Neutrophil/lymphocyte ratio and overall survival in patients with breast cancer: a cohort study in a Latin-American hospital

Introduction

Breast cancer represents 16% of all female cancers, causing 2.3 million cases and 685,000 deaths worldwide in 2020.^1,2 In Latin America, this pathology is the most important cause of death by neoplasms among women, with nearly 92,000 deaths,³ while in Peru it represents the most common primary neoplasm among women.⁴ Therefore, it is fundamental to determine the factors associated to overall survival (OS), such as age, comorbidities, metastasis, and laboratory analytical determinations such as neutrophil-lymphocyte ratio (NLR).^5,6

While research exists on an international level that associates an elevated NLR with lower OS in patients with breast cancer, these were developed in populations with early-stage diagnosis, as opposed to countries with middle and low incomes, where a high proportion are diagnosed in advanced stages. Furthermore, differences exist in breast cancer phenotype, prevalence of oncological infections,⁷ and the presence of social risk factors such as marked poverty and reduced access to healthcare services between Latin-American countries⁸ and others.⁹ This highlights the need for studying the OS associated with breast cancer and its prognostic indices in this context.

Since in developing countries, the rate of clinical oncologists per million inhabitants is below the recommended value,⁷ simple, accessible, true, and low-cost methods, such as NLR, are required to potentially improve OS in these patients. Some research in Latin America has studied the NLR as an adverse prognostic factor in patients with different types of cancer, with few related to breast cancer.^10,11 However, these studies were conducted only in patients with triple-negative breast cancer, which, although more aggressive, is not the most frequent of those presented in the region; therefore, it is necessary to perform studies including other cancer subtypes such as Luminal A, Luminal B or HER 2. In this sense, the general objetive of this study is to determine the association between NLR and OS in patients with breast cancer in the gynecology department of Hospital Nacional Hipólito Unanue (HNHU), Lima, Peru between 2012 and 2014. The specific objectives were to evaluate the association between OS and stage, molecular subtype and age. We hope to find an association between the NLR and the OS, so that the higher the NLR, the lower the OS.

Methods

Results

Of the 324 initial medical records, we excluded 35 not found and 48 incomplete medical records, leaving a total of 241 medical records for analysis (Figure 1). A follow-up of patients was performed in a maximum period of 8 years, with a mean follow-up of 3.5±1.8 years total, 2.7±1.8 years in the exposed group, and 3.9±1.7 years in the non-exposed group. Seventy-five patients (31.1%) were diagnosed in 2012, 84 (34.9%) in 2013, and 82 (34%) in 2014, with 130 patients dying during the study (Table 1). The age range was 27 to 85 years old (mean: 56.1 years±11.6), the most frequent molecular subtype of breast cancer was HER2 (39.2%), while the most frequent clinical stage was II (42.7%). Furthermore, 144 (59.8%) patients had NLR≥3. The other general population characteristics are presented in Table 2.

Figure 1. Flow diagram of number of patients with breast cancer in each stage of the study.

When evaluating OS according to the year of follow-up, we observed that patients with NLR≥3 had a lower OS (Figure 2A), compared to those who had an NLR<3. This association was statistically significant (p<0.001). When evaluating OS by molecular subtype, no statistically significant differences are observed (p=0.528) (Figure 2B). While in the clinical stage (Figure 2C), OS decreases in stage IV (p<0.001). Furthermore, we observed that the age group >55 years (Figure 2D) presented a lower OS within the follow-up years (p=0.017), compared to patients up to 55 years of age.

Figure 2. Kaplan-Meier survival estimate for mortality in patients with breast cancer in a Latin-American hospital.

(A) According to NLR. (B) According to molecular subtype. (C) According to stage. (D) According to age.

In the multivariate analysis, the NLR≥3 (p<0.001) was a risk factor for mortality, adjusted for confounder variables of age and clinical stage. Additionally, age (p=0.016) and clinical stage IV (p=0.001) were also risk factors for mortality (Table 3).

