Volume 115
Number 5
1 September 1991
Annals of Internal Medicine ARTICLES
Timing of Breast Cancer Excision during the Menstrual Cycle Influences Duration of Disease-free Survival Ruby T. Senie, PhD; Paul Peter Rosen, MD; Philip Rhodes, MS; and Martin L. Lesser, PhD
• Objective: To study disease-free survival at 10 years in relation to timing of breast tumor excision during the menstrual cycle. • Design: A prospective study of consecutively treated patients with primary breast cancer. • Setting: Memorial Sloan-Kettering Cancer Center, New York. • Patients: Two hundred and eighty-three premenopausal patients treated by mastectomy and axillary dissection. • Main Results: When the tumor was excised during the follicular phase, approximated by setting the putative day of ovulation on day 14 after the onset of last menses, a higher recurrence risk (43%) was observed compared with excision later in the menstrual cycle (29%, P = 0.02). The rate peaked among patients treated between days 7 and 14 and was lowest between days 20 and 30. Multivariate analysis using the Cox regression model to control for tumor size, nodal status, estrogen receptor status, adjuvant chemotherapy, and family history indicated that the hazard rate of breast cancer recurrence after excision during the follicular phase was 1.53 (95% CI, 1.02 to 2.29). Stratification by nodal status indicated that the effect of phase was statistically significant only among patients with positive nodes (hazard ratio, 2.10; CI, 1.19 to 3.70). • Conclusions: Our results support the hypothesis that the risk for recurrence may be affected by the hormonal milieu of the menstrual cycle; these findings must be confirmed, however, by a prospective study in which cycle phase at time of tumor excision is biochemically documented.
Annals of Internal Medicine. 1991;115:337-342. From the Centers for Disease Control, Atlanta, Georgia; Memorial Sloan-Kettering Cancer Center, New York, New York; and North Shore University Hospital-Cornell University Medical College, Manhasset, New York. For current author addresses, see end of text.
I n a study of 41 patients with breast cancer, Hrushesky and colleagues (1) observed significantly better diseasefree and total survival after tumor resection between days 7 and 20 of the menstrual cycle compared with days closer to onset of menses. However, a significant association between prognosis and timing of tumor resection was not confirmed in three subsequent reports of somewhat larger patient samples (2-4). To address this question further, we evaluated two schemes for subdividing the menstrual cycle: the intervals defined by Hrushesky and colleagues and the hormone-dependent phases determined by the putative time of ovulation, the fourteenth day after the last menstrual period, approximating the division to follicular and luteal phases of the cycle (5). Table 1. Demographic and Pathologic Characteristics of 283 Premenopausal Patients Characteristic
Number (%)
Age at diagnosis, y 46 Family history of breast carcinoma None Primary only: mother, sister, daughter Secondary only: other relatives Both primary and secondary relative Parity Nulliparous Parous Tumor size Microscopic < 2 cm 2.1 to 3.9 cm > 4 cm Axillary lymph node status Negative Positive Estrogen receptor status* Positive Negative Borderline Contralateral breast disease Benign Concurrent bilateral Subsequent bilateral No biopsy done [
33 (12) 128 (45) 122 (43) 186 (66) 30(11) 50 (18) 17(6) 49 (17) 234 (83) 34 (12) 121 (43) 79 (28) 49 (17) 166 (59) 117(41) 79 (37) 88 (41) 47 (22) 142 (50) 37 (13) 9(3) 95 (34)
Analysis was based on data available for 214 patients.
1 September 1991 • Annals of Internal Medicine • Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 06/08/2017
337
Methods
Data Analysis
The premenopausal patients included in our study were a subset of 1254 consecutive patients with breast carcinoma treated at Memorial Sloan-Kettering Cancer Center between October 1976 and June 1978 (6, 7) who were recruited for the study of risk factors associated with histologic type of tumor. Epidemiologic information was obtained through personal interviews conducted by one of the investigators with 1216 (97%) patients during their postoperative hospitalization. Tumor size, status of the axillary lymph nodes, and other prognostic factors were assessed by another investigator, who reviewed the histologic slides (8-10). The cohort was followed closely for 10 years through annual contact with the patients and their physicians. Recurrence of breast cancer was confirmed by clinical or pathologic assessment, or both. Disease-free months were calculated from date of diagnosis at time of tumor excision to date of first recurrence, date of death due to other causes, or date of last contact for patients free of recurrent disease. The interval between onset of last menstrual period and tumor excision was used in the analysis for the 27% who had two separate surgical procedures, tumor excision before mastectomy and axillary dissection. Survival analyses have been done on 928 of the 1216 patients after the following exclusions: primary treatment other than mastectomy with axillary dissection (45 patients, 4%), indeterminate size of the primary tumor (91 patients, 7%), distant metastases at diagnosis (63 patients, 5%), previous treatment for contralateral breast carcinoma (82 patients, 7%), and lack of follow-up (7 patients, < 1%). Interview data and medical records were reviewed to determine date of last menstrual flow for all patients with regular cycles classified as premenopausal (10). Of the 315 identified patients, 14 were considered inappropriate for inclusion: 2 were less than 6 months postpartum, and 12 reported hormone therapy during the year before diagnosis. The date of the last menstrual period was not available for 18 (6%) patients, resulting in complete data available for 283 (94%) of 301 patients considered appropriate for inclusion. Two schemes were used to stratify the study sample at the time of diagnostic surgery. In the first analyses, patients were grouped as described by Hrushesky and colleagues: midcycle (days 7 through 20) or perimenstrual (days 0 through 6 and days 21 through 40) (1). The second stratification method relied on the hormone-dependent phases determined by the putative time of ovulation 14 days after last menstrual period, approximating the end of the follicular phase. The luteal phase was assumed to commence with the fifteenth day.
