text in 
English (pdf)
Article in xml format
Article references
Automatic translation
Send this article by e-mail
Cited by SciELO 
Cited by Google 
Similars in
SciELO 
Similars in Google 
Permalink
Introduction
Overweight is recognized as one of the main prevent able risk factors for cancer. Up to 30% of mortality due to this disease could be avoided if excessive weight gain were addressed. Regrettably, this phenomenon shows epidemic behavior with a rising prevalence, especially in developing countries.
One of the prototypes of the impact of overweight on malignancies is breast cancer. This association may be due to mechanisms such as increased oxidative stress, peripheral conversion of sex hormones 1, adipokines and inflammato ry cytokines involved in the inhibition of cellular apoptosis, migration and increased proliferation of malignant cells 2.
From a clinical perspective, several cohort studies have researched the association between overweight/obesity and breast cancer prognosis. Most have reported a concerning association between overweight and decreased survival and increased distant recurrence 3-5. Strikingly, a low BMI (< 18.5 kg/m2) also showed decreased overall and breast cancer-specific survival 6.
Considering that weight behavior (increasing or decreasing) and cancer are health phenomena which are determined by genetic and environmental factors, it is pertinent and justifiable to research their relationship in different populations and territories. Especially when most of the body of evidence is derived from white and Asian women, and the results in Latin American women are still scant.
Methods
This was a retrospective, longitudinal cohort study with descriptive objectives which was approved by the ethics committees of the participating institutions. Convenience non-probability sampling was used with data obtained from the records of two oncology reference centers in northeastern Colombia. The medical records inclusion period was from 2005-2015 (10 years) and the following selection criteria were considered: a) women over the age of 18, b) having a diagnosis of invasive breast cancer, c) being classified as stage IIIB, d) luminal A subtype (hormone receptor-positive and HER2/neu-negative), e) having received multimodal treatment (mastectomy, chemotherapy, radiation and hor mone therapy with tamoxifen), f) no clinical suspicion of residual (local) disease postoperatively or recurrence at the beginning of follow-up (first year), and g) having at least two recorded weights at least six months apart, after begin ning cancer treatment.
The main objective was to describe the weight behavior in selected patients during a 24 month follow up period after finishing multimodal cancer treatment. In addition, the relationship with the outcomes of distant recurrence (metastasis) and mortality would be explored, recognizing the scope of the available sample.
The results analysis was carried out using the STATA 14.0® software package and Microsoft Excel®, applying measures of central tendency, dispersion, comparison of proportions (x 2) and analysis of variance. Statistical signifi cance was established as p < 0.05.
Results
A total of 1,662 breast cancer records were reviewed; of these, 74 patients fulfilled the strict selection criteria. The sociodemographic characteristics are described in Table 1.
Table 1 Sociodemographic characteristics.
| Characteristic (n) | General percentage % |
|---|---|
| Age at diagnosis (74) | 52.01 years (SD 10.76, 26-78) |
| Residence (72) | % |
| Rural | 12.50 (9) |
| Urban | 87.50 (63) |
| Socioeconomic level (48) | % |
| One | 4.17 (2) |
| Two | 20.83 (10) |
| Three | 37.50 (18) |
| Four | 33.33 (16) |
| Five | 2.08 (1) |
| Six | 2.08 (1) |
| Marital status (57) | % |
| Single | 33.33 (19) |
| Married | 8.77 (5) |
| Common law | 57.89 (33) |
| Age at menarche (42) | 13.54 years (SD 1.64, 10-18) |
| Age at menopause (9) | 48.55 years (SD 4.69, 41-55) |
| Term pregnancies (70) | 2.97 (SD 2.28, 0-14) |
| Breastfeeding (30) | % |
| No | 16.67 |
| Yes | 83.33 |
| Birth control (44) | % |
| No | 52.27 |
| Yes | 47.73 |
| Smoking (33) | % |
| No | 72.3 (24) |
| Yes | 27.27 (9) |
The average follow up after finishing cancer treatment was 2.33 years. There were nine tumor recurrence events, five with bone metastasis, three pulmonary and one brain, which, in turn, was associated with the only recorded death.
No patients had a BMI <18 kg/m2; on the contrary, 67.56% were classified from the first recorded weight as overweight/obese. At the end of follow up, the overweight subgroup increased, mainly due to migration of obese patients who lost weight. However, no patient had enough weight loss to be reclassified in the WHO underweight group (Table 2).
