Transcript Document

Romero IL, McCormick A, McEwen et al. 2012.
Obstetrics & Gynecology, vol. 119:61-67
What does the descriptive epidemiology
teach us?
Changes in Cancer Mortality Rates, 2006-2010. U.S.
Trend in Ovarian Cancer Incidence and Mortality
Rates, California, 1988-2011
18
16
14
12
AGE ADJUSTED
RATE/100,000
10
8
6
4
2
0
YEAR DX OR DEATH
Incidence Rate
Mortality rate
Age Adjusted Incidence and Mortality Rates of
Ovarian Cancer in California, 2007-2011, by
Race/Ethnicity
Incidence Rate
16
Mortality Rate
14
12
10
8
6
4
2
0
NH White
NH Black
Hispanic
Asian/PI
Age Specific incidence Rates of Ovarian cancer
In California, 2007-2011
Age Specific Rate
60
50
40
RATE/100,000 30
20
10
0
AGE DX
Age Specific incidence rate of Ovarian Cancer in
California, 1988-2011, by Race/Ethnicity
70
60
50
40
RATE/100,000
30
20
10
0
AGE DX
NH White
NH Black
Hispanic
Asian/PI
Age Adjusted Ovarian Cancer Incidence Rates in California,
2007-2011, by County of Residence
Age Adjusted Ovarian Cancer Mortality Rates, 2007-2011,
in California, By County of residence
What does the analytic
epidemiology teach us?
What “risk factors” alter risk of
ovarian cancer?
Risk Factors for Ovarian Cancer by
Strength of Evidence
What is the distribution of stage at
diagnosis for ovarian cancer and
why is this important for survival?
Ovarian Cancer Stage Distribution, SEER
Data, 2004-2010 (N=33,841)
Percent
70
61
60
50
40
PERCENT DX
30
20
15
18
10
6
0
Localised
Regional
Distant
Unstaged
Five Year Relative Survival, Ovarian Cancer,
U.S., SEER Data, 2004-2010, by Stage at
Diagnosis
100
90
80
70
PERCENT
SURVIVING
5 YEARS
60
50
40
30
20
10
0
Localized
Regional
Distant
STAGE AT DX
Unstaged
Five Year Relative Survival, Ovarian Cancer, U.S.
SEER Data, 2006-2010
100
100
90
80
75.4
70
63.9
60
55.5
49.5
PERCENT SURVIVING 50
45.2
40
30
20
10
0
Diagnosis
1-Year
2-Year
3-Year
YEARS SINCE DIAGNOSIS
4-Year
5-Year
Type II Diabetes
• It is estimated that two in five women born in the
United States in the year 2000 will have type II diabetes
diagnosed during their lifetime.
• Available data suggest that ovarian cancer patients with
type II diabetes have decreased survival.
• It is biologically plausible that hyperinsulinemia and
hyperglycemia induced by type II diabetes promotes
tumorigenesis.
• Hypergylcemia provides a nutrient rich
microenvironment for rapidly dividing cancer cells,
which have elevated metabolic demands and consume
glucose at a higher rate than normal cells.
Proportions of the California Female Population,
Age > 45 years, Ever Diagnosed with Diabetes
(2011-2012 CHIS data)
25
20
15
PERCENT OF THE
POPULATION
10
5
0
SJV
California
AREA OF RESIDENCE
All Race/Ethnic Groups
Hispanics Only
Metformin
• Metformin is the most commonly prescribed
drug for the treatment of type II diabetes.
Metformin reduces both insulin and glucose
levels.
• In ovarian cancer preclinical studies,
metformin inhibits proliferation of cancer cell
lines in a dose-dependent fashion (Gotlieb,
2008) and in a time dependent manner
(Rattan, 2009).
Frequency of Metformin prescription, U.S., 2012
(Lindsley, ACS. Chem Neuroscience, 2013, 4,
1133-1135).
Epidemiology of Ovarian Cancer
With Reference to
the Role of Metformin
CCRA Meeting, November 7, 2014
Paul K Mills, Ph.D., Kristine McLane, B.S.,
Cynthia Cortez, B.S., Soe Naing, MD, Maria
Arambula, MD and Theresa Gipps, MD
Institutional Review Board and
Funding
• Approval for the study was received from CMC
IRB: Approval #2012026, March, 2012
• Funding generously provided from: Central
California Faculty Medical Group, Maria
Arambula, MD, P.I.