Statistical procedures for univariate analyses included the chi-square test applied to categorical data, the Student /-test for continuous variables, and nonparametric analyses for variables not normally distributed. Statistical techniques designed for censored data were used to evaluate the duration of disease-free survival (11). Product-limit survival curves were calculated by the method of Kaplan and Meier (12). The equality of the distributions was tested by the Mantel-Cox procedure (13). The results of analyses performed on all 283 patients were not altered when the study sample was restricted to 249 cases with invasive breast cancer. Cox proportional hazards modeling was used to estimate the relative rate of recurrence associated with timing of diagnostic surgery while simultaneously adjusting for the effects of tumor size, number of positive axillary lymph nodes, estrogen receptor status, and the presence or absence of a family history of breast cancer (14). We used the programs of BMDP statistical package (BMDP Statistical Software, Los Angeles, California) to do all survival analyses. We also calculated and plotted running averages per 1000 person-months of disease-free survival to produce smoothed rates of recurrence by the day of the menstrual cycle at the time of diagnostic surgery. Confidence intervals (CIs) of 95% are given where appropriate. Results Table 1 presents demographic and pathologic characteristics of the study sample. The age at diagnosis ranged from 24 to 55 years (mean, 43.4 years ± 6.0 [SD] years). The mean number of days from the onset of last menses to surgery was 15 ± 9.7 days. The tumor was excised within 28 days of the last menstrual period for 90% of the women. A carcinoma of microscopic size (not grossly measurable) was identified in 34 (12%) patients. The mean size of the 249 measurable tumors was 2.8 ± 1.9 cm, and 41% of the patients had axillary lymph node metastases. Histologic evaluation showed that 62% of the tumors were infiltrating duct; 12%, medullary carcinoma; 9%, infiltrating lobular; and 17%, other histologic types. Of the 214 (76%) tumors analyzed for estrogen-receptor status, 37% were positive; 4 1 % , negative; and 22%, borderline. Ten years after diagnosis, 177 (63%) patients were alive and free of recurrent disease, 8 (3%) had died of other causes without recurrent breast carcinoma, 74 (26%) had died of breast carcinoma after developing metastases, and 24 (8%) were alive with recurrent carcinoma. Midcycle Compared with Perimenstrual Analysis
Figure 1. Disease-free survival according to menstrual interval at time of tumor excision. The number of patients at risk at time 0, at 5 years, and at 10 years is shown for each set of survival curves. 338
No statistically significant differences were noted when the demographic and pathologic characteristics presented in Table 1 were assessed by menstrual interval, although a greater percentage of diagnostic surgery was done during the perimenstrual interval (54%) than during midcycle (46%). Ten years after diagnosis, the risk for recurrence did not differ significantly; 40% developed recurrence after surgery during midcycle compared with 32% of the perimenstrual subset (P = 0.17). Although Figure 1 indicates a higher recurrence rate for women with tumor excision during midcycle, the differences in survival were not statistically significant. Multivariate analysis
1 September 1991 • Annals of Internal Medicine • Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 06/08/2017
Figure 2. Disease-free survival. Left. Disease-free survival according to menstrual phase at the time the tumor was excised. The number of patients atriskat time 0, at 5 years, and at 10 years is shown for each set of survival curves. Hazard ratio, 1.59 (95% CI, 1.06 to 2.38). Right. Disease-free survival according to menstrual phase and axillary lymph node status. Negative nodes: hazard ratio, 1.07 (CI, 0.58 to 1.99); P > 0.2. Positive nodes: hazard ratio, 1.97 (CI, 1.14 to 3.41); P = 0.02. controlling for tumor size, number of positive nodes, and family history indicated that survival did not differ by timing of tumor excision during midcycle compared with the perimenstrual interval. Follicular Compared with Luteal Analysis When the second scheme of subdivision was used, the patients were equally divided by follicular and luteal phases. Women in the luteal subset were slightly older at diagnosis (mean age, 44.1 years compared with 42.7 years) and reported a family history of breast cancer more frequently (41% compared with 28%; P = 0.05). More women who had tumor excision during the follicular phase were found to have axillary lymph node metastases compared with those diagnosed during the luteal phase (46% compared with 37%); however, among the 117 women with positive nodes, the mean number of involved nodes did not differ significantly by phase. Seventy-eight patients received combination adjuvant chemotherapy (cytoxan, methotrexate, and fluorouracil). The proportion of patients with positive nodes treated with adjuvant chemotherapy did not differ significantly by phase (68% of follicular and 54% of luteal); three negative-node patients in each subset also received chemotherapy. Among those treated, the number of months of therapy did not differ by phase. Estrogen-receptor status, assessed in 214 patients, was similar for the two subsets: 35% of patients in follicular and 39% of those in luteal phase had a positive result; 44% of patients in follicular and 39% of those in luteal phase had a negative result; the remaining were classified as borderline. No significant differences in specific binding by phase were observed using the Mann-Whitney nonparametric test or by r-test analysis after log transformation. Other pathologic features, including tumor size, histologic type, nuclear grade, and bilateral disease, were similar for the two subsets.
The estimated 10-year risk for recurrence associated with tumor excision during the first 14 days after onset of last menstrual flow (compared with later in the cycle) was significantly higher (43% compared with 29%, P = 0.022). Figure 2, left, indicates statistically significant differences in time to first recurrence by phase (P = 0.023). The hazard ratio (1.59; CI, 1.06 to 2.38) for surgery during the follicular phase uncontrolled for other prognostic factors is shown in Table 2. Because the presence of positive nodes differed by phase, the patients were subdivided by both menstrual phase and nodal status. As noted in Figure 2, right, the effect of phase was limited to patients with positive nodes. The hazard ratio of surgery during the follicular phase among patients with positive nodes was 1.97 (CI, 1.14 to 3.41) but was not statistically significant among patients with negative nodes (hazard ratio, 1.07; CI, 0.58 to 1.99). The plot of smoothed rates of recurrence per 1000 person-months of disease-free survival (Figure 3) indicates highest rates of recurrence when the tumor was excised between days 7 and 14 and lowest rates when the tumor was excised between days 20 and 30. The plot shows the likelihood of recurrence in relation to day of the menstrual cycle without imposing predetermined intervals and without controlling for other potential prognostic factors. Cox regression models were used to assess the effect of phase for the total sample and were subdivided by nodal status. Estimates of the hazard ratios, 95% CIs, and their significance levels are shown in Table 2. After controlling for tumor size, number of positive axillary nodes, adjuvant chemotherapy, and family history, the menstrual phase at time of tumor excision remained significantly associated with disease-free survival (hazard ratio, 1.53; CI, 1.02 to 2.29). When the effect of phase was compared by nodal status, a sizable although
