Introduction

Globally, breast cancer (BC) is the most commonly diagnosed cancer among women (age-standardised rate of 46.8 per 100,000), and ranks second in cancer related mortality (16.8 per 100,000). In India, BC accounts for 13.6% of all malignancies and is responsible for 10.7% of deaths due to cancer in females1. BC is a multifactorial disease influenced by a number of factors, including obesity, late menopause, early menarche, and advanced age at first pregnancy2. Notably, most of these factors relate to prolonged exposure to endogenous and exogenous estrogen3. It is also believed that prenatal exposure to elevated levels of estrogen in utero might affect the risk of developing BC later in life4. The second to fourth fingers/digits ratio (2D:4D) has been proposed as a biomarker to measure prenatal exposure to the sex hormones5,6. The 2D:4D is higher when prenatal exposure to estrogen is higher and lower when prenatal exposure to testosterone is more7, as observed in mouse model as well8. Additionally, factors such as the conceiving woman’s age, body mass index, race9, and stress level10 affects in utero levels of sex hormone. Epidemiological studies suggests 2D:4D ratio to be positively associated with the BC risk4,11,12,13,14,15, which presents an area for further exploration.

Given the significant burden of BC worldwide, its risk factors have been widely studied across various populations. This might not be true for 2D:4D ratio, which hasn’t been extensively explored as a risk factor for BC. Even though there are few studies related to 2D:4D and risk of BC, majority of them have only compared means of 2D:4D amongst cases and controls, and there is inconsistency in results observed for 2D:4D ratio of left hand, right hand, as well as difference in ratio of right and left 2D:4D ratio (DR−L)4,11,12,13,14. Furthermore, to date there are few to no studies from the south Asian countries. There is also lack of studies showing the strength of association of BC risk with 2D:4D, and DR−L. Moreover, none of the studies have observed the importance of 2D:4D measurements in BC risk after stratification of menopausal status. This gap highlights the need for studies that evaluate the strength of the association between BC risk and 2D:4D ratios including DR−L measurements, while considering the menopausal status.

The current study aims to fill these gaps by exploring the association between the 2D:4D ratios (for both the left and right hands) and the risk of developing BC. We conducted a large-scale, multi-centric case-control study within the Indian population, after considering the important confounders. Additionally, we investigated whether this association varies between premenopausal and postmenopausal women, contributing valuable insights to the understanding of BC risk factors.

Materials and methods

Study design and study participants

For studying the risk factors of BC, hospital-based case-control study was conducted. The study was conducted across different hospitals of the Tata Memorial Centre, India: Tata Memorial Hospital, Mumbai; Advanced Centre for Treatment, Research and Education in Cancer, Navi Mumbai; Mahamana Pandit Madan Mohan Malaviya Cancer Centre/Homi Bhabha Cancer Hospital, Varanasi; as well as Nargis Dutt Memorial Cancer Hospital, Barshi. The enrolment of study participants was done from the year 2019 to 2023.

All females aged 21–67 years who came to the hospital (patients or visitors) and who were residents of India for at least a year were eligible for the study. Cases were histopathologically confirmed primary BC patients who were diagnosed no more than 6 months before the time of enrolment. Controls were selected from the visitors who came along with the patients (this includes friends, family, neighbours etc.) and were frequency-matched to cases for age (± 10 years) and current residential zone (North, East, West, South, and Central). In order to be eligible for the study the visitor controls had to be cancer-free with no prior history of cancer and having no first degree relative (mother, sister or daughter) who was diagnosed with BC or any other gynaecological cancer. Any female who was pregnant or breastfeeding at the time of enrolment were excluded from the study.

For the purpose of data collection, a structured questionnaire was utilised. Interviewers were trained to collect information regarding socio-demographic factors (income, education, religion, and so on), residential history, reproductive history (including menopausal status, parity, and so on) and anthropometric measurements. Interviewers visited various out-patient departments and wards to find eligible study participants, for cases the interviewers went to the BC disease management group (DMG), while for controls all other DMGs excluding breast as well as other gynaecological DMGs were visited. All the eligible participants were approached and explained about the study and those who agreed to participate were provided with a written informed consent.

