Probing polypharmacy, ageing and sex effects on physical function using different tests

Ageing, sex and polypharmacy affect physical function.


| INTRODUCTION
Physical function in old age is crucial for healthy living, impacting capacity to complete daily tasks 1 and reducing risk of frailty 2 and falls. 3][6] Decline in physical function, including grip strength and balance, may signal neuromuscular and musculoskeletal morbidities, sarcopenia and frailty. 7Reduced grip strength is associated with adverse global health outcomes, including nutritional decline, falls and cognitive impairment, [7][8][9] and is a component of the frailty phenotype. 10Additionally, adverse drug reactions to polypharmacy in older adults may present nonspecifically as falls, physical function impairment and cognitive decline. 11,12Balance improvement is key to reducing falls 3 and is assessed through various tests, including the widely used Short Physical Performance Battery. 13olypharmacy is associated with increased prevalence and incidence of frailty 14,15 and sarcopenia 16 in epidemiological studies.Clinical observational studies show a bidirectional relationship between polypharmacy and impaired physical function. 4The Drug Burden Index (DBI), which calculates an individual's exposure to medications with anticholinergic and sedative effects, is clinically used to identify drug regimens that increase the risk of impaired physical function. 179][20] However, these studies are confounded by participants' use of different medication combinations, disease states and various lifestyle and genetic factors that also contribute to physical decline.Polypharmacy models in mice with greater genetic, dietary and environmental homogeneity are vital in investigating causation and mechanistic changes in old age 21 and assessing sex differences. 22nimal studies have used grip strength, wire hang and balance beam tests to assess physical function.These tests cover different abilities, offering a comprehensive evaluation of age and sex differences, and intervention effects in ageing.The automated grip strength is an instantaneous measure of maximal force generated when the mouse resists releasing its grip from the metal bar, 23 requiring type 2 fast twitch muscle fibres.The wire hang assesses upper body strength and endurance (requiring type 1 slow twitch muscle fibres) by measuring how long mice can hang from a wire against their body weight 24 and prolonging testing duration can reduce ceiling effects. 25The balance beam assesses balance, locomotor skills and motor coordination, demonstrating differences between young and old mice. 26Additional behavioural measures associated with physical ability can be observed during wire hang and balance beam tests. 25,27Behaviours while completing physical assessment tasks change with age, sex and interventions and may assist in identifying geriatric syndromes like frailty. 28ranslational research can help overcome the limitations of clinical studies in understanding polypharmacy effects in older adults and help identify interindividual variability.Previously in controlled studies of inbred mice, we found that polypharmacy regimens with low and high DBI reduce physical function 5,29 and increase frailty depending on the measure of frailty used alike clinical studies. 30In contrast, a zero DBI polypharmacy regimen lacking anticholinergic or sedative medications did not impair function. 5Our more recent longitudinal study in young and old male and female mice demonstrated that high DBI polypharmacy, but not control, impaired grip strength, gait speed and rotarod latency, with some age and sex interactions. 31owever, this study did not include balance beam and wire hang measures nor their adjuvant behavioural assessments.In the same mice randomised to receive control or high DBI polypharmacy for 6-8 weeks prior to assessment, the current study assesses changes in the balance beam and wire hang, including behavioural assessments, and compares them to the commonly assessed grip strength, aiming the following: Young (2.5-month-old) and old (21.5-month-old) male and female mice underwent pretreatment baseline assessments and then received 6-8 weeks of polypharmacy or control diet prior to follow-up testing.Age groups were selected to correspond with young adults (2 to 6 months old mice) and older adults (20 + months old mice).At pretreatment baseline, male (n = 12 aged 2.5 months, n = 19 aged 21.5 months) and female (n = 12 aged 2.5 months, n = 20 aged 21.5 months) mice were assessed on the grip strength, balance beam and wire hang tests.Following pretreatment testing, these mice were randomised into control (no medication) or high DBI polypharmacy (medication regimen detailed below), which they received for 6 to 8 weeks before being reassessed at approximately age 4.5 months (n = 6 for all young mice subgroups) or approximately age 24.5 months (n = 6-8 for all old mice subgroups).In cases where mice died before the second testing timepoint, only their pretreatment testing values were included in statistical analysis (n = 2 old female control; n = 1 old male control; n = 3 old female high DBI polypharmacy; n = 2 old male high DBI polypharmacy).One old female mouse receiving polypharmacy died before posttreatment balance beam testing but had been assessed for wire hang and grip strength posttreatment, and all assessed test values were analysed.

