Multiple risk factors for MBC have been described and studied and these fall into two main groups, non-modifiable and modifiable. The former include age and ethnicity and in common with the majority of solid tumours increasing age is a major risk factor for MBC. Anderson et al. analysed data from the Surveillance, Epidemiology and End Results (SEER) to compare 5494 MBC with 835,895 FBC cases and reported that the median age at diagnosis was 67 for the former and 61 in the latter group [1]. When age specific incidence rates were plotted there was a striking difference between males and females. There was a rapid rise up to age 50 in women with an inflexion at age 50 (Clemmesen’s hook) and then a slower rate of increase. For MBC the rate rose steadily so that with advancing age there was narrowing of the male/female incidence ratio.

Kreiter et al. analysed 104 different populations including North America, Europe, Russia, Asia and Australasia [2]. Age-adjusted incidence rates were calculated for both MBC and FBC in each population to compare 5-year age group incidence rates for both males and females. They found a worldwide correlation between incidence rates of FBC and MBC which was highly significant when MBC was compared with that in women aged ≥50 and also those <50. Clemmesen’s hook was also seen in males, albeit to a lesser extent but at age 60 rather than 50 for females.

There are wide geographical variations in male/female incidence ratios, with the lowest in Western countries and the highest in Africa. Using data from Cancer in Five Continents, Ly et al. compared international incidence rates in MBC and FBC based on 8681 and 1140,00 cases respectively [3]. The lowest ratios were found in Israel (70) and Iceland (84) with the highest ratios in the UK (153) and Japan (169).

The situation regarding risk for ethnic groups within one country is more complex. When Anderson et al. compared age-adjusted incidence trends rates in SEER data from 1973–2005 they stratified the FBC by age <50 and age ≥50 and found that the trend was stable in males but increasing in females [1]. The male/female incidence ratio was higher among the black cases.

In the 2013 US Census, 78% of the population described themselves as white, 17% Hispanic and 13% were black. There has been a series of publications from the US describing the ethnic background of MBC cases based largely on SEER or Veterans Administration data [4–7]. These are outlined in Table 3.1 and suggest that there is no over-representation of African American MBC cases.

There is however a lot of anecdotal information suggesting an increase in risk of MBC in Africa. Series reported from Sub-Saharan Africa are outlined in Table 3.2 and show an increased ratio of MBC/FBC cases compared with the western world with 1.9–8.9% of mammary malignancy occurring in males [8–21]. Amir et al. examined incidence rates in Tanzania both before and during the AIDS epidemic which started in 1983. After 1982 there was a fall in the number of MBC cases but a rise in FBC. There was a significant widening in the male-female ratio which fell from 0.09:1 to 0.03:1.

In North Africa a different pattern emerges as is shown in Table 3.3. El-Gazayerli and Abdul-Aziz from Alexandria reported a high incidence of MBC which may have resulted from endemic bilharziasis (schistosomiasis) which was known to have affected seven out of eight of the traceable MBC cases [22]. After penetration of the skin S. mansoni migrates to the liver, feeding off erythrocytes, growing up to 1 cm in length and damaging the liver with ensuing hyperestrogenisation. A similarly large proportion of MBC was reported from Sudan, where 15% of the population suffer from schistosomiasis [23]. Similarly in Ethiopia almost 14% of breast cancer cases were male [24]. A subsequent much larger Egyptian study reported that MBC comprised only 0.89% of total breast cancer cases [25]. MBC rates in Morocco, Tunisia and Libya were slightly elevated compared with Europe [26–28].

Jewish ancestry does constitute a significant risk factor for MBC. Steinitz et al. compared 187 cases reported to the Israel Cancer Registry with 194 in the U.S. Third National Cancer Survey together with contemporaneous FBC rates [29]. In both countries MBC rates were related to those of FBC and the rate of MBC in Israel was higher than in the USA. When Mabuchi et al. carried out a case-control study of 52 MBC with 52 controls in 5 US cities, they found a significantly increased risk of cases being Jewish [30].

Following the isolation of the breast cancer susceptibility gene BRCA2, Couch et al. analysed 50 MBC cases and found pathogenic mutations in 7 (14%), of whom 4 were of Ashkenazi Jewish origin [31]. Brenner et al. studied 131 Israeli MBC cases and reported that there was an overrepresentation of Ashkenazi compared with Sephardic Jews with an almost doubling of risk [32]. As part of a large series of 10,000 cases, Frank et al. looked for mutations in 76 MBC cases and found them present in 21 (28%) of whom 11 39%) were Ashkenazim [33].

