2.1. Study design, population, and sample

We conducted a secondary analysis of data from a cross-sectional study [27]. The primary study included 453 older adults recruited by census sampling, representing 95% of persons aged 60 years or older residing in 12 high Andean communities, evaluated during 2013–2020. Previous studies have been conducted using this database [27,28]. Residents of the following high Andean communities located above 1,500 meters above sea level (masl) were recruited from the departments of: (a) Amazonas (La Jalca and Leimebamba), (b) Ancash (Atipayán, Llupa, Macashca and San Pedro de Chaná), (c) Ayacucho (Ayahuanco), (d) Huancavelica (Vilca), (e) Huancayo (Chacapampa), (f) Huánuco (Pampamarca), (g) Lima (Viñac), and (h) Puno (Paucarcolla) [28]. For this secondary analysis, we excluded participants with missing data on the variables of interest, resulting in a final sample of 417 older adults. For the calculation of statistical power, the study by Launay et al. [29], was used, in which depressive symptoms and the prevalence of falls (a variable associated with balance disorders) were compared. The proportion of falls in the group with depressive symptoms was 29.6%, while the proportion of falls in the group without depressive symptoms was 14.8%, considering a 95% confidence level, providing a statistical power of 99.9% for a sample of 417 older adults.

2.2. Procedure

The data were obtained by means of surveys and anthropometric measurements taken by trained medical personnel, including general practitioners, geriatricians, and medical students, who visited the inhabitants in their homes up to three times to invite them to voluntarily participate in the study, after reading the informed consent form. A form was filled out in which sociodemographic data, medical and personal history (tobacco and alcohol consumption, comorbidities, and body mass index (BMI), neuropsychiatric evaluation (depressive symptoms and cognitive impairment), and functional evaluation (Edmonton test, and functional reach test) were collected.

In the primary study, we excluded older adults with severe cognitive impairment (greater than or equal to eight points) assessed by the Pfeiffer questionnaire, as well as individuals unable to understand or communicate in Spanish or for whom a translator was not available, those with hearing or visual impairments that hindered the application of the survey, persons presenting any physical disability that did not allow the execution of physical performance tests, and finally, individuals without the variables of interest for the elaboration of the present study.

2.3. Variables

2.4. Statistical analysis

Data analysis was performed with the statistical package Stata/SE v17.0 (StataCorp, TX, US). For descriptive analysis, numerical variables were presented using means and standard deviation, as well as medians with their respective interquartile ranges, depending on the normality of the distribution of the variable. On the other hand, relative and absolute frequencies were used to summarize the categorical variables. Similarly, we used the Pearson’s Chi-square test or the Fisher’s exact test for comparison between categorical variables. Likewise, the Student’s-t test or the Mann Whitney U test was used to compare the differences of a numerical variable between groups according to the presence of depressive symptoms and balance disorders.

Two generalized linear models (crude and adjusted) of the Poisson family with logarithmic link function with robust variance were constructed with the aim of assessing the association between depressive symptoms and balance disorders in older adults. In this way, we estimated crude (cPR) and adjusted (aPR) prevalence ratios with their respective 95% confidence intervals (95%CI). We pre-specified a minimal sufficient adjustment set under an epidemiological criterion guided by prior literature [4–6,8,41–43]. A variable qualified as a confounder if it was a plausible common cause of both depressive symptoms and balance impairment, preceded both processes, and was not on the causal pathway nor a descendant of either. Accordingly, the adjusted model included sex, age, educational level, comorbidities, BMI, social support, and neurocognitive disorder [4–6,8,41–43]. Covariates were selected a priori, based on prior literature and not by statistical significance in bivariate analysis; bivariate results are presented only to describe the sample. We assessed multicollinearity in the adjusted model using variance inflation factors (VIFs).

2.5. Ethical approval

The primary study was evaluated and approved by the Institutional Research Ethics Committee (IREC) of the Naval Medical Center (Memorandum N°49). The study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki and included obtaining informed consent from all older adults participating in the study. Neither the integrity nor the confidentiality of the data provided was violated. Likewise, the present secondary analysis did not violate participant confidentiality and was evaluated and approved by the IREC of Universidad Científica del Sur (No. 011-DACMH-DAFCS-U. CIENTIFICA-2022).

