Arch. Gyn/ik. 221, 83-91 (1976) Archiv fi3r

Gyn/ikologie �9 J. E. Bergmann-Verlag 1976

Altered Carbohydrate Metabolism in Breast Cancer and Benign Breast Affections

B. R. Muck 1, S. Trotnow 1, H. Egger I and G. Hommel 2

1Universit/its-Frauenklinik Erlangen-N/irnberg (Dir.: Prof. Dr. K. G. Ober) Universit/itsstr. 21/23, D-8520 Erlangen, Federal Republic of Germany 2Institut ffir Medizinische Statistik und Dokumentation der Universitfit Erlangen-Nfirnberg (Dir.: Prof. Dr. L. Horbach)

Glucoseintoleranz bei Frauen mit gut- und bfsartigen Brustbefunden

Summary. Breast diseases in 792 women were studied by biopsy and histological evaluation. In all subjects glucose tolerance was examined by OGTT (100 g glucose). The diabetes frequency of 22% in 326 women with breast cancer was compared with the frequency in women with fibroadenoma (n = 101), papilloma (n = 80), fibrocystic disease (n -- 107), lipoma, granuloma, fibrosis (n = 88), papilloma with proliferation (n = 32), mastopathy with proliferation (n = 33) and carcinoma in situ lobulare (n = 11). The statistical evaluation was done with an electronic data processing system. We used matched pairs according to age, height and weight. Diabetogenic factors like age and overweight were thus al- lowed for. These comparative statistics showed a frequency of diabetes twice or three times higher in women with breast cancer. This result cannot be regarded ! as a consequence of age, overweight and menopause. In groups with fibroadeno- ma, fibrocystic disease and lipoma, we found glucose intolerance in 1-3%, whereas the group with proliferation (including carcinoma in situ) showed an incidence of 7%. The remarkably high incidence rate of 14% in women with papilloma can be explained by the higher age and the more frequent obesity in this collective.

Zusammenfassung. Bei 792 Frauen wurde ein Brustbefund histologisch ab- gekl~rt. Die Glucosetoleranz wurde bei allen Patienten pr/ioperativ durch eine orale Belastungsprobe mit 100 g Glucose fiberprfift. Die H/iufigkeit pathologi- scher Glucosetoleranz von 22% bei 326 Frauen mit Brustkrebs wird mit der bei Frauen mit Fibroadenom (n = 101), PapiUom (n = 80), Mastopathia cystika fibrosa (n = 107), Lipom, Granulom, Fibrose (n = 88), proliferierender Papillo- matose (n = 32), proliferierender Mastopathie (n = 33) sowie lobul~irem Carci- noma in situ (n = 11) verglichen. Die Auswertung erfolgte mittels einer elektroni- schen Informationsverarbeitungsanlage durch Bildung yon matched pairs unter Berficksichtigung diabetogener Faktoren wie Alter, GrSl3e und Gewicht. Im sta- tistischen Vergleich wird deutlich, dab die 2-3fach hShere Diabetesh~iufigkeit

84 B.R. Muck et al.

bei F r auen mit Brustkrebs nicht ausschliel31ich durch Alter, f0bergewicht oder Menopause bedingt ist. W/ihrend sich bei F r a u e n mit F ib roadenom, Mas topa- thia cys t ika f ibrosa oder Lipom, Granulom, F ibrose eine Diabetesh/iufigkeit yon 1 - 3 % findet, haben F rauen mit proliferierenden Befunden (einschliel3tich eines Ca in situ) zu 7% eine Glucoseintoleranz. D ie / ibe r ra schend h6here Diabetesra te yon 14% bei F rauen mit Papi l lomen erkl~irt sich durch das h6here Alter und die h~iufigere Adiposi tas dieses Kollektivs.

Much clinical emphasis has been placed on the identification of individuals tending to develop breast cancer by discovering factors which might influence the develop- ment of cancer o f the breast [6, 13, 14, 17, 19].

In an earlier prospective s tudy we repor ted an altered ca rbohydra te metabol ism related to breas t cancer [11]. 22% of women with breas t cancer showed glucose intolerance as compared to 3% with histologically benign breas t diseases. This com- municat ion will review some factors that appear to increase or - equally impor tant - decrease the risk of developing breas t cancer by analysing the above mentioned study. Our question was" do we find different incidence rates of diabetes in the different groups of breas t affections with regard to the histological characterist ics, statist ically compared to groups with breast cancer? In the following we will discuss whether some of these statist ically significant variables can account for an epidemio- logical pat tern of breas t cancer.