1 September 1991 • Annals of Internal Medicine • Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 06/08/2017
339
not statistically significant difference was observed; the hazard ratio for patients with positive nodes was 2.10 (CI, 1.19 to 3.70) and for negative-node patients, 0.98 (CI, 0.52 to 1.83). The inclusion of estrogen-receptor status in the model restricted the study population to 214 cases; however, results were not significantly altered (hazard ratio, 1.68; CI, 1.07 to 2.63). Discussion Studies of mice indicated a lower rate of subsequent pulmonary metastases when mammary tumors were excised during the estrus cycle, close to the time of ovulation (15). A similar association of survival differences in relation to timing of tumor excision was observed among 41 premenopausal patients. Disease-free and total survival were greater after surgery in midcycle (around ovulation) than during the perimenstrual interval (1). Although their findings were statistically significant, the authors cautioned that their results were based on a small sample and needed to be confirmed by others. In three subsequent reports of larger samples in which the same menstrual cycle divisions were applied, time of diagnostic surgery was not associated with statistically significant differences in disease-free or overall survival (2-4). Comparisons across studies were complicated by differences in stage of disease, treatment protocols, and duration and completeness of follow-up as well as rates of survival. When we assessed time to first recurrence by the menstrual categories, perimenstrual and midcycle, we obtained results similar to those of others (2-4), which were contrary to the findings of Hrushesky and colleagues (1). Because our study group was significantly larger with 98 recurrences, our study had sufficient power to observe effects even smaller than those suggested by Hrushesky and colleagues. Indeed, our results indicate that the effect was in the opposite direction to results previously reported (1). After categorizing our study sample in relation to the putative timing of ovulation, approximating the follicular and luteal phases, however, statistically significant differences in duration of disease-free survival were apparent. Our cohort of 283 patients was treated at one institution with a standard surgical protocol during an 18-month period and was carefully followed for 10 years. Among the 98 (35%) women who had relapses, the rate was 50% higher, especially among women with positive ax-
illary nodes, when surgery had been done within 14 days of last menstrual period. The menstrual cycle is characterized by an estrogen peak in the absence of progesterone before ovulation with a second estrogen surge later in the cycle coinciding with a rapid secretion of progesterone (16). Our results revealed an elevated hazard ratio among women whose tumor was excised during the follicular phase, in the presence of unopposed estrogen that potentially enhanced micrometastases. The significantly better prognosis among women with tumor excision during the luteal phase, when progesterone levels were highest, may be related to the findings of in-vitro studies that revealed an inhibition of human breast carcinoma cell proliferation by synthetic progestin (17). In 1896, Beatson (18) reported the influence of endogenous hormones on breast cancer survival; the precise mechanisms of action, however, are still not fully understood. In our series and in others (19, 20), estrogenreceptor status of the tumor did not vary with menstrual phase at time of tumor excision. However, cyclical patterns of cell division and deletion of normal breast tissue have been found to correspond to endogenous hormonal surges of the natural menstrual cycle (21, 22). Natural killer-cell activity has been found to vary in relation to the estrus cycle of tumor-free mice (23). Studies of healthy women have noted cyclical reductions in natural killer-cell activity before ovulation with a return to higher levels later in the menstrual cycle (24, 25). Other investigators (26) have reported a significant decrease in phagocytic activity of mononuclear cells early in the menstrual cycle. Therefore, diminished immune function before the putative day of ovulation may be one mechanism associated with the observed pattern of recurrence. One limitation of our study and others (1-4) is the reliance on date of last menstrual period to estimate the menstrual phase at time of tumor excision without precise determination of the hormonal milieu. Because we set the putative time of ovulation at the fourteenth day after the onset of the last menstrual flow and because the usual cycle length had not been recorded, we may have miscoded some women in the midrange of days between onset of last menses and tumor excision (27, 28). Biochemical data were not available to determine if ovulation had occurred on or after the fifteenth day after the last menstrual cycle and before tumor excision for the 142 patients classified to the luteal subset; at the
Table 2. Effect of Follicular Phase on Disease-free Survival at 10 Years in 283 Premenopausal Patients (Using Cox Regression Models) Patients and Nodal Status All women Nodal statust Negative Positive
Controlled* (95% CI)
Hazard Ratio
Crude (95% CI)
1.59
(1.06 to 2.38)
0.025
1.53
(1.02 to 2.29)
0.04
1.07 1.97
(0.58 to 1.99) (1.14 to 3.41)
>0.2 0.016
0.98 2.10
(0.52 to 1.83) (1.19 to 3.70)
>0.2 0.01
P Value
Hazard Ratio
P Value
* Controlling for tumor size, number of positive nodes, chemotherapy, and family history. t P values comparing the effects of follicular phase in 117 women having positive nodes with 166 women having negative nodes (crude P = 0.15; controlled P = 0.08).
340
1 September 1991 • Annals of Internal Medicine • Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 06/08/2017
among 83 positive-node patients who had tumors excised during mensturation (days 0 to 2 after onset of menses) or later in the cycle (days 13 to 32) compared with the 43 positive-node patients who were treated between days 3 and 12 after onset of last menses. As in our study, their results were not influenced by other prognostic factors, including estrogen-receptor status. Presented in part (34) at the 13th Annual San Antonio Breast Cancer Symposium on 2 November 1990, San Antonio, Texas. Acknowledgments: The authors thank the women for their willingness to be interviewed during a stressful period of hospitalization. They also thank their surgical colleagues for their assistance during the inception of this study and throughout the 10 years of follow-up. Requests for Reprints: Ruby T. Senie, PhD, Centers for Disease Control, 1600 Clifton Road, Atlanta, GA 30333. Current Author Addresses: Dr. Senie and Mr. Rhodes: Centers for Disease Control, 1600 Clifton Road, Atlanta, GA 30333. Dr. Rosen: Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10021. Dr. Lesser: North Shore University Hospital, 300 Community Drive, Manhasset, NY 11030.
Figure 3. Smoothed recurrence rates. Recurrence of breast carcinoma per 1000 months of disease-free survival according to the day of the menstrual cycle at time of tumor excision.
