This study is an extension of a case-control study among all Swedish residents born in Sweden and aged 50 to 74 years between 1 October 1993 and 31 March 1995 18192021. The study was approved by the ethical review board at the Karolinska Institute, and by the five ethical review boards in other regions in Sweden. Women with incident primary invasive breast cancer were identified via the six Swedish Regional Cancer Registries. The women were contacted via their doctors and asked for written consent to be approached with a mailed questionnaire. The study identified 3,979 women with a diagnosis of invasive breast cancer, of whom 84% (3,345) participated. The primary reasons for non-participation were patient's refusal or doctor's refusal because of the patient's poor health. The mean interval from diagnosis to data collection was 4.3 months (standard deviation 1.5 months).

Controls were frequency matched by the expected age distribution among cases and identified through the Swedish National Population Register holding data on national registration number, name, address, and place of birth of all Swedish residents. The response rate among controls was 82% (3,455/4,188). Women previously diagnosed with invasive cancer (other than non-melanoma skin cancer) were excluded from the study (112 cases and 91 controls). Menopause was defined as the age at last menstrual period or age at bilateral oophorectomy, if one year or more prior to data collection. Premenopausal women (198 cases and 152 controls), women below the age of 55 years with unknown age at menopause considered premenopausal (202 cases and 101 controls), and women with missing information on body mass index (BMI; 14 cases and 45 controls) or age at first birth (5 cases and 1 control) were excluded.

In a second phase of the study, we retrieved information about histology and various other tumour characteristics from the medical records of all participating cases. Following a decision of the ethical review board of the University of Lund, written informed consent to retrieve this information was sought from cases in that region (n = 563), among whom 58 women did not provide informed consent. The medical records for 31 study participants could not be found.

Information from the medical records led us to exclude a further 58 cases with non-invasive breast cancer, 35 cases with previous cancer, one case with a cancer diagnosis other than breast cancer, and 19 cases diagnosed before or after the study period. The final study comprised 2,643 breast cancer cases and 3,065 controls.

Data on sociodemographic, anthropometric, reproductive and menstrual factors, use of oral contraceptives, medical history, lifetime physical activity and smoking habits, as well as recent (one year before data collection) dietary habits and alcohol use were collected by means of a postal questionnaire. Detailed information on use of menopausal hormone therapy, including timing and type of hormones for each treatment episode, was also requested and a colour chart displaying all preparations ever marketed in Sweden was included with the questionnaire to aid recall. In addition, approximately 50% of both cases and controls were contacted by telephone to complete missing or ambiguous responses, mainly on use of menopausal hormone therapy. Of all eligible controls, 14% did not return the questionnaire, but agreed to a telephone interview covering the most important items, including use of menopausal hormone therapy.

We obtained information on tumour histology from the pathology report in the medical record. Histology was classified after the invasive component as ductal (n = 1,888), lobular (n = 308), tubular (≥90% tubular component, n = 93), mixed lobulo-ductal (n = 58), medullary (n = 29), mucinous (n = 61), papillary (n = 10), adenocarcinoma not otherwise specified (n = 122), or unspecified tumours (n = 38). Information on histology was missing in 36 cases because the medical record was not identified, or the woman had not had a breast cancer operation. We analysed ductal, lobular, and tubular cases and all other cases were excluded from the analyses. Adenocarcinoma not otherwise specified was not analysed as we considered it an undefined subtype, and the other histology subtypes were too few to be analysed separately.

We used unconditional logistic regression to estimate odds ratios (ORs) with associated 95% confidence intervals (95% CIs) separately for ductal, lobular, and tubular cancer cases compared with controls. To formally test whether effect estimates were different for lobular or tubular cancer compared with ductal cancer we fitted models comparing lobular or tubular cases with ductal cases. For the tests comparing lobular or tubular cancer with ductal cancer, a two-sided p-value < 0.05 was considered a significant difference.

