Higher levels of tumor necrosis factor β are associated with frailty in socially vulnerable community-dwelling older adults

Frailty is an age-related state characterized as a syndrome in which individuals may become more vulnerable to adverse health outcomes with a higher risk for mortality when exposed to stressors [1‐3]. Social vulnerability is defined as the degree to which a person’s overall social situation leaves them susceptible to health problems, which include physical, mental, psychological and functional problems. Social vulnerability is particularly important for older adults as a condition related to cultural, social and economic aspects that may influence the access to goods and services. This is an important indicator to explain the high exposure of this population to the development of pathological and degenerative conditions as it increases the frailty rate [4‐8].

On the other hand, knowledge regarding the cellular and molecular mechanisms related to this syndrome is limited, especially in vulnerable populations. Changes related to the immune-endocrine axis are extensively described during aging and are frequently associated to increases in morbidity and mortality [9]. These changes are characterized by a low-grade, controlled, asymptomatic, systemic, and chronic progressive increase in the pro-inflammatory status, also called inflammaging [10]. This pro-inflammatory state generated by age, sex, lifestyle, socioeconomic background, comorbidities and affective, cognitive or sensory impairments has been postulated as a potential driver of frailty pathogenesis, producing even higher levels of systemic inflammatory markers [11, 12]. Associations between frailty and endocrine and immune changes have been demonstrated, suggesting that alterations in these systems may accelerate the development of age-related diseases and consequently, frailty. [13, 14]. Considering that frailty is characterized by the loss of resilience, poor social conditions may reflect chronic stressors, which increase risk of disturbances in metabolic and immunological parameters, leading to a higher risk of developing frailty [15‐17].

Previous studies have pointed out several candidates for blood-based biomarkers for frailty [18]. Those include inflammatory molecules, such as C-reactive protein (CRP) [19, 20], Interleukin-6 (IL-6) [19‐22] and Tumor Necrosis Factor-α (TNF- α) [19, 21, 23], clinical parameters such as hemoglobin [19, 24] and serum albumin [25], hormones, such as dehydroepiandrosterone sulfate (DHEA-S) [26], testosterone [27], Insulin-like Growth Factor-1 (IGF-1) [28] and vitamin D [29], among others. To the best of our knowledge, however, no study has described the differences of clinical parameters and pro-inflammatory cytokines so far among different frailty statuses of older adults living in socially vulnerable populations.

As a result, we performed a cross-sectional study to investigate potential blood-based biomarkers for frailty under a social vulnerability context, considering demographic, psychological and clinical characteristics. The results of this study could contribute to a greater understanding of frailty syndrome in socially vulnerable populations, as well as to the biology of frailty with regards to the panel of expressed inflammatory cytokines. Moreover, identifying potential blood-based biomarkers will enable the development of interventions and specific policies for this public.

The mean age of the study cohort was 70.1 ± 7.7 years, mostly female (56.2%) with a low education level (3.8 ± 2.3 years) (data not shown). In this sample, 34 subjects (9.8%) were non-frail, 197 (56.8%) were classified as pre-frail and 116 (33.4%) subjects fulfilled the criteria for physical frailty. Considering age, as expected, frail individuals were significantly older than their non-frail counterparts (Table 1).

Results showed significant differences between groups concerning the sex, indicating that women were more frequent in the non-frail group, while men were most frequent in pre-frail and frail categories. Regarding age, as expected, older people (> 80 years) were frailer compared to those in the age groups of 60–69 and 70–79 years. The Geriatric Depression Scale (GDS) indicated that prevalence of depressive symptoms was also higher in pre-frail and frail, compared to non-frail individuals. Frail individuals also presented insufficient levels of physical activity evaluated by the International Physical Activity Questionnaire (IPAQ) (Table 1), which is consistent to the frailty condition.

Considering the results of metabolic parameters and biomarker evaluation, we observed that urea and TNF-β were significantly higher in frail participants compared to non-frail individuals. Levels of creatinine and IL-6 were higher among pre-frail and frail individuals, compared to non-frail ones. Cytokines IL-1α and TNF-α levels were higher in pre-frail, compared to non-frail, but no statistically significant differences for these cytokines were found comparing non-frail with frail individuals (Table 2).

We performed a univariate regression analysis to track the main variables that seem to contribute to the frailty phenotype considering the three groups (frail vs pre-frail vs non-frail) evaluated. Variables were considered to be part of the ordinal model when they presented a statistical significance above 20% (p < 0.20). Thus, we considered the following variables for our model: sex, age, HDL and VLDL cholesterols, urea, creatinine, triglycerides IL-1β, IL-6 and TNF-β (Table 3).

Given these results, an ordinal regression model was conducted in order to predict which variables were significantly associated with frailty. Results showed that age and TNF-β levels were significantly associated with frail phenotype. Older people with increased risk to becoming frail were those older than 80 years (OR: 2.5; 95% CI: 1.09–5.61) and who presented higher levels of TNF-β (OR: 2.5; 95% CI: 1.30–4.9) (Table 4).

