12Breast cancer

General recommendations

1.

A patient who presents with symptoms suggestive of breast cancer should be referred to a team specialising in the management of breast cancer. [D]

2.

In most cases, the definitive diagnosis will not be known at the time of referral, and many patients who are referred will be found not to have cancer. However, primary healthcare professionals should convey optimism about the effectiveness of treatment and survival because a patient being referred with a breast lump will be naturally concerned. [C]

3.

People of all ages who suspect they have breast cancer may have particular information and support needs. The primary healthcare professional should discuss these needs with the patient and respond sensitively to them. [D]

4.

Primary healthcare professionals should encourage all patients, including women over 50 years old, to be breast aware14 in order to minimise delay in the presentation of symptoms. [D]

Specific Recomendations

5.

A woman’s first suspicion that she may have breast cancer is often when she finds a lump in her breast. The primary healthcare professional should examine the lump with the patient’s consent. The features of a lump that should make the primary healthcare professional strongly suspect cancer are a discrete, hard lump with fixation, with or without skin tethering. In patients presenting in this way an urgent referral should be made, irrespective of age. [C]

6.

In a woman aged 30 years and older with a discrete lump that persists after her next period, or presents after menopause, an urgent referral should be made. [C]

7.

Breast cancer in women aged younger than 30 years is rare, but does occur. Benign lumps (for example, fibroadenoma) are common, however, and a policy of referring these women urgently would not be appropriate; instead, non-urgent referral should be considered. However, in women aged younger than 30 years with: [C/D]

  • a lump that enlarges, [C] or
  • a lump that has other features associated with cancer (fixed and hard), [C] or
  • in whom there are other reasons for concern such as family history. [D]

an urgent referral should be made

8.

The patient’s history should always be taken into account. For example, it may be appropriate, in discussion with a specialist, to agree referral within a few days in patients reporting a lump or other symptom that has been present for several months. [D]

9.

In a patient who has previously had histologically confirmed breast cancer, who presents with a further lump or suspicious symptoms, an urgent referral should be made, irrespective of age. [C]

10.

In patients presenting with unilateral eczematous skin or nipple change that does not respond to topical treatment, or with nipple distortion of recent onset, an urgent referral should be made. [C]

11.

In patients presenting with spontaneous unilateral bloody nipple discharge, an urgent referral should be made. [C]

12.

Breast cancer in men is rare and is particularly rare in men under 50 years of age. However, in a man aged 50 years and older with a unilateral, firm subareaolar mass with or without nipple distortion or associated skin changes, an urgent referral should be made. [C]

Investigations

13.

In patients presenting with symptoms and/or signs suggestive of breast cancer, investigation prior to referral is not recommended. [D]

14.

In patients presenting solely with breast pain, with no palpable abnormality, there is no evidence to support the use of mammography as a discriminatory investigation for breast cancer. Therefore, its use in this group of patients is not recommended. Non-urgent referral may be considered in the event of failure of initial treatment and/or unexplained persistent symptoms. [B (DS)]

Introduction

Pathology

Breast carcinoma develops from the epithelial cells within the terminal duct/lobular unit (186). It is categorised as either ‘invasive’ or ‘in situ’. Before malignant cells breach the basement membrane the cancer is ‘in situ’, but once that membrane has been breached the cancer is ‘invasive’.(186). Breast cancers can be classified as either ‘ductal’ or ‘lobular’ on the basis of carcinoma type. The terms ‘ductal carcinoma in situ’ (DCIS) and lobular carcinoma in situ’ (LCIS) are widely used but carry no more relevance than invasive cancer.(186)

Staging breast cancer

Staging is used to classify cancers on their anatomic extent. Tumour staging is based on size and the whether there is fixation of the cancer to surrounding tissue(186). The TNM staging system (Table 33) was developed from work in the 1940s by Pierre Denoix and is now the most widely used system of cancer classification.(187)

Table 33. TNM classification and stage grouping for breast tumours ((186)).

Table 33

TNM classification and stage grouping for breast tumours ((186)).

Incidence

Breast cancer is the most common malignancy in women, accounting for almost 30% of female cancers. A general practitioner can expect to encounter one new case of breast cancer approximately every 11 months. It is estimated that more than 75% of cases present symptomatically and not through screening programmes. In 2001 there 40,740 cases in women.

Figure 15. 2001 Registrations of Malignant Neoplasm of the Breast in England and Wales. (77).

Figure 152001 Registrations of Malignant Neoplasm of the Breast in England and Wales. (77)

Breast cancer in males is rare occurring approximately 100 times less than in women(128). The distribution of incidence by age is shown in Figure 16.

Figure 16. 2001 Registration rates of Malignant Neoplasm of the Breast in Males in England and Wales. (77).

Figure 16

2001 Registration rates of Malignant Neoplasm of the Breast in Males in England and Wales. (77).

Mortality

Despite increased incidence rates, mortality among women from breast malignancies has been falling since 1990 decreasing from approximately 38 to 32 per 100,000 population between 1971 and 1999. In 2002, there were 11,476 deaths among women and 81 among men (Figure 17).

Figure 17. 2002 Mortality rates per 100,000 population from Malignant Neoplasm of the Breast in England and Wales. (78).

Figure 17

2002 Mortality rates per 100,000 population from Malignant Neoplasm of the Breast in England and Wales. (78).

Review of cancer referral audits

The review (CRD, 2004) identified 72 clinical audits. The proportion of two week referrals in accordance with the symptoms listed in the Department of Health (2000) guidelines ranged from 65% to 99% (20 audits). The proportion of patients found to have cancer who had been referred under the two week system ranged from 0% to 34% (37 audits). The proportion of patients referred as ‘urgent’ but not under the two week system ranged from 4% to 20% (five audits). The proportion of patients found to have cancer who had been referred non-urgently ranged from 0% to 10%. Of the patients found to have cancer, between 4% and 83% had been referred under the two week system (nine audits). The proportion of two week referrals considered by the consultant to be appropriate or warrant an urgent appointment ranged from 18% to 96%.

Demographic information

(ONS, 2001) (17)

Breast cancer is the most common cancer in women worldwide, although cervical cancer is more frequent in some developing countries. It accounts for about 30% of all malignancies in women in England and Wales and recorded rates are higher in women in western, developed, countries. Breast cancer in men is extremely rare.

In 1997 there were 33,100 new registrations of breast cancer in women in England and Wales (Table 34), almost 30% of all cancers in women, and more than twice as many as for the second most common site, colorectal cancer. Worldwide, the highest recorded incidence rates occur in the USA and other western, developed countries. Rates in Japan, China and India are only about a quarter of those in the USA.

Table 34. Breast Cancer incidence and mortality, England and Wales, 1997 (ONS, 2001. (17)).

Table 34

Breast Cancer incidence and mortality, England and Wales, 1997 (ONS, 2001. (17)).

Before the introduction of screening, incidence rates rose with age from the late 20s, but slowed at around 45–54 years, the age of the menopause. The effect of breast screening has been to raise the incidence in women aged 50–54, because many women were being screened for the first time with cancers being detected at an earlier stage. Rates in women aged 55–64 also rose during the early years of screening, but have since returned to levels expected based on the earlier trends. Incidence in women aged 65–69 has fallen in recent years: many cancers in these women will have been detected at earlier ages during screening; their rates in 1995–97 were lower than those in women aged 50–64.

As the incidence of breast cancer is high and survival is relatively good compared with many other cancers, there are large numbers of women alive who have been diagnosed with breast cancer. About 81% (75,000) of those diagnosed in 1990–92, and 62% (168,000) of those diagnosed in 1983–92 were still alive at the beginning of 1993.

One-year survival rates for patients in England and Wales diagnosed in 1991–93 was 92%; five-year survival was 74%. Women aged under 40 at diagnosis had worse survival than those aged 40–49. In the late 1980s, mortality in England was not only higher than in most western European countries, it was among the highest in the world. However, survival has improved steadily over time, and in all regions. Five-year survival rose by 14% points between the early 1970s and the late 1980s and by a further 6% for patients diagnosed in 1991–93. The five-year survival from breast cancer in the UK is now 75.9%, (www.cancerresearchuk.org/aboutcancer/statistics/survival). and for screen-detected cancers five-year survival is 94.1% (http://www.cancerscreening.nhs.uk/breastscreen/publications/ba00-01.html).

12.1. Signs and Symptoms

Women

12.1.1. Key Clinical Question

Which symptoms, signs and other features raise a suspicion of cancer in women consulting in primary care and those that make cancer less likely as a diagnosis?

12.1.2. Evidence Question

In women attending primary care services with breast symptoms, which symptoms and signs and other features when compared with the “gold standard” are predictive of a diagnosis of cancer; and which symptoms and signs are not?

12.1.3. Evidence Statements

The incidence of breast cancer in women in England and Wales rises sharply with age and is rare in women aged under 30 (III).