Use of menopausal hormone therapy was categorized according to type as medium potency estrogen alone (mainly estradiol or conjugated estrogens), medium potency estrogen with progestin, progestin alone (without concomitant estrogen), low potency oral estrogen (oral estriol without progestin), or local estrogen (cream or pessary, without progestin) for each treatment episode. Medium potency estrogen-progestin therapy was also classified according to regimen, that is, as sequential (progestin taken less than 16 days per 28 days) or continuous use (19 or more days per 28 days). We censored all exposure after a reference date, defined in cases as date of diagnosis minus 3 months, and in controls as the date of questionnaire arrival minus mean time from diagnosis to questionnaire arrival in cases, minus an additional 3 months.

We studied the effect of ever use, duration and recency of use. Analyses of the associations between use of specific types of menopausal hormone therapy and the risk of breast cancer were performed both for non-exclusive use (all users of that specific therapy), and restricted to exclusive users of that particular type, with never users of any menopausal hormone therapy as the reference group. No more than five women reported exclusive current use of progestin alone therapy, so we were unable to evaluate this therapy. Only 31 women had simultaneously used progestin with low potency oral estrogen or local estrogen, so this combination was not analysed separately, but these women were excluded from all analyses on exclusive use.

In the analyses on menopausal hormone therapy, women with an unknown age at menopause were excluded (296 cases and 301 controls) to avoid bias leading to underestimation of the associations as first pointed out by Pike and colleagues 22.

We adjusted all estimates for age. We also assessed the potential confounding effects of a range of other variables, including number of births, age at first birth, age at menopause, surgical menopause (bilateral oophorectomy), menopausal symptoms, BMI (weight(kg)/height (m)²) one year before data collection, height, socioeconomic status, smoking, and ever use of medium potency estrogen-progestin therapy. Confounding was defined as changing the point estimates more than 10% from the age-adjusted values. Only covariates confounding the effect of at least one of the exposures in the same table were included in the models. In addition, age at first birth and parity were adjusted for each other.

The analyses of the influence of age at menopause on the risk of different breast cancer subtypes were restricted to women with a known age at menopause that had not used menopausal hormone therapy apart from local estrogen before menopause in order to reduce confounding. The influence of BMI one year before data collection and the risks of different breast cancer subtypes was evaluated only among never users of menopausal hormone therapy apart from local estrogen – as BMI and menopausal hormone therapy are known to interact biologically to cause breast cancer 20. This interaction was studied in another approach, where study participants were cross-classified by their use of medium potency estrogen-progestin therapy and BMI, and the risks of ductal and lobular cancer were calculated. Analyses were performed using the SAS System, version 9.1 (SAS Institute, Carey, NC, USA).

Use of medium potency estrogen alone was similarly associated with increased risks of ductal and lobular cancer (Table 2). Due to few cases, tubular cancer was not possible to evaluate. Exclusive ever use of medium potency estrogen alone was similarly associated with the risks of ductal (OR 2.0, 95% CI 1.5–2.9) and lobular (OR 2.4, 95% CI 1.3–4.6) breast cancer. Use for more than five years entailed ORs of 2.9 (95% CI 1.6–5.2) and 3.1 (95% CI 1.0–9.5) for ductal and lobular breast cancer, respectively (Table 2). Ductal and lobular cancers were both associated with current use, and ductal cancer in addition with past use, but without significant differences between ductal and lobular cancer.

Use of medium potency estrogen-progestin was associated with increased risks of all three subtypes of breast cancer, with the highest risks among ≥5 years and current exclusive use (Table 2). These risks were significantly higher among lobular and tubular cancer compared with ductal cancer. Women who used medium potency estrogen-progestin for ≥5 years (of whom > 80% were current users) had an OR of 2.3 for ductal cancer (95% CI 1.6–3.3), OR 5.6 for lobular cancer (95% CI 3.2–9.7), and OR 6.5 (95% CI 2.8–14.9) for tubular cancer compared with those who never used menopausal hormone therapy (Table 2). Among past users, no clear association with any cancer subtype was discerned.