It can be difficult to precisely allocate individuals into a social class and proxy approaches have been adopted including years of education and job title. Nevertheless, a consistent effect of sociodemographic status (SES) on risk can be seen. In a relatively small case control study of 21 cases and 82 controls D’Avanzo & La Vecchia reported that there was a significant relation between years of education and risk [34]. Compared with men who had <7 years of education, those who had studied for ≥12 years had an odds ratio of 2.6. There was a threefold increase in risk for men of high versus low social class.

Using a 1% sample of all the US deaths in 1986, Cocco et al. compared 178 MBC with 1042 age matched controls and determined socioeconomic status based on longest held occupation [35]. There was a significant association of increased risk with SES with odds ratios of 1, 1.5 and 2.3 in low, medium and high categories respectively.

Hansen investigated members of the Danish National Pension Fund and identified 230 MBC cases [36]. For each, 56 controls of similar age were selected and among the variables studied were exposure to vehicle fuel and fumes together with social class. The latter was derived from job title with definitions from the Danish Institute of Social Sciences: Group 1 corporate managers and academics, Group 2 proprietors and small business managers, Group 3 nurses and technicians, Group 4 skilled workers, Group 5 unskilled workers. There was a higher but not statistically significant increase in risk for groups 1 and 2 compared with group 5.

Individuals with Klinefelter’s syndrome (KS), comprise 1 in 1000 of newborn boys. They have at least one X chromosome added to the normal XY karyotype (most frequently 47XXY). This is associated with testicular dysgenesis, aspermia, gynaecomastia, and a variety of psychological disturbances. Wang et al. analysed the endocrine profile of 19 KS patients [37]. Plasma secretion rates of total and free testosterone were low, plasma estradiol, LH and FSH levels were elevated and there was increased peripheral conversion of testosterone to estradiol. In contrast, secretion and binding of estradiol was normal. Fluctuations in blood levels of LH, FSH, testosterone and estradiol were less marked than in healthy men.

In 1971 Harnden et al. who had collected 150 MBC cases from Birmingham and Edinburgh reported that 5 (3.3%) were chromatin positive [38]. Hultborn et al. studied 93 unselected MBC patients from the western Sweden and used fluorescence in situ hybridisation (FISH) on intact nuclei lymph nodes and reported a 7.5% prevalence rate of KS in MBC [39]. This suggested that there was a 50-fold increased risk in KS compared with XY males. Median age at diagnosis was 72 years in both groups of patients.

Others conducted follow-up studies of known cases of KS. Price et al. reported mortality data of a cohort of 466 KS followed for 25 years [40]. There were 2 fatal cases of MBC which represented a significant increase in risk, similar to that in the female population. From the Danish Cytogenetic Register Hasle et al. collected a cohort of 696 KS cases in which 39 cancers were diagnosed but no breast cancers [41]. Swerdlow et al. reported 163 deaths in a cohort of 646 KS patients with an RR of 1.63 (1.40–1.91) [42]. This was predominantly due to diabetes and vascular disease but there was however an increased risk of lung and breast cancer in patients with KS. In a large UK cohort comprising 3518 men with KS incidence and mortality were compared with that of men in the national population [43]. The standardised mortality ratio (SMR) for all cancers was 1.2. There was a higher mortality rate for lung cancer (SMR 1.5) and for MBC (SMR 57.80. Among those with 47XXY mosaicism the SMR was 222.8. In contrast there was a reduced risk of prostate cancer (SMR 0).

The relationship between gynaecomastia and MBC remains problematic, in part because of the difficulty defining gynaecomastia. It means male breast enlargement but the majority of cases have enlargement because of increased subcutaneous fat (pseudogynaecomastia) rather than an increase in the glandular issue of the breast. Cakan & Kamat defined gynaecomastia in boys as being glandular tissue >5mm [44]. In contrast Daniels & Layer, reporting a series of 175 males >16 years with gynaecomastia had a threshold of 20 mm for making the diagnosis [45]. Such measurements are largely lacking in the epidemiological literature.

Simon et al. classified gynaecomastia based on size and described three grades [46]. Grade 1 is visible breast enlargement without redundant skin. Grade 2A is moderate breast enlargement but no skin redundancy and grade 2B moderate enlargement with minor redundant skin. Grade 3 gynaecomastia was defined as gross enlargement with skin redundancy and breast ptosis. Again, this classification has not been used epidemiologically.