In studies of risk factors associated with a diagnosis of breast cancer, age is the only factor consistently reported in association with breast symptoms and a diagnosis of cancer (III).

Women with breast symptoms commonly consult general practitioners. In one study, the typical general practitioner was consulted by one woman with breast symptoms every two weeks (III).

Among women presenting in general practice with breast problems, the most common presenting features are a lump and/or pain (III).

Women who attend primary care with the following features have an increased likelihood of having breast cancer: Palpable mass Skin or nipple change (III)

The likelihood of having a diagnosis of breast cancer is highest in women who present to primary care with a palpable mass. However, the absence of a palpable mass does not rule out the possibility of cancer (III).

There is little or no research evidence on the characteristics of breast lumps among women presenting in primary care and the likelihood of cancer. Benign lumps are said to be more likely to be smooth and well demarcated, whereas less mobile lumps with poorly defined margins are more likely to be malignant (IV).

Guidelines
(Austoker and Mansel, 2003) (188)

These guidelines quoted Barclay et al (1991) and Cochrane et al (1997). Cochrane et al (1997) reported that of 2332 new patients presenting to a breast clinic, 147 had symptomatic carcinomas. The symptoms and signs reported by the general practitioners in patients referred with carcinoma were:

  • lumps 90% painful lumps 21%
  • nipple discharge 3.4%
  • nipple change 10.2%
  • skin contour change 4.8%
  • any family history 6.1%.

The guidelines recommended urgent referral for patients with a discrete lump in the appropriate age group, or definite signs of cancer such as: ulceration, skin nodule, skin distortion (<3 months). Nipple discharge or pain in the absence of a lump were said to be much less common presentations of breast cancer.

(All Wales Minimum Standards, 2000) (189)

Standard 10 stipulated that there should be a mechanism to provide general practitioners with rapid access to an appropriate specialist, urgent referrals being seen within ten working days of receipt of the referral by the hospital. The Standards did not include guidance on the presenting symptoms or signs.

(Steering Committee on Clinical Practice Guidelines for the Care and Treatment of Breast Cancer, 1998) (190)

This publication is a Canadian evidence-based guideline to assist decisions in excluding or confirming the presence of cancer when a breast lump is detected. The guidelines were based on published evidence supplemented by expert opinion. Articles were identified through a database search using MEDLINE (from 1966) and CANCERLIT (from 1985) to January 1996. A non systematic review of breast cancer literature continued to January 1997. The guidelines made recommendations on how to establish a reliable diagnosis using the minimum of procedures. Evidence graded I–III was used as far as possible, but when experimental evidence was weak or lacking, the opinion of respected authorities (level IV) was employed. The conclusions arising from the review are outlined below.

Most lumps are not caused by cancer, but the possibility of malignancy must always be considered. The first step is to obtain a clinical history and carry out a physical examination. When necessary, this is followed by further diagnostic procedures (mammography, fine needle aspiration [FNA], ultrasonography) and, if uncertainty still remains, by tissue biopsy (core or open surgical). The clinical history should establish how long the lump has been noted, whether any change has been observed and whether there is a history of biopsy or breast cancer. Risk factors for breast cancer should be noted, but the guidelines advised that their presence or absence should not influence the decision to investigate a lump further.

The presence of certain factors increases the likelihood of breast cancer. These include a history of a biopsy of either breast showing atypical hyperplasia, lobular carcinoma in situ (LCIS) or ductal carcinoma in situ (DCIS), a history of resected carcinoma or radiation treatment for Hodgkin’s disease in childhood, or a strong family history of breast cancer (level III evidence). Although known risk factors, including ageing, all increase the risk of breast cancer, they do not substantially influence the probability that any particular lump will be malignant. The fact remains that most women in whom breast cancer is diagnosed have no identifiable risk factors and breast cancer does not develop in most women with common risk factors.

The physical examination of the breast should aim to identify those features that distinguish malignant from benign lumps. Breast examination should be accompanied by a thorough examination of the axilla and supraclavicular areas to check for nodal involvement. Premenopausal women are best examined one week after the onset of the last menstrual period when engorgement of the breast is at a minimum (level IV evidence).

Paget’s like lesions of the nipple are frequently caused by breast cancer. The condition may resemble a benign dermatitis that is sometimes moist and eczematous or sometimes dry and psoriatic and usually accompanied by thickening of the nipple-areolar complex. These features usually reflect centrifugal spread of cancer cells from the ductal epithelium into the overlying skin of the nipple. Biopsy is indicated when the condition fails to respond rapidly to topical treatment.

Smooth, well demarcated lumps are usually benign (level IV evidence). These are either cysts or fibroadenomas. Lesions that are less smooth and less mobile, with poorly defined margins, increase the suspicion of carcinoma.

Nipple discharge is not a common feature of cancer. Persistent unilateral discharge may be due to cancer in 4% to 21% of cases. The discharge may be watery, sanguineous, serosanguineous or serous. A non-bloody discharge is unlikely to be caused by cancer, and even a sangineous discharge is often not due to cancer. Also, a bilateral discharge is unlikely to be caused by cancer.

Breast cancer may or may not be painless. Although breast cancers are usually painless, the cancer may be accompanied by discomfort. Thus, the presence or absence of pain and tenderness should not influence the investigation of a suspicious lump.

(SIGN, 1998)(191)

The SIGN guidelines recommended referral of patients who presented with any new discrete lump, a new lump in pre-existing nodularity, asymmetrical nodularity that persist at review after menstruation, an abscess or breast inflammation which does not settle after one course of antibiotics, or a cyst persistently refilling or recurrent cyst (if the patient has recurrent multiple cysts and the general practitioner has the necessary skills, then aspiration is acceptable). It was also recommended that pain in association with a lump, or that was intractable or unilateral in a post-menopausal women should be an indication for referral, and nipple discharge is also an indication for referral in women over the age of 50 and also under 50 if the discharge is blood stained, persistent single duct or sufficient to stain clothes.

Secondary studies
(Centre for Reviews and Dissemination, 2002)(192)

This Service Guidance Evidence Review did not find any studies of the effectiveness of routine physical breast examination in self-presenting well women in the primary care setting. The review identified two large randomised controlled trials, a non-randomised trial, two cohort studies and three case control studies but no reliable evidence to suggest that breast self-examination (BSE) among asymptomatic women reduces mortality rates from breast cancer. In fact some evidence suggested that BSE can do harm through increased rates of biopsy for benign lesions (grade of evidence I [systematic review of randomised controlled trials] and III [systematic review of non-randomised controlled trials]).

(Levine et al, 2001) (193)

In this systematic review undertaken by the US Agency for Healthcare Research and Quality, studies published from 1994 to 1999 were searched using Medline and Current Contents databases. The review included observational studies, randomised and non-randomised trials, and uncontrolled case series. The first question addressed in the review was ‘What are the recommendations for evaluation of breast symptoms, mammographic findings and other suspicious findings based on menstrual status, use of hormone replacement therapy (HRT), pregnancy, age, and family history?’

Information about the association of symptoms and signs and a diagnosis of breast cancer could only be drawn from those studies that reported individual rather than aggregated data. Patients who presented with palpable masses were much more likely to be diagnosed with cancer than those with non-palpable masses, nipple discharge or breast pain. Ten studies reported the number of patients with palpable masses who developed cancer. Of a total of 2027 patients with masses, 303 (14.9%) had cancer. Six studies reported patients with ‘lesions’ as clinical findings; of 1094 with lesions, 358 (32.7%) were cancer. Four studies reported on nipple discharge, and among the total of 570 patients with discharge, 18 (3.2%) had cancer. Only two studies reported the incidence of cancer in association with breast pain, the proportions being seven of 216 (3.2%) in one study, and four of 221 (1.8%) in another. However, it should be noted that the reviewed studies included samples of women after referral.

Primary studies

There were few studies of the symptoms and signs associated with breast cancer among women presenting to primary care. Most studies involved only a small number of practices and patients, and consequently the numbers of women with cancer were usually too few to draw any meaningful conclusions about the predictive value of symptoms and signs in primary care. Since general practitioners encounter around one new patient with breast cancer per year, studies of presentation in primary care would require the participation of a large number of general practitioners.

The gold standard used in several studies was referral rather than subsequent diagnosis. One study provides more detail (Barton et al, 1999(194)), and this is described at greater length. There are several studies of the symptoms and signs of women attending specialist services, and we have included two of these only to highlight the different patient features found among a specialist service in contrast to primary care. Considerable caution is needed, therefore, in extrapolating from studies undertaken in specialist clinics to patients presenting to primary care.

Studies of patients presenting in primary care
(Newton et al, 1999) (195)

In this case series, data were collected prospectively from 508 women consulting 248 general practitioners in Sheffield over a four week period between January and July 1995. The general practitioners used a standard pro-forma to record information about women consulting primarily for a breast problem. The pro-formas were not completed for women who had a breast examination as part of a consultation for any other reason.