| Treatment details
Mice completed pretreatment baseline testing at ages 2.5 or 21.5 months and were then randomly allocated to treatment groups stratified by age and sex.Before treatment allocation, all mice received the same food sourced from specialty feeds (Standard Meat Free Mouse and Rat Feed, Specialty Feeds, Western Australia, Australia), which was also the control diet.Polypharmacy treated mice received medications at therapeutic doses, based on previous studies of the minimal oral chronic efficacious dose when administered as monotherapy and estimates of ad libitum food and water intake in mice.The high DBI polypharmacy regimen comprised five medications: simvastatin (20 mg/kg/day), metoprolol (350 mg/kg/day), oxybutynin (27.2 mg/kg/day) and citalopram (15 mg/kg/day) delivered via food pellets mixed within the baseline diet by Specialty Feeds (Western Australia, Australia) and oxycodone (5 mg/kg/day) delivered in drinking water.Delivery via food and water has been demonstrated to be effective through tested plasma drug levels. 32xybutynin, oxycodone and citalopram contributed to the high DBI score (DBI approximately 1.6; DBI > 1 considered high) of the regimen.Selection of drug classes in the regimen was according to frequency of use in the Australian population, while selection of specific medications within the drug classes considered similar pharmacokinetics and pharmacodynamics in mice and humans and potential for drug-drug interactions. 5

| Assessment of physical function and activity
All outcomes were measured at pretreatment (baseline), and 6-8 weeks after allocation to polypharmacy or control groups (during treatment).To reduce cohort effects on experimental outcomes, all five cohorts included mice from both sexes, young and old, and polypharmacy and control groups.Assessments of grip strength, wire hang and balance beam are detailed below.Prior to all experiments, mice were acclimatised in the behavioural observation room for 30 min.Researchers were blinded to mouse ID and treatment/ age/sex groups only for balance beam experiments during test day, but not during grip strength and wire hang experiments.

| Automated grip strength device (TSE grip strength metre)
The TSE grip strength device (AIMEDICAL International Pty Ltd, South Australia, Australia) was used to measure the grip strength.These experiments were done under white light.Mice were held at the base of the tail, elevated from the bench and allowed to grip on the 'Grasping grip' '2-Paw-measurement' handle with their forepaws.Mice were then pulled horizontally away (180 angle) from the bar until the mouse ceased gripping it.The maximum force generated was displayed by the device.Each mouse was tested in five trials per testing day (one testing day at baseline, one during treatment), with an intertrial period of 20 min.The data are displayed as the mean of the five trials ± standard error of mean (SEM).Additional analysis was conducted using maximum grip strength (supporting information Figure S2).

| Wire hang
The wire hang experiment was performed as described previously, 5,33 with modifications detailed below.All experiments were conducted under white light.The wire hang apparatus consists of a 2 mm thick circular wire, elevated 40-50 cm above the bench by two vertical wooden planks and positioned parallel to the horizontal bench.Mice were slowly lowered onto the wire, ensuring that both forearm paws had gripped around the middle of the length of the wire before the researcher let go.The latency to fall was measured.The maximum duration was 60 s for trials 1-2 and 300 s for trial 3. Wire hang time was calculated into two scores to evaluate whether an extended time minimises ceiling effects and compare the scores in differentiating the effects of the independent variables.The WH60s was calculated by averaging the three trials of up to 60 s for each mouse and capping the value at 60s when mice remained hanging after 60 s during trial 3. The WH300s score used wire hang time from trial 3, with a maximum duration of 300 s.
Mice that could not grip the wire longer than 10 s were immediately repositioned on the wire no more than three times, after which the highest value was recorded for their performance in that trial.Three trials were conducted with an intertrial interval period of at least 20 min per mouse.Body weight was measured during the first trial of the experiment.
Compared with the previous study, 5 the experiment included three modifications.Firstly, the stopwatch was paused when mice reached down off the wire and landed, and mice were placed back onto the wire as this is a different behaviour from ungrasping the wire.Secondly, duration of the third and final trial was increased from 60 to 300 s to reduce the ceiling effect.Finally, observations of mouse behaviours during the wire hang trials were calculated into one cumulative behaviour score, which were recorded as a pilot by G.G. when conducting the wire hang with ageing male mice for a previous study 5