Sirtori & Veronesi reported 218 cases of gynaecomastia of whom 108 had a biopsy [47]. No histological abnormality was found in 47 (44%) suggesting that these were pseudogynaecomastia cases. Among the true cases, the most obvious change was in the stroma with increased cellularity and hyalinisation in the fibrous component and a change in the loose periductal connective tissue to the mantle firm found in the female breast. There were variations in the proportion of the two components and sometimes fibrosis was total with adjacent ducts resting thereupon. Cysts were present in 105 (97%) of specimens but lobule formation was seen in only one case. The ductal tissue showed both atrophy and of proliferation with elongation and occasional twisting. Ducts comprised either two or three layers with marked proliferation in the basal layer. Mitotic activity was observed in 89 (82%) sometimes giving rise to true papillomas.

Gynaecomastia is common and shows temporal changes in incidence. It is found in up to 90% of neonates as a result of transplacental transfer of steroid hormones. This may persist for several months. There is another peak at puberty and up to 60% of 14-year old boys have gynaecomastia possibly because of relative delay in full testosterone synthesis or a temporary surge in aromatase activity, or variation in estrogen sensitivity. With increasing age and decreasing free testosterone there is another increase in incidence, correlated with body mass index, so that in autopsy studies up to 50% of the males had gynaecomastia [48]. Apart from the effects of aging and obesity, multiple diseases have been implicated including hyperthyroidism, liver disease, Klinefelter’s syndrome and tumours of the testis and adrenal gland. Individuals who have starved and then refeed are also at risk, as are those undergoing chemotherapy, suffering from HIV and men with chronic illness.

In addition to this age-related increase in incidence of gynaecomastia the situation is exacerbated by the pharmacological interventions to which the aging body is prone. Many of the reports were anecdotal, so Deepinder & Braunstein carried out an assessment of the quality of the evidence, assigning a good, fair or poor score to each of the publications [49]. What emerged was that most of the reported associations were derived from poor quality evidence. Those drugs that were definitely associated with gynecomastia are summarised in Table 3.4. Approximately 10% of patients treated with spironolactone develop gynaecomastia and up to one in five of men given cimetidine have this problem.

In a Scandinavian study of risk factors for MBC, Ewertz et al. collected 282 cases from Denmark, Norway and Sweden between 1987 and 1991 [50]. They used self-questionnaires on 156 MBC cases who agreed to participate and with 468 male controls matched for country and year of birth. Significant risk factors that emerged included family history (OR 3.3), obesity, BMI > 30 (OR = 2.1), and diabetes (OR = 2.6). They asked in detail about gynaecomastia and found that many cases confused signs of breast cancer with gynaecomastia and suggested that this might explain the strong association previously reported.

Because gynaecomastia may be associated with absolute or relative estrogen excess Olsson et al. followed a cohort of 446 affected men to determine whether this was associated with increased risk of malignancy [51]. All had histologically confirmed gynaecomastia. Before being diagnosed with gynaecomastia, eight had prostate carcinoma, two unilateral MBC and one Hodgkin’s disease. There was a total of 8375.2 person years of follow-up and during this time 68 cancers were diagnosed compared with an expected 66.07 (SIR = 1.03 (95% CI 0.80–1.30). There was a significantly increased risk of testicular cancer; SIR = 5.82 (95% CI 1.20–17.00) and squamous cell carcinoma of the skin; SIR = 3.21 (95% CI 1.71–5.48). No new cases of MBC were diagnosed but the diagnostic surgery may have substantially reduced this risk.

Satram-Hoang et al. conducted a pilot case-control study with 44 MBC and 77 age- and ethnicity- matched controls. Subset of the male breast cancer cases (n = 44) and age- and ethnicity-matched controls (n = 77) [52]. Of the MBC cases, 20% were overweight/obese with comorbidity present in 55%.There was a significantly greater proportion of MBC with gynaecomastia, family history of cancer, antibiotic use and smoking compared with the controls. The authors conceded that the study was underpowered and a large collaborative effort was needed.