Referral rates increased according to patient age: 16–39 32.6%, 40–49 38.7%, 50–64 40.6%, 65+ 50.0% (Table 35). The mean number of consultations was 2.05 over the four week period, suggesting that a general practitioner would see 15.8 women with new breast problems in one year. However, this figure excludes women who consulted for primarily other problems but also had a breast problem.

Table 35. Presenting features among 508 women consulting with breast problems (Newton et al, 1999 (195)).

Table 35

Presenting features among 508 women consulting with breast problems (Newton et al, 1999 (195)).

(Nichols et al, 1980) (196)

In this case series, 193 general practitioners were recruited in Southampton to record in a booklet all women seen with breast symptoms over four weeks. There were 331 consultations recorded by 323 women for breast conditions (mean: 3.5 per general practitioner). Of those consultations 241 were for new episodes (Table 36).

Table 36. Presenting features among 323 women consulting with breast problems (Nichols et al 1980).

Table 36

Presenting features among 323 women consulting with breast problems (Nichols et al 1980).

(Bywaters, 1977)(197)

This study involved six general practitioners in one UK practice recording 451 consultations for breast problems by 180 women. Details of consultations were recorded and a list was created of women consulting with breast complaints between October 1972 and December 1974. The presenting features are summarised in Table 37.

Table 37. Presenting features among 180 women consulting with breast problems (Bywaters 1977).

Table 37

Presenting features among 180 women consulting with breast problems (Bywaters 1977).

28 of the 180 had cancer (18 new cases −10%); All these were aged 30 or over. Of 57 patients seen with a discrete lump, 32 (56.1%) were referred immediately.

(Roberts et al, 1987)(198)

This was a study to ascertain the effects of a recent health campaign on the number of general practitioner consultations for breast problems. The study involved giving each patient consulting with breast problems a questionnaire; women having a breast examination associated with contraceptive care or routine cervical cytology tests were not included. 262 women returned questionnaires from five UK general practices over 18 months. Their symptoms and referral rates are shown in Table 38.

Table 38. Presenting features among women consulting at primary care with breast problems (Roberts et al, 1987 (198)).

Table 38

Presenting features among women consulting at primary care with breast problems (Roberts et al, 1987 (198)).

In addition, the study suggested that public health campaigns had little measurable affect on consultation rates.

Studies of referred patients
(Seltzer, 2004)(199)

This study reviewed data on 10 000 consecutive new surgical referrals for breast complaints in the US. Female patients referred between 1987 and 1999 completed a comprehensive medical history form. The aim of the study was to demonstrate those situations which are likely to yield a cancer diagnosis.

Across all ages, 9% of patients presenting with lump yielded cancer; 16% of those presenting with pain; 4% of those presenting with discharge; 11% of those found by mammogram and 5% presenting withg miscellaneous complaints.

(Campbell, 2004)(200)

This study reviewed prospective audit data from patients referred to a symptomatic breast unit in the UK. The patients with a breast lump were significantly more likely to have breast cancer than patients without a lump (OR = 5.0765, CI = [3.06662–8.4047], p < 0.001). The likelihood of breast cancer increased with age (OR = 1.0808, CI = [1.0712–1.0906], p < 0.001). Pain was the least likely to indicate the presence of cancer (OR = 0.1351, CI = [0.0664–0.22749], p < 0.001), as was breast lumpiness (OR = 0.3192, CI = 0.1718–0.5930], p < 0.003), nipple discharge (OR = 0.5337, CI = [0.1821–1.5647], p > 0.05), HRT use (OR = 0.6995, CI = [0.4431–1.1042], p < 0.05) and signs of cancer (OR = 0.6842, CI = 0.4156–1.1265], p < 0.003). Family history was not found to be statistically significant within their model.

(Patel et al, 2000)(201)

This study was prospective case series involving new patient referrals from general practitioners to a specialist breast clinic in Glasgow. Of the 321 patients referred, 10% had breast cancer and 90% had either benign disease or no pathology. The study concluded that one third of the referrals were inappropriate (Table 40).

Table 40. Features at presentation among 940 women with breast cancer (Barclay et el 1991).

Table 40

Features at presentation among 940 women with breast cancer (Barclay et el 1991).

Table 39. Features among 321 women referred to a breast clinic (Patel et al 2000)(201).

Table 39

Features among 321 women referred to a breast clinic (Patel et al 2000)(201).

(Barclay et al, 1991)(202)

In this case series, information was collected about women referred to breast or surgical outpatient clinics in Dundee between 1979 and 1989. During this period, 940 women presented with new breast cancers and 3,500 were referred with benign conditions. The features at presentation among the patients with cancer are shown in Table 40. The median age of those with benign disease was 35 years, but for those with cancer the median age was 57 years. The majority (91%) of referrals to the breast unit for benign disease occurred in patients under 55 years.

Among those with cancer, a visible abnormality was noted in the left breast in 362 patients, and the right breast in 320 patients. The most common observed abnormalities were asymmetry (68%), nipple abnormalities (43%) and skin changes (7%).

Of those diagnosed with breast disorders, 15% reported a family history of breast cancer, compared with only 18% of the 940 who had cancer reporting family history.

(Barton et al, 1999) (194)

This US population-based retrospective cohort study was undertaken at a large health maintenance organisation in New England over a ten year period. The study sought to determine 1) how often women presented with breast symptoms to primary care providers 2) how these symptoms were evaluated, and 3) how often symptoms led to a diagnosis of breast cancer. The study population was 2400 women aged 40–69 years, sampled in a random age stratified manner and from people who had been continuously enrolled in the health maintenance organisation (HMO) from July 1983 to June 1993. For this sample, information was abstracted on all breast related encounters and diagnoses of cancer subsequent to presented symptom(s) were recorded.

Patient symptoms were classified as 1) mass (a single lump or nodule); 2) pain (a report of pain or tenderness in either breast or bilaterally), 3) skin or nipple change (including nipple discharge) 4) multiple lumps or nodules often described by clinicians as ‘fibrocystic’ or ‘diffuse cystic change’, or 5) other symptoms (such as increasing breast size). Clinicians’ diagnostic interpretations were classified as normal (even if fibrocystic), abnormal-benign (no further follow up required), indeterminate (record of firm or fixed lumps, or follow up by surgeon recommended, or suspicion of cancer noted). The meaning of such terms as benign or normal had to be inferred because clinicians did not use a standard taxonomy to describe their examination findings nor a standard metric to convey level of concern.

Over the ten year period, 372 (16%) of the HMO population presented with a breast symptom (22.8 presentations per 1000 person years). Women younger than 50 years of age presented nearly twice as often as older women (P=0.0001). Rates did not differ by ethnic group. Women with a family history of breast cancer were more likely to present with breast symptoms than those without a family history (22% compared with 14%; P=0.001).

The most common symptom was pain, followed by a mass, skin or nipple change, lumpiness and other symptoms. Two symptoms were noted in 59 episodes (13%); the most frequent combinations were pain and mass (31 episodes [7%]) and pain and skin or nipple changes (14 episodes [3%]). In 69 episodes, no specific symptom was documented. Presenting symptoms and signs varied by age. A mass was the most common feature among women in their 40s, and pain was the most common feature among women in all other age groups. Pain was unilateral in 91% of episodes and bilateral in 9% of episodes.

On physical examination, the clinicians found a mass in 184 episodes (34%), skin changes or nipple discharge in 43 episodes (8%), fibrocystic changes in 112 episodes (21%) and other findings in 32 episodes (6%). More than one finding was documented in 45 episodes and no specific findings were documented in 214 episodes (40%). Of the 196 episodes in which a patient reported a mass, the clinician confirmed the mass in 160 (82%). Of the 343 episodes in which mass was not one of the patient’s symptoms, the clinician documented a mass in 24 (7%).

Clinicians interpreted physical findings as normal in 33% of episodes, abnormal-benign in 27%, indeterminate in 35%, and suspicious for cancer in 6%. Breast cancer was diagnosed in 23 of the 372 women who presented with breast symptoms (6.2%); 21 had invasive disease (six with stage 1 disease, 14 with stage 2 disease, and one with stage 3 disease) and two had ductal carcinoma in situ.

Of the 23 women with cancer, 11 (6.4%) presented while in their 40s, six (4.4%) while in their 50s, three (4.4%) while in their 60s, and three (8.3%) in their 70s. Clinicians had found a mass in 22 (96%), skin findings in two (9%), fibrocystic changes in three (13%) and other findings in three (9%).

The 23 women with symptomatic breast cancer had higher tumour stages at diagnosis than 58 women whose breast cancer was detected by screening mammography during the study period (P=0.02). The likelihood of breast cancer varied by symptom or sign. A report of a mass was associated with a 10.7% chance of breast cancer and a likelihood ratio of 65, whereas a report of pain led to a diagnosis of cancer in 1.8% of episodes, with a likelihood ratio of 10. A mass accompanying any other symptom or sign increased the risk for cancer. At the same time, each symptom or sign alone was associated with a significantly higher risk for cancer than in the population at large.