Digit/finger measurement

The interviewers recorded the index and ring finger length measurements of both hands for all the study participants. The measurements were taken using calibrated digital Vernier calipers.Detailed guidelines were prepared to make sure that the measurements were recorded in a similar manner across all study centres16. Participants were made to keep their hand on a flat surface, after removing any jewellery/accessories that could hinder the measurement process, with their palms facing upwards. As the Vernier calliper had two jaws, one fixed and other movable, the interviewer kept the fixed jaw at the tip of the finger to be measured and extended the movable jaw towards the basal crease of the finger (nearest crease where the finger meets the palm) (Supplementary figure S1). The measurements were then used to derive the 2D:4D of both hands, by dividing the length of index finger to ring finger. The DR−L was calculated by subtracting 2D:4D of left hand from the 2D:4D of right hand.

Statistical analysis

Unpaired t-test (for continuous variables) and chi squared test (for categorical variables) were used to check whether there is a difference between cases and controls in terms of baseline characteristics. In order to examine the confounders for our models we used the Stata post-estimation command “chest” after fitting all the potential confounders in a multivariate unconditional logistic regression model17. Based on the results (Supplementary figure S2-S4) we decided to estimate odds ratio (OR) and their corresponding 95% confidence intervals (95% CI) for the association of right and left 2D:4D ratios and difference between right and left 2D:4D ratios with the risk of BC using univariate logistic regression model, since none of the risk factors affected the risk estimates much. Due to unavailability of cut-off values in the literature, to evaluate the dose-response with increase in 2D:4D ratio, we created quintiles for left and right 2D:4D based on cut-off values of the control population. While for DR−L, tertiles were created due to sample size issue in the quintiles. The analysis was also performed separately for premenopausal, and postmenopausal women. Correlation between age at diagnosis and 2D:4D ratio as well the DR−L was also estimated using Pearson’s correlation coefficient. All the analysis was performed on Stata version 15.018.

Ethics approval

This study was approved by the Institutional Ethics Committee of the Tata Memorial Hospital, Mumbai (Approval ID: 3302). The research was conducted in accordance with the Declaration of Helsinki.

Consent to participate

Informed consent was obtained from all individual participants included in the study.

Results

We performed analysis on a total of 3361 study participants (1638 BC cases and 1723 controls). Comparison between cases and controls for baseline characteristics has been presented in Table 1. The mean age of cases was 48 years and for controls it was 47 years. More than half of the cases were postmenopausal, whereas there were higher proportion of premenopausal women in the controls. In both cases and controls, majority of the study participants received more than or equal to five years of education (72.16% and 81.08% respectively). Irrespective of the disease status, majority of the study participants resided in urban parts of the country. Around 85% of the cases reported no family history of any cancer, as opposed to 56% amongst the controls. Cases and controls were not very different in terms of average body mass index, age at menarche, age at first full-term pregnancy, age at last pregnancy, parity, number of abortions, number of miscarriages and birth residential zone.

The distribution of participants based on their ring finger length (right and left hand) was similar for cases and controls, although controls seemed to have higher peak for right ring finger length. The distribution of participants for ring finger of right hand was slightly left-skewed. For the index finger length of the right hand, the data followed almost normal distribution, with the peak slightly higher for the controls as compared to cases. Similarly, for the length of index finger of the left hand, the peak was higher for controls (supplementary figure S5-S8).

Table 2 presents the association of BC with 2D:4D ratio of left and right as well as DR−L. We observed significantly inverse risk of BC amongst premenopausal women in the fifth quintile (0.71, 95% CI: 0.52–0.96) of 2D:4D ratio of the right hand. While for postmenopausal women there was positive association of BC with right 2D:4D in the fifth quintile (OR: 1.21, 95% CI: 0.00–1.63.00.63), although not significant. For the left 2D:4D, we observed non-significant slight increased risk of BC in the highest category amongst the premenopausal women (OR: 1.12, 95% CI: 0.82–1.52). Decreased risk was observed for BC with increasing DR−L (OR: 0.81, 95% CI: 0.68–0.95), which was consistent among premenopausal women but not in postmenopausal women. We also observed a slight negative correlation between age at diagnosis of BC and right 2D:4D ratio (correlation coefficient=−0.01, p-value = 0.75), as well as for DR−L (correlation coefficient =−0.01, p-value = 0.61) (results not shown).

