Physiological Genomics

Genetic influence on exercise-induced changes in physical function among mobility-limited older adults

Thomas W. Buford, Fang-Chi Hsu, Tina E. Brinkley, Christy S. Carter, Timothy S. Church, John A. Dodson, Bret H. Goodpaster, Mary M. McDermott, Barbara J. Nicklas, Veronica Yank, Julie A. Johnson, Marco Pahor


To date, physical exercise is the only intervention consistently demonstrated to attenuate age-related declines in physical function. However, variability exists in seniors' responsiveness to training. One potential source of variability is the insertion (I allele) or deletion (D allele) of a 287 bp fragment in intron 16 of the angiotensin-converting enzyme (ACE) gene. This polymorphism is known to influence a variety of physiological adaptions to exercise. However, evidence is inconclusive regarding the influence of this polymorphism on older adults' functional responses to exercise. This study aimed to evaluate the association of ACE I/D genotypes with changes in physical function among Caucasian older adults (n = 283) following 12 mo of either structured, multimodal physical activity or health education. Measures of physical function included usual-paced gait speed and performance on the Short Physical Performance Battery (SPPB). After checking Hardy-Weinberg equilibrium, we used using linear regression to evaluate the genotype*treatment interaction for each outcome. Covariates included clinic site, body mass index, age, sex, baseline score, comorbidity, and use of angiotensin receptor blockers or ACE inhibitors. Genotype frequencies [II (19.4%), ID (42.4%), DD (38.2%)] were in Hardy-Weinberg equilibrium (P > 0.05). The genotype*treatment interaction was statistically significant for both gait speed (P = 0.002) and SPPB (P = 0.020). Exercise improved gait speed by 0.06 ± 0.01 m/sec and SPPB score by 0.72 ± 0.16 points among those with at least one D allele (ID/DD carriers), but function was not improved among II carriers. Thus, ACE I/D genotype appears to play a role in modulating functional responses to exercise training in seniors.

  • aging
  • exercise
  • genetics
  • disability
  • ACE gene

as the number of older men and women continues to rise worldwide, maintaining physical independence among older adults is an important public health challenge. The capacity to perform basic physical tasks is a central tenet of health-related quality of life (40) and a key predictor of adverse health outcomes including hospitalization, postsurgical morbidity, and mortality (1, 19, 46, 55). To date, physical exercise is the only intervention consistently demonstrated to attenuate functional decline among older adults, as numerous studies have reported that regular exercise improves performance on a variety of functional tasks (12, 16, 34, 38, 42). However, variability exists in the responsiveness of older adults to exercise training even under well-controlled, experimental conditions (31). Furthermore, there is an established body of literature documenting response heterogeneity across a variety of subject populations, training protocols, and dependent outcomes (9, 11, 13, 15, 26, 56). Accordingly, many seniors do not experience clinically significant improvements in physical function despite good adherence to exercise programs (13, 35).

While these studies highlight the fact that there is large variation in the way individuals respond to regular training, a comprehensive understanding of the factors that contribute to poor exercise responsiveness is far from complete. One potential source of this variability in the responsiveness to physical exercise lies within the human genome. Indeed, there is a small but established literature documenting the influence of genetic variability on physiological adaptations following long-term engagement in various modes of exercise (8, 51). To date, the best-characterized genetic source of variability in exercise responsiveness is the presence (insertion, I allele) or absence (deletion, D allele) of a 287 bp fragment in intron 16 of the angiotensin-converting enzyme (ACE) gene locus (50). ACE functions both systemically and within tissue-specific renin-angiotensin systems (RAS) to catalyze the conversion of angiotensin 1 into angiotensin 2 and to catalyze the breakdown of bradykinin, a potent vasodilatory substance. Comparatively, the I allele is associated with lower systemic ACE activity and an increased half-life of bradykinin, while the D allele is associated with higher ACE activity higher concentrations of angiotensin 2 and lower bradykinin concentrations (2, 39, 50). These genotype-dependent differences in RAS modulation have important implications for modifying adaptive responses to exercise. Lower bradykinin concentrations associated with the D allele are indicative of less efficient substrate use, while greater conversion of angiotensin 1 to angiotensin 2 has been demonstrated to augment overload-induced skeletal muscle hypertrophy (22, 57). As such, the mode of exercise appears to be an important factor in determining the influence of the I/D polymorphism on adaptations to training. Accordingly, although the data are not entirely consistent, longitudinal intervention studies and cross-sectional studies from young athletes have predominantly reported that individuals homozygous for the I allele (II) display strong responses to aerobic training while the DD genotype is associated with better responsiveness to resistance training (45, 48). In summary, the ACE I/D polymorphism appears to modulate exercise responsiveness in young adults in a manner that is dependent on the mode of exercise performed.

However, few studies to date have evaluated the influence of the ACE I/D polymorphism on changes in physical function among older adults. Moreover, this question is further complicated by the fact that many exercise regimens designed to improve physical function among older adults are multimodal (i.e., aerobic + resistance training) as professional guidelines suggest (3, 4, 43). Studies are needed to evaluate the influence of the I/D polymorphism on functional responses of older adults to multimodal exercise interventions. Accordingly, the objective of the present study was to evaluate the association of I/D genotypes with changes in physical function among older adults following 12 mo of either structured, multimodal exercise training or health education. This study is an ancillary study to the Lifestyle Interventions and Independence for Elders Pilot (LIFE-P) study, a multisite, randomized controlled trial (RCT) that compared, in older adults at risk for physical disability, the effects of a 12 mo structured physical activity (PA) intervention to those of a successful aging (SA) health education program. As published previously, LIFE-P demonstrated an overall beneficial effect of the PA intervention on the physical function of older adults (34). Two of the primary functional outcomes of the trial were usual-paced gait speed measured during a 400 m test and performance on the short physical performance battery (SPPB), a battery of short-duration mobility tasks validated for use among older adults (25). These tests have high clinical relevance as they have proven reliable and valid for predicting adverse health outcomes among seniors (24, 44, 46, 55). The present study evaluates the effects of the interventions on gait speed and SPPB score among Caucasian participants in LIFE-P according to ACE I/D genotype.



Details about specific study inclusion and exclