Falls in elderly patients and their association with genetic factors

Introduction
Modern society and the global trend towards population aging pose a number of new challenges to medicine. The main organizational goal that determines the quality of medical care is patient safety, and therefore the maximum possible prevention of adverse events, defined as adverse reactions, injuries or other unintentional damage that lead to disability or dysfunction of varying severity, temporary or permanent loss of ability to work, and / or an extension of the stay in a medical institution [1]. Thus, in the social and medical sphere, one of the significant problems is falls. According to statistics from the World Health Organization (WHO), falls are the second cause of death from injuries (excluding intentional ones) [2]. It is noted that the average cost of treating patients after a fall is 3611 and 1049 US dollars per person in Finland and Australia, respectively [2]. Falls are most often observed in children, the elderly and, together with the latter, the senile age categories [2, 3]. Moreover, for the elderly, they not only occur more frequently (in a third during the year, and at the age of over 75 years - more than half), but also with a higher probability end in injuries or even death [2-6].
There is a growing interest in the topic among both foreign and domestic authors, despite which it is emphasized that the problem is still underestimated [7]. At present, a multifactorial structure of the development of falls has been identified. Internal (features of the body) and external (environmental factors) are distinguished, as well as non-modifiable and, what is more interesting for researchers and practitioners, modifiable causes of the development of falls [1, 4, 5, 8]. Internal factors that increase the risk of falling include age (especially old age), genetic characteristics, gender, frailty and decreased strength in the lower limbs, balance and gait disorders, urinary dysfunction, COPD, diabetes, heart failure, osteopenia or osteoporosis, multiple comorbid pathology, cognitive impairment, Parkinson's disease, sensory disorders (vision and hearing), fear of falling. External factors include environmental parameters, including the lack of mobility aids (canes, walkers, etc.), improperly selected or poor-quality footwear, poor lighting, and the characteristics of the surface on which movement occurs - stairs, slippery or uneven surfaces, slope, etc. Age, genetic characteristics, and gender are assumed to be non-modifiable factors, while physical function and environmental factors are modifiable [8–14].
Among the modifiable risk factors, adverse reactions to pharmacotherapy, polypharmacy (the simultaneous use of five or more drugs) and the use of certain groups of drugs as a factor associated with the development of falls are separately identified [15, 16].
Thus, drugs that increase the risk of falls (abbreviated as FRIDS in foreign English-language literature) mainly include psychotropic drugs (antidepressants, hypnotics, antipsychotics, anticonvulsants), narcotic analgesics, hypoglycemic drugs, and drugs used to treat cardiovascular diseases (antiarrhythmics, diuretics, hypotensives) [4, 15, 17]. In addition, in the age of chronic diseases, there is an increasing need for a personalized approach to ensure treatment safety and minimize side effects; in these conditions, research and activities to identify individual, including genetic, risk factors for the development of adverse drug reactions are especially relevant. A 2017 study also noted that for many of the above drugs, the dosage can be optimized based on pharmacogenetic variants. Although it is noted that there is only scant empirical information on the impact of genomic variations on the risk of falls in the elderly and the need for further molecular and clinical studies is emphasized [18].
It has also been established that CYP3A4 expression in patients is the main determinant of clonazepam plasma concentrations in relation to dose and body weight; thus, prospective analysis of CYP3A4 expression in patients may identify poor metabolizers with a higher risk of adverse effects, including dizziness, drowsiness, and falls. Adjustment of clonazepam therapy taking into account CYP3A4 expression may help prevent adverse effects caused by inappropriate dosing in patients [19].
Polymorphisms of a number of genes were assessed, namely CYP3A4, CYP3A5, CYP3A7, CYP2C9, CYP2C19. An increased risk of falls in elderly people taking benzodiazepine tranquilizers and carrying CYP2C9*3 and CYP2C9*2 polymorphisms (by 36 and 18%, respectively) was found when compared with patients without the presented polymorphisms. At the same time, an inverse relationship with falls was found when using benzodiazepine tranquilizers in people with the "wild" variant of the CYP2C19 gene polymorphism. An increased probability of falls is noted when using benzodiazepine tranquilizers in people with heterozygous and minor CYP2C19*2 polymorphisms [20]. Thus, the identification of new pharmacogenetic factors in the development of falls for the purpose of their further prevention may serve as a promising area of ​​research.

Objective
To assess the possibility of preventing falls in elderly patients with comorbid cardiovascular pathology by optimizing pharmacotherapy based on pharmacogenetic parameters.