Discussion

2D:4D is a proposed biomarker for prenatal sex hormone exposure. Research indicates a potential link between 2D:4D and cancers of prostate19, brain20, gastric21, and breast12,14. In the present study, an inverse association between right 2D:4D, difference in 2D:4D between right and left hand, with the risk of developing BC was observed amongst the premenopausal women. However, for postmenopausal women, non-significant positive association was observed for right hand 2D:4D. For the left 2D:4D ratio, we found slightly increased risk of BC only in the premenopausal women belonging to the highest category of 2D:4D, although not significant.

Some of the existing studies report only right hand 2D:4D4 to be positively related with risk of BC, while some others suggest higher left hand 2D:4D increases BC risk12, others found both right and left hand mean 2D:4D to be higher in cases than in controls13,14. For DR−L, two studies have observed inverse relation with BC risk12,14, while others reported no significant relation (Supplementary Table 1). However, none of the studies have stratified participants by menopausal status and only study by Muller et al. calculated the hazard ratio (HR) for 2D:4D, DR−L and the risk of BC, where they reported a HR of 1.12 for left 2D:4D, while for right hand 2D:4D they did not find any association with BC. Similar to our study results, they found DR−L to be inversely associated with BC (HR = 0.90), and speculated the role of in utero testosterone exposure12. Previous studies have also observed negative correlation between age at diagnosis of BC and right 2D:4D ratio as well as difference in the right and left ratio4,12,13, consistent with the trends observed in the current study.

The suggested theory for relation between BC and digit ratio is that the 2D:4D is positively correlated with in utero estrogen exposure and that higher intrauterine estrogen influences development of BC in offspring22. It is also believed that genes like HOX family genes, WNTs, and FGFR genes, that are thought to be involved in differentiation of 2D:4D ratio during the foetal stage, are also involved in the mammary gland initiation and breast carcinogenesis23,24. The ethnic variation related to 2D:4D ratio has also been observed, indicating involvement of genetic polymorphism in the phenotype25.

Moreover, the prenatal exposure to sex hormones has been shown to be independently associated with the risk of BC. The exposure to diethylstilboestrol, a synthetic estrogen consumed during pregnancy has been associated with increased risk of BC after the age of 40 years, but no increase in risk was observed for younger women26. Maternal preeclampsia which is associated with high maternal concentration of androgens and low concentration of oestrogen has also been shown to have inverse association with BC risk27.

The association of 2D:4D ratio and difference in 2D:4D ratio for right and left hand with BC development was not consistent across the available studies. In our study significant association was observed only for premenopausal but not for postmenopausal women, future studies are required to confirm these findings using larger samples. Even though prenatal exposure to sex hormones may be a strong risk factor, 2D:4D which is used as proxy measure for this exposure may not be a very strong risk factor for BC, since neither of the previously available literature reports, nor our study could find a strong strength of association. In order to measure this weak association (with lowest confidence interval of HR reported being 1.02)12, one will need a huge sample size of more than 150,000 cases and more than 160,000 controls (supplementary figure S9).

Strengths and limitations

Our study has the strength of being a large-scale study, with information available for potential confounders, with the study participants coming from various parts of the country. However, we cannot rule out the possibility of measurement bias since we took the measurement only once as Vernier caliper measurements require proper setup and due to time constraints during hospital enrolment, we were unable to perform the measurement twice. Moreover, we measured the digit lengths of the participants directly and did not use scanned copies or photocopies, which would have provided repeatable measurements28. However, any misclassification that might have happened will be non-differential since we used the same procedure for both cases and controls. Furthermore, our sample size may not have the power to assess the risk stratified by menopausal status.

In conclusion, we observed that among premenopausal women exclusively, there was an inverse association between right 2D:4D, as well as DR−L and the BC risk. In contrast, non-significant positive association was observed for postmenopausal women for right 2D:4D. For the left 2D:4D ratio, we found a somewhat increased risk of BC only in the premenopausal, although not significant. Future studies with substantially larger sample sizes or with a reduction in random measurement error achieved with repeated measures of digit length and smaller numbers of measurers are warranted. Additionally, longitudinal follow up is indicated to determine whether a clinically meaningful threshold or cut-off exists in 2D:4D ratios for BC risk prediction and risk stratification.

Table 1 Characteristics of the study population for selected parameters (N = 3361).
Full size table
Table 2 Association between right and left 2D:4D ratio, and difference between right and left 2D:4D ratio with breast cancer.
Full size table