School of SOCIAL AND COMMUNITY MEDICINE University of BRISTOL Pre- and post-diagnostic lifestyle factors and mortality in women with breast cancer Mona.

School ofSOCIAL AND COMMUNITY

MEDICINE

University ofBRISTOL

Pre- and post-diagnostic lifestyle

factors and mortality in women with breast

cancer Mona Jeffreys

School of Social and Community Medicine


MEDICINE


Breast cancer

Most common cancer in women in UK Affects approximately 50,000 women in

UK each year 85.1% of patients survive for five years or

more (diagnosed 2005-2009)


MEDICINE


Breast cancer incidence

http://www.cancerresearchuk.org/cancer-info/cancerstats/types/breast/incidence


MEDICINE


(Un)modifiable risk factors

Age BRCA genes (family history) Breast density Benign breast disease Birthweight, growth, height


MEDICINE


Modifiable risk factors

Reproductive factors Nulliparity / late age at first birth, early menarche,

late menopause, breastfeeding

Exogenous oestrogens (OC, HRT)

Shiftwork Diet

Saturated fat, phyto-oestrogens, fibre

Lifestyle Body weight, alcohol, physical activity, smoking


MEDICINE


Breast cancer survival

http://www.cancerresearchuk.org/cancer-info/cancerstats/types/breast/survival/


MEDICINE


Relative survival

Survival “adjusted for” background mortality (excess mortality “due” to cancer)

Avoids need for cause of death Observed : expected survival ratio Compares “observed” survival in the cancer

population to “expected” survival in general population


MEDICINE


Relative survival (cont’d)

Life-table methods (census) May not be valid when “observed” and

“expected” groups have a different underlying expected mortality Compare ethnic groups Cohort study with low response rate


MEDICINE


Lifestyle determinants of survival


MEDICINE


Lifestyle determinants of survival

Good evidence of lower survival in overweight women


MEDICINE


Obesity and all cause mortality

Protani, BCRT 2010


MEDICINE


Obesity and breast-cancer specific mortality

Protani, BCRT 2010


MEDICINE


But is it all confounding?

Women’s Intervention Nutrition Study RCT of 2437 women, aged 43-79 Early stage breast cancer Intervention: dietary fat reduction Associated weight loss (2kg difference at 3 years) Lower recurrence in intervention group (9.8% vs

12.4%, HR 0.76 (95% CI = 0.60 to 0.98)


MEDICINE


But is it all confounding?

Women’s Healthy Eating and Living RCT of 3088 women, aged 18-70 Early stage breast cancer Intervention: High F&V, fibre and low fat No change in body weight or energy intake No difference in breast cancer events (HR 0.96, CI:

0.80 to 1.14) or mortality (HR 0.91; CI: 0.72 to 1.15) Interaction with PA: reduced mortality in women with

high F&V and high PA, irrespective of obesity (HR 0.56; CI: 0.31 to 0.98)


MEDICINE


Physical activity and breast cancer outcomes

Patterson 2010, http://dx.doi.org/10.1016/j.maturitas.2010.01.004


MEDICINE


Diet and breast cancer outcomes



MEDICINE


Diet and breast cancer outcomes (cont’d)



MEDICINE


New Zealand Breast Cancer Study

Nationwide multi-ethnic, age- and ethnicity-matched population-based case-control study

Over-sampling of Māori and Pacific women


MEDICINE


Methods

Cases identified from Cancer Registry April 2006 to April 2007 Māori and Pacific cases to April 2008

Controls from Electoral Roll General Māori Additional methods for Pacific controls

• GP and community-based


MEDICINE


Response rates (cases)

1,799 cases

• 302 (81%) Māori,• 70 (46%) Pacific, • 1,427 (78%) non-

Māori /non-Pacific


MEDICINE


Exposures

BMI and WHR Smoking Alcohol Physical activity Diet

Servings of F&V, meat (red/white), fish, milk, cream, cheese


MEDICINE


Timing of exposure

Lifetime “In the last year, on average...”