Although younger women presented more frequently with breast symptoms or signs, cancer rates did not vary significantly by age group. The study indicated that 4.3% of breast symptom or sign episodes led to a diagnosis of breast cancer, but it should be noted that the incidence of cancer may be lower in this study than in an unscreened population because of the use of screening mammography in the study population. A mass was the feature most often associated with breast cancer. Only two of 23 women (8.7%) who were found to have cancer presented with pain as the only feature.

It should be noted in interpreting these findings that the study did not include women younger than 40 years of age, and that a relatively high proportion (18%) had a family history of breast cancer.

(Chalabian and Dunnington, 1998) (203)

This study involved 66 graduating primary care physicians, assessing the link between observed breast examination skills during an objective structured clinical examination (OSCE) and ability to detect lumps in silicone models. The correlation detected between lump detection and examination skills, although statistically significant, was only 0.34. No relationship was found between breast model sensitivity and specificity. Although the authors commented that thorough clinical breast examinations are imperative as they can identify 10% of breast cancers not visible on mammograms (204), no specific manoeuvres or techniques could be recommended.

(Khan and Apkarian, 2002a) (205)

In this study, a modified version of the McGill Pain Questionnaire was administered to 271 women with breast pain but without breast cancer. 134 women had cyclic breast pain and 152 non-cyclic. Cyclical breast pain tended to be a diffuse, heavy ache, most prominent towards the end of the cycle, although may also be severe during menstruation. It may occur in one breast, but commonly in both. However, there are very few studies of women with breast pain in primary care, and the significance of pain as an indicator of cancer is difficult to determine.

(Khan and Apkarian 2002) (206)

This study was a retrospective case controlled investigation into the relationship between breast mastalgia and cancer studying a population of 5463 women aged over 30 attending a New York breast care centre. Of those women, 861 were diagnosed with breast cancer, of whom 141 (16.4%) reported breast pain (mastalgia). Of the 4602 women who did not have cancer, 1391 (30.2%) reported mastalgia. Breast pain was reported as an incidental complaint at first visit to the centre by 1532 (28%) of all the women in the study.

This investigation found that within their study population, women who experienced breast pain were less likely to be diagnosed with breast cancer than those without, regardless of age or other risk factors. Additionally the study found that risk factors associated with breast cancer (age, age of menarche, age at first full term pregnancy, age at menopause, family history, alcohol use) were associated with a decreased frequency of breast pain, with the exception of exogenous hormone use.

Risk Factors
Evidence Statements

Epidemiological studies have reported a number of risk factors as being associated with an increased probability of developing breast cancer. Such risk factors include: age; family history of breast cancer; age of having first child and use of hormone replacement therapy. (III)

In a woman who presents to a medical practitioner with a palpable breast lump, the presence or absence of any given risk factor has no significant effect on the likelihood of that woman having breast cancer. (III/DS)

There is no evidence that information on risk factors is of use in selecting those symptomatic women who should be referred (III)

Guidelines
(NICE: The Classification and care of women at risk of familial breast cancer 2004) (207)

This evidence based guideline is limited to women over 18 who have not been previously diagnosed with breast cancer. The evidence searches were wide ranging and papers were graded according to NICE specifications, while quality of studies was assessed using modified SIGN checklists.

The guideline states that although most breast cancer occurrences are random, in 16–19% of cases a family history of the disease is identifiable. The probability of a 20 year old woman developing breast cancer by 80 increases with the incidence of breast cancer within her family. With no affected relatives the risk is 7.8%, with one 13.3%, and with two 21.1%(207).

The evidence used in assessing the specific risk factors of breast cancer evaluated by the guideline was of varying quality and a summary of the findings and subsequent recommendations follow.

Family history

Risk increases with the proximity of the relationship to an affected relative, the number of affected relatives and with the decrease in age of those relatives at the time of developing breast cancer. The high risk genes BRCA1 or BRCA2 account for only a small amount of this increased risk. However, the risk of carrying one of these mutated genes is related to the strength of the family history, and risk of breast cancer is increased by their occurrence (BRCA1 60–80% risk, BRCA2 40–80% risk).

Hormone Replacement Therapy (HRT)

The risk of breast cancer is increased and continues to increase in association with the duration of HRT use. Increased risk reduces once treatment is stopped and risk returns to same level as a woman who has never taken HRT after five years. Thus, it was recommended that treatment time is restricted to short term (no definition of short term was given) in women with familial risk, and alternative treatments should be considered and the woman informed of the increased risk.

Oral Hormonal Contraceptives

Evidence concerning ever-use, current use, duration, and cessation of oral contraceptive use is contradictory and inconsistent. Ever–use was not associated with increased risk in breast cancer in women of any age. Findings on current use and duration of use were inconclusive and contradictory as some studies suggested an increase in risk and some did not. A 16% increased risk was observed within the first four years after stopping oral contraceptive use and a 7% increase between five and nine years. After ten years no increased risk was observed. A statistically significant increase in risk was found in women using oral contraceptives prior to their first full term pregnancy (72%). No specific increase in risk was recorded among those with familial risk taking oral contraceptives. One study identified carriers of the BRCA1 mutation gene as having a 20% increased risk when using oral contraceptives, but no increased risk in carriers of the BRCA2 mutation gene.

Breastfeeding

Breastfeeding has a protective affect against breast cancer, which is proportionate to the total duration of breastfeeding. There is a 4% reduction in risk for every 12 months of breastfeeding and the risk is similar in women with familial risk. It was recommended that women be advised to breastfeed.

Alcohol consumption

Risk increases with alcohol consumption by 7.1% per 10g daily intake and is unaffected by familial risk. It is recommended that information is provided to women with familial risk.

Smoking

Evidence reviewed reached different conclusions ranging from no association of smoking with increased risk of breast cancer, to significant increases in both current or former smokers, with additional particularly high risks in premenopausal women or those who began smoking very early. The guideline concludes that as scientific studies have produced inconsistent findings a relationship is merely speculative.

Weight and physical activity

No specific link between diet and familial risk of breast cancer was found, although moderate exercise was thought to confer a decrease in risk of cancer. However, high BMI was associated with an increase of risk in postmenopausal breast cancer. Thus it was recommended that women are informed of the increase in risk associated with being overweight.

Menstrual/reproductive factors

Menstrual and reproductive factors carry the same risks among women with or without a family history of breast cancer. In both groups of women, older age at first birth and earlier menarche were associated with increased risk. Risk decreases with the number of live births. It was recommended that the practitioner should provide information about hormonal risk factors.

Secondary studies
(Levine et al, 2001) (193)

This review undertaken by the Agency for Healthcare Research and Quality is outlined in the section dealing with symptoms and signs above. Age was the only risk factor consistently reported in association with symptoms and cancer diagnosis. The influence of family history varies depending on the age of the patient and the closeness of the affected relative(s), the ages at which the relatives developed cancer, the number of relatives with breast cancer, and the number with other gynaecological or other cancers. Women whose mother or sister had breast cancer before the age of 40 had the highest risk (relative risk 2.2, 95%CI 1.5–4.2). HRT was reported as not significantly increasing the risk among women who have a family history.

Risk of breast cancer increases with duration of oestrogen exposure. Women who had an early menarche are at increased risk (before age 12 RR 1.1–1.3), as are those with a late menopause (after age 55 RR 2.0). Women who delay their first child until after age 30 have an increased risk (RR 1.3–1.9). The impact of pregnancy is not well understood, since there is an increased risk for up to 10 years after delivery.

The review did not consider the impact of smoking, diet, alcohol, lactation or genetic factors on risk of breast cancer.

(Collaborative Group on Hormonal Factors in Breast Cancer, 2002) (208)

The authors analysed individual data from 47 epidemiological studies in 30 countries to estimate the association between breastfeeding patterns and childbearing with breast cancer. For women who had never breastfed, the relative risk of breast cancer declined by 3% for each year younger they were when their first child was born. The relative risk of breast cancer decreased by 4.3% for every 12 months of breastfeeding (not necessarily consecutively) in addition to a decrease of 7% for each birth. The size of the decline in the relative risk of breast cancer associated with breastfeeding did not differ significantly for women in developed and developing countries, and did not vary significantly by age, menopausal status, ethnic origin, the number of births or age when the first child was born. It is estimated that the cumulative incidence of breast cancer in developed countries would be reduced by more than half, from 6.3 to 2.7 per 100 women by age 70, if women had the average number of births and lifetime duration of breastfeeding that had been prevalent in developing countries until recently.

Primary studies
(McPherson et al, 2000)(78)

This paper reviews the risk factors for breast cancer in the UK, the findings are summarised in Table 44 below.

Table 41. Established and probable risk factors for breast cancer.

Table 41

Established and probable risk factors for breast cancer.