Discussion

In our study, we found that NLR≥3 (p<0.001), age>55 years (p=0.039), and clinical stage (p<0.001), were associated with a lower OS. In the multivariate analysis, the factors significantly associated to a lower OS were NLR≥3, age>55 years, and clinical stage IV. Although there have been international studies, we must consider that in all of them the population characteristics, such as risk factors, prevalence of infections, genetic susceptibility, and existing polymorphisms, are different from the Latin American context,^9,17,18 which makes it difficult to generalize its results in this context. This study is especially important given that a broad follow-up of 8 years for the NLR evaluation is a prognostic factor of OS in patients in a Latin American hospital. In this regard, this type of cancer is considered a health problem of high priority,¹⁹ and the lack of resources requires cost-effective tools of easy access to orient the prognosis and need for treatment by a specialist.

Regarding OS, we found that an NLR>3 was related to a lower OS. Similar results were found by a Peruvian study, only in breast cancer with a triple-negative molecular subtype, while in our study we evidenced the use of this ratio in other molecular subtypes. Similar findings have been found in studies carried out in other regions of the world.²⁰ This is probably because a greater NLR reflects a greater number of neutrophils, whose high infiltration has been associated with the aggressiveness of the disease and resistance to therapy, in addition to promoting pro-cancerous factors and suppressing the cytotoxic activity of lymphocytes.^21–23 On the other hand, a decrease in lymphocyte count may decrease the effectiveness of the anti-tumoral immune response, which is why the combination of these conditions reflected in the NLR seems to be an excellent prognostic tool. Furthermore, it can be obtained through an accessible exam such as blood count, which is why its incorporation into the established prediction of risk justifies a larger study.²⁴

We observed that patients with NLR≥3 had a lower OS, however, as of the fifth follow-up year the survival curves in both NLR categories tended to approach each other. This is similar to other inflammatory indices such as the combination of the NLR with the lymphocyte platelet index (NLR/LPR) studied by Hirahara et al.²⁵ They also observed that survival was greater as this rate decreased (p<0.0001); however, from approximately the fifth follow-up year, we observed that survival functions were similar. This indicates that NLR or other hematological rates are likely not useful in predicting OS after a certain number of years. Likewise, in a Peruvian study,¹¹ the survival functions with an NLR<2.5 and NLR≥2 were different during a 2-year follow-up in patients with triple negative breast cancer. It is probable that if it they had a longer follow-up, the behavior of the curves would have been similar to our study. Breast cancer mortality after 5 years could be high due to diverse factors foreign to what NLR might reflect, such as tumoral growth, comorbidities, and age-related functional impairment.

We found that over half of patients had an NLR≥3, similar to that found in other studies.²⁶ On the other hand, the stages that were most frequently found during diagnosis were stages II and III, unlike the United States, where the most frequent stages were I and II.²⁷ This could reflect a late diagnosis, typical of developing countries with deficits in their health system.⁷ Conversely, the average age at the time of diagnosis is comparable to that of other international reports.²⁸ The most frequent molecular subtype was HER2 with a 39.2%, unlike the Asiatic indigenous population, where the most frequent was the luminal A type (33%). This correlates with greater risk and worse results in cancer among the indigenous communities.²⁹

Furthermore, OS decreased noticeably in stage IV, probably since cancer diagnosis in advanced stages is more biologically aggressive and presents high recurrence rates.³⁰ Likewise, we found that NLR was associated with the disease staging, which could imply a confusing role of that variable when evaluating the predictive capacity of the NLR. Additionally, we found that the greater the age, the lower the OS. This finding is related to the study by Tao et al.³¹ where older patients with breast cancer had 17% greater mortality than younger patients. This age group also had a lower OS, since breast cancer at an older age may have late diagnosis and insufficient treatments.³²

By studying the entire population of a referral hospital that accepts patients from all over the country, these results could be generalized to the urban context of Peru and other cities in Latin America, since the sociodemographic characteristics and the environment where the study was made is similar to that of other Latin American cities. On the other hand, although there was a long follow-up of eight years, during this period there was no important change in the treatment that significantly affected the survival of the patients evaluated, so the results are still useful.

Conclusions

We conclude that NLR≥3 is a risk factor for mortality, adjusted by age and clinical stage. Additionally, age and clinical stage IV could also be risk factors for mortality. A total of 59.3% of breast cancer patients had an NLR≥3, while the most frequently diagnosed stage was stage II with 42.74%, and the most frequent molecular subtype was HER2 with 39.2%.