Within 1 year After 1 year

Number (%) 1,237 (71.5%) 492 (28.5%)

Median 7.9 months 22.3 months

Range 4 to 12 months 12 to 39 months


MEDICINE


Ascertainment of outcome

Linked to death register Covers all of New Zealand

• Requires ethical approval• Censored at 28/02/2009

Not valid for Pacific women• “Going home to die”

Informed of date of death but not cause All cause mortality


MEDICINE


Statistical methods

Kaplan Meier curves Log rank tests

Cox regression Followed from time of diagnosis to dead /

censoring Adjusted for age at diagnosis, menopausal status,

interview method, extent of disease at diagnosis (stage)


MEDICINE


Basic results

Māori nMnP

Total 302 1,427

Deaths 21 (7.0%) 109 (7.6%)

Follow-up (years) 2.92 (1.1 to 3.9 yrs) 3.29 (11 mths to 3.9 yrs)

Median age 54.3 (47.0 to 62.8) 57.9 (49.1 to 67.6)

Interviewed after 1 year 152 (50%) 340 (24%)


MEDICINE


Follow-up by ethnicity0

.70

0.8

00

.90

1.0

0

0 1 year 3 years2 years 4 yearsFollow-up time

non-Maori/non-Pacific Maori

Kaplan-Meier survival estimates

Log rank testP=0.56


MEDICINE


Effect of BMI on survival0

.90

0.9

20

.94

0.9

60

.98

1.0

0

1 year 2 years 3 years 4 years0analysis time

BMI <25 kg/m2 BMI 25-29.9 kg/m2BMI >=30 kg/m2


Log rank testP=0.21


MEDICINE


Effect of BMI on survivalBMI (kg/m2)

<25 25 to 30 >=30

Adjusted* 1.00 0.79 (0.50 to 1.24) 0.73 (0.44 to 1.21)

Pre-diagnostic 1.00 0.86 (0.52 to 1.44) 0.87 (0.49 to 1.54)

Post-diagnostic 1.00 0.49 (0.15 to 1.56) 0.39 (0.14 to 1.13)

* Adjusted for age, menopausal status, type of interview, ethnicity and extent of disease


MEDICINE


Effect of WHR on survival0

.75

0.8

00

.85

0.9

00

.95

1.0

0

0 1 year 2 years 3 years 4 yearsanalysis time

WHR tertile 1 WHR tertile 2WHR tertile 3


Log rank testP=0.085


MEDICINE


Effect of WHR on survivalWaist – hip ratio

Tertile 1 Tertile 2 Tertile 3

Adjusted* 1.00 1.20 (0.74 to 1.97) 1.35 (0.82 to 2.22)





MEDICINE


Effect of alcohol on survival


MEDICINE



Newcomb, JCO 2013


MEDICINE



Newcomb, JCO 2013


MEDICINE


Measurement of alcohol

How often did you have a drink containing alcohol? (frequency)

How many drinks containing alcohol did you have on a typical day when you are drinking? (amount)


MEDICINE


Alcohol frequency and amount

Alcohol frequency

Never <=1/mth Up to 1/wk 2-3/wk 4+/wk

Alcohol amount

None 334 0 0 0 0

1-2 drinks 0 354 221 209 257

3-4 drinks 0 20 50 62 61

5-6 drinks 0 16 15 17 17

7-9 drinks 0 10 3 3 8

10 or more 0 8 8 9 3

Nil, slight, moderate, heavyUsed “slight” as the reference group


MEDICINE



0.8

00

.85

0.9

00

.95

1.0

0

0 1 year 2 years 3 years 4 yearsanalysis time

Nil SlightModerate Heavy


Log rank testP=0.14


MEDICINE


Effect of alcohol on survival: overall

Alcohol intake

None Slight Moderate Heavy

Adjusted* 0.95 (0.57 to 1.58) 1.00 0.93 (0.58 to 1.49) 1.96 (0.98 to 3.94)