12.1.2. Men

12.1.2.1. Key Clinical Question

Which are the symptoms, signs and other features that raise a suspicion of cancer in a man presenting with a breast abnormality, and those that make cancer less likely as a diagnosis?

12.1.2.2. Evidence Question

In men attending primary care services with breast symptoms, which symptoms and signs and other features when compared with the “gold standard” are predictive of a diagnosis of cancer; and which symptoms and signs are not?

12.1.2.3. Evidence Statements

A subareolar mass is the most common presenting sign in men with breast cancer. Less common signs include nipple retraction, local pain, nipple ulceration, discharge or bleeding (III).

In men, breast cancer is more common, but not confined to, those over 50 years of age (III).

There are several risk factors for breast cancer in men, but their significance in estimating the likelihood of cancer among men presenting with symptoms is unclear (III).

Secondary studies
(Giordano et al, 2002) (209)

This is an up to date systematic review. The authors sought articles published between 1942 and 2000, and used CancerLit, Medline and study bibliographies to identify articles. They included studies on the epidemiology, risk factors, genetics and pathology of breast cancer in men. The review reports the following conclusions.

The incidence of breast cancer in men has remained stable in the past 40 years, and the median age at diagnosis is 68 (compared to 63 in women). However, the disease has been reported in males from ages 5 to 93 years. The incidence increases exponentially with age. Breast cancer in men may be hormonally driven, as in women. The risk factors include: testicular abnormalities (undescended testis, congenital inguinal hernia, orchidectomy, orchitis, testicular injury); infertility; Klinefelter syndrome; positive family history; benign breast conditions (nipple discharge, breast cysts, breast trauma); radiation exposure; increasing age; Jewish ancestry. The rate of gynaecomastia in men with breast cancer is similar to the rate in the general population.

Approximately 90% of all breast tumours in men are invasive carcinomas, the remaining 10% being non-invasive (most being ductal carcinoma in situ). Approximately 85% (ranging between 50–97% in different studies) of affected men present with a painless subareolar mass. Other common signs include nipple retraction (10–51%), local pain (4–20%), nipple ulceration (4–17%), nipple discharge (1–12%), and nipple bleeding (2–9%). Men are more likely than women to have a delay between the onset of symptoms and diagnosis. Mammography is reported as being helpful in distinguishing a benign from a malignant lesion, and fine needle aspiration has been found to be sensitive and specific.

No primary studies are included in this evidence review as the systematic review of Giordano et al (209) is recent and comprehensive.

12.2. Investigations

12.2.1. Key Clinical Question

Should any investigations be undertaken in primary care, before referral?

12.2.2. Evidence Question

In women attending primary care services with breast symptoms, which investigations when compared with the “gold standard” are predictive of a diagnosis of cancer; and which are not?

What investigations to diagnose a suspicious breast lump are available to primary care practitioners in the UK?

12.2.3. Evidence Statements

Evidence from studies in Britain and Sweden indicate that decisions on whether to refer women presenting with breast symptoms are commonly made at the first consultation, and without recourse to investigations (III).

There is no evidence that laboratory tests have a role in initial investigation of women presenting with breast lumps in primary care (III).

In some countries, some primary care physicians undertake FNA for cytological examination. However, success in obtaining a satisfactory sample is dependent on the skill of the physician. There is no evidence on the role of FNA in primary care in the UK (IV).

There is no evidence from the UK to suggest that a policy of investigation with mammography and/or FNA accelerates referral to secondary care of patients with cancer. It is possible that use of these investigations would delay referral (IV).

Women presenting to primary care with breast pain and in whom cancer is not suspected but who are referred for a mammogram are unlikely to have a suspicious mammogram. (III)

Background

Established management of women suspected of having breast cancer includes the triple assessment of physical examination, mammography and percutaneous biopsy (also referred to as fine needle aspiration – FNA).

We found very few studies of the role of investigations in women presenting with breast symptoms in primary care. The majority of studies of investigations involved women who had been referred, and since the findings cannot be extrapolated to the population of symptomatic women before referral, these studies have been excluded.

Guidelines
(Austoker and Mansel, 2003).(188)

These guidelines did not suggest any primary care investigations before referral in patients presenting with a breast lump, breast pain, or severe cyclical mastalgia. In the case of nipple discharge in women less than 50 years of age, a test for blood was advised if the discharge is from multiple ducts. Referral was recommended when the test is positive. Other investigations, including triple assessment, were restricted to patients who had been referred, the investigations being carried out by the specialist.

(All Wales Minimum Standards, 2000) (189)

Standard 11 requires that all diagnostic tests are carried out in one visit. The standard related to patients referred to and attending specialist services.

(Steering Committee on Clinical Practice Guidelines for the Care and Treatment of Breast Cancer, 1998) (190)

These guidelines were based on a systematic review of evidence (Medline from 1966, Cancerlit from 1985, through to 1996). However, the studies cited were not confined to those involving patients in primary care. Mammography was found to be unlikely to give useful information in younger women, although is more useful from aged mid-30s. The overall level of sensitivity of mammography was reported as possibly no higher than 82% (level III evidence), and therefore a normal mammogram cannot exclude cancer. The guideline indicated that fine needle aspiration can be carried out in office settings, and that cytologic examination should be ordered if the obtained fluid is bloody. Success in obtaining satisfactory samples, however, is operator dependent. The false negative rate in one reviewed study had been 15.2%. When physical examination, mammography and cytology are combined, the diagnosis is likely to be confirmed in 99% of cases in which all three tests are positive; cancer will be found in 0.5% of cases if all tests are negative.

(Royal Australian College of General Practitioners 1997) (210)

These guidelines are reported as based on a review of evidence, although there is insufficient information to judge the extent and quality of the review. The guidelines encourage the use by general practitioners of imaging and fine needle aspiration. Ultrasound is recommended in place of mammography in women under age 35.

Secondary Studies
(Kerlikowske, 2003)(211)

A review of papers found on Medline between January 1966 and March 2003 to determine the most accurate and least invasive means to evaluate an abnormal mammography result and palpable breast abnormality.

This study found that a diagnostic mammography is most helpful in deciding whether a nonpalpable breast lesion should be biopsied but not whether a palpable breast abnormality should be. For palpable masses, fine needle aspiration biopsy or core-needle biopsy were preferred. However in order to determine whether a lesion is a simple cyst and therefore benign, core needle biopsy or needle localisation with surgical biopsy was usually preferred.

Primary studies
(Duijm et al, 1998a) (212)

In a study of 987 women with a painful breast referred to the radiology department of a Netherlands hospital between 1992–1996, follow up was undertaken for two years. The gold standard was a recorded diagnosis of breast cancer during follow up. 84.1% of the sample had been referred by general practitioners. The findings were compared with a control sample of 987 asymptomatic women undergoing a screening mammogram. Four (0.4%) of the women with pain were diagnosed with cancer, in comparison with seven (0.7%) of the controls. Mammograms were classified as suspicious or malignant in only 1.2% of the symptomatic cases.

(Mansson et al, 2001) (213)

This study was undertaken in four primary health care centres in Sweden 1995–1997, and investigated the diagnostic actions of general practitioners in relation to colorectal, pulmonary, breast and prostate cancer. The total patient population in the area served by the health centres was 9556, and 125 women were recorded as presenting with breast problems. In most, no laboratory test had been performed, although 80 mammographies were undertaken, with a yield of three cancers. Seven breast cancers were diagnosed in total, six at the first consultation; one was interpreted as a benign tumour, and six were referred to a surgeon. Two patients had haemoglobin tests, one ESR, and four various other tests not related to breast cancer (e.g. urine dipslide). The study did not indicate whether these laboratory tests served a useful role in the initial assessment of the patients with breast cancer.

(Mansson and Bengtsson, 1992) (214)

The primary care records of all 62 women with a diagnosis of breast cancer between 1981 and 1983 in Kungsbacka in Sweden were reviewed. Information was collected about the investigations ordered before diagnosis. The article does not report the number of women who underwent laboratory investigations, but notes that 12 (19%) were found to have an elevated erythrocyte sedimentation rate, eight (13%) had anaemia, and six (10%) had a leucocytosis. However, in another report from this study (Mansson et al, 1999), it was reported that 59 (95%) had a haemoglobin estimation and 57 (92%) an erythrocyte sedimentation rate estimation. The authors concluded that haematology and erythrocyte sedimentation tests did not assist in the diagnosis of breast cancer.

12.3. Delay and Diagnostic Difficulties

12.3.1. Key Clinical Question

What influence do age, gender, social class and ethnicity have on the differential at presentation?

What diagnostic difficulties do primary care professionals themselves report in determining whether a woman/man who presents with breast symptoms/signs may or may not need urgent referral with suspected cancer?

12.3.2. Evidence Question

In women attending primary care services with breast symptoms, which psychosocial and socio-demographic factors are associated with delayed presentation of breast cancer?