time of interview, however, all patients reported regular cycles. Although the proportion of nonovulatory cycles has been observed to increase with age (5), studies indicate that women with regular menses ovulate in more than 95% of cycles, even when they are close to menopause (29). Variations in the duration and regularity of the menstrual cycle have been documented between women and during the menstrual years of individual women (28, 30). A prospective study is therefore required to confirm our findings of a relation between disease-free survival and phase of the menstrual cycle at time of tumor excision. Biochemical assays combined with the measurement of basal body temperature enable precise determination of menstrual phase. The time of ovulation can now be reliably and simply estimated by detection of the midcycle luteinizing-hormone surge with a 5-minute enzyme immunoassay used primarily to identify a woman's fertile period (31, 32). Because breast cancer cannot currently be prevented, the search continues for ways to improve survival. Any prognostic differences related to time of surgery have clinically significant implications for patients and their surgeons. We encourage other investigators to study this question further. Additional research should also evaluate changes in immunologic function and other factors associated with cyclical hormonal patterns at time of surgery for malignant neoplasms in other organ systems. Addendum The study by Badwe and colleagues (33) at Guy's Hospital, London, of 249 patients with operable breast cancer, reported after the preparation of this article and before its publication, confirms our findings. Superior recurrence-free and overall survival was observed
References 1. Hrushesky WJ, Bluming AZ, Gruber SA, Sothern RB. Menstrual influence on surgical cure of breast cancer. Lancet. 1989;2:949-52. 2. Powles TJ, Jones AL, Ashley S, Tidy A. Menstrual effect on surgical cure of breast cancer [Letter]. Lancet 1989;2:1343-4. 3. Gelber RD, Goldhirsch A. Menstrual effect on surgical cure of breast cancer [Letter]. Lancet 1989 ;2:1344. 4. Ville Y, Lasry S, Spyratos F, Hacene F, Brunet M. Menstrual status and breast cancer surgery [Letter]. Breast Cancer Res Treat. 1990; 16:119-21. 5. Vollman RF. The Menstrual Cycle. Philadelphia: W.B. Saunders; 1977. 6. Senie RT, Rosen PP, Lesser ML, Snyder RE, Schottenfeld D, Duthie K. Epidemiology of breast carcinoma II: factors related to the predominance of left-sided disease. Cancer. 1980;46:1705-13. 7. Rosen PP, Lesser ML, Senie RT, Kinne DW. Epidemiology of breast carcinoma III: relationship of family history to tumor type. Cancer. 1982;50:171-9. 8. Senie RT, Rosen PP, Lesser ML, Kinne DW. Breast self-examination and medical examination related to breast cancer stage. Am J Public Health. 1981;71:583-90. 9. Lesser ML, Rosen PP, Kinne DW. Multicentricity and bilaterality in invasive breast carcinoma. Surgery. 1982;91:234-40. 10. Senie RT, Lobenthal SW, Rosen PP. Association of vaginal smear cytology with menstrual status in breast cancer. Breast Cancer Res Treat. 1985;5:301-10. 11. Kalbfleisch JD, Prentice RL. The Statistical Analysis of Failure Time Data. New York: John Wiley; 1980. 12. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. Journal of the American Statistical Association. 1958; 53:457-81. 13. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep. 1966; 50:163-70. 14. Cox DR. Regression models and life-tables. Journal of the Royal Statistical Society [B]. 1972;34:187-220. 15. Ratajczak HV, Sothern RB, Hrushesky WJ. Estrous influence on surgical cure of a mouse breast cancer. J Exp Med. 1988;168:73-83. 16. Regulation of the menstrual cycle. In: Speroff L, Glass RH, Kase NG; eds. Clinical Gynecologic Endocrinology and Infertility. Baltimore: Williams & Wilkins; 1989:3-119. 17. Vignon F, Bardon S, Chalbos D, Rochefort H. Antiestrogenic effect of R5020, a synthetic progestin in human breast cancer cells in culture. J Clin Endocrinol Metab. 1983;56:1124-30. 18. Beatson GT. On the treatment of inoperable cases of carcinoma of the mamma: suggestions for a new method of treatment with illustrative cases. Lancet. 1896;2:104-7. 19. Axelrod DM, Menendez-Botet CJ, Kinne DW, Osborne MP. Levels of estrogen and progesterone receptor proteins in patients with breast cancer during various phases of the menses. Cancer Invest. 1988;6:7-14. 20. Heise E, Gorlich M. Estradiol receptor in human breast cancers throughout the menstrual cycle. Oncology. 1982;99:340-4. 21. Anderson TJ, Battersby S, King RJ, McPherson K, Going J J . Oral contraceptive use influences resting breast proliferation. Hum Pathol. 1989;20:1139-44. 22. Vogel PM, Georgiade NG, Fetter BF, Voge FS, McCarty KS J r . The
1 S e p t e m b e r 1991 • Annals
of Internal
Medicine
• V o l u m e 115 • N u m b e r 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 06/08/2017
341
23.
24. 25. 26. 27. 28.
correlation of histologic changes in the human breast with the menstrual cycle. Am J Pathol. 1981;104:23-34. Hrushesky WJ, Gruber SA, Sothern RB, Hoffman RA, Lakatua D, Carlson A, et al. Natural killer cell activity: age, estrous- and circadian-stage dependence and inverse correlation with metastatic potential. J Natl Cancer Inst. 1988;80:1232-7. White D, Jones DB, Cooke T, Kirkham N. Natural killer (NK) activity in peripheral blood lymphocytes of patients with benign and malignant breast disease. Br J Cancer. 1982;46:611-6. Sulke AN, Jones DB, Wood PJ. Variation in natural killer activity in peripheral blood during the menstrual cycle. Br J Med [Clin Res]. 1985;290:884-6. Stratum JA, Miller RD, Kent DR, Thrupp LD, Richards C, DiSaila PJ. Depressed mononuclear cell phagocytic activity associated with menstruation. J Clin Lab Immunol. 1984;15:127-31. Aksel S. Hormonal characteristics of long cycles in fertile women. Fertil Steril. 1981;36:521-3. Chiazze L Jr, Brayer FT, Macisco JJ Jr, Parker MP, Duffy BJ. The length and variability of the human menstrual cycle. JAMA. 1968; 203:377-80.
342
29. Metcalf MG. Incidence of ovulation from the menarche to the menopause: observations of 622 New Zealand women. N Z Med J. 1983;96:645-8. 30. Treloar AE, Boynton RE, Behn BG, Brown BW. Variation of the human menstrual cycle through reproductive life. Int J Fertil. 1967; 12:77-126. 31. Brown JB, Blackwell LF, Holmes J, Smyth K. New assays for identifying the fertile period. Suppl Int J Gynecol Obstet. 1989;1: 111-22. 32. Bieglmayer C, Fischl F, Janisch H. Evaluation of a simple and fast self-test for urine luteinizing hormone. Fertil Steril. 1990;55:842-6. 33. Badwe RA, Gregory WM, Chaudary MA, Richards MA, Bentley AE, Rubens RD, et al. Timing of surgery during the menstrual cycle and survival of premenopausal women with operable breast cancer. Lancet. 1991;337:1261-4. 34. Senie RT, Rosen PP, Rhodes P, Lesser ML. Prognosis of primary breast cancer patients in relation to time of diagnostic surgery during the menstrual cycle. Breast Cancer Res Treat [Abstract]. 1990;16:146.