3.1. Descriptive and bivariate analysis according to the prevalence of balance disorders in the study sample

Data from 417 older adults residing in 12 high Andean communities in Peru were analyzed. A total of 255 (61.1%) women were included, with a mean age of 73.2 ± 6.9 years, residing at a median altitude of 3,400 masl (3,315−3,445), and 220 (52.8%) and 204 (48.9%) of the participants had depressive symptoms and balance disorders, respectively (Table 1). Likewise, the mean functional reach was 19.9 ± 6.5 cm. Table 1 shows the characteristics according to the prevalence of balance disorders in older adults. We found that marital status (p = 0.007), altitude of residence (p = 0.013), tobacco consumption (p = 0.002), alcohol consumption (p < 0.001), polypharmacy (p = 0.001), BMI (p = 0.002), social support (p = 0.016), and neurocognitive disorder (p < 0.001) presented statistically significant differences according to the presence of balance disorders.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Descriptive and bivariate analysis of the variables of interest according to the prevalence of balance disorders in older adults in 12 high Andean communities of Peru (n = 417).

https://doi.org/10.1371/journal.pone.0353158.t001

3.2. Descriptive and bivariate analysis according to depressive symptoms in the study sample

The presence of depressive symptoms was significantly associated with age (p < 0.001), older adults younger than or equal to 70 years of age (p = 0.003), living alone (p = 0.045), tobacco consumption (p < 0.001), polypharmacy (p = 0.019), urinary incontinence (p < 0.001), social support (p < 0.001), and neurocognitive disorder (p < 0.001) (Table 2).

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Descriptive and bivariate analysis of the variables of interest according to the prevalence of depressive symptoms in older adults in 12 high Andean communities of Peru (n = 417).

https://doi.org/10.1371/journal.pone.0353158.t002

3.3. Crude and adjusted prevalence ratios to estimate the association between depressive symptoms and balance disorders in the study sample

We found that the presence of depressive symptoms was associated with a higher prevalence of balance disorders (aPR = 1.66; 95%CI: 1.28–2.15; p < 0.001) after adjusting for sex, age, educational level, comorbidities, BMI, social support, and neurocognitive disorder (Table 3).

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Crude and adjusted regression models to estimate the association between depressive symptoms and balance disorders in older adults in 12 high Andean communities of Peru.

https://doi.org/10.1371/journal.pone.0353158.t003

4.1. Main findings

We found that five out of ten older adults had depressive symptoms or balance disorders. The presence of depressive symptomatology was associated with a higher prevalence of balance disorders in older adults living in 12 Peruvian high Andean communities. The difference between the crude and adjusted prevalence ratios reflects negative confounding. In the crude analysis, the reference group of older adults without depressive symptoms was disproportionately burdened by neurocognitive disorder and lack of social support, two factors independently associated with worse balance, affecting the crude prevalence ratio and attenuating the true association. Once these imbalances were accounted for in the adjusted model, the association between depressive symptoms and balance disorders appeared with its actual magnitude. These findings support the development of targeted interventions for older adults residing at high altitude. Non-pharmacological approaches with evidence in older adult populations include balance and resistance training programs [44] and cognitive-behavioral therapy. Regarding pharmacological interventions, selective serotonin reuptake inhibitors (SSRIs) may improve depressive symptoms; however, long-term antidepressant use has also been associated with balance and gait impairment in older adults [45,46]. Therefore, a personalized, multi-component approach to treatment is recommended.

4.2. Comparison with previous studies

We found that depressive symptoms are associated with an increased likelihood of balance disorders in older adults from 12 high Andean communities, which is consistent with findings reported by a previous systematic review [11] and a study in India [23]. However, a study performed in Turkey did not find the present association [22], while a study in India reported that higher depression scores were associated with worse balance performance [13], which is directionally consistent with our findings. It should be noted that no study exploring this association in populations residing at high altitudes were found. However, previous studies have reported the prevalences and averages of depressive symptoms and balance disorders in residents of high-altitude communities. Previous studies in the Andes [47] and Himalayan regions [7,24,47] reported a lower prevalence of depressive symptoms (1.8% to 26.9%) than that found in our study (52.8%) in adults and older adults. These differences might be explained using different depressive symptom assessment instruments (the PHQ-9, the PHQ-2 and the 10-question GDS), collection time and sample size [7,24,47]. On the other hand, previous studies on balance disorders in the Himalayas and Karakoram [19,48] conducted in older adults, reported means between 13.7 cm to 26.5 cm in the functional reach test, similar to the values described in our study.

4.3. Interpretation of the findings

Hypobaric hypoxia affects both brain bioenergetics and serotonin metabolism, both of which contribute to the development of depression [6]. Likewise, the presence of motor abnormalities affecting gait, posture and balance are described as part of the pathophysiology of depression, as well as its association with a worse response to treatment [11]. In addition, hypobaric-normobaric hypoxia also influences inspired, arterial and alveolar oxygen partial pressure, central nervous system integration and sensorimotor function, all of which are involved in the control of standing balance [12]. Therefore, residents at high altitudes may have a higher prevalence of depressive symptoms and balance disorders as found in the present study. In addition, it has been reported that individuals presenting both conditions present greater executive dysfunction, as well as a greater risk of falls [49], which is important in the older adult population, since falls represent one of the main causes of morbidity and mortality [50]. Moreover, executive dysfunction may suggest that the older adult is in a mild stage of Alzheimer’s disease [51,52]. Another important aspect of evaluation is vascular diseases, because the burden of lesions in the white matter can cause simultaneous changes in gait, balance, and mood [53].