Material and Methods

From January 1st, 1971 until December 31st, 1973 we studied the frequency of diabetes in the breast cancer population at the Universit/its-Frauenklinik Erlangen. Patients with diagnoses of benign lesions were taken as a control group.

All patients (n = 792) included in this study had a histological diagnosis of cancer or of some benign lesion. The tumor or suspect area of tissue was removed under intubation anaesthesia in all subjects.

Following an overnight fast the carbohydrate metabolism was evaluated previous to surgery by a standard oral glucose tolerance test (OGTT), using 100 g of glucose [17]. The test was considered positive if the following criteria were present [18]:

1. The blood-sugar level rose above 180 rag/100 ml. 2. The blood-sugar level was more than 140 mg/100 ml at the two-hour point. The blood glucose was determined by the hexokinase method. Our histological classification is

demonstrated in Table 1. Each slide of the benign findings (n = 466) was re~4ewed under the aspect of searching for proliferation.

Electronic data processing: According to the histological findings, we formed matched pairs allow- ing for the patients' age, height and weight. The program was called Nohomm Q 2. 2 x 2 contingency tables were assessed for the frequency of diabetes (test of McNemar) [19].

Results

The frequency of diabetes is demons t ra ted in Figures 1 - 4 . Significant differences in the various groups are marked with " + " (p = 0.05) (test of McNemar) .

Pathological glucose tolerance curves after O G T T were found in 29% of the women with breas t cancer (n = 326) and in 5.7% of the patients with benign breas t

Altered Carbohydrate Metabolism in Breast Cancer

Table 1. Histological classification of the breast diseases

85

Number 88

107 101 80 11 3

32 33 11

326

n = 792

of cases Lipoma, granuloma, fibrosis Fibrocystic disease Fibroadenoma Papilloma Sclerosing adenosis Adenoma of the nipple Proliferating duct papiltomatosis Fibroeystic disease accompanied by marked epithelial proliferation Lobular and intraductal carcinoma in situ Cancer

% 30 25 20 15 10 5 0 5 10 15 20 25 30 % Test of McNemar I I I , I I I I , , , f t . . .

Total collective (n = 5 1 1 ) ~ ~ \ \ \ \ \ \ \ \ \ \ \ ' ~ (n = 326)Matched pairs +

Benign diseases with proliferation :::::: n = 75 +

Benign diseases without proliferation 185 +

Fibroadenoma ~ ~ 54 +

Fibro-cystic disease ~ 90 +

Lipoma, Granuloma ~ ~ 69 +

Papilloma ~ ~ 72 G

D. without proliferation/without papilloma ~ ~ 144 +

D. with proliferation and papilloma 137 +

t

Benign diseases ~ manifest ~ Cancer subclinical Diabetes

Fig. 1. Frequency of diabetes in women with benign and malignant breast diseases (Matched pairs according to age, height and weight)

diseases (n = 510) (Fig. 1). The statistical comparison of the matched pairs (accord- ing to age, height and weight) shows that patients with breast cancer are more frequently diabetics than women with f ibroadenoma, fibro-cystic disease, lipoma, granuloma and benign diseases with epithelial proliferation (including carc inoma in situ lobulare). Glucose intolerance is significantly correlated to breast cancer (Fig. 1). N o significant difference can be found in patients with papil loma (Fig. 1). This is a surprising result. A further analysis o f this fact (7% manifest and 7% subclinical diabetics; n = 72) demonstrates that the above mentioned higher frequency is attri- butable to a more frequent obesity and increasing age (Fig. 2 and 4). Fur thermore

Tes

t of

McN

emar

+ + + Q

+ q-

+

%

40 3

0 20

!

I I

Mat

ched

pai

rs

n =

44

42

66

48

45

113

99

8O

Nor

mal

wei

ght

10

0 10

20

30 4

0 I.

I I

I I

% Fib

road

enom

a

Pap

illo

ma

Fib

ro-c

ysti

c di

seas

e

Lip

oma,

Gra

nulo

ma

Ben

ign

dise

ases

wit

h ep

ithe

lial

pro

life

rati

on

Ben

ign

dise

ases

wit

hout

epi

thel

ial

prol

ifer

atio

n ~

N~

.