1 September 1991 • Annals of Internal Medicine • Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 06/08/2017
Number 5
1 September 1991
Annals of Internal Medicine ARTICLES
Timing of Breast Cancer Excision during the Menstrual Cycle Influences Duration of Disease-free Survival Ruby T. Senie, PhD; Paul Peter Rosen, MD; Philip Rhodes, MS; and Martin L. Lesser, PhD
• Objective: To study disease-free survival at 10 years in relation to timing of breast tumor excision during the menstrual cycle. • Design: A prospective study of consecutively treated patients with primary breast cancer. • Setting: Memorial Sloan-Kettering Cancer Center, New York. • Patients: Two hundred and eighty-three premenopausal patients treated by mastectomy and axillary dissection. • Main Results: When the tumor was excised during the follicular phase, approximated by setting the putative day of ovulation on day 14 after the onset of last menses, a higher recurrence risk (43%) was observed compared with excision later in the menstrual cycle (29%, P = 0.02). The rate peaked among patients treated between days 7 and 14 and was lowest between days 20 and 30. Multivariate analysis using the Cox regression model to control for tumor size, nodal status, estrogen receptor status, adjuvant chemotherapy, and family history indicated that the hazard rate of breast cancer recurrence after excision during the follicular phase was 1.53 (95% CI, 1.02 to 2.29). Stratification by nodal status indicated that the effect of phase was statistically significant only among patients with positive nodes (hazard ratio, 2.10; CI, 1.19 to 3.70). • Conclusions: Our results support the hypothesis that the risk for recurrence may be affected by the hormonal milieu of the menstrual cycle; these findings must be confirmed, however, by a prospective study in which cycle phase at time of tumor excision is biochemically documented.
Annals of Internal Medicine. 1991;115:337-342. From the Centers for Disease Control, Atlanta, Georgia; Memorial Sloan-Kettering Cancer Center, New York, New York; and North Shore University Hospital-Cornell University Medical College, Manhasset, New York. For current author addresses, see end of text.
I n a study of 41 patients with breast cancer, Hrushesky and colleagues (1) observed significantly better diseasefree and total survival after tumor resection between days 7 and 20 of the menstrual cycle compared with days closer to onset of menses. However, a significant association between prognosis and timing of tumor resection was not confirmed in three subsequent reports of somewhat larger patient samples (2-4). To address this question further, we evaluated two schemes for subdividing the menstrual cycle: the intervals defined by Hrushesky and colleagues and the hormone-dependent phases determined by the putative time of ovulation, the fourteenth day after the last menstrual period, approximating the division to follicular and luteal phases of the cycle (5). Table 1. Demographic and Pathologic Characteristics of 283 Premenopausal Patients Characteristic
Number (%)
Age at diagnosis, y 46 Family history of breast carcinoma None Primary only: mother, sister, daughter Secondary only: other relatives Both primary and secondary relative Parity Nulliparous Parous Tumor size Microscopic < 2 cm 2.1 to 3.9 cm > 4 cm Axillary lymph node status Negative Positive Estrogen receptor status* Positive Negative Borderline Contralateral breast disease Benign Concurrent bilateral Subsequent bilateral No biopsy done [
33 (12) 128 (45) 122 (43) 186 (66) 30(11) 50 (18) 17(6) 49 (17) 234 (83) 34 (12) 121 (43) 79 (28) 49 (17) 166 (59) 117(41) 79 (37) 88 (41) 47 (22) 142 (50) 37 (13) 9(3) 95 (34)
Analysis was based on data available for 214 patients.
1 September 1991 • Annals of Internal Medicine • Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 06/08/2017
337
Methods
Data Analysis
The premenopausal patients included in our study were a subset of 1254 consecutive patients with breast carcinoma treated at Memorial Sloan-Kettering Cancer Center between October 1976 and June 1978 (6, 7) who were recruited for the study of risk factors associated with histologic type of tumor. Epidemiologic information was obtained through personal interviews conducted by one of the investigators with 1216 (97%) patients during their postoperative hospitalization. Tumor size, status of the axillary lymph nodes, and other prognostic factors were assessed by another investigator, who reviewed the histologic slides (8-10). The cohort was followed closely for 10 years through annual contact with the patients and their physicians. Recurrence of breast cancer was confirmed by clinical or pathologic assessment, or both. Disease-free months were calculated from date of diagnosis at time of tumor excision to date of first recurrence, date of death due to other causes, or date of last contact for patients free of recurrent disease. The interval between onset of last menstrual period and tumor excision was used in the analysis for the 27% who had two separate surgical procedures, tumor excision before mastectomy and axillary dissection. Survival analyses have been done on 928 of the 1216 patients after the following exclusions: primary treatment other than mastectomy with axillary dissection (45 patients, 4%), indeterminate size of the primary tumor (91 patients, 7%), distant metastases at diagnosis (63 patients, 5%), previous treatment for contralateral breast carcinoma (82 patients, 7%), and lack of follow-up (7 patients, < 1%). Interview data and medical records were reviewed to determine date of last menstrual flow for all patients with regular cycles classified as premenopausal (10). Of the 315 identified patients, 14 were considered inappropriate for inclusion: 2 were less than 6 months postpartum, and 12 reported hormone therapy during the year before diagnosis. The date of the last menstrual period was not available for 18 (6%) patients, resulting in complete data available for 283 (94%) of 301 patients considered appropriate for inclusion. Two schemes were used to stratify the study sample at the time of diagnostic surgery. In the first analyses, patients were grouped as described by Hrushesky and colleagues: midcycle (days 7 through 20) or perimenstrual (days 0 through 6 and days 21 through 40) (1). The second stratification method relied on the hormone-dependent phases determined by the putative time of ovulation 14 days after last menstrual period, approximating the end of the follicular phase. The luteal phase was assumed to commence with the fifteenth day.