What diagnostic difficulties do primary care professionals themselves report in determining whether a woman/man who presents with breast symptoms/signs may or may not need urgent referral with suspected cancer?

12.3.3. Evidence Statements

Delay

There is strong evidence of an association between older age and delay by patients, and strong evidence that marital status is unrelated to delays by patients (III).

There is an association between socioeconomic status and survival (III)

There is moderate evidence for an association with delay by patients with five other factors:

  • fewer years of education
  • presenting with breast symptoms other than a lump
  • not disclosing the breast symptom to another not attributing the breast symptom to breast cancer (III).

Younger age and presentation with a breast symptom other than a lump were strong risk factors for delays by health professionals. There is moderate evidence that ethnicity does not influencing delay by providers. (III)

Diagnostic Difficulties

Primary care professionals report that detection of the possibility of breast cancer is often straightforward, but in some cases is difficult (III).

A past history of benign breast conditions, young age, and presentation without a palpable lump are features that can make the detection of possible cancer more difficult (III).

Primary care professionals’ referral decisions are influenced by their own and their patients’ anxiety. Past experience of a delayed or missed diagnosis can lower the professional’s referral threshold (III).

Delay

The following section addresses the influence that socio-demographic and psychosocial factors have on the women’s decision to seek help when confronted with symptoms and signs suspicious of breast cancer. The four that will be considered are:

  • Psychosocial factors
  • Socio-economic status
  • Age
  • Ethnicity.
(Sainsbury et al, 1999)(215)

An retrospective analysis of 36,222 patients with breast cancer listed on the Yorkshire Cancer Registry between 1976 to 1995, in order to investigate whether delay in referral from primary care influences survival. Patients were grouped according to time taken from family-physician referral to treatment (<30 days/30–59 days/60–89 days and 90> days).

Results demonstrated no evidence that delay up to three months (90 days) adversely influenced survival. From 1976 to 1995 the time from family-physician referral varied very little with a median of 10 vs.13 days. However the time from first visit to until the patient received treatment doubled for the same time period going from 7–13 days. Of the women included in the study, those who presented early and were in less than 30 days actually had significantly worse outcomes (p<0.001).

Secondary studies
(Ramirez et al, 1999) (216)

The authors undertook a systematic review of 23 papers to assess the quality and strength of evidence on risk factors for delays by patients and providers. There was strong evidence for an association between older age and delay by patients, and strong evidence that marital status was unrelated to patient delays. There was moderate evidence for an association between patient delay and five other factors: fewer years of education, non-white ethnic origin, presenting with breast symptoms other than a lump, not disclosing the breast symptom to another, and not attributing the symptom to breast cancer. Younger age and presentation with a breast symptom other than a lump were strong risk factors for delays by providers. There was moderate evidence against non-white ethnic origin influencing delay by providers.

Primary studies
A. Papers that explore the influence of more than one factor
(Grunfeld et al, 2002) (217)

This study investigated the influence that women’s age and socio-economic status play on delayed presentation. 996 women, randomly selected though the postal address file were interviewed by the authors to elicit their knowledge of breast cancer risk, breast cancer symptoms, and their perceptions of the management and outcomes associated with breast cancer. Older women were particularly poor at identifying symptoms of breast cancer, risk factors associated with breast cancer and their personal risk of developing the disease. Professional women and women classified as intermediate had a greater knowledge of risk factors than women from lower socio-economic groups. 32% of professional and intermediate women reported reduced risk compared to 10–15% of partly skilled and unskilled women, and women who were unskilled or had never worked identified significantly fewer symptoms than the other socio-economic groups.

(Grunfeld et al, 2003) (218)

This study primarily investigated the influence of psychosocial factors but in relation to women’s age. The authors recruited a sample of 546 women as the second phase of a previous study (Grunfeld et al, 2002 (217)). All women completed a postal questionnaire about beliefs regarding the symptoms, causes and outcomes associated with breast cancer, attitudes towards help seeking and beliefs about one’s ability to seek help. The inability to correctly identify a range of potential breast cancer symptoms was a significant predictor of intention delay in seeking help across all age groups. For women aged 35–54, negative attitudes towards medical help seeking for breast symptoms and a negative belief in one’s ability to seek help were additional predictors of intention not to seek help. Holding negative beliefs about the consequences of breast cancer (i.e. that the disease could be potentially disabling or disfiguring) was found to be an important additional predictor of delay in help seeking among women aged over 65 years.

(Nosarti et al, 2000) (219)

This paper examined the influence exerted by women’s symptoms, psychosocial, socio-economic status and ethnicity. The authors interviewed 692 women referred to a London breast clinic to identify factors associated with delay in presentation. Sixty per cent of women with a breast lump presented to their doctor within 27 days from symptom discovery, compared to 34% of those without a lump. Of patients with breast tenderness or pain, 76% presented to their doctor within 27 days from symptom discovery, compared to 62% of those without pain. Thirty-five per cent of the women delayed presentation 4 weeks or more (median 13 days). The most common reason was that they thought their symptom was not serious. Others thought their symptom would go away or delayed presenting because they were scared. Delay was associated with psychiatric morbidity but not age. Median system delay was 18 days. Patients who thought they had cancer and those so diagnosed were seen more promptly (median 14 days). Most socio-demographic factors, including socio-economic status and ethnicity, were non-contributory to delay.

(Nichols et al, 1981) (220)

In this UK study, women with breast symptoms referred to a specialist outpatient department were interviewed to ascertain the interval between first noticing a breast symptom and consulting a doctor. The largest component of delay was patient delay, with 20% of women delaying longer than 12 weeks. Long delays were related to age and symptoms other than lumps.

B. Papers that explore the influence of psychosocial factors
(Burgess et al, 2001) (221)

The authors interviewed 46 women in the UK with newly diagnosed breast cancer to explore the factors that influence general practitioner consultation by women with breast cancer symptoms. The main factors that influenced help seeking behaviour were: the identification the woman made of their symptoms as suggestive or not of breast cancer; their attitudes to requesting an appointment with a general practitioner; their beliefs about the consequences of cancer treatment; the effect of competing events and difficulties that could be prioritised over and above their personal health; and influences or experiences that functioned as triggers to action.

(Burgess et al, 2000) (222)

In this UK study, 158 women were interviewed five months after diagnosis to examine the influence of adverse life experiences and mood disorders on delayed presentation of breast cancer. The study did not identify statistically significant associations between these factors and delay, and suggested that neither adverse life events nor mood disorders in the year before symptom discovery increased the risk of patients with symptoms of breast cancer delaying their presentation to their general practitioner.

C. Papers that explore the influence of socio-economic status
(Malik and Gopalan, 2003)(199)

This is a prospective study of 138 recently diagnosed (within three months) breast cancer patients who had initially presented with breast lump in Pakistan. The majority (85%) of the patients discovered the lump accidentally, 10% were identified by a family physician and 5% as part of regular self examination. These patients took an average of 8.7 weeks to inform members of their family and 17.2 weeks until their first physician visit.

The initial perceptions of the lump included milk clots, trauma, infection benign growth, other and cancer (however only 17% perceived it as cancer). f those patients included in the study 73 (52.9%) were recorded to have delayed seeking medical advice. The reasons given were; antecedent use of complimentary/alternative therapies (34%), lack of significance attached to the lump (23%), fear of surgery (22%), conflicting personal commitments (7%), fear of cancer (5%) and other reasons (8%).

(MacLeod et al, 2000b) (223)

This was a UK population-based review of the case records of 417 women under 75 with breast cancer. Women living in deprived areas (according to the Carstairs Index) were more likely to present with large, locally advanced cancers or with metastatic disease than those living in affluent areas. There were no major differences in pathological prognostic factors at presentation between socio-economic groups. Although stage at presentation accounts for some of the differences in survival between affluent and deprived women, other unidentified factors adversely affect survival in deprived women.

(Thomson et al, 2001) (224)

The authors analysed two datasets relating to breast cancer patients in Scotland (23,866 women). Survival differences of 8.7% at five years and 10.2% at ten years between affluent and deprived women were observed across all age groups. No differences were observed in tumour size or nodal status at presentation between the deprivation groups. Although deprived women were more likely to have oestrogen receptor negative tumours, this difference explained only about a third of the difference in survival between affluent and deprived women. Women aged under 65 with non-metastatic disease were more likely to have breast conservation than mastectomy if they were affluent (45%) than deprived (32%); the affluent were also more likely to receive endocrine therapy (65%) than the deprived (50%). However, differences in treatment between affluent and deprived women do not seem to account for their different survival.

(Carnon et al, 1994) (225)

The authors carried out a retrospective analysis of data from a cancer registry within the catchment areas of two large hospitals in Glasgow, and attempted to explain socio-economic differences in survival from pathology and biochemistry records for 1361 women diagnosed with breast cancer. They could find no significant relation between socio-economic deprivation and four pathological prognostic factors at presentation: tumour size, negative nodes, tumour grade, and low oestrogen receptor concentration.