Ovarian Cancer-Metformin Study: Methodology
• Retrospective Cohort Study Design
• Inclusion Criteria: Epithelial Ovarian
cancer patients (ICD-O-3 =56.9)
diagnosed 2001-2010 at CRMC, Fresno,
California
• Exclusion criteria: Non-invasive
pathology, non-epithelial malignancies,
non ovarian primary cancer that
metastasized to ovary.
Ovarian Cancer-Metformin Study:
Predictors, Outcomes and Analysis.
• Primary Predictor Variable: Metformin
Exposure
• Primary Outcome Variable: Time to recurrence
and time to death.
• Analyses: Simple descriptive analysis of
means, standard deviations and proportions
with use of Student’s t test and chi-square to
compare groups. Survival analysis using Kaplan
–Meier and Cox Proportional hazards
regression.
• Statistical significance set at p<=0.05, twosided tests.
Data Sources, Data Collection, Coding
and Quality Control
• Four primary data sources: Epic (post-2009),
Last Word, EMR and paper medical records.
• Data dictionary developed, standardized data
abstracting and coding techniques were
developed by two trained research
coordinators (Kristine M and Cynthia C.).
• Results recorded on excel spreadsheets,
analysis performed using IBM/SPSS, version 21.
Results
Characteristics of the Study Cohort,
Histology
Histology
Number (N=220)
Percent
Serous
98
44.5
Mucinous
23
10.5
Endometrioid
24
10.9
Clear Cell
7
3.2
Adenocarcinoma, NOS 53
24.1
Carcinoma, NOS
15
6.8
Total
220
100
Characteristics of the Cohort, Demographics,
Labs and Follow-up (means, medians and SD)
Variable
Mean
SD
Age at Dx (N=220)
59.1
14.6
BMI (N=199)
29.7
6.8
Blood Glucose at Dx (N=186)
120.3
37.7
CA-125 at Dx (N=142)
1,361.3 (mean)
288.5 (median)
3394
Follow-up, Dx to Last f.u.
845 days (median)
Follow-up, Dx to death (N=122)
497.5 days (median)
Characteristics of the Study Cohort,
Race/Ethnicity
Race/Ethnicity
Number (N=220)
Percent
Non-Hispanic White
145
65.9
Non-Hispanic Black
7
3.2
Hispanic
56
25.5
Asian/Pacific Islander
12
5.5
Total
220
100
Characteristics of the Cohort, Stage at
Diagnosis
Stage at Diagnosis
Number (N=220)
Percent
Localized
49
22.3
Regional
17
7.7
Remote
88
40
Distant Metastasis
62
28.2
Unknown
4
1.8
Total
220
100
Characteristics of the Cohort, Smoking
Smoking Status
Number (N=220)
Percent
Never Smoker
149
67.7
Current Smoker
22
10
Past Smoker
38
17.3
Unknown
11
5
Total
220
100
Characteristics of the Cohort, Treatment
Related Variables
Variable
Number (N=220)
Percent
No Surgery
47
21.3
Class I-II
52
23.6
Class II-IV
41
18.6
Unknown
80
36.4
None
61
27.7
Rec’d
77
35.0
Unknown
82
37.3
None
60
27.3
Rec’d.
77
35.0
Unknown
83
37.7
ASA Class
Platinum Chemo. Rx
Taxane Chemo Rx
Characteristics of the Cohort, Diabetes
Status
Number
(N=220)
Percent
Non-Diabetic
159
72.3
Diabetic
48
21.8
Unknown Diabetes Status
13
5.9
Total
220
100
Characteristics of the Cohort, Metformin
Use/Exposure
Metformin Use
Number
(N=220)
Percent
No Metformin Exposure
179
81.4
Metformin taken at or after Ca Dx
18
8.2
Metformin taken only prior to Ca Dx
4
1.8
Unknown Status
19
8.6
Total
220
100
Characteristics of the Cohort, Vital Status
at End of Follow-up
Vital Status at End of
Follow-up
Number
(N=220)
Percent
Alive
90
40.9
Known Dead
122
55.4
Dead, Ovarian Ca
29
13.2
Dead, Other Cause
10
4.5
Dead, Unknown Cause
83
37.7
Unknown Vital Status
8
3.6
Total
220
100
Comparisons of the Study Groups
Variable
Non-Diabetic
(n=159)
Diabetic, no
Diabetic,
Metformin Use Metformin Use
(N=18)
(N=18)
p-value
Age at Dx
(mean, SD)