Statistical procedures for univariate analyses included the chi-square test applied to categorical data, the Student /-test for continuous variables, and nonparametric analyses for variables not normally distributed. Statistical techniques designed for censored data were used to evaluate the duration of disease-free survival (11). Product-limit survival curves were calculated by the method of Kaplan and Meier (12). The equality of the distributions was tested by the Mantel-Cox procedure (13). The results of analyses performed on all 283 patients were not altered when the study sample was restricted to 249 cases with invasive breast cancer. Cox proportional hazards modeling was used to estimate the relative rate of recurrence associated with timing of diagnostic surgery while simultaneously adjusting for the effects of tumor size, number of positive axillary lymph nodes, estrogen receptor status, and the presence or absence of a family history of breast cancer (14). We used the programs of BMDP statistical package (BMDP Statistical Software, Los Angeles, California) to do all survival analyses. We also calculated and plotted running averages per 1000 person-months of disease-free survival to produce smoothed rates of recurrence by the day of the menstrual cycle at the time of diagnostic surgery. Confidence intervals (CIs) of 95% are given where appropriate. Results Table 1 presents demographic and pathologic characteristics of the study sample. The age at diagnosis ranged from 24 to 55 years (mean, 43.4 years ± 6.0 [SD] years). The mean number of days from the onset of last menses to surgery was 15 ± 9.7 days. The tumor was excised within 28 days of the last menstrual period for 90% of the women. A carcinoma of microscopic size (not grossly measurable) was identified in 34 (12%) patients. The mean size of the 249 measurable tumors was 2.8 ± 1.9 cm, and 41% of the patients had axillary lymph node metastases. Histologic evaluation showed that 62% of the tumors were infiltrating duct; 12%, medullary carcinoma; 9%, infiltrating lobular; and 17%, other histologic types. Of the 214 (76%) tumors analyzed for estrogen-receptor status, 37% were positive; 4 1 % , negative; and 22%, borderline. Ten years after diagnosis, 177 (63%) patients were alive and free of recurrent disease, 8 (3%) had died of other causes without recurrent breast carcinoma, 74 (26%) had died of breast carcinoma after developing metastases, and 24 (8%) were alive with recurrent carcinoma. Midcycle Compared with Perimenstrual Analysis
Figure 1. Disease-free survival according to menstrual interval at time of tumor excision. The number of patients at risk at time 0, at 5 years, and at 10 years is shown for each set of survival curves. 338
No statistically significant differences were noted when the demographic and pathologic characteristics presented in Table 1 were assessed by menstrual interval, although a greater percentage of diagnostic surgery was done during the perimenstrual interval (54%) than during midcycle (46%). Ten years after diagnosis, the risk for recurrence did not differ significantly; 40% developed recurrence after surgery during midcycle compared with 32% of the perimenstrual subset (P = 0.17). Although Figure 1 indicates a higher recurrence rate for women with tumor excision during midcycle, the differences in survival were not statistically significant. Multivariate analysis
1 September 1991 • Annals of Internal Medicine • Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 06/08/2017
Figure 2. Disease-free survival. Left. Disease-free survival according to menstrual phase at the time the tumor was excised. The number of patients atriskat time 0, at 5 years, and at 10 years is shown for each set of survival curves. Hazard ratio, 1.59 (95% CI, 1.06 to 2.38). Right. Disease-free survival according to menstrual phase and axillary lymph node status. Negative nodes: hazard ratio, 1.07 (CI, 0.58 to 1.99); P > 0.2. Positive nodes: hazard ratio, 1.97 (CI, 1.14 to 3.41); P = 0.02. controlling for tumor size, number of positive nodes, and family history indicated that survival did not differ by timing of tumor excision during midcycle compared with the perimenstrual interval. Follicular Compared with Luteal Analysis When the second scheme of subdivision was used, the patients were equally divided by follicular and luteal phases. Women in the luteal subset were slightly older at diagnosis (mean age, 44.1 years compared with 42.7 years) and reported a family history of breast cancer more frequently (41% compared with 28%; P = 0.05). More women who had tumor excision during the follicular phase were found to have axillary lymph node metastases compared with those diagnosed during the luteal phase (46% compared with 37%); however, among the 117 women with positive nodes, the mean number of involved nodes did not differ significantly by phase. Seventy-eight patients received combination adjuvant chemotherapy (cytoxan, methotrexate, and fluorouracil). The proportion of patients with positive nodes treated with adjuvant chemotherapy did not differ significantly by phase (68% of follicular and 54% of luteal); three negative-node patients in each subset also received chemotherapy. Among those treated, the number of months of therapy did not differ by phase. Estrogen-receptor status, assessed in 214 patients, was similar for the two subsets: 35% of patients in follicular and 39% of those in luteal phase had a positive result; 44% of patients in follicular and 39% of those in luteal phase had a negative result; the remaining were classified as borderline. No significant differences in specific binding by phase were observed using the Mann-Whitney nonparametric test or by r-test analysis after log transformation. Other pathologic features, including tumor size, histologic type, nuclear grade, and bilateral disease, were similar for the two subsets.
The estimated 10-year risk for recurrence associated with tumor excision during the first 14 days after onset of last menstrual flow (compared with later in the cycle) was significantly higher (43% compared with 29%, P = 0.022). Figure 2, left, indicates statistically significant differences in time to first recurrence by phase (P = 0.023). The hazard ratio (1.59; CI, 1.06 to 2.38) for surgery during the follicular phase uncontrolled for other prognostic factors is shown in Table 2. Because the presence of positive nodes differed by phase, the patients were subdivided by both menstrual phase and nodal status. As noted in Figure 2, right, the effect of phase was limited to patients with positive nodes. The hazard ratio of surgery during the follicular phase among patients with positive nodes was 1.97 (CI, 1.14 to 3.41) but was not statistically significant among patients with negative nodes (hazard ratio, 1.07; CI, 0.58 to 1.99). The plot of smoothed rates of recurrence per 1000 person-months of disease-free survival (Figure 3) indicates highest rates of recurrence when the tumor was excised between days 7 and 14 and lowest rates when the tumor was excised between days 20 and 30. The plot shows the likelihood of recurrence in relation to day of the menstrual cycle without imposing predetermined intervals and without controlling for other potential prognostic factors. Cox regression models were used to assess the effect of phase for the total sample and were subdivided by nodal status. Estimates of the hazard ratios, 95% CIs, and their significance levels are shown in Table 2. After controlling for tumor size, number of positive axillary nodes, adjuvant chemotherapy, and family history, the menstrual phase at time of tumor excision remained significantly associated with disease-free survival (hazard ratio, 1.53; CI, 1.02 to 2.29). When the effect of phase was compared by nodal status, a sizable although