(Schrijvers et al, 1995) (226)

The authors explored the association between deprivation and survival from breast cancer in 29,676 women aged 30 and over. There was a clear gradient in survival that increased slightly with time since diagnosis, with better survival for women from more affluent areas. At all ages, women in the most deprived category had a 35% greater risk of death than women from the most affluent areas after adjustment for stage at diagnosis, morphology and type of treatment. In younger women (30–64 years), the survival gradient by deprivation category cannot be explained by these prognostic factors. In older women (65–99 years), part of the unadjusted gradient in survival can be explained by differences in the stage of disease: older women in the most deprived category were more often diagnosed with advanced disease. Other factors, so far unidentified, are responsible for the gradient in breast cancer survival by deprivation category.

(Quinn et al, 2001) (17)

Data from National Statistics provide some information about incidence and survival according to level of deprivation. In 1993, there was a negative gradient in the incidence of breast cancer by Carstairs deprivation category, the rate being about 30% higher in the most affluent groups. In contrast, mortality was not related to deprivation, implying that survival is better in the more affluent groups. The gap in survival between deprived and affluent groups in the 1980s was 6% at one year after diagnosis, and 9% at five years.

(MacLeod et al, 2000) (227)

The authors reviewed hospital and general practice case records of 821 women with invasive breast cancer. Women living in affluent areas did not receive better NHS care for breast cancer than women in deprived areas. Admissions to hospital for problems not related to breast cancer were more common in those living in deprived areas, as also were the number of consultations with their general practitioners in the two years following diagnosis.

D. Papers that explore the influence of age
(Kroman et al, 2000) (228)

The authors undertook a retrospective cohort study in Denmark based on 10,356 women who were less than 50 years old when diagnosed with breast cancer to investigate the effect of young age on prognosis, and the influence of tumour staging and treatment on such association. Young women with low risk disease who did not receive adjuvant treatment had a significantly increased risk of dying than the women who did, and the risk was increased with decreasing age at diagnosis. This increased risk remained when women were grouped according to presence of node negative disease and by tumour size.

E. Papers that explore the influence of ethnicity

We have not found any relevant papers that exclusively investigated the influence of ethnicity in delayed presentation of women with breast cancer since the publication of the systematic review by Ramirez et al (1999) (216). Most recent identified studies that explore this factor have studied the experiences of African-American women. Caution is required when extrapolating results from these studies to England and Wales because of the different characteristics of the UK and US health care systems.

(Velikova, 2004)(229)

This retrospective UK study examined population based data on 16,879 women with breast cancer diagnosed between 1986 and 1994 with an aim to evaluate patient and provider delays of South Asian patients. Of those included in the study, 120 (0.7%) were South Asian and the standardised incidence rate ratio of South Asian with non-South Asian was 0.56 (95% CI 0.46–0.66).

Asian women were significantly younger than non-Asian at the time of diagnosis with a greater proportion being diagnosed before 50 years of age. The mean age at diagnosis of Asian and non-Asian was 49.7 years compared to 62 years respectively. A significantly higher proportion of South Asian patients presented with tumours larger than 2cm. Asian patients had a longer period of delay between symptom onset and presentation to a general practitioner with a median of 61 days compared to 31 days for non-Asian women which could not be explained. However no significant difference in delay was recorded between general practitioner visit and first hospital visit.

(Coates, 1992)(230)

This study collected retrospective data over 410 black women and 325 white women who were newly diagnosed with invasive breast cancer in 1985 or 1986 in the US in order to evaluate racial differences in delayed presentation.

The study found that black women were diagnosed more commonly at later disease stage. They were twice as likely to be diagnosed with Stage IV breast cancer and one and a half times as likely to be diagnosed with Stage III than white women. Additionally black women were only half as likely to be diagnosed with Stage I breast cancer. Black women were also found to be twice as likely as white women to be diagnosed with tumours larger than 5cm.

There was a low but statistically significant (15%) difference in the rate with which black women obtained initial consultation compared to white women and the median time between symptom recognition and consultation was 16 days for black women and 14 days for white women. The study concluded that although there were significant differences in delay, the differences were small and therefore unlikely to account for differences in survival rates.

(Bassett et al, 1986) (231)

This study used data from the Western Washington cancer surveillance system, and examined the influence of social class and race as predictors of survival in breast cancer in 1506 women in the first 11 years after diagnosis. Although survival was poorer among African-Americans, in regression analysis, the difference between them and whites was largely explained by socio-economic status.

F. Paper that explore the influence of where people live
(Robertson, 2004)(232)

This study evaluated data from 1097 patients with breast cancer and 1223 with colorectal cancer in the UK between January 1997 and December 1998 to asses delay in diagnosis in those living further away from treatment centres.

The geometric mean time from presentation to treatment was 42 days. However, it was found that women living further away were treated faster than those living closer (P=0.011) although multilevel modelling discovered that this may be attributable to then receiving earlier treatment at hospitals other than the cancer centres. This study also found that older people were treated more quickly but that deprivation was not a significant factor. Under multilevel model evaluation only one organisational variable remained significant: that treatment was quicker for those referred to general hospitals than for those referred to cancer centres, and quicker still for those referred to private hospitals.

Diagnostic Difficulties

In a comparison of survival of women with breast cancer in 12 countries in Europe, the lowest five year survival rates were in Spain, the UK, Estonia and Poland (55–64%) (Sant el al, 1998 (233)). In the period 1985–1989, one year and five year survival rates in the UK had improved, but were still below the European average (by 3–4% and 6–9% respectively), although were higher than in Slovakia, Poland or Estonia (Quinn et al, 1998 (234)). Variation in survival between regions in the same country were observed, a finding that may in part be related to socio-economic indicators.

However, survival rates in the UK have continued to improve, and recent UK data indicate that five-year survival is now 75.9% among women who present with symptoms [(www.cancerresearchuk.org/aboutcancer/statistics/survival). (http://www.doh.gov.uk/nhsperformanceindicators/hlpi2002/NationalDocument.pdf)], and 94.1% among women who have cancer detected at screening (http://www.cancerscreening.nhs.uk/breastscreen/publications/ba00-01.html).

No relevant, good quality systematic reviews were identified.

Primary Studies
(Ruston, 2004)(33)

This study draws information from 85 women newly referred to four specialist breast clinics and their referring general practitioners in the UK in order to understand the referral decision-making process. The data was collected through semi-structured interviews with the patients and then separately with their matched doctor.

The study reported that the general practitioners felt under pressure from a ‘cloud of medical litigation’ that surrounds breast cancer and symptoms associated with it to refer all cases. Only 25 of the 85 cases reported trying to deal with the patient in primary care. There were three main categories identified where general practitioners would refer, the first that in the professional opinion of the practitioner the symptoms were indicative of cancer and urgent referral required. The second was that the nature of the lump was ‘sinister’ and referral decision was affect by patient anxiety, family history and medico-legal concerns over the implications of not referring the patient. The third category was that the practitioner felt that the symptoms were probably benign and referral was based on patient anxiety and concern over medico-legal consequences.

(The Bridge Study Group, 2002) (235)

The BRIDGE study evaluated the effects on patient management of breast disease guidelines issued to all general practitioners in the UK in January 1996. The practices in the BRIDGE study were randomised to receive either the breast lump or the breast pain guideline. During the study, general practitioners and practice nurses in the participating 34 practices were invited to take part in discussion seminars. The views of the participants were sought on the management of women with breast symptoms, the problems encountered, and influences on decisions about treatment. The transcripts of the recorded discussions were analysed to identify primary health care professionals’ views about patients presenting with breast problems. Referral decisions emerged as an overarching theme, which set the context for discussions with participants about the nature of clinical presentation.

The “easy” presentation was characterised by a single problem of the breast, where the clinical findings did not conflict with the history, in a woman with no or few preceding breast problems. The “difficult” presentation usually concerned a woman who had presented on numerous previous occasions, and who may have had previous investigation or surgery. Many practitioners expressed considerable uncertainty in establishing diagnoses for patients with breast symptoms on clinical grounds alone. For example, there was a reluctance to make an essentially histological diagnosis on the basis of palpation.

Doctors reported high levels of anxiety running through these consultations, not all confined to the patient. This sometimes resulted in cautious management strategies, perhaps with negative consequences for patients who were exposed to radiation during mammography, but it calmed the general practitioner’s own anxieties. The high level of patient and doctor anxiety about breast symptoms appeared to be a pervasive context for managing women presenting with these conditions. These levels of anxiety reflected underlying perceptions of risk, mainly of breast cancer. There are medico-legal issues about the liability for a delayed or missed diagnosis of breast cancer. Other comments however, suggested that both doctors and patients overestimated the predictive value of symptoms for breast cancer and also did not relate presentation and diagnosis to the overall natural history of the condition.