| Balance beam
Balance beam assessment occurred over three consecutive days.Each cohort of mice was divided into groups to ensure a maximum of 15 min between the first and last mouse completing a testing trial and to ensure representation of every age/sex/treatment subgroup within each group.Mice were acclimated and tested in a room with dimmed white light, ensuring the higher-intensity white light from the lamp at the start of the beam motivated escape traversing the beam into a less illuminated square box at the end containing tissues.During each of the 3 days mice completed six trials, with three trials on the 12-mm-wide beam, followed by three trials on the 6-mm-wide beam.Reduced beam thickness increased difficulty of crossing, while the lengths of both beams were kept constant at 1 m.The first trial on day 1 involved redirecting the mice during pauses or attempts to turn back, which was not necessary on subsequent trials or days.All other procedures performed on day 1 were repeated on days 2 and 3. Cohort 1 was an exception at pretreatment assessment as mice were trained on the 12-mm beam on all 3 days but tested on the 6-mm beam only on day 3.As this cohort showed the same learning and no difference in performance at treatment testing compared with other cohorts and was equally representative of both sexes and both treatment groups; their data were included.On day 3, G.G. was blinded to treatment group, as mice were taken out of their home cage and placed in temporary clear cages with coded identifiers to blind treatment/sex/age group.Mice were timed manually with a stopwatch (Balance beam latency; average time crossing the 6-mm beam over three trials on day 3) and scored for proportion (0/25/50/75/100%) of beam crossed by walking (1-%dragged), while counting slips and falls.

| Statistical analysis
Increasing time to cross the balance beam indicated lower performance, whereas an increase in grip strength, wire hang and behavioural measures indicated higher performance.
The mixed-model linear regression analyses were conducted separately for all dependent variables using SPSS, followed by a subsequent Benjamini-Hochberg adjustment for multiple comparisons in Microsoft Excel.Figures displaying these relationships as well as heatmap correlations matrix were prepared using GraphPad Prism 9.5.0 software.Pearson's correlation analyses were also conducted on SPSS software and scatterplot graphs prepared on Microsoft Excel.