That was provided by Brinton et al. who accessed 26 million hospital discharge records from the US Veterans Administration covering 1969 to 1996 and determined the relationship between confirmed medical conditions and risk of subsequent MBC [53]. There were 4,501,578 males aged between 18 and 100 years, and among these 642 MBC cases were diagnosed. The relative risk of various medical conditions are summarised in Table 3.5 which shows that gynaecomastia and Klinefelter’s syndrome were associated with the highest relative risks. Other conditions that had been previously reported to increase risk of MBC such as thyroid diseases, smoking-related conditions, liver cirrhosis, prostatic hyperplasia, and fractures did not affect risk in this study. Although the link with diabetes disappeared after adjustment for obesity, that with gynaecomastia persisted. This risk was highest for those diagnosed with MBC in the following 2–5 years but was also evident in those diagnosed >5 years later. The problem remains concerning the lack of precision since these cases will have been a mixture of true and pseudo gynaecomastia. Prospective studies of radiologically or histologically defined cases are going to be necessary,

Various studies have shown a very different pattern of steroid receptor expression in gynaecomastia and MBC. Sasano et al. investigated ER, PR, and AR in 30 gynaecomastia specimens and 15 MBC samples [54]. All the MBC showed strong cytoplasmic staining for ER/PR compared with only 11/30 (37%) of gynaecomastia. There was however nuclear ER/PR expression in ductal cells of all the gynecomastia specimens. AR was present in 13/15 (87%) of MBC and all cases of gynecomastia. There was a significantly higher proportion of ER/PR/AR-positive cells in gynecomastia compared with MBC. Interestingly, there was a positive association between AR and ER/PR in gynecomastia but a significant inverse association in MBC. Increased aromatase expression in the stromal cells is considered to contribute to the increased in situ estrogen concentration and the development of male breast carcinoma.

Ferreira et al. measured prolactin receptor (PRLR) expression in 30 gynaecomastia and 30 MBC specimens [55]. Additionally they determined ER, PR and AR status and male breast carcinoma (MBC). There was PRLR positivity in 20% of the gynaecomastia cases and in 60% of MBC specimens. The PRLR staining was mostly in luminal cell borders of gynaecomastia whereas it was heterogeneous and predominantly cytoplasmic in the MBC. The quantitative and qualitative differences in PRLR expression in gynaecomastia and MBC also suggests that the former is not necessarily a forerunner of the latter.

Kornegoor et al. examined tissue microarrays from 46 gynaecomastia specimens using IHC for a range of epitopes including ER, PR, HER2, AR, CK5, CK14, cyclin D1, and Bcl-2 [56]. They found a consistent IHC staining pattern of one myoepithelial and two epithelial cell layers with a distinctive immunohistochemical staining pattern. There were vertically oriented cuboidal/columnar cells in the intermediate luminal layer which were ER/PR +ve and expressed Bcl-2 and cyclin D1. The cells of the inner luminal layer were smaller, usually ER/PR−ve, Bcl-2−ve, CK5+ve and often CK14+ve. In contrast in the DCIS cases that they examined the cells were ER+ve and CK5/CK14 −ve. This suggests that there are different cell compartments in gynaecomastia so that it might not be an obligate precursor of MBC.

There is a substantial body of evidence that subnormal testicular function is an important risk factor for MBC. In 1963 Schottenfeld et al. published the first case-control study with 53 MBC cases who were interviewed and two control groups with gynaecomastia (126) or colon cancer (154), matched for age, time of diagnosis and who had been treated in the same hospital as each MBC case [57]. Participants were asked among many questions whether they had had mumps and the age at which this had happened. Four of the cases reported mumps and three of them were diagnosed post-puberty. One of the gynaecomastia controls reported mumps but none of the colon cancer controls had been affected.

Keller collected 181 MBC cases from Veterans Administration Hospitals and by random dialling of other veterans selected two groups of controls matched for age and location, with one group having a non-malignant diagnosis (181) and the other comprising males with bladder or renal cancer (181) [58]. Of the cases, 19 (11%) had a clinical diagnosis of atrophic, enlarged or inflamed testes compared with 22 (12%) of the cancer controls. Nicolis et al. reported a case of MBC who had mumps orchitis at age 16 and was subsequently found to have aspermia [59]. At the age of the age of 41 he developed type II diabetes and was diagnosed with MBC aged 47.