There was variation between general practitioners about the effects of their past experiences on current practice. Some were open about the fact that adverse previous experiences had had a major impact on subsequent referring behaviour. For example, a young woman with cyclic breast pain, who later had cancer, reduced a general practitioner’s referral threshold. Others highlighted a change in clinical practice resulting from having previously missed a diagnosis. For instance a lump was only suspected as being cancerous when a patient returned with the same complaint, and a lymph node was detected in the axilla after a more thorough examination. There was particular concern about “atypical” presentations, especially those in younger women or those that had culminated in a patient’s death. A case many years previously sometimes continued to have a strong effect on a clinician’s practice.

Risk factors were mentioned frequently, especially a family history of breast cancer. A positive family history was seen as a factor likely to raise anxiety in a woman presenting with a breast problem, and make it more difficult for the general practitioner to reassure her.

The availability and use of investigations in specialist clinics may undermine attempts to rationalise referrals. General practitioners do not deny the need to assess patients, but on occasions they view it as legitimate to arrange referral purely for reasons of reassurance. These general practitioners may be resistant to changing their clinical practice as they feel that they are making ‘safe’ choices.

Management of breast cancer is often complex and is an area in which general practitioners do not feel they have special skills. A single, and often atypical, case may have a profound influence on the way general practitioners manage their patients. Decision making about referral is often a consequence of a negotiation between patient and doctor. Attempts to modify clinical management of women presenting with breast symptoms must take account of these contextual issues, especially the high levels of patient and doctor anxiety.

(Watson et al, 2002) (71)

This cluster randomised controlled trial of educational interventions on general practitioner management of familial breast and ovarian cancer involved 688 general practitioners in 170 UK practices. Group A were provided an information pack and in-practice educational session, group B were mailed an information pack, and group C received no intervention at all. All general practitioner referral letters between March 1999 and December 2000 were audited and classified as appropriate or inappropriate referral.

The appropriateness of referrals improved among general practitioners who either received the guidelines alone (68.7% of referrals appropriate), or reinforced with an educational session (75.0% appropriate). In the group that did not receive the guideline or any other intervention, only 52.6% of referrals were judged appropriate.

(Burgess et al, 1998) (236)

In an interview study of 185 patients referred to a London breast clinic, referral did not occur at the first general practitioner consultation in 32 (17%). Delayed referral was observed more frequently among patients who were not aware of a lump at the time of presentation to the general practitioner (accounting for 44% of all cases of general practitioner delay). Patients experiencing general practitioner delay were younger (49 years vs. 55 years).

(McLeod et al, 1999) (237)

In this New Zealand study, 30 general practitioners were interviewed in depth to identify the key issues relating to the early detection and diagnosis of breast cancer in primary care. Following the interviews, a postal survey of a national random sample of 639 active general practitioners was undertaken, of whom 524 (82%) returned completed questionnaires.

The general practitioners reported that they were limited in their management of symptomatic women by the availability of services such as mammography and fine needle aspiration, and access to specialist breast surgeons or clinics. In some isolated rural communities, distance to services was a limiting factor. Some general practitioners used investigations to confirm the presence of a lump, or the nature of a lump. In the postal survey, 137 (27%) general practitioners personally aspirated cysts and 39 (8%) personally performed fine needle aspiration for diagnostic purposes. Most considered referral should occur either when a lump was palpated or after abnormal test results, although would refer women over aged 50 more promptly. In younger patients, recall and review were more likely.

Risk was viewed as associated with family history, although the definition of family history varied between respondents. There was a tendency to over estimate the impact of a first degree relative with breast cancer on the risk of cancer.

The key area of difficulty was reported as being the management of young women with lumpy breasts. Concern about the possibility of missing a malignant lump had to be balanced with the risk of causing unnecessary worry. Some general practitioners requested more information on the management of breast pain and nipple discharge.

12.4. Support and Information needs

12.4.1. Key Clinical Question

What are the relevant patient vulnerability factors? These factors concern the psychological and social factors that influence the patient’s ability to manage the consequences of referral for suspected cancer.

12.4.2. Evidence Question

In women attending primary care services with breast symptoms, which patient vulnerability factors, when compared with patients without vulnerability factors, are associated with the need for psychosocial support; and which are not?

12.4.3. Evidence Statement

There is little evidence about the support and information needs of women at referral. Before diagnosis, women are anxious and focused on quick referral and diagnosis (III).

General recommendations about the support and information needs of patients undergoing referral for suspected cancer are included in Chapter Seven. This section is confined to a consideration of the particular needs of women being referred with suspected breast cancer. There are very few studies of the needs of women suspected of having breast cancer at the time of referral, although many more studies have been undertaken relating to the time of diagnosis and after diagnosis. We discuss below a review that drew on studies undertaken at or after diagnosis, and also include information from the small number of studies that do consider the stage of referral.

Secondary studies

(Centre for Reviews and Dissemination, 1996) (232)

An Effective Health Care bulletin by the Nuffield Institute for Health and NHS Centre for Reviews and Dissemination (1996) offers a review of relevant trials that explore the information and communication needs of patients with breast cancer, as well as the psychosocial support required.

Information giving

The most common complaints by patients were about poor communication and inadequate information. Focus groups of patients revealed that they wanted information in both verbal and written forms about their cancer, treatment options, the likelihood of treatment success and possible side effects. Patients who are given more complete information showed greater satisfaction without an increase in anxiety.

Studies of consultations suggest that patients and their doctors may disagree about the adequacy of information given. Patients often feel they are not given sufficient information, while doctors tend to overestimate the amount of information they provide. Younger, better educated women, and those with better prognoses, tend to get more detailed information. Patients are likely to get more complete information when it is given in a structured way. They consistently find audiotapes of their consultation and information booklets about treatment helpful (grade of evidence range I-IIC).

Participation in decision-making

The fact that women want to be properly informed does not, however, imply that they want to be responsible for the final treatment decisions. The degree to which women wish to take an active role in decision-making varies between individuals and is affected by age, education and other social and cultural factors.

One study exploring the effects of choice between mastectomy and breast conservation surgery suggested that offering such a choice could cause distress (grade IIA). Other studies reported that a significant proportion of women found the process of making a choice problematic (grade IIA and IIC).

Psychosocial support

The bulletin identified 13 studies that assessed the effects of a range of psychotherapeutic interventions and also two critical reviews of the literature. These studies showed that psychotherapeutic counselling and educational interventions can improve quality of life and may possibly improve immune function and increase life expectancy. In general, interventions that focussed on past problems, as in the psychoanalytic model, were not found to be effective, whereas those that dealt with the woman’s current problems were more likely to be helpful. A more definitive statement about the impact of psychosocial interventions was not possible because of the poor quality of the studies, which were often small and poorly controlled. The multiplicity of types of intervention and outcomes made comparisons between studies difficult.

Cognitive/behavioural interventions

Cognitive/behavioural interventions, including psychotherapy, relaxation training, systematic desensitisation, guided imagery, pain control training, biofeedback and physical exercise, have mainly been used to reduce side-effects of cancer therapy such as nausea. They have been assessed in 21 RCTs. 16 of these studies demonstrated some degree of benefit, while the rest were equivocal.

Effectiveness of follow-up policies

The bulletin also reviews trials that explore the effectiveness of different follow-up strategies. Two RCTs from Italy and one from Britain compared general practitioner-based with hospital follow up Results from both trials suggested that patients followed up by their general practitioners experience the same quality of life as those cared for by specialist clinics, and that general practitioner follow-up was acceptable to both patients and general practitioners.

The provision to women of a contact number for the breast care nurse has been shown to lead to better quality of life and lower levels of psychological and physical morbidity than either routine care or support from a local voluntary agency.

(Centre for Reviews and Dissemination 2002) (192)

The Service Guidance Evidence review did not identify trials of interventions to improve communication between professionals and patients leading up to referral.

Primary Studies

We have very little evidence on need for information and support of women who are referred. There are studies of the reasons for delay in presentation of symptoms, and in reaction to investigation and diagnosis (Oktay, 1998), but the needs of women who are referred have not been adequately studied.

(Breakthrough Breast Cancer 2002) (238)

A qualitative study involving individual and group interviews was undertaken and did consider this question. Women had different levels of knowledge about breast cancer. The pre-diagnosis stage was distressing because of fear; women were extremely sensitive to what was said to them and how health professionals behaved. The focus at this stage was on quick referral for testing and diagnosis. Although no recommendation from the study dealt specifically with initial presentation and referral, it was recommended that women be given clear expectations of services. Highlighted in the study as particularly beneficial was 24-hour access to information, advice and psychosocial support pre-diagnosis and beyond and, in particular, encouragement to use such services.

Footnotes

14

Breast awareness means knowing what your breasts look and feel like normally. Evidence suggests that there is no need to follow a specific or detailed routine such as Breast Self Examination, but women should be aware of any changes in their breasts. See http:​//cancerscreening​.org.uk/breastscreen/breastawareness​.html for further information.