| Polypharmacy, age and sex effects on different tests
3.1.1| Polypharmacy reduced grip strength in young and old, male and female mice, while reducing balance and wire hang performance in only old female mice.Sex differences in control mice were found in wire hang and balance beam, but not the grip strength Performance on all three tests was impaired in old mice compared with young.Young mice showed greater average and maximum grip strength ( p < 0.05; Figure 1a, supporting information Figure S2), and greater wire hang times both in the three trials of 60 s (WH60s; p < 0.05; Figure 1b) and in one trial of 300 s (WH300s; p < 0.05; Figure 1c).Similarly, in the wire hang cumulative behaviour score, young mice scored higher than the old showing greater functional movements ( p < 0.05; Figure 1d).On the balance beam, irrespective of sex and treatment subgroups, time to cross was greater in old mice ( p < 0.05; Figure 1e), and young mice were more likely to walk rather than drag their hindlimbs across the 6-mm beam (p < 0.05; Figure 1f).However, number of slips from the balance beam were only significantly lower in young females compared with older females ( p < 0.05; Figure 1g), and there was no significant difference between young and old in other subgroups of mice.This may be explained by differences in the proportion of distance spent walking, since slipping usually occurs when walking rather than dragging across the beam.
Among control mice, females generally performed better than males within each age group.On the WH60s, old female control mice hung significantly longer than old males ( p < 0.05; Figure 1b), and the WH300s, young female control mice hung significantly longer than young males ( p < 0.05; Figure 1c).WH300s was the only test to detect a significant difference between young male and female mice.On the balance been, compared with old females, old male mice took longer to cross (p < 0.05; Figure 1e), walked less and dragged more (p < 0.05; Figure 1e) and slipped less ( p < 0.05; Figure 1f).Consistently, on the wire hang, old males demonstrated poorer cumulative functional behaviour than the old females ( p < 0.05; Figure 1d), and this significant sex difference was apparent in all WH behaviours when analysed individually (p < 0.05; supporting information Figure S1a-d).All tests except the grip strength (Figure 1a) showed significant differences between old males and females.However, the grip strength was the only test that showed significant differences between control and polypharmacy treated mice in all subgroups (p < 0.05), irrespective of whether the average of 5 trials (Figure 1a) or the maximum score in five trials (supporting information Figure S2) was used.
Sex differences were not observed in mice receiving polypharmacy, suggesting the effects of the polypharmacy outweighed the effects of sex.However, in WH60s, WH300s, wire hang cumulative behaviour, balance beam time and balance beam %walked, there was a significant reduction in performance in only old female mice receiving high DBI polypharmacy (p < 0.05; Figures 1b-f) and in no other age and sex subgroups.

| Correlations between different physical function measures
3.2.1 | Performances in grip strength, wire hang and balance beam all correlated with each other and showed different relationships across subgroups.Correlations following high DBI polypharmacy differed to those before high DBI polypharmacy Subgroup analyses showed strong correlations between different performance measures.During treatment, both old male control and high DBI polypharmacy subgroups, showed strong positive correlations between WH60s, WH300s and several WH behaviours (Figure 2a,b, p < 0.05), resembling pretreatment patterns (supporting information Figure S3).Old female controls showed a negative correlation between balance beam latency and %walked measures, while these measures were positively correlated in old females receiving polypharmacy (Figure 2c,d, p < 0.05); indicating that those walking across a greater percentage of the beam while receiving polypharmacy took longer to cross the beam than the mice that dragged more, while the opposite was true in controls.Unlike young male controls, young males receiving high DBI polypharmacy showed a significant positive correlation between grip strength and balance beam latency and slips, and a negative correlation between grip strength and balance beam percentage walked (Figure 2e,f, p < 0.05).In young females receiving high DBI polypharmacy, three wire hang measures (WH300s, tail use type, and limbs used) negatively correlated with balance beam latency (p < 0.05).For measures where young female mice performed at ceiling levels (Figures 1b,c,f, supporting information Figure S1a,c), there was a perfect positive correlation (r = 1) (Figure 2g,h).At the sampled population level, only the grip strength showed a significant polypharmacy-related decline in all subgroups of mice, while all other assessments of wire hang and balance beam demonstrated a decline in only old female mice receiving polypharmacy.However, when assessing the performance of individual mice across different physical assessments, more high DBI polypharmacy mice showed decline and less showed improved performance compared with their corresponding nonmedicated control subgroups (Table 1 and supporting information Table S1).Among all age and sex groups, old female HDBI polypharmacy mice were most likely to consistently decline, while across other subgroups numerous mice declined in one test and not the other, with no clear pattern for which test they would show decline on (Table 1 and supporting information Table S1).These differences between individual mice were not always reflected in raw subgroup standard error (SE) values (supporting information Table S2).In young female mice performing at ceiling levels in the wire hang, receiving high DBI polypharmacy resulted in increased SE (supporting information Table S2).In contrast, old female high DBI polypharmacy mice showed significantly decreased performance in the %walked when crossing the balance beam with a reduction in SE during treatment compared with pretreatment (supporting information Table S2).