In Mabuchi’s study there were 52 MBC cases and 52 controls, matched for age, ethnicity and marital status [30]. They reported a significant association of MBC with mumps diagnosed at age ≥20 years, Of the cases, seven had been employed in hot environments – steel mills, blast furnaces and rolling mills whereas none of the controls had worked at such sites. Casagrande et al. compared 75 non-Hispanic white MBC patients with 69 neighbourhood controls aged within 5 years of each case. In both groups 3 men had been diagnosed with mumps at age ≥20 years [60]. Undescended testis was reported by three cases and one control.

Olsson & Ranstam reported that testicular damage was reported by 6/95 (6%) of MBC cases compared with 3/383 (1%) lung cancer controls and 2/69 (3%) non-Hodgkins lymphoma controls [61]. In a Franco-Swiss collaboration, 91 MBC cases were compared with 255 cancer controls matched for age and year of diagnosis [62]. Of the controls, 91 had colon cancer. 91 haematological malignancy and 73 had basal cell carcinoma. Working in a hot environment was reported by 5 (5%) cases and 5 (2%) of the controls.

Thomas et al. conducted standardised personal interviews of 227 MBC cases and 300 age-matched controls selected by controls by random digit dialling for those aged <65 and Insurance records of older men [63]. There was no excess of mumps in the cases 6 (3%) versus 10 (3%). Nevertheless, undescended testis was reported by seven cases and one control giving a significantly elevated relative risk of 11.6. Additionally, more of the cases had prior inguinal herniorraphy, orchidectomy and testicular injury. In a relatively small study of 21 MBC cases and 82 hospital controls, D’Avanzo reported the relative infertility of the cases [34]. Of the controls 25 (31%) had no children compared with 15 (67%) of the MBC cases. In another small study from Athens, Petridou et al. reported a high ambient working temperature in 6/23 (26%) of cases compare with 22/76 (29%) of controls (Table 3.6) [64].

The recognition that small studies had inadequate statistical power prompted collaborative research with cohort studies which yielded more definitive results. Brinton et al. from the National Cancer Institute examined the U.S. Veterans Affairs database of 26 million discharge records between 1969 and 1996 [53]. They found 643 MBC cases diagnosed >1 year after discharge out of 4,501,578 veterans and were able to make adjustment for age, ethnicity, year of diagnosis, length of follow-up, and number of consultations. A past history of orchitis or epididymitis was associated with a significant increase in relative risk of MBC (RR =1.30, 95% CI 1.05–1.60) (Table 3.7).

The Male Breast Cancer Pooling Project analysed data from 11 case–control and 10 cohort studies [65]. There were 2405 MBC cases with 1190 in case–control and 1215 from cohort studies with 52,013 controls. For individuals with cryptorchidism the odds ratio was elevated but without statistical significance (OR = 2.18; 95% CI = 0.96 to 4.94) and the same was true for orchitis (OR = 1.43; 95% CI = 1.02 to 1.99). For those who had never fathered children there was a significant increase in risk (OR = 1.29; 95% CI = 1.01 to 1.66).

Although there are variations in the findings, the majority of studies do suggest that testicular damage, whether traumatic, viral, vascular or ambient temperature-related does lead to an increased risk of MBC. There is the possibility that testosterone replacement in high risk cases might improve quality of life and possibly reduce the likelihood of development of MBC.

In 1968, Symmers described 2 unfortunate cases of birth males who had undergone gender realignment and taken estrogens, oral contraceptives and implants to promote breast enlargement [66]. They both led a demimondaine existence and developed MBC at the age of 30, having been taking estrogens for 5 years. Sadly both succumbed rapidly to aggressive disease. Kanhai et al. investigated the histological changes in 14 castrated males receiving estrogens and compared these with alterations in two men taking antiandrogens for prostate cancer [67]. In the latter group they reported some development of acini and lobules. Among the transsexuals given progestins and estrogens they observed full development of acini/lobules with pseudolactational changes. This suggests that a combination of progestin and estrogen are required to mimic normal female breast morphology in the male.

There have now been several other reports linking estrogen therapy with MBC in transsexuals and these are summarised in Table 3.8 [66, 68–75]. It is striking that these cancers occurred in relatively young individuals with the oldest case of invasive cancer being 58. All tumours were invasive ductal cancers (IDC) and there was a disproportionately large number of ER−ve carcinomas (50%). When Gooren et al. reported follow-up of a cohort of 2307 male to female transsexuals in 2013 there was only one case of MBC and they concluded that cross-sex hormone administration did not increase the risk of breast cancer development. By 2015 they had observed two further cases of what was a relatively rare complication of administering estrogens to males.