B.D

. w

itho

ut p

roli

fera

tion

/wit

hout

pap

illo

ma

B.D

. w

ith

prol

ifer

atio

n an

d pa

pill

oma

Ben

ign

dise

ases

~

man

ifes

t ~

Can

cer

subc

lini

cal

Ove

rwei

ght

%

40

30

20

10

0 10

20

30

40

%

t__a

__.a

.__L

_._[

,

~ '

--

Mat

ched

pai

rs

Tes

t of

McN

emar

n =

10

Q

30

24

Q

21

O

30

O

72

45

57

+

Dia

bete

s

Fig.

2.

Fre

quen

cy o

f di

abet

es i

n w

omen

wit

h be

nign

and

mal

igna

nt b

reas

t di

seas

es (

Mat

ched

pai

rs a

ccor

ding

to

age,

hei

ght

and

wei

ght)

Altered Carbohydrate Metabolism in Breast Cancer 87

% 20 15 10 5 5 10 15 20 % i | | |

i , , ' | Matched pairs Test of McNemar

Fibroadenoma ~ Papilloma n = 30 Q

Fibro-cystic disease Papilloma 54 2~

Lipoma, granuloma ~ Papilloma 44 @

Disease with proliferation ~ Papilloma 45 @

Disease without proliferation Papilloma 66

Benign diseases ~ manifest [ ~ Papilloma B subclinical

Diabetes

Fig. 3. Frequency of diabetes in women with benign breast diseases (Matched pairs according to age, height and weight)

2s %

20

�9 Cancer/Fibroadenoma

o Cancer/Papilloma

Cancer/Fibro-cystic disease

+ Cancer/Lipoma, Granuloma

�9 Cancer/Diseases with epithelial proliferation 15

o

10 �9 ~ @

5

I I I I I I

30 40 50 60 70 80 Years

Fig. 4. Age distribution in women with benign and malignant breast diseases (Matched pairs according to age, height and weight)

Figure 2 shows that the coincidence of diabetes and breast cancer is especially to be found significantly increased in women with normal weight. The same applies to normal weight women with papilloma. In comparison to other benign diseases, women with papilloma are not more frequently diabetics (Fig. 3). A very interesting result is the fact that normal weight women with cancer as well as those women with benign diseases plus epithelial proliferation show no significant difference in diabetes frequency (Fig. 2). Obese women with breast cancer don' t have a significantly higher glucose intolerance rate compared to obese women with benign findings (except for the group with proliferation and papilloma). Investigating each group, however, we find more diabetics with cancer (Fig. 2).

88 B.R. Muck et al.

% 25

20 �9 Cancer/Benign diseases with proliferation

A Cancer/B.D. with proliferation and papilloma

15 o Cancer/B.D. without proliferation and without papilloma

10

5-

I I I 1 ! !

30 40 50 60 70 80 years

Fig. 5. Age distribution in women with benign and malignant breast diseases (Matched pairs according to age, height and weight)

The age distribution of the matched pair groups is demonstrated in Figures 4 and 5. In all groups women with an age of about 50 years showed the highest frequency of breast diseases. More women are seen with a papilloma, lipoma or granuloma with increasing age. In these cases a second frequency peak appears at about 70 years. In contrast, fewer seventy years old patients are seen with fibroade- noma (Fig. 4). Only 13% of the patients with fibroadenoma were in the postmeno- pause, whereas in the other groups we see a rate of 4 0 - 5 0 % (Table 2). The age of menopause was the same in all groups, about 50 years.

Discussion

Our data show that the incidence of diabetes in breast cancer patients is excessive. There is no doubt about a positive correlation. In previous studies the glucose in- tolerance was not significantly related to breast cancer; and those studies which de- scribed glucose intolerance in this connection attributed this factor to age and obe- sity of patients [4, 15, 18]. Our results, however, show that the decreased glucose tolerance in women with breast cancer is not to be regarded as a consequence of overweight and age. The menopausal status does not contradict this conclusion, since the age of menopause was about the same (49-50 yrs.) in the different groups. Glucose intolerance was shown only in 1 - 3 % of the women with fibroadenoma, fibro-cystic disease, and lipoma/granuloma groups (Fig. 1). On the other hand dia- betes incidence was found in 14% of the women with papilloma and in 7% of the women with benign diseases and epithelial proliferation. The higher incidence rate in the group with papilloma can be attributed to the increasing age and weight (Fig. 2, Table 2). This fact is not contrary to our second hypothesis. Increased carbohydrate metabolism disorder is also correlated to women with benign breast disease and

Tab

le 2

. F

requ

ency

of

diab

etes

in w

omen

wit

h a

regu

lar

men

stru

al c

ycle

and

in

wom

en a

fter

men

opau

se

Wom

en w

ith

a re

gula

r m

enst

rual

cyc

le

Wom

en a

fter

men

opau

se

8.