Pre-diagnostic 0.79 (0.43 to 1.49) 1.00 0.94 (0.56 to 1.57) 1.39 (0.57 to 3.34)

Post-diagnostic 1.61 (0.64 to 4.06) 1.00 0.57 (0.12 to 2.73) 4.82 (1.43 to 16.29)



MEDICINE


Effect of alcohol on survival: frequency

Alcohol intake (frequency)

Never ≤1/mth Up to 1/wk 2+/wk

Adjusted* 1.08(0.61 to 1.90)

1.00 1.49(0.82 to 2.72)

1.18(0.70 to 1.98)

Adjusted** 1.19(0.66 to 2.16)

1.00 1.63(0.86 to 3.09)

1.26(0.73 to 2.19)

Pre-diagnostic**

0.82 (0.41 to 1.64)

1.00 1.05(0.50 to 2.21)

0.95(0.52 to 1.71)

Post-diagnostic**

4.88(1.22 to 19.61)

1.00 8.25(1.93 to 35.22)

3.80(0.82 to 17.63)

Merged categories 2-3 per week and 4+ per week* Adjusted for age, menopausal status, type of interview, ethnicity and extent of disease** Also adjusted for alcohol amount


MEDICINE


Effect of alcohol on survival: amount

Alcohol intake (drinks per day)

None 1-2 3-4 5-6 7+

Adjusted* 0.98(0.61 to 1.58)

1.00 0.97(0.48 to 1.97)

1.33(0.48 to 3.74)

2.36(0.79 to 7.07)

Adjusted** 1.19(0.66 to 2.15)

1.00 0.93(0.45 to 1.90)

1.36(0.49 to 3.83)

2.39(0.80 to 7.18)

Pre-diagnostic**

0.83 (0.41 to 1.66)

1.00 1.06(0.49 to 2.30)

0.94(0.22 to 3.93)

1.87(0.40 to 8.80)

Post-diagnostic**

4.53(1.15 to 17.80)

1.00 0.54(0.07 to 4.33)

2.85(0.56 to 14.53)

3.41(0.54 to 21.30)

Merged categories 7-9 and 10+ drinks per day* Adjusted for age, menopausal status, type of interview, ethnicity and extent of disease** Also adjusted for alcohol frequency


MEDICINE


Smoking

Have you ever smoked, now or in the past?

Are you a current smoker? Categorised into never, current, ex-

smokers


MEDICINE


0.8

50

.90

0.9

51

.00

0 1 Year 2 Years 3 Years 4 Yearsanalysis time

Never smoker Ex-smokerCurrent smoker


Effect of smoking on survival

Log rank testP=0.66


MEDICINE


Effect of smoking on survival

Smoking

Never smoked Ex-smokers Current smokers

Adjusted* 1.00 1.05 (0.69 to 1.58) 1.43 (0.76 to 2.68)





MEDICINE


Physical activity

Measured using Godin questionnaire Frequency of mild, moderate and strenuous

Analysed in quartiles Lowest quartile designated as “sedentary”


MEDICINE


0.7

50

.80

0.8

50

.90

0.9

51

.00

0 1 Year 2 Years 4 years3 Yearsanalysis time

active sedentary


Effect of physical activity on survival

Log rank testP=0.146


MEDICINE



Effect of physical activity on survival

Physical activity

Active Sedentary

Adjusted* 1.00 1.12 (0.74 to 1.69)

Pre-diagnostic 1.00 0.93 (0.57 to 1.52)

Post-diagnostic 1.00 2.03 (0.90 to 4.60)


MEDICINE


Fruit and vegetable intake

How many servings of vegetables (excluding potatoes) did you usually eat each week?

How many servings of fruit did you usually eat each week?


