Current issues of ACP Journal Club are published in Annals of Internal Medicine


Psychological stress was not associated with relapse of breast cancer

ACP J Club. 1992 Nov-Dec;117:87. doi:10.7326/ACPJC-1992-117-3-087

Source Citation

Barraclough J, Pinder P, Cruddas M, et al. Life events and breast cancer prognosis. BMJ. 1992 Apr 25; 304:1078-81.



To determine whether psychosocial stress is associated with the relapse of breast cancer after surgery.


A cohort of women with newly diagnosed breast cancer was followed for 42 months.


Home interviews of women identified through the surgical records of 2 university hospitals in England between February 1986 and August 1987.


Selection criteria were age ≤ 70 years and a diagnosis of breast cancer without evidence of spread beyond the axilla. Patients had either mastectomy or local excision and radiotherapy. 204 of 246 consecutive eligible patients (83%) of the participating surgeons entered the study; 3 patients dropped out.

Assessment of prognostic factors

Sociodemographic and clinical information (including menopausal status, tumor size, and axillary node dissemination) was obtained from records. During interviews done 4, 24, and 42 months after surgery, life events (e.g., bereavements), social difficulties (e.g., unemployment), and health during the previous 18 months were assessed using the Bedford College life events and difficulties schedule. Major depression was defined by DSM-III criteria.

Main outcome measure

Local recurrence or distant metastasis of cancer confirmed histologically or radiologically.

Main results

47 patients (23%) had relapse of breast cancer; 26 of these women died (13%). Relapse of breast cancer was not associated with major depression before or after surgery (relative risks [RRs] 1.26, 95% CI 0.49 to 3.26 before surgery and 0.85, CI 0.41 to 1.79 after surgery respectively) or with having some period without a confidant (RR, 0.93, CI 0.42 to 2.09 before surgery and 0.86, CI 0.38 to 1.93 after surgery respectively). Of 103 women who reported any severe event or difficulty other than with their own physical health after surgery, the RR for relapse, adjusted for age and disease severity, was 0.88 (CI 0.48 to 1.64); of 54 women with any severe event or difficulty before surgery, the RR was 0.43 (CI 0.20 to 0.93). In contrast with psychosocial stress, severe events or difficulties related to the patient's health were associated with relapse if they occurred after surgery (RR 3.06, CI 1.29 to 7.31). Axillary node involvement was associated with relapse (RR 4.5, CI 2.41 to 8.42) as was tumor size > 4 cm in diameter (RR 2.23, CI 1.00 to 4.99).


Psychosocial stress was not associated with relapse of breast cancer in the 3.5 years after surgical intervention.

Sources of funding: Medical Research Council; Wessex Medical School Trust; Marks and Spencer plc.

For article reprint: Dr. J. Barraclough, Sir Michael Sobell House, Churchill Hospital, Oxford OX3 7LJ, United Kingdom.


The relation between psychologic, social, and behavioral factors and cancer risk, recurrence, and survival is intriguing. Environmental exposure, socioeconomic considerations, the physiologic consequences of stress, and emotional makeup all affect carcinogenesis, perhaps by neurologic, immune, and endocrine mechanisms.

The study by Barraclough and colleagues appears, superficially, to be negative. Although it is conceded that "severe" personal health events and difficulties were associated with increased risk for relapse from breast cancer, psychosocial stresses were not contributory. This conclusion may be premature because the sample size was relatively small and the duration of follow-up was relatively brief.

Cancer mortality is higher where poverty, unemployment, disintegration of family life, and other social stresses occur. Individuals with a "fighting spirit" have a better survival rate compared with those with stoic acceptance of illness, helplessness, and hopelessness. Lippman (1) has postulated that the rate of tumor progression is affected by hormonal influences that may be disturbed by strong emotional states through a psychoneuroendocrine axis. Temoshok (2), exploring the relation between personality and cancer risk, described the "type C" personality, in which there is a tendency to repress emotions. These individuals may become cancer prone.

Although most studies of the effect of stress on disease vulnerability (3) indicate that the effect of psychosocial influences is small, growing awareness of the effects of environmental and behavioral influences on cancer makes this an area for much research interest.

Alan J. Lippman, MD
Newark Beth Israel Medical CenterNewark, New Jersey, USA

Alan J. Lippman, MD
Newark Beth Israel Medical Center
Newark, New Jersey, USA


1. Lippman ME. In: Levy SM; ed. Behavior and Cancer. San Francisco: Jossey-Bass; 1985:134-47.

2. Temoshok L. Cancer Surv. 1987;6: 545-67.

3. Holland J. In: Holland JC, Rowland JH; eds. Handbook of Psychooncology. Oxford: Oxford University Press; 1989: 705-26.