C?

Nor

mal

M

anif

est

Subc

lini

cal

Mat

ched

pai

rs a

ccor

ding

to

age,

N

orm

al

Man

ifes

t Su

bcli

nica

l ~r

o

gluc

ose

diab

etes

di

abet

es

heig

ht,

wei

ght

gluc

ose

diab

etes

di

abet

es

tole

ranc

e to

lera

nce

1%

38 (

70.3

%)

1 (1

.9%

) 3

(5.6

%)

Can

cer

54

10 (

19.0

%)

~ 2

(3.7

%)

90=

�9

o.

�9

~

~'~

=4

l:a"

~ I :~

�9

~-~

~ 5

.~a

~ o

46 (

81.5

%)

27 (

37.5

%)

33 (

46.0

%)

42 (

47.0

%)

51 (

57.0

%)

28 (

41.o

%)

38 (

54.0

%)

23 (

32.9

%)

30 (

40.0

%)

60 (

32.4

%)

75 (

40.6

%)

59 (

41.0

%)

73 (

50.7

%)

43 (

32.3

%)

66 (

48.2

%)

Q

Q

Fib

road

enom

a 7

(12.

9%)

1 (1

.9%

)

1 (1

,4%

) 1

(1.4

%)

Can

cer

29 (

40.0

%)

11 (

15.3

~ 1

(1.4

%)

3 (4

.2%

) Pa

pill

oma

72

29 (

40.0

%)

4 (5

.6%

2 (2

.2%

) 3

(3,3

%)

Can

cer

27 (

30.0

%)

5 (5

.5%

Q

1

(1.1

%)

Fib

rocy

stic

dis

ease

90

36

(40

.0%

) 2

(2,2

%

2 (2

.9%

) 2

(2.9

%)

Can

cer

27 (

39.0

%)

8 (1

1.6%

Q

1

(1.5

%)

Lip

oma,

gra

nulo

ma

69

29 (

42.0

%)

2 (2

.9%

3 (4

.0%

) 4

(5.3

%)

Can

cer

38 (

50.7

%)

5 (6

.7%

1

(1.3

%)

1 (1

.3%

) B

enig

n di

seas

es w

ith

prol

if.

75

40 (

53.4

%)

2 (2

.6%

4 (2

.2%

) 6

(3.2

%)

Can

cer

89 (

48.1

%)

22 (

11,9

%)

2 (1

.1%

) 5

(2.7

%)

Ben

ign

dise

ases

wit

hout

pro

lif.

185

92 (

49.7

%)

9 (4

.9%

)

3 (2

.1%

) 6

(4.2

%)

Can

cer

63 (

43.7

%)

11

(7.6

%)

1 (0

.7%

) 2

(1.4

%)

Ben

ign

dise

ases

wit

hout

64

(44

.4%

) 4

(2.8

%)

prol

if,

wit

hout

pap

illo

ma

144

5 (3

.8%

) 5

(3.8

%)

Can

cer

58 (

43.6

%)

24 (

17.5

%)

2;

5 (3

.8%

) B

enig

n di

seas

es w

ith

prol

if.

137

56 (

42.1

%)

7 (5

.1%

) an

d pa

pill

oma

Q

3 (4

.2%

) 2

(2.8

%)

1 (1

.1%

) 0 2

(2.9

%)

Q

2 (2

.6%

) 1

(1.3

%)

4 (2

.2%

) 2

(1.1

%)

2 (1

.4%

)