MEDICINE


Effect offruit and vegetable intake on

survivalFruit intake

(>14 servings per week)Vegetable intake

(>21 servings per week)

No Yes No Yes

Adjusted* 1.00 1.42 (0.96 to 2.10) 1.00 1.41 (0.91 to 2.19)

Pre-diagnostic 1.00 1.43 (0.92 to 2.22) 1.00 1.39 (0.84 to 2.29)

Post-diagnostic 1.00 1.27 (0.54 to 3.00) 1.00 1.44 (0.55 to 3.73)



MEDICINE


Summary

Suggestion of poorer survival in women who were Had lower BMI, but higher WHR Never and heavy alcohol drinkers Sedentary

No clear relationship with F&V intake Smoking


MEDICINE


Other dietary measures

No clear relationship with intake of Milk Cream Cheese Meat (red/white) Fish


MEDICINE


Dairy intake and cancer survival


MEDICINE


Dairy intake and cancer survival

Kroenke, JNCI 2013


MEDICINE


Limitations of NZ study

Poor measurement of some exposures Particularly dietary measures What time frame are women actually reporting on?


MEDICINE


Effect of childhoodfruit and vegetable intake on

survivalFruit intake(days/week)

Vegetable intake(days/week)

<=4 5+ <=4 5+

Adjusted* 1.00 1.09 (0.74 to 1.62) 1.00 0.48 (0.28 to 0.83)



MEDICINE


Effect of childhoodfruit and vegetable intake on

survivalFruit intake(days/week)

Vegetable intake(days/week)

<=4 5+ <=4 5+

Adjusted* 1.00 1.09 (0.74 to 1.62) 1.00 0.48 (0.28 to 0.83)

Pre-diagnostic 1.00 1.39 (0.87 to 2.19) 1.00 0.44 (0.24 to 0.82)

Post-diagnostic

1.00 0.47 (0.20 to 1.08) 1.00 0.61 (0.20 to 1.84)



MEDICINE


Determinants of high F&V reporting

Fruit intake(>14 servings per

week)

Vegetable intake(>21 servings per week)

Childhood F&V 1.86 (1.42 to 2.46) 1.65 (1.28 to 2.13)

Ever smoked 0.71 (0.57 to 0.88) 1.00 (0.79 to 1.28)

Ethnicity 0.87 (0.65 to 1.16) 0.83 (0.59 to 1.16)

Unrelated: area deprivation, childhood SEP, age, menopausal status, extent of disease


MEDICINE


Determinants of high F&V reporting

Fruit intake(>14 servings per

week)

Vegetable intake(>21 servings per week)

Childhood F&V 1.86 (1.42 to 2.46) 1.65 (1.28 to 2.13)

Ever smoked 0.71 (0.57 to 0.88) 1.00 (0.79 to 1.28)

Ethnicity 0.87 (0.65 to 1.16) 0.83 (0.59 to 1.16)

Interviewed after 1 year

1.15 (0.91 to 1.45) 1.50 (1.16 to 1.95)

Unrelated: area deprivation, childhood SEP, age, menopausal status, extent of disease


MEDICINE


Limitations of NZ study

Poor measurement of some exposures Particularly dietary measures What time frame are women actually reporting on?

Pre- and post-diagnostic comparisons are not of the same women Ideally have a measure of change

Survivor bias Conditional survival

Limited statistical power for some analyses


MEDICINE


Where next?

Re-analysis with longer follow-up Physical activity intervention trial

acceptability of intervention and randomisation

Consider weight loss intervention effect on lean body mass


MEDICINE


Acknowledgements

Co-PIs: Lis Ellison-Loschmann, Fiona McKenzie, Riz Firestone

Co-Investigators: Neil Pearce, Michelle Gray, Ate Moala, Soo Cheng

Funders: New Zealand Lottery Grants Board Massey University Research Fund Health Research Council of New Zealand Cancer Society of New Zealand

School of SOCIAL AND COMMUNITY MEDICINE University of BRISTOL Pre- and post-diagnostic lifestyle factors and mortality in women with breast cancer Mona.

Documents

communitybased slide

smoking slide

height slide

cheese slide

cancer population

breast cancer events

common cancer

cancer registry