1 September 1991 • Annals of Internal Medicine • Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 06/08/2017
339
not statistically significant difference was observed; the hazard ratio for patients with positive nodes was 2.10 (CI, 1.19 to 3.70) and for negative-node patients, 0.98 (CI, 0.52 to 1.83). The inclusion of estrogen-receptor status in the model restricted the study population to 214 cases; however, results were not significantly altered (hazard ratio, 1.68; CI, 1.07 to 2.63). Discussion Studies of mice indicated a lower rate of subsequent pulmonary metastases when mammary tumors were excised during the estrus cycle, close to the time of ovulation (15). A similar association of survival differences in relation to timing of tumor excision was observed among 41 premenopausal patients. Disease-free and total survival were greater after surgery in midcycle (around ovulation) than during the perimenstrual interval (1). Although their findings were statistically significant, the authors cautioned that their results were based on a small sample and needed to be confirmed by others. In three subsequent reports of larger samples in which the same menstrual cycle divisions were applied, time of diagnostic surgery was not associated with statistically significant differences in disease-free or overall survival (2-4). Comparisons across studies were complicated by differences in stage of disease, treatment protocols, and duration and completeness of follow-up as well as rates of survival. When we assessed time to first recurrence by the menstrual categories, perimenstrual and midcycle, we obtained results similar to those of others (2-4), which were contrary to the findings of Hrushesky and colleagues (1). Because our study group was significantly larger with 98 recurrences, our study had sufficient power to observe effects even smaller than those suggested by Hrushesky and colleagues. Indeed, our results indicate that the effect was in the opposite direction to results previously reported (1). After categorizing our study sample in relation to the putative timing of ovulation, approximating the follicular and luteal phases, however, statistically significant differences in duration of disease-free survival were apparent. Our cohort of 283 patients was treated at one institution with a standard surgical protocol during an 18-month period and was carefully followed for 10 years. Among the 98 (35%) women who had relapses, the rate was 50% higher, especially among women with positive ax-
illary nodes, when surgery had been done within 14 days of last menstrual period. The menstrual cycle is characterized by an estrogen peak in the absence of progesterone before ovulation with a second estrogen surge later in the cycle coinciding with a rapid secretion of progesterone (16). Our results revealed an elevated hazard ratio among women whose tumor was excised during the follicular phase, in the presence of unopposed estrogen that potentially enhanced micrometastases. The significantly better prognosis among women with tumor excision during the luteal phase, when progesterone levels were highest, may be related to the findings of in-vitro studies that revealed an inhibition of human breast carcinoma cell proliferation by synthetic progestin (17). In 1896, Beatson (18) reported the influence of endogenous hormones on breast cancer survival; the precise mechanisms of action, however, are still not fully understood. In our series and in others (19, 20), estrogenreceptor status of the tumor did not vary with menstrual phase at time of tumor excision. However, cyclical patterns of cell division and deletion of normal breast tissue have been found to correspond to endogenous hormonal surges of the natural menstrual cycle (21, 22). Natural killer-cell activity has been found to vary in relation to the estrus cycle of tumor-free mice (23). Studies of healthy women have noted cyclical reductions in natural killer-cell activity before ovulation with a return to higher levels later in the menstrual cycle (24, 25). Other investigators (26) have reported a significant decrease in phagocytic activity of mononuclear cells early in the menstrual cycle. Therefore, diminished immune function before the putative day of ovulation may be one mechanism associated with the observed pattern of recurrence. One limitation of our study and others (1-4) is the reliance on date of last menstrual period to estimate the menstrual phase at time of tumor excision without precise determination of the hormonal milieu. Because we set the putative time of ovulation at the fourteenth day after the onset of the last menstrual flow and because the usual cycle length had not been recorded, we may have miscoded some women in the midrange of days between onset of last menses and tumor excision (27, 28). Biochemical data were not available to determine if ovulation had occurred on or after the fifteenth day after the last menstrual cycle and before tumor excision for the 142 patients classified to the luteal subset; at the
Table 2. Effect of Follicular Phase on Disease-free Survival at 10 Years in 283 Premenopausal Patients (Using Cox Regression Models) Patients and Nodal Status All women Nodal statust Negative Positive
Controlled* (95% CI)
Hazard Ratio
Crude (95% CI)
1.59
(1.06 to 2.38)
0.025
1.53
(1.02 to 2.29)
0.04
1.07 1.97
(0.58 to 1.99) (1.14 to 3.41)
>0.2 0.016
0.98 2.10
(0.52 to 1.83) (1.19 to 3.70)
>0.2 0.01
P Value
Hazard Ratio
P Value
* Controlling for tumor size, number of positive nodes, chemotherapy, and family history. t P values comparing the effects of follicular phase in 117 women having positive nodes with 166 women having negative nodes (crude P = 0.15; controlled P = 0.08).
340
1 September 1991 • Annals of Internal Medicine • Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 06/08/2017
among 83 positive-node patients who had tumors excised during mensturation (days 0 to 2 after onset of menses) or later in the cycle (days 13 to 32) compared with the 43 positive-node patients who were treated between days 3 and 12 after onset of last menses. As in our study, their results were not influenced by other prognostic factors, including estrogen-receptor status. Presented in part (34) at the 13th Annual San Antonio Breast Cancer Symposium on 2 November 1990, San Antonio, Texas. Acknowledgments: The authors thank the women for their willingness to be interviewed during a stressful period of hospitalization. They also thank their surgical colleagues for their assistance during the inception of this study and throughout the 10 years of follow-up. Requests for Reprints: Ruby T. Senie, PhD, Centers for Disease Control, 1600 Clifton Road, Atlanta, GA 30333. Current Author Addresses: Dr. Senie and Mr. Rhodes: Centers for Disease Control, 1600 Clifton Road, Atlanta, GA 30333. Dr. Rosen: Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10021. Dr. Lesser: North Shore University Hospital, 300 Community Drive, Manhasset, NY 11030.
Figure 3. Smoothed recurrence rates. Recurrence of breast carcinoma per 1000 months of disease-free survival according to the day of the menstrual cycle at time of tumor excision.