Materials and methods
We studied 172 medical histories of people over 65 years old with comorbid cardiovascular pathologies, undergoing treatment in a multidisciplinary hospital in Moscow for the period from 2017 to 2020, which were divided into 2 groups - a study group with falls (n = 40 - 23.3%) and a control group without falls (n = 132 - 76.7%) in the anamnesis over the past year.
The study considered patients who met the following criteria - people over 65 years old, without severe cognitive impairment with comorbid cardiovascular pathology, with regular intake of more than 5 drugs, including beta-blockers and / or benzodiazepine tranquilizers or Z-drugs (zopiclone, zolpidem and zaleplon).
Patients with severe pathologies or decompensated somatic and mental status were excluded from the study - with manifestations of moderate and severe dementia, with severe neurological pathology, pronounced manifestations of respiratory or cardiac failure (from 3 functional class), with severe senile asthenia, in the presence of an active oncological process, with bad habits (drinking alcohol, drugs, smoking), and, of course, those who refused to participate in the study. Medical documentation, including prescription sheets, were assessed for the therapy received and the presence of risk factors for falls. If all the necessary information was missing, the missing data were clarified directly with the patients themselves. All of them had their blood sampled for genotyping for CYP2D6*4, CYP2C19*2, CYP2C19*3, CYP2C19*17, CYP3A5*3, CYP3A4*22 polymorphisms using the polymerase chain reaction (PCR; real-time PCR). The demographic data of the groups are presented in Table 1.

Table 1

Demographic data of the study groups

To the control group (n=132), we added literature data on the prevalence of CYP2D6*4 (n=93) [21] (we did not find similar data in Russian publications on other studied polymorphisms), with which a comparative assessment of the obtained results was also carried out.
In each of the groups, subgroups were identified according to the drugs used. The entire study consisted of several substages performed sequentially:
1) communication with patients, assessment of medical documentation and blood sampling for genotyping;
2) genotyping using the PCR method and compilation of a common database with the obtained information;
3) assessment of the prevalence of the presented genetic polymorphisms within the groups;
4) analysis of their prevalence in subgroups according to the drugs used; 5) comparison of the obtained data with each other and, in relation to the CYP2D6*4 polymorphism, with data on its prevalence taken from literary sources (n=93).

All data obtained through the statistical package IBM SPSS STATISTICS version 20 were analyzed pairwise using contingency tables; methods such as Pearson's χ2 and Fisher's exact test were used to determine statistical reliability. To confirm the obtained calculations, mathematical models were additionally constructed using logistic regressions with step-by-step filter selection. Values ​​of p <0.05 were considered reliable.

Results
Genotyping data from 172 elderly patients with comorbid cardiovascular pathology were analyzed. The main group consisted of 40 patients who had fallen in the last year, and the comparison group consisted of 132 patients without falls.
The demographic composition of the main group was represented by 33 (82.5%) women and 7 (17.5%) men; the average age was 81.2±7.3 years. The comparison group included 100 (75.8%) women and 32 (24.2%) men, with an average age of 86.9±4.5 years. When checking the Hardy–Weinberg equilibrium in the distribution of alleles and genotypes in the population, its observance was revealed in both presented groups (p ≥ 0.05), which indicates the correspondence of the frequency distribution of alleles and genotypes in the studied groups in the general population and in our data sample (Table 2).

Table 2

Distribution of Hardy–Weinberg equilibrium in groups

The results of genotyping for carriage of polymorphisms CYP2D6*4, CYP2C19*2, CYP2C19*3, CYP2C19*17, CYP3A5*3, CYP3A4*22 are presented in Table 3.

Table 3

Results of the distribution of the carriage of polymorphic variants of the studied genes in the groups of patients with falls and without

Taking into account several polymorphisms of the CYP2C19 gene, additional division was made by phenotypic variants of the gene in the group with and without falls. In total, we obtained 6 phenotypic variants of the gene in each of the groups (Table 4).

Table 4

Results of distribution of phenotypic variants of the CYP2C19 gene in the main and control groups (p=0.715)