Frailty is a geriatric syndrome associated with disability and mortality outcomes in older people. In a recent multicenter, a population-based cross-sectional study showed a prevalence around 38% of frailty in Brazilian community-dwelling older people [38]. We identified in our sample, in which all individuals were living in a high social vulnerability context, that 33.4% of aged subjects investigated fulfilled criteria for physical frailty. A possible explanation for these differences may be related to the fact that in our study, most of the participants were in the 60 to 69 year old age group (53.3%; data not shown) and few participants were above 80 years old (13.5%; data not shown), while in a study carried out by Pereira et al. [38] the age group ranging from 60 to 69 years old represented only 36.4% of the sample. More importantly, in the study by Pereira et al. [38], the authors did not evaluate frailty using Fried’s score, but instead used the frailty index (FI), which probably contributed to the difference found in both studies, regarding frailty prevalence. Accordingly, Collard et al. [39], in a systematic review of 21 studies (n = 61,500), described that the prevalence of frailty in a community-dwelling population has a high variability, ranging from 4 to 59.1%. Moreover, in another study, Moreira and Lourenço [40] found a prevalence of 9.1% of frailty when a different scale was applied to track this phenotype, while Sousa et al. [41] found a prevalence of 17.1% for Brazilian older people using Fried’s phenotype scale. These variations may be highlighted by the different operationalization of frailty status used in the studies, resulting in widely different frailty prevalences.

The frail condition in our sample was also more frequent in men, which is different from previously published epidemiological data that showed higher rates of frailty in women [42, 43]. It has been demonstrated that men frequently have more acute illnesses, while women have more comorbidities and disabilities [44, 45]. This may consequently lead to men becoming more vulnerable to frailty in advanced ages. Additionally, it is important to consider that Brazilian population social inequalities in health behavior are very relevant to explain the particularity of some results found for our sample. It has already been demonstrated that in low-income, less schooled people and those without private health insurance, Brazilian men were found to have more risk behavior, including smoking, sedentary lifestyles and bad quality of food, with a low intake of fruit and vegetables [46]. Furthermore, another investigation found that around 1.1% among men and 0.3% among women in this social condition have never visited a physician and 35% and 52% of men and women respectively, had not visited a physician in the last year [47]. These findings help to elucidate the underutilization of health care services among individuals of the lowest economic class, but especially among men, which may characterize them as a more vulnerable group for frailty. On the other hand, our data is in accordance with other research, indicating that frailty increases with age and low levels of physical activity [39].

The aging process frequently results in a decline of the immune function, known as immunosenescense, which was related to the role of inflammation in frailty [48]. These alterations are characterized by around 2-fold increased levels of cytokine production when compared to young individuals’ levels. Increases in adipose tissue, low levels of physical activity and the senescent cellular process itself were suggested as mainly factors to drive this phenomenon that results in a low-grade chronic inflammatory state [49]. However, it is difficult to establish the relationship between inflammation and frailty, considering that both conditions linearly increase with advancing age, thus avoiding a clear answer to determine whether inflammation is a cause or a consequence of the frailty phenotype. Nevertheless, in a recent meta-analysis involving 32 cross-sectional studies and 23,910 older people, it was demonstrated that frail and pre-frail conditions are related to increased levels of serum cytokines, when compared to the levels of non-frail participants [50]. Our results corroborate the literature, demonstrating that an increased pro-inflammatory state is also present in frail older adults living in a context of social vulnerability.

Using an ordinal regression model, we found that TNF-β was significantly associated to the frail phenotype. Abnormalities in TNF-α or TNF-β expression have been implicated in several diseases [51‐53]. A TNF-β signaling pathway seems to be involved in the support of efficient immune responses against pathogens, which is a key point in the immunosenescense [54]. An important link between the lack of adaptability of immunological function during aging and increased risk for frailty in the vulnerable oldest old individuals was found in our study and may indicate that imbalances in immune functions seem to trigger physical frailty in this population. Similarly, in a systematic and meta-analysis review, Soysal and co-workers [50] demonstrated that three large prospective studies failed to find any association between higher inflammatory levels and frailty. The authors attribute this result to the paucity of data, to an over-adjustment of the analyses and/or to the fact that frail people were prone to diseases that might have increased the levels of inflammatory cytokines during the follow-up period and the lack of an adjustment for inherent changes to these markers in their analyses.

As an important limitation of this work, we can point out that this is a cross-sectional study, which does not follow individuals over time. A follow-up study using the same approaches to investigate frailty in this population is currently underway in our group. On the other hand, the statistical power of our sample size allows the expansion of our findings to other already recognized socially vulnerable populations. Notwithstanding, the knowledge of the characteristics related to socioeconomic and biological conditions configures an important tool to raise preventive strategies to decrease frailty risk in the most vulnerable populations.

Taken together, the results presented here could help to implement preventive healthcare strategies by evaluating frailty and at the same time measuring a set of inflammatory biomarkers, paying special attention to TNF-β plasmatic levels. This could enable health teams to plan and improve care actions for this population, especially in the community and in primary health services.