Table 2 Patient distribution by weight subgroups according to the WHO classification.
| WHO Group | Beginning of follow up % | End of follow up % | Group tendency |
|---|---|---|---|
| Underweight | 0 (0) | 0 (0) | Stable |
| Normal weight | 32.43 (24) | 31.08 (23) | Decrease |
| Overweight | 44.59 (33) | 51.35 (38) | Increasse |
| Obesity | 22.97% (17) | 17.57 (13) | Decrease |
In general, weight remained steady, comparing the first and the last recorded weight, with an absolute difference of means of -0.17 kg. However, in the patients with metastasis, there was a difference adjusted for
Table 3 Anthropometric measurements during follow up.
| Anthropometric measurements (74) | Initial average | Final average | Nonrecurrence group (65) | Recurrence group (9) | p |
|---|---|---|---|---|---|
| Height (mt) | 1.57 | 1.57 | 1.57 | 1.57 | 0.84 |
| Weight (kg) | 67.02 | 66.85 | 67.10 | 61.80 | 0.87 |
| BMI (kg/m2) | 26.93 | 26.88 | 27.95 | 26.77 | 0.90 |
The patients' follow up appointments were every four to six months. Therefore, a comparative analysis of periods was proposed using ANOVA, without achieving statistical significance in the comparison of means between patients with or without distant recurrence. However, a clinical ten dency towards weight loss was identified in patients with recurrence during the third follow up period (18-24 months), compared to their counterparts without metastasis (Table 4).
Table 4 Difference in weight averages at the various measurement times every four to six months.
| Time of follow up | General group average | Average of subgroup with-out recurrence | Average of subgroup with recurrence | ANOVA |
|---|---|---|---|---|
| Beginning of follow up | ||||
| Initial weight | 67.02 (74) | 67.10 (65) | 66.45 (9) | 0.87 |
| Initial BMI | 26.93 | 26.95 | 26.77 | 0.90 |
| Period 1 | ||||
| Weight | 67.11 (74) | 67.15 (65) | 66.83 (9) | 0.93 |
| BMI | 26.96 | 26.95 | 26.77 | 0.97 |
| Period 2 | ||||
| Weight | 67.91 (73) | 68.15 (64) | 66.17 (9) | 0.64 |
| BMI | 27.33 | 27.43 | 26.65 | 0.61 |
| Period 3 | ||||
| Weight | 67.57 (70) | 68.23 (61) | *63.08 (9) | 0.21 |
| BMI | 27.10 | 27.36 | 25.37 | 0.15 |
| Period 4 | ||||
| Weight 4 | 67.03 (62) | 67.72 (53) | 62.95 (9) | 0.28 |
| BMI 4 | 26.85 | 27.11 | 25.36 | 0.23 |
| Period 5 | ||||
| Weight | 68.24 (40) | 68.58 (34) | 66.3 (6) | 0.70 |
| BMI | 27.12 | 27.25 | 26.43 | 0.68 |
| Period 6 | ||||
| Weight | 68.12 (34) | 68.46 (28) | 66.55 (6) | 0.75 |
| BMI | 26.95 | 27.05 | 26.53 | 0.81 |
| Period 7 | ||||
| Weight | 69.52 (27) | 71.38 (22) | 61.34 (5) | 0.10 |
| BMI | 27.20 | 27.78 | 24.68 | 0.18 |
| Period 8 | ||||
| Weight | 72.85 (18) | 73.07 (16) | 71.05 (2) | 0.81 |
| BMI | 28.22 | 28.42 | 26.63 | 0.60 |
| End of follow up | ||||
| Final weight | 66.85 (74) | 67.54 (65) | 61.88 (9) | 0.18 |
| Final BMI | 26.88 | 27.05 | 25.71 | 0.40 |
Clinical weight variation was established if at least one of three criteria was met: 1) a weight gain or loss of at least 5% of the initial weight, 2) a change of plus or minus two BMI points from the initial score, or 3) a change in WHO weight group at any time during follow up (Table 5).
Table 5 Frequency of weight variability according to the various definitions accepted in the study.
| Variability criterion | General (74) % | Without recurrence (65) % | With recurrence (9) % | p value |
|---|---|---|---|---|
| 5% of the initial weight | 0.32 | |||
| Absent | 40.54 (30) | 33.78 | 6.75 | |
| Present | 59.45 (44) | 54.05 | 5.40 | |
| 2 BMI points | 0.46 | |||
| Absent | 55.41 (41) | 47.29 | 8.10 | |
| Present | 44.59 (33) | 40.54 | 4.05 | |
| WHO group | 0.13 | |||
| Absent | 75.68 (56) | 68.91 | 6.75 | |
| Present | 24.34 (18) | 18.91 | 5.40 | |
| At least one criterion | 0.97 | |||
| Absent | 33.78 (25) | 29.72 | 4.05 | |
| Present | 66.22 (49) | 58.10 | 8.10 | |
| Absolute weight difference | -0.17 kg | + 0.50 kg | -5.06 kg | 0.01 |
Discussion
This is the first longitudinal study to date in a Colombian population which seeks to describe the behavior of weight in women with breast cancer. The reason for only choos ing luminal A type patients was that tumors with hormone receptors are known to be more susceptible to extra-ovarian estrogen production 7,8 and part of the heterogeneity in the results of many studies is due to not recognizing this differential susceptibility between intrinsic subtypes.