| DISCUSSION
This is the first randomised-controlled study to characterise the effects of polypharmacy on the balance beam, to assesses adjuvant behavioural measures and also the first to characterise changes on the wire hang in female mice.Previous studies of high DBI polypharmacy in mice demonstrated impaired physical function.This study examined age, sex and polypharmacy effects, comparing findings from different measures of physical function, assessed correlations between tests and interindividual variability between and within subgroups.Old mice showed reduced physical function in all measures compared with young mice, attesting to the suitability of these tests for ageing studies.Control mice showed sex differences on all wire hang and balance beam assessments.Old female control mice displayed longer WH60s endurance, more wire hang cumulative behaviours, faster balance beam crossings with higher proportion walked, albeit with more slips than males.Similarly, young female controls outperformed young males in the wire hang 300 s.Old mice were far from the 60-s wire hang ceiling, so there was no further discrimination with the addition of the 300-s wire hang test in old mice.Neither grip strength, wire hang nor balance beam detected significant sex differences among the mice receiving polypharmacy, suggesting polypharmacy outweighed sex effects on performance.Grip strength detected polypharmacy effects in all subgroups, while only old female mice receiving high DBI polypharmacy declined in all other physical function assessments.Grip strength, wire hang time and behaviour and balance beam crossing time and behaviour were correlated, but these correlations differed between subgroups.Interindividual variability in the pattern of decline across the battery of tests within age, sex and treatment groups suggests that other factors contribute to performance even in this highly controlled model.
Grip strength assesses maximal force exertion and is highly translatable to clinical studies.The current study showed no sex difference in grip strength, unlike prior mouse studies. 34,35These discrepancies may arise from methodological variations; females perform better when all four limbs are assessed simultaneously, 35 and males when only the front two limbs are assessed. 34Repeating grip strength testing over 3 days 34 might have revealed sex differences in the current study.Endurance tests may also be influenced by task-specific factors, as the same mice that showed sex differences in wire hang endurance in the current study showed no sex differences in rotarod  S1.Raw values are summarised in the supporting information Table S2.
endurance, 31 which combines more muscle groups including knee extensors and flexors.Considering rotarod endurance was assessed in a short timeframe, 31 testing duration may also impact sex differences.When provided unlimited daily treadmill access, young 2-month-old female mice show greater distance coverage, with significantly faster velocities than males. 36This sex difference reduced with ageing, showing no significant difference between old male and female mice at 24 months. 36ex differences in strength and endurance may be due to differences in myosin-heavy chain muscle fibres and sex hormones.Compared with male mice, females have increased type 1 slow twitch fibres required for endurance and decreased type 2 fast twitch fibres required for strength. 37As for sex-dependent factors, male mice given oestrogen perform similarly to females, while postovariectomy females perform similarly to males. 37However, unlike the effect of oestrogens in male mice, studies in old female mice that showed grip strength improvement are oestrogenindependent and are related to the presence of young ovarian tissue. 38ranslating sex-specific findings from mice to humans has limitations.Female mice, unlike human females who undergo menopause, exhibit a continuous decline in fertility, retain some primordial follicles and maintain gonadotropin levels. 39Both species experience changes in luteinising hormone release and eventual termination of hormone cycling. 38,39Sex differences in physical function are mostly similar in mice and humans.Females performing better in endurance assessments are task-specific, with sex differences in the fatiguability of elbow flexors more significant than knee extensors. 40Furthermore, human males also show higher grip strength than females throughout the lifespan, including studies with young adults 41 and older adults averaging in their late 60s 42 and early 70s. 43educed grip strength as part of sarcopenia assessments in older adults was associated with loss of type 2 muscle fibres within the vastus lateralis quadriceps muscle of males, but not females. 44ehavioural measures revealed no additional differences between groups compared with time-based assessments but provided further information from the tests.Increased limb dragging, indicating beam crossing difficulty, 27 increased in old female mice receiving high DBI polypharmacy.In contrast, control old females dragged less but slipped more than males.Similarly in the wire hang, control male mice exhibited functional behaviours less than females indicating reduced physical ability, 25 which significantly reduced while receiving high DBI polypharmacy.However, correlations between tests on mice showed differences in behavioural strategies with polypharmacy.Old females showed more correlations between wire hang timebased measures and wire hang behaviours, potentially indicating these behaviours helped increase hanging time after receiving polypharmacy.While before polypharmacy a faster balance beam crossing time was correlated with greater %walked, during polypharmacy, old females that walked more than dragged crossed the beam slower.
Mechanisms underlying polypharmacy effects are unknown and likely multifactorial.Females have a higher prevalence of adverse drug reactions that are not merely due to weight differences and cannot be relieved through dosing adjustments, as preclinical studies suggest sex differences in the physiology of pharmacokinetic responses are an underlying mechanism. 45In this study, primarily old females receiving high DBI polypharmacy declined in physical function, not young females despite their lower body weights, 31 suggesting sex differences in these physiological processes interact with ageing.Among the medications in this polypharmacy regimen, statins and are most characterised for muscular effects.In a previous study, simvastatin administered in 68-week-old male mice for 2 weeks decreased grip strength and running mill activity. 46Despite the same dose of 20 mg/kg/day being delivered in our previous study featuring simvastatin monotherapy, 5 we did not replicate these findings in male mice.One possible explanation for these mixed findings are the variable activity levels of the mice being tested.Mice with higher levels of physical activity prior to statin initiation did not show reduced physical function compared with sedentary and less active mice. 47Statin effects on physical function in humans also report mixed findings, demonstrating increased muscle fatiguability impacting endurance and decreased rate of maximal force generation impacting strength, 48 and also no difference from nontreatment in their strength or endurance effects. 49,50eta-blockers also show mixed findings.Betablockers in rodents reduce endurance exercise performance, more so with beta-2 blockade than beta-1. 51onversely, in our previous mouse study, metoprolol monotherapy (beta-1 selective blockade) increased physical activity of young mice, but old mice receiving these beta-blockers as a constituent of the polypharmacy regimen showed reduced physical activity, suggesting either the findings in old mice were driven by one of the other 4 polypharmacy constituents or differed with age. 29Beta-blockers in high doses, compared with low doses and no medication, increase muscle fatiguability in humans 52 and also show no effects on knee extensors and handgrip muscle strength in healthy participants. 535][56] Age and sex differences in pharmacokinetics and pharmacodynamics may contribute to the differences in polypharmacy effects seen between groups. 57espite being genetically identical and receiving the same food, water and environmental conditions, there was individual variability in the performance of inbred mice in the same subgroup.The old females receiving high DBI polypharmacy showed the most consistent functional impairment.However, declining in one test did not guarantee a decline in all tests, and mice did not consistently decline in the same test as each other.This highlights the importance of using a battery of tests when evaluating the effects of polypharmacy, both in preclinical and clinical studies, to capture individual differences in drug effects.
Future studies can improve on the several limitations of the current study.Firstly, it was not feasible to blind grip strength and wire hang experiments, or conduct experiments during the dark phase.These experiments were conducted during the 12-h light phase of the light-dark cycle, when mice are less active.Additionally, this study does not investigate mechanisms underlying medication effects.Our studies on mice receiving the same polypharmacy regimen found no difference in inflammatory markers 58 but altered microbiome composition following high DBI polypharmacy, 59 which may contribute to some of the observed changes.