57.3 (14.3)
65.8(14.5)
64.1( 15)
0.02
NHW
111 (.69)
8 (.44)
8 (.44)
.005
NHB
4 (.02)
1 (.06)
1 (.06)
Hispanic
37 (.23)
9 (.50)
5 (.28)
Asian/PI
7 (.04)
0
4 (.22)
BMI (Kg/M2)
28.8 (5.9)
30.6 (7.1)
28.9 (6.1)
0.64
Gravidity
(mean)
2.6 (2.2)
3.4 (3.1)
5.3 (4.2)
0.02
Parity (mean)
2.2 (1.9)
2.9 (2.5)
4.9 (3.7)
0.004
Race/Ethnicity
Comparisons of Study Groups
Variable
No Diabetes
(N=159)
Diabetic, no
Metformin
(N=18)
Diabetic with
Metformin
(N=18)
P-value
I
38 (23.9)
3 (16.7)
3 (16.7)
0.16
II
10 (6.3)
4 (22.2)
3 (16.7)
II
67(42.10
9 (50.0)
6 (33.3)
IV
43 (27.0)
2 (11.1)
6 (33.3)
FIGO Stage
Histology
0.64
Serous
74 (46.5)
6 (33.3)
7 (38.9)
Mucinous
18 (11.3)
2 (11.1)
1(5.6)
Endometriod
18 (11.3)
3 (16.7)
1 (5.6)
Clear cell
6 (3.8)
1 (5.6)
0 (0.0)
Carcinoma, NOS
43 27.0)
6 (33.3)
9 (50.0)
Treatment Related Variables in Study Groups
Variable
Non-Diabetics
(N=159)
Diabetics, no
Metformin (N=18)
Diabetes, with
Metformin (N=18)
P-Value
I-II
38 (23.9)
5 (27.8)
7 (38.9)
0.28
II-IV
34 (21.4)
3 (16.7)
2 (11.1)
No Surgery
25 (15.7)
7 (38.9)
5 (27.8)
Not recorded
62 (39.0)
3 (16.7)
4 (22.2)
Received
63 (39.6)
3 (16.7)
6 (33.3)
None
44 (27.7)
8 (44.4)
3 (16.7)
Not recorded
52 (32.7)
9 (50.0)
11 (61.1)
Received
62 (39.0)
4 (22.2)
6 (33.3)
None
44 (27.7)
7 (38.9)
3 (16.7)
Not recorded
53 (33.3)
9 (50.0)
11 (61.1)
ASA Class
Platinum
0.11
Taxane
0.31
All Cause Mortality Proportions Among Three
Exposure Groups (N=122 deaths)
0.8
0.7
0.6
0.5
% Dead at End
0.4
of F.U.
0.3
0.2
0.1
0
No Diabetes
DM/No Metformin Diabetes/Metformin
Median Survival Time, Date of Diagnosis until date of Recurrence
or Death, CRMC, Epithelial Ovarian Cancer Patients, 2001-2010
Comparison Groups
1000
900
P=0.549
893
800
765
700
600
NO. OF
500
DAYS
579
400
300
200
100
0
Non Diabetic (N=134)
Diabetes, no Metformin (N=15)
Diabetes, Metformin (N=15)
Life table estimates of progression–free survival among three
groups of epithelial ovarian cancer patients, CRMC, Fresno,
2001-2010. (log rank test p=0.549)
P=0.05
Cox Regression estimates of survival without recurrence outcomes among
epithelial ovarian cancer patients, adjusted for age at diagnosis. The two
groups are ovarian cancer patients with type II diabetes using Metformin (n=16)
and ovarian cancer patients with or without type II diabetes not using
Metformin (n=149). (p = 0.393)
Cox Regression Model Hazard Ratios for
Progression-Free and Overall Survival
Variable
Progression
Free Survival
Overall
Survival
DM/Metformin Status
Hazard ratio
Hazard Ratio
Non-Diabetic
1.00
1.00
Diabetic, no metformin
1.18 (.63-2.20)
O.83(.40-1.67)
Diabetic, metformin
0.85 (.45-1.58)
0.92(.46-1.79)
Limitations/Conclusions
• This is not a randomized clinical trial and sample size is
limited.
• Data quality issues are important, and it is important to
recognize the limitations of medical records as a source for
research data.
• Among 220 epithelial ovarian cancer patients at CRMC, Fresno,
approximately 72% had never been diagnosed with diabetes, 8
% had diabetes but no metformin exposure and, 8% had
exposure to the drug metformin.
• Metformin users were older, more likely to be Hispanic and to
be of higher parity and gravity than non metformin users.
• Survival among ovarian cancer patients exposed to metformin
was not statistically different from those not exposed, although
there was a suggestive improvement in survival in metformin
users in younger aged women.
The French Lilac plant, Galega officinalis