2010) 1993-1998
Newcomb (2010) 2003-2006
Rennert
(2010) 2000-2006
Vestergaard (2011) 1996-2006
 No associations for other cancers
•
Cardwell
(2011) 1996-2006
• QResearch
– EMIS
– 660 GP
practices
• Clinical Practice
Research Datalink
– Main (EHR)-IT
– 643 GP practices
Patient level data
from GP clinical records
Anonymised
Longitudinal data for 20+ years
Validated externally and internally
Industry independent
Methods: Study design
•
•
•
•
•
Nested case control study
Study period Jan 1997 – July 2011
Aged 50 – 100
At least 2 years of records
Up to 5 controls matched by
–
–
–
–
Age
Sex
Practice
Calendar year
Methods: Assessment of Exposure
•
•
•
•
excluding 6 months before the index date
use: at least 1 script
short-term (<12m), long-term users
different types of bisphosphonates
Methods: Confounding factors
• BMI, ethnicity, smoking, alcohol
• Osteoporosis
• Morbidities (rheumatoid arthritis, diabetes, benign
breast disease, GI disorders, hypertension)
• Family history of cancers
• Other medications (NSAIDs, corticosteroids,
•
acid-lowering drugs,
•
vitamin D, HRT)
Methods: Statistical analysis
• Multiple imputations
• Conditional logistic regression
• Combined analysis
– Odds ratios + 95% CI
– 1% significance level
QRESEARCH
Cases:
4.2% users
CPRD
exposed
total
1304
24,489
Breast
460
26,554
Prostate
1035
17,961
Lung
Bladder
274
7,464
Melanoma
241
Ovary
Pancreas
exposed
total
1324
25,444
376
23,176
1114
19,059
Bladder
280
7,084
4,998
Melanoma
178
4,621
204
3,427
Ovary
170
3,088
178
3,593
Pancreas
196
3,485
Uterus
99
2,248
Uterus
96
2,149
Cervix
32
822
Cervix
35
739
Breast
Prostate
Lung
Cases:
4.2% users
Proportion of cases and controls exposed to bisphosphonates by index year
QResearch
CPRD
8
7
6
%
5
4
3
2
1
1998
2001
2004
2007
2010
1998
2001
2004
Index year
controls
bisphosphonates
cases
controls
etidronate
cases
controls
alendronate
cases
controls
risedronate
cases
© QRESEARCH 2011 version 31, CPRD March 2012
2007
2010
Any use of bisphosphonates and risk of cancer
CPRD
QResearch
OR and 95% CI
OR and 95% CI
Cancer site,
Cancer site,
OR (95%CI)
OR (95%CI)
Breast, 0.89 (0.82 to 0.97)
Breast, 0.95 (0.88 to 1.03)
Prostate, 0.90 (0.79 to 1.02)
Prostate, 0.84 (0.73 to 0.96)
Lung, 0.97 (0.88 to 1.08)
Lung, 1.12 (1.01 to 1.23)
Bladder, 0.96 (0.80 to 1.14)
Bladder, 0.94 (0.79 to 1.12)
Melanoma, 1.05 (0.87 to 1.28)
Melanoma, 0.95 (0.77 to 1.19)
Pancreas, 0.81 (0.64 to 1.01)
Pancreas, 0.78 (0.63 to 0.97)
Ovary, 1.19 (0.96 to 1.47)
Ovary, 0.84 (0.67 to 1.04)
Uterus, 1.07 (0.79 to 1.44)
Uterus, 0.95 (0.71 to 1.27)
Cervix, 0.78 (0.48 to 1.27)
Cervix, 1.21 (0.76 to 1.93)
.5
.75
1
1.25
1.5
.5
.75
1
1.25
1.5
Any use of bisphosphonates and risk of cancer, combined analysis
Adjusted OR's and 95% CI's
Cancer site
(N of cases)
OR (95%CI)
Breast (49933)
0.92 (0.87 to 0.97)
P=0.004
P-trend=0.2
Prostate (49730)
0.87 (0.79 to 0.96)
P=0.003
P-trend=0.005
Lung (37020)
1.04 (0.97 to 1.12)
Bladder (14548)
0.95 (0.84 to 1.08)
Melanoma (9620)
1.01 (0.87 to 1.17)
Pancreas (7079)
0.79 (0.68 to 0.93)
P=0.003
P-trend=0.02
Ovary (6516)
1.00 (0.86 to 1.16)
Uterus (4393)
1.00 (0.81 to 1.24)
Cervix (1560)
0.98 (0.70 to 1.37)
.5
.75
1
1.25
1.5
Prostate cancer: long-term use of bisphosphonates
OR and 95%CI
OR and 95%CI
QResearch
CPRD
OR (95%CI)
OR (95%CI)
up to 6 mths 1.11 (0.90 to 1.36)
up to 6 mths 0.86 (0.68 to 1.08)
7mths < 3yrs 0.85 (0.72 to 1.00)
7mths < 3yrs 0.83 (0.69 to 0.99)
3yrs < 6yrs 0.77 (0.60 to 1.00)
3yrs < 6yrs 0.87 (0.66 to 1.14)
6years+
6years+
0.89 (0.59 to 1.33)
.5
.75
1
1.25
0.77 (0.49 to 1.22)
.5
.75
1
1.25
Prostate cancer: types of bisphosphonates
CPRD
QResearch
OR and 95%CI
OR and 95%CI
Alendronate
Alendronate
P=0.004
Ptrend=0.009
overall use, 0.81 (0.70 to 0.93)
1yr or more, 0.75 (0.62 to 0.92)
Etidronate
overall use, 0.88 (0.75 to 1.03)
1yr or more, 0.86 (0.70 to 1.05)
Etidronate
overall use, 1.06 (0.87 to 1.30)
overall use, 0.88 (0.70 to 1.10)
1yr or more, 1.04 (0.81 to 1.33)
1yr or more, 1.02 (0.77 to 1.34)
Risedronate
Risedronate
overall use, 0.98 (0.77 to 1.23)
overall use, 0.87 (0.66 to 1.13)
1yr or more, 1.09 (0.81 to 1.46)
1yr or more, 0.83 (0.58 to 1.19)
.75
1
1.25
1.5
.75
1
1.25
1.5
Summary of findings
Use of bisphosphonates is not associated with
increased risk of any of the most common
cancers
• Decreased risk of breast, prostate and
pancreatic cancers had no duration relationship
in either database
• Decreased risk of prostate cancer associated
with alendronate use only in QResearch
Limitations and Strengths
• Residual confounding as no
information
– on stage of cancer
– on any cancer-related tests
– on bone density test
• Information on prescriptions
only
• Missing data
• Large sample size and
representative population
• Data electronically collected –
unlikely misclassification bias
• Data collected before diagnosis
– no recall bias
• All cases used
– no selection bias
• Data in the last 6 months before
the diagnosis was excluded as
might be misleading
• Based on the most recent data
Thank you
Questions?

similar documents