2 (1

.5%

/ 3

(2.3

%)

g ,--t

90 B.R. Muck et al.

breast cancer may be increased in patients with such metabolic defects, or the tumor may adversely affect the tissue metabolism of its host. In previous studies we could demonstrate that similar somatic, clinical and metabolic characteristics exist in breast cancer patients as well as in women with endometrial cancer [9-12] . We found glucose intolerance in 50% of the women with endometrial cancer and in 22% of breast cancer patients. According to these findings a uniform endocrine hormone constellation in breast and endometrial cancer patients can be discussed [5, 16]. Based on this assumption glucose intolerance can be regarded as a consequence of this hormone metabolism disorder. Carter reported elevated growth-hormone levels and prolonged insulin secretion in women with breast cancer [1]. These changes were also not attributable to age and obesity. Therefore measurements of metabolic and hormonal characteristics seem to be more promising than epidemiologic vari- ables in identifying those women with a high risk of developing mammary cancer. These metabolic abnormalities may either characterize host susceptibility to breast cancer or be effects of the tumor. However much more needs to be known as to when the abnormal glucose tolerance pattern appears in women showing this crite- ria. Is it a characteristic of the individual, or is it a change coinciding with cellular changes in the breast that eventually becomes recognizable as cancer? In the first instance the GTT would serve to assign women to different categories of mammary cancer risk. In the second instance it would be a periodic screening procedure for identifying women moving from the low-risk into the high-risk category. In both cases the high-risk women would have to be screened periodically in search of presymptomatic characteristics like epithelial proliferation or carcinoma in situ.

The proposal of a clear-cut hypothesis explaining the development of breast cancer is difficult because the present study as well as other studies have not been able to reveal factors that play a decisive role in the prevention of breast cancer [2, 18]. This may partly be due to the fact that we are dealing with endogenous factors which are difficult to discover with epidemiological surveys, this partly being due to the fact that breast cancer may be related to more than one internal factor.

References

1. Carter, A. C., Lefkon, B. W., Farlin, M., Feldmann, E. B.: Metabolic parameters in women with metastatic breast cancer. J. Clin. Endocrinol. Metab. 40, 260 (1975)

2. Dunn, J. E.: Epidemiology and possible identification of high-risk groups that could develop cancer of the breast. Cancer 23, 775 (1969)

3. Geigy, J. R.: Documenta Geigy. Wissenschaftliche Tabellen, 7. Aufl. Basel (1969) 4. Glicksmann, A. S., Rawson, R. W.: Diabetes and altered carbohydrate metabolism in patients with

cancer. Cancer 13, 559 (1960) 5. Kaiser, R.: Endokrine Schutzmechanismen gegen Endometrium- und Mammakarzinome. DMW

94, 2467 (1969) 6. Maass, H.: Das Mammacarcinom. Geburtsh. u. Frauenheilk. 28, 823 (1968) 7. Mehnert, H., F6rster, H.: Stoffwechselkrankheiten, p. 216. Stuttgart: Thieme 1970 8. Mehnert, H., Haslbeek, M., F6rster, H.: Zur Pr/ifung der oralen Glucosetoleranz. DMW 45, 1763

(1972) 9. Muck, B. R.: Glucosetoleranz bei 122 Frauen mit Endometriumcarcinom. 9. Kgrs. d. Dtsch.

Diabetes-Gesellschaft, Travemfinde (1974) 10. Muck, B. R., Trotnow, S.: Endometriumcarcinom und pathologische Glucosetoleranz. Arch.

Gyn/ik. 216, 301 (1974)

Altered Carbohydrate Metabolism in Breast Cancer 91

11. Muck, B. R., Trotnow, S., Hommel, G.: Cancer of the breast, diabetes and pathological glucose tolerance. Arch. Gyn/ik. 218, 299 (1975)

12. Muck, B. R., Trotnow, S., Egger, H.: Diabetes und Mammatumoren. VII. Tagg. Deutschsprecben- der Hochschullehrer in der Gyn/ikologie und Geburtshilfe, Mfinchen (1975)

13. Pauli, K. H., Trotnow, S.: Zur Epidemiologie des Mammaearcinoms. Arch. Gyn/ik. 213, 271 (1973)

14. Waard, F. de: The epidemiology of breast cancer: review and prospects. Int, J. Cancer 4, 577 (1969)

15. Werner, W.: Diabetes mellitus und Carcinom. Z. Krebsforsch. 60, 399 (1955) 16. Wynder, E. L.: Current concepts of the aetiology of breast cancer (eds. Forrest, A. P. M., Kunkler,

P. B.), p. 32. Baltimore: Williams and Wilkins 1968 17. Wynder, E. L.: Identification of women at high risk for breast cancer. Cancer 24, 1235

(1969) 18. Wynder, E. L., Bross, I. J., Hirayama, T.: A study of the epidemiology of cancer of the breast.

Cancer 13, 559 (1960) 19. Zippin, C., Petrakis, N. L.: Identification of high risk groups in breast cancer. Cancer 28, 1381

(1971)

Received December 19, 1975