Non-exclusive users of both sequential and continuous medium potency estrogen-progestin had a median duration treatment of four years, compared with two years for exclusive users (data not shown). This was reflected in the higher point estimates for non-exclusive compared with exclusive ever use for lobular and tubular breast cancer (Table 2). Exclusive use of sequential medium potency estrogen-progestin therapy was not significantly differently associated with ductal, lobular or tubular cancer. Exclusive use of continuous medium potency estrogen-progestin therapy, on the other hand, was significantly more strongly associated with lobular compared with ductal cancer, when measured as ever use, <5 years use, and current use. Continuous use ≥5 years was associated with high point estimates for lobular (OR 5.9, 95% CI 2.5–14.0) and tubular (OR 8.2, 95% CI 2.4–27.5) cancer, but not significantly different from ductal cancer (OR 2.7, 85% CI 1.5–4.7). For lobular cancer the different risk estimates for exclusive ever use of continuous versus sequential use did not reach statistical significance (p = 0.36).

Non-exclusive and exclusive ever use of low potency oral estrogen was associated with an increased risk for lobular (OR 2.0, 95% CI 1.3–3.2) but not ductal (OR 1.2, 95% CI 0.9–1.5) or tubular (OR 1.0, 95% CI 0.3–2.8) breast cancer (Table 2). The increased risk was confined to <5 years of use and past users.

We found no associations between non-exclusive or exclusive use of local low potency estrogen therapy and either ductal or lobular breast cancer (Table 2). The estimates for tubular cancer were non-significantly above unity, with no trend of increased estimates for longer duration.

The risks associated with parity, age at menarche and age at menopause did not vary significantly between the three histological subtypes (Table 3). Increasing number of births was associated with a decreased risk of ductal but not of lobular cancer. Age at first birth was slightly more associated with lobular than with ductal cancer, but the difference was not significant.

Height was similarly associated with the three subtypes. Recent BMI and adult weight gain, evaluated among never users of menopausal hormone therapy apart from local estrogen, were associated with similar ORs for ductal and lobular cancer (Table 3). The association between recent BMI and tubular cancer was close to unity, but not statistically different from the estimates for ductal cancer.

Having a mother or sister with breast cancer seemed more strongly associated with lobular (OR 2.4, 95% CI 1.8–3.4) than with ductal (OR 1.8, 95% CI 1.5–2.2) breast cancer (p for heterogeneity = 0.06). Previous operation due to benign breast disease was associated with slightly higher risk for lobular (OR 1.9, 95% CI 1.4–2.6) compared with ductal (OR 1.5, 95% CI 1.2–1.8) cancer. The OR for tubular cancer was near unity, but not statistically different from ductal cancer. Use of oral contraceptives (99% of all users stopped v5 years before the reference date) was not associated with any of the three histological subtypes of breast cancer among these postmenopausal women (Table 3). Recent high alcohol consumption was associated with tubular breast cancer (OR for ≥10 g/day 3.1, 95% CI 1.4–6.8). For lobular cancer a tendency of higher ORs with increasing alcohol consumption was seen, but no category had a significantly increased OR (Table 3). We found no indication of an association between alcohol consumption and ductal cancer. Recent smoking was borderline associated with risk of ductal cancer (OR 1.2, 95% CI 1.0–1.4), and the OR for tubular cancer was significantly lower compared with ductal cancer (OR 0.8, 95% CI 0.5–1.2) but not significantly different compared with controls (Table 3).

We stratified exclusive ever use of medium potency estrogen-progestin into three categories of BMI (Table 4). The increased risk for both ductal and lobular breast cancer with medium potency estrogen-progestin use seemed to be confined to women with BMI ≤27. Among women with BMI < 22, exclusive ever users of medium potency estrogen-progestin had OR 2.0 (95% CI 1.2–3.3) for ductal cancer and OR 3.4 (95%CI1.4–8.3) for lobular cancer compared with never users of menopausal hormone therapy