| CONCLUSION
All physical function assessments detected changes with ageing.In males, only the grip strength detected polypharmacy effects, while old females receiving high DBI polypharmacy showed decline across all physical assessments.Overall, the behavioural measures were useful adjuvants to assessing performance with time, providing more information during the same physical assessment.Even in this highly controlled model, marked interindividual variability was seen in response to the same intervention on different tests.Future studies can utilise these physical function assessments to demarcate biological mechanisms and understand different sources of interindividual variability.Such biomarkers could be tested in humans as predictors of functional impairment with polypharmacy.
AUTHOR CONTRIBUTIONS Sarah N. Hilmer contributed to study design, data interpretation, acquired project funding and supervised data acquisition and analysis.Gizem Gemikonakli contributed to study design, animal care and experiments, data acquisition, analysis and interpretation and drafted the manuscript.John Mach contributed to study design, data analysis and interpretation and supervised data acquisition.Trang Tran and Harry Wu contributed to animal care, experiments, and data acquisition.All contributed to finalising the manuscript.
and formally evaluated in the current study.Details of behaviours during the wire hang and how they were scored are as follows: Tail use type-0 = Circling perpendicular to floor, no attempt to wrap around wire; 0.5 = Tried to wrap tail but unsuccessful; 1 = Wrapped tail around wire.Reached down-Number of times reached down vertically (90 ) from wire throughout each trial were counted and scored: 0-Did not reach down; 0.5-Reached down 1-2 times in a trial; 1.0-Reached down > 2 times in a trial.Limb use-0 = Two limbs used throughout trial; 0.5 = Alternate between two limbs and four limbs throughout trial; 1 = All four limbs used throughout trial.Travel across wire-0 = No sideways travel across wire; 0.5 = Travelled sideways but did not reach end of wire; 1 = Travelled sideways across and reach end of wire at least once during trial.