Although residents of these communities have lived at high altitude for decades, evidence suggests that compensatory physiological adaptations do not fully restore normal arterial oxygen partial pressure in chronic high-altitude residents [6]. The resulting persistent relative reduction in cerebral oxygen availability may continue to alter serotonin metabolism and sensorimotor function [6,12], possibly contributing to the elevated prevalence of both depressive symptoms and balance disorders observed in our sample. Future studies incorporating biomarkers of altitude adaptation are needed to confirm these mechanisms in chronically adapted Andean populations.

4.4. Relevance of the findings

People with depressive symptoms present neurocognitive changes associated with motor coordination [14], with deficits in balance and reduced postural control [11,54,55]. They also show a more stooped appearance, characterized by alterations in body posture and presenting a greater degree of thoracic kyphosis and forward inclination of the head [11]. Stooped posture is also characteristic of aging, which is commonly described as hyperkyphosis and is accompanied by increased postural sway [21]. Postural kyphosis increases the risk of fractures and mortality, and is also associated with deterioration of health, quality of life and physical performance [56,57]. In this sense, it is important to detect depressive symptoms in a timely manner, particularly in older adults, to prevent serious complications associated with balance disorders.

Cultural factors and language barriers can modulate early detection and management of depressive problems in rural populations [58]. The domains and subdomains of intrinsic capacity are interrelated and justify a comprehensive clinical approach to optimize functional ability [59]. The findings of our study demonstrate how we can assess the risk of having balance disorders through the detection of depressive problems. It remains for future research to demonstrate whether the management of depressive problems contributes to improving balance disorders in high Andean populations and vice versa. Given the cross-sectional design of the present study, causality cannot be established; longitudinal studies are needed to confirm the directionality of this association

4.5. Limitations and strengths

The present study has some limitations. First, some variables were collected by self-reporting, and therefore, may be affected by recall bias. Nevertheless, 61.6% of the participants did not present cognitive impairment and were interviewed together with a family member to corroborate the information. Future studies should incorporate objective clinical records and validated informant-based instruments to minimize recall bias. Second, the sampling employed was non-probabilistic; however, we included more than 95% of the total potentially eligible older adults in the 12 high Andean communities. Third, the cross-sectional design of the study did not allow the estimation of causal associations between the two main variables. However, our findings provided an important approximation of the association in older adults living at high altitudes to subsequently design longitudinal studies. Longitudinal cohort studies are needed to establish temporal relationships and assess the directionality of this association. Fourth, the GDS-5 and the FRT were administered in Spanish, which may have introduced misclassification in participants whose primary language is Quechua, as cultural and linguistic factors can influence the interpretation of questionnaire items and the execution of physical performance tests. Future studies should incorporate culturally and linguistically adapted versions of these instruments. Additionally, the FRT cut-off of 20.32 cm was originally derived from sea-level populations; its applicability to high-Andean communities warrants further validation. Fifth, altitude of residence varied across communities and, while included as a descriptive variable, was not incorporated as a confounder in the regression models, as it operates at the community level rather than as an individual-level confounder. Future studies with a larger number of communities should consider multilevel modeling approaches that explicitly partition individual- and community-level variance. Sixth, as a secondary analysis, direct measurement of altitude adaptation biomarkers, such as hemoglobin concentration, arterial oxygen saturation, or serotonin metabolism markers, was not feasible. Future studies should incorporate objective physiological measurements to confirm whether hypoxia-related mechanisms are active in chronically adapted Andean populations. The strengths of our study lie in the high initial participation rate and the use of multiple instruments for screening and assessment of geriatric syndromes in older adults in the high Andes.

4.6. Conclusions

Depressive symptoms were associated with a higher prevalence of balance disorders in older adults residing in 12 high Andean communities in Peru. Approximately half of the older adults studied presented depressive symptoms or balance disorders, underscoring the significant burden of these conditions in this underserved population. To our knowledge, this is the first study to evaluate this association in rural high-Andean communities, filling an important gap in the geriatric literature for high-altitude settings. These findings have practical implications for primary care: the GDS-5 and the Functional Reach Test are brief, low-cost, validated tools that can be feasibly administered in settings with limited access to health services, and their combined use may help identify older adults at risk for balance disorders and falls. Future epidemiological studies with larger probabilistic samples and longitudinal designs are needed to confirm the directionality of this association and assess the effectiveness of early screening programs in improving functional outcomes and quality of life in older adults living at high altitude.