time of interview, however, all patients reported regular cycles. Although the proportion of nonovulatory cycles has been observed to increase with age (5), studies indicate that women with regular menses ovulate in more than 95% of cycles, even when they are close to menopause (29). Variations in the duration and regularity of the menstrual cycle have been documented between women and during the menstrual years of individual women (28, 30). A prospective study is therefore required to confirm our findings of a relation between disease-free survival and phase of the menstrual cycle at time of tumor excision. Biochemical assays combined with the measurement of basal body temperature enable precise determination of menstrual phase. The time of ovulation can now be reliably and simply estimated by detection of the midcycle luteinizing-hormone surge with a 5-minute enzyme immunoassay used primarily to identify a woman's fertile period (31, 32). Because breast cancer cannot currently be prevented, the search continues for ways to improve survival. Any prognostic differences related to time of surgery have clinically significant implications for patients and their surgeons. We encourage other investigators to study this question further. Additional research should also evaluate changes in immunologic function and other factors associated with cyclical hormonal patterns at time of surgery for malignant neoplasms in other organ systems. Addendum The study by Badwe and colleagues (33) at Guy's Hospital, London, of 249 patients with operable breast cancer, reported after the preparation of this article and before its publication, confirms our findings. Superior recurrence-free and overall survival was observed
References 1. Hrushesky WJ, Bluming AZ, Gruber SA, Sothern RB. Menstrual influence on surgical cure of breast cancer. Lancet. 1989;2:949-52. 2. Powles TJ, Jones AL, Ashley S, Tidy A. Menstrual effect on surgical cure of breast cancer [Letter]. Lancet 1989;2:1343-4. 3. Gelber RD, Goldhirsch A. Menstrual effect on surgical cure of breast cancer [Letter]. Lancet 1989 ;2:1344. 4. Ville Y, Lasry S, Spyratos F, Hacene F, Brunet M. Menstrual status and breast cancer surgery [Letter]. Breast Cancer Res Treat. 1990; 16:119-21. 5. Vollman RF. The Menstrual Cycle. Philadelphia: W.B. Saunders; 1977. 6. Senie RT, Rosen PP, Lesser ML, Snyder RE, Schottenfeld D, Duthie K. Epidemiology of breast carcinoma II: factors related to the predominance of left-sided disease. Cancer. 1980;46:1705-13. 7. Rosen PP, Lesser ML, Senie RT, Kinne DW. Epidemiology of breast carcinoma III: relationship of family history to tumor type. Cancer. 1982;50:171-9. 8. Senie RT, Rosen PP, Lesser ML, Kinne DW. Breast self-examination and medical examination related to breast cancer stage. Am J Public Health. 1981;71:583-90. 9. Lesser ML, Rosen PP, Kinne DW. Multicentricity and bilaterality in invasive breast carcinoma. Surgery. 1982;91:234-40. 10. Senie RT, Lobenthal SW, Rosen PP. Association of vaginal smear cytology with menstrual status in breast cancer. Breast Cancer Res Treat. 1985;5:301-10. 11. Kalbfleisch JD, Prentice RL. The Statistical Analysis of Failure Time Data. New York: John Wiley; 1980. 12. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. Journal of the American Statistical Association. 1958; 53:457-81. 13. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep. 1966; 50:163-70. 14. Cox DR. Regression models and life-tables. Journal of the Royal Statistical Society [B]. 1972;34:187-220. 15. Ratajczak HV, Sothern RB, Hrushesky WJ. Estrous influence on surgical cure of a mouse breast cancer. J Exp Med. 1988;168:73-83. 16. Regulation of the menstrual cycle. In: Speroff L, Glass RH, Kase NG; eds. Clinical Gynecologic Endocrinology and Infertility. Baltimore: Williams & Wilkins; 1989:3-119. 17. Vignon F, Bardon S, Chalbos D, Rochefort H. Antiestrogenic effect of R5020, a synthetic progestin in human breast cancer cells in culture. J Clin Endocrinol Metab. 1983;56:1124-30. 18. Beatson GT. On the treatment of inoperable cases of carcinoma of the mamma: suggestions for a new method of treatment with illustrative cases. Lancet. 1896;2:104-7. 19. Axelrod DM, Menendez-Botet CJ, Kinne DW, Osborne MP. Levels of estrogen and progesterone receptor proteins in patients with breast cancer during various phases of the menses. Cancer Invest. 1988;6:7-14. 20. Heise E, Gorlich M. Estradiol receptor in human breast cancers throughout the menstrual cycle. Oncology. 1982;99:340-4. 21. Anderson TJ, Battersby S, King RJ, McPherson K, Going J J . Oral contraceptive use influences resting breast proliferation. Hum Pathol. 1989;20:1139-44. 22. Vogel PM, Georgiade NG, Fetter BF, Voge FS, McCarty KS J r . The
1 S e p t e m b e r 1991 • Annals
of Internal
Medicine
• V o l u m e 115 • N u m b e r 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 06/08/2017
341
23.
24. 25. 26. 27. 28.
correlation of histologic changes in the human breast with the menstrual cycle. Am J Pathol. 1981;104:23-34. Hrushesky WJ, Gruber SA, Sothern RB, Hoffman RA, Lakatua D, Carlson A, et al. Natural killer cell activity: age, estrous- and circadian-stage dependence and inverse correlation with metastatic potential. J Natl Cancer Inst. 1988;80:1232-7. White D, Jones DB, Cooke T, Kirkham N. Natural killer (NK) activity in peripheral blood lymphocytes of patients with benign and malignant breast disease. Br J Cancer. 1982;46:611-6. Sulke AN, Jones DB, Wood PJ. Variation in natural killer activity in peripheral blood during the menstrual cycle. Br J Med [Clin Res]. 1985;290:884-6. Stratum JA, Miller RD, Kent DR, Thrupp LD, Richards C, DiSaila PJ. Depressed mononuclear cell phagocytic activity associated with menstruation. J Clin Lab Immunol. 1984;15:127-31. Aksel S. Hormonal characteristics of long cycles in fertile women. Fertil Steril. 1981;36:521-3. Chiazze L Jr, Brayer FT, Macisco JJ Jr, Parker MP, Duffy BJ. The length and variability of the human menstrual cycle. JAMA. 1968; 203:377-80.
342
29. Metcalf MG. Incidence of ovulation from the menarche to the menopause: observations of 622 New Zealand women. N Z Med J. 1983;96:645-8. 30. Treloar AE, Boynton RE, Behn BG, Brown BW. Variation of the human menstrual cycle through reproductive life. Int J Fertil. 1967; 12:77-126. 31. Brown JB, Blackwell LF, Holmes J, Smyth K. New assays for identifying the fertile period. Suppl Int J Gynecol Obstet. 1989;1: 111-22. 32. Bieglmayer C, Fischl F, Janisch H. Evaluation of a simple and fast self-test for urine luteinizing hormone. Fertil Steril. 1990;55:842-6. 33. Badwe RA, Gregory WM, Chaudary MA, Richards MA, Bentley AE, Rubens RD, et al. Timing of surgery during the menstrual cycle and survival of premenopausal women with operable breast cancer. Lancet. 1991;337:1261-4. 34. Senie RT, Rosen PP, Rhodes P, Lesser ML. Prognosis of primary breast cancer patients in relation to time of diagnostic surgery during the menstrual cycle. Breast Cancer Res Treat [Abstract]. 1990;16:146.
1 September 1991 • Annals of Internal Medicine • Volume 115 • Number 5
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19733/ by a University of California San Diego User on 06/08/2017