F
I G U R E 1 Legend on next page.

3. 3 |
Assessing variability within subgroups and the performance of the same mouse on different tests 3.3.1 | Individual variability between mice within the same subgroup highlights polypharmacy-related changes in physical function that are not reflected at the sampled population level

F I G U R E 1
Bar graphs of pretreatment and during treatment, old and young, male and female mice randomised to receive control or high DBI polypharmacy in (a) grip strength, (b) wire hang latency over three trials (60s max); (c) wire hang trial 3 latency (300 s max); (d) wire hang cumulative behaviour; (e) balance beam (6 mm) average time to cross on day 3 over three trials; (f) balance beam (6 mm) average %walk day 3 over three trials; (g) balance beam (6 mm) average slips on day 3 over three trials.Number of mice by subgroup at both pretreatment and treatment timepoints: old males n = 10-8, old females n = 10-6, young males n = 6 and young females n = 6.Significant differences <0.05 are indicated as '*' when comparing performance of mice receiving control and high DBI polypharmacy from pretreatment baseline to after 6-8 weeks of treatment, '#' when comparing males and females of the same age and treatment group and '%' when comparing young and old mice of the same sex and treatment group.Error bars show standard error of the mean.Linear mixed-model analysis with Benjamini-Hochberg adjustment for multiple comparisons were used for statistical analysis, computed using SPSS software.Baseline performance indicated by checkered columns; performance following treatment indicated by filled columns.F I G U R E 2 Heatmaps display performance during control or polypharmacy treatment, showing Pearson's r correlations between all physical function measures.Number of mice by subgroup at both pretreatment and treatment timepoints: old males n = 10-8, old females n = 10-6, young males n = 6 and young females n = 6.Significant differences <0.05 are indicated as '*'.Correlation matrix prepared and Pearson's correlation coefficient analysed in GraphPad Prism 9. 60 s = latency averaged over three trials of 60 s; 300 s, latency in trial three lasting 300 s maximum; CB, cumulative behaviour; BB, balance beam 6 mm; GS, grip strength; LU, limbs used; OF, old female; OM, old male; RD, reached down; TAW, travel across wire; TUT, tail use type; WH, wire hang; YF, young female; YM, young male.T A B L E 1 Summary by mouse showing change from baseline in five different assessments of physical function.Change from baseline to treatment-allocated: # Àdecline; ¼ Àsame; " Àimproved Change from baseline to treatment-allocated: Wire hang cumulative behaviour score was composed by cumulating the score across four behaviours exhibited while completing the wire hang: Limbs used, Reached down, Tail use type and Travel across wire.BB %walked was calculated as (1 -% of beam dragged), averaged over three trials.Mean and standard error (SE) values have been calculated as part of mixed-model linear regression analysis in SPSS and are therefore estimates.Changes in maximum grip strength from baseline are summarised in the supporting information Table