Population data in Colombia (ENSIN 2010), similar to what is recorded in developed countries, show an alarming prevalence of overweight/obesity, mainly in women, with more than 60% of women over the age of 40 being over weight/obese 9.
This study found a prevalence of overweight/obesity of 67.56% from the beginning of follow up, and it remained practically unchanged despite multimodal treatment (68.92%). Interestingly, no woman was underweight at the beginning or end of the study. This suggests that, in this population, overweight may be a common characteristic among women diagnosed with breast cancer, as has been described in other studies 10.
The evidence suggests that the relationship between overweight and the risk of breast cancer is different in pre and postmenopausal women. There are data which suggest an inverse or protective relationship of overweight in young women, before menopause 11. However, these conflicting results may be explained by a greater presentation of basal tumors in young women, whose biological susceptibility to the effects of fat mass are not well understood. Once again, this study overcame this difficulty by homogenizing the patients by molecular subtype, and this could explain why, despite 51% of the women in this cohort being under 50 years old, and most being premenopausal, there were no statistical differences in the variables of interest compared with patients who reported being menopausal. Perhaps the effect of weight in women with luminal A breast cancer is independent of the patients' fertility status.
It is important to compare the results reported by Cuello-López et al. who, in 2017, presented a cross-sectional study analyzing weight as a prognostic variable in Colombian women 12, which showed that postmenopausal patients at an oncology center in Medellín had a positive associa tion between BMI and the luminal A subtype. This same study found a low prevalence (1.53%) of women with breast cancer who were underweight, while there was a high frequency (65.4%) of overweight/obesity, similar to what was described in our study. Furthermore, Gutiérrez et al. described a group of 45 patients in Mexico 13, all at a locally advanced stage (III) with a high prevalence (86.6%) of overweight/obesity, which coincides with our findings and those of other studies in the hemisphere, with a high prevalence of overweight consistently reported over 50% in populations with similar characteristics to those of the patients presented in the current study.
One of the strengths of this study was the analysis of the temporal behavior of weight, evaluating its tendency at various patient observation times. Therefore, one of the main data of interest was weight variability defined by sev eral criteria (BMI, weight in kg, and WHO group). Weight behavior patterns have been identified in the literature ac cording to body mass index prior to treatment. Thus, women with BMIs closer to normal had a greater tendency to gain weight, compared with obese women, who even tended to lose weight 14. In this cohort, the stability of weight between the first and last record is remarkable; however, it must be noted that there was weight variability between these periods, but the tendency was to stabilize, returning to the initial weight.
There are studies with significant population samples, such as the study by Chen et al. in China, which report a general tendency to gain close to two kg in the first 18 months after chemotherapy 15. The current study found a subtle but generalized tendency to lose weight (-0.17 kg) which differs from the described behavior as well as from Western studies such as the Women's Healthy Eating and Liv ing Study (WHEL). Once again, this could be due to ethnic differences, which justifies the performance of this type of studies to record the populational differences.
The rate of recurrence was 12.1%, which is significantly greater than that reported in other studies, with a risk of metastasis less than 9% during the longest observation pe riods (60 months)16,17. The design of this study does not identify possible causes associated with the relative increase in the number of metastatic events. However, patients with overweight/obesity, like most of the participants in this cohort, are known to have worse outcomes 8,18,19. For example, neoplasms in overweight women are larger, have a worse proliferation index and more lymphatic metastases20.
It should be noted that there was a statistical difference in the degree of weight loss between patients with and without distant recurrence, a phenomenon which occurred even before the clinical diagnosis of metastasis. In most cases, weight loss was clinically significant after the first 18 months of follow up. This gives validity to the longitudinal assessment of weight in cancer patients as a classic indicator of active cancer.
To conclude, it should be recognized that this is a study limited by its small sample size, short follow up period, lack of other more specific anthropometric measurements and lack of randomization. It is hoped that this paper will lay the groundwork for prospective and analytical studies interested in continuing to analyze the relationship between weight and cancer in the Latin American population, with a more robust methodology which will improve the power and scope of the results.
Conclusions
Overweight and obesity are a very prevalent charac teristic in patients with locally advanced luminal A breast cancer. There was no conclusive evidence of increased risk of metastasis or death associated with excess weight in this population. On the contrary, weight loss was a statistically significant characteristic of patients with distant recurrence during follow up, although it was not able to be defined as a causal factor.