When comparing the groups with each other using the Pearson χ2 criterion, no significant differences were found between the samples for CYP3A4*22 and CYP3A5*3 polymorphisms in correlation with falls, in the case of CYP3A4*22 polymorphisms in correlation with falls (p=0.67), and in the case of CYP3A5*3 polymorphisms (p=0.33). Evaluation of the CYP2C19*2 allelic variant between the groups did not yield significant differences (p=0.61). Calculations for the CYP2C19*17 polymorphism in comparison between the groups in correlation with falls similarly showed no statistically significant differences (p=0.47). Phenotypic variants of the CYP2C19 gene did not show significant variability when compared in the groups (p=0.715). No significant differences in CYP2D6*4 allelic variants in the groups were found in correlation with falls (p=0.88).
Calculation of differences in CYP2C19*3 polymorphisms is technically not feasible, since in 100% of cases, all 172 individuals had the same “wild” genotype — GG.
When testing the binary logistic regression model among all genotypes (5 polymorphisms in the analysis, since among the CYP2C19*3 polymorphism, only one “wild” genotype was found: “GG”), an effective mathematical model was not obtained, due to the fact that with a statistically significant model and high specificity, the sensitivity of the analysis was at a low level of 12%. However, a small group of four carriers of the heterozygous genotype ("AG") of the CYP2C19 gene suffered falls in their entirety (10% of all falls), p=0.003 (according to calculations using Fisher's exact test), which can be considered a putative predictor of falls, which, however, requires confirmation in a study with a larger sample. At the same time, we took population data on CYP2D6 genotyping from literary sources, such as Fedorinov D.S. et al. An analysis of the carriage of genetic polymorphisms was conducted in 93 Russian people taking beta-blockers, as a result of which the GG genotype was identified in 73 people (78.5%), the GA genotype was observed in 3 patients (3.2%), and another 17 people (18.3%) were carriers of the CT genotype (Fig. 1).

Thus, when comparing elderly patients with comorbid cardiovascular pathology and a history of falls with literature population data on CYP2D6 genotyping, we obtained a significant difference between the groups (p <0.001) (Table 5).

Table 5

Results of comparison of the distribution of the carriage of allelic variants of the CYP2D6 gene between groups of patients with falls and population literature data

The results of the pharmacogenetic study did not reveal any statistically significant difference between patients from our sample with and without a history of falls in the polymorphisms CYP2D6*4, CYP2C19*2, CYP2C19*3, CYP2C19*17, CYP3A5*3, CYP3A4*22 and phenotypes of the CYP2C19 gene. However, a comparison with the literature population data on CYP2D6 gene polymorphisms revealed a statistically significant higher frequency of carriage of the heterozygous genotype (GA) in the subgroup of patients taking beta-blockers (p <0.001). A wider range of data was found when assessing genotypes in subgroups of patients according to the therapy received and the presence of concomitant pathology. Thus, a predominance of heterozygous and minor genotypes (CT/TT) for CYP2C19*17 was noted in patients who fell and were taking beta-blockers in ophthalmic and oral forms simultaneously (p=0.05). It is noteworthy that the prevalence of the heterozygous genotype, characterized as an "intermediate" metabolizer, among patients who fell in the subgroups taking enteric-coated acetylsalicylic acid for the CYP2C19*2 polymorphism (GA/AA) (p=0.044), in those taking inhaled glucocorticoids for the CYP2C19*17 (CT) polymorphism (p=0.047) and phenotypic variants of the CYP2C19 gene (p=0.029), and in those taking alpha-blockers for the CYP3A5*3 (GA) polymorphism (p=0.026), which, given the lack of specificity of the enzymes encoded by these genes to the above-mentioned groups of drugs, suggests an indirect effect of "less efficient metabolism" on the development of adverse drug reactions, including falls.
Similarly, in patients with falls and diabetes mellitus (as a comorbidity) in the anamnesis, the heterozygous genotype (GA) for CYP3A5*3 is predominant (p=0.006), which correlates (p=0.01) with the data of the subgroup of patients receiving hypoglycemic therapy.
In the study by Ham AC et al., which they published in 2017, an increased likelihood of falls was found when prescribing benzodiazepine tranquilizers in patients with heterozygous or minor polymorphisms of the CYP2C19 gene (p <0.001), which confirms the optimality of having a “wild” genotype [20]. The data of our study indirectly overlap with the results of the work of Erdman V.V. et al., who suggested the need for alleles that ensure efficient metabolism of xenobiotics as a survival factor for elderly patients in order to achieve longevity [22].

Conclusion
Thus, closer attention to patients with heterozygous and minor genotype variants for the purpose of fall prevention is quite justified. In this case, it is necessary to consider the possibility of analyzing the probability of falls, ensuring the prescription (especially starting) of drugs in smaller doses if possible and more careful monitoring of health status. It is promising to discover and develop new, more reliable methods for identifying patients with increased risks of falls. In addition, it is necessary to conduct additional larger-scale and narrowly focused studies aimed at identifying genetic factors that contribute to an increased risk of falls.
Due to the large number of parameters assessed, our study has a number of limitations in the form of heterogeneity of the overall sample and the small number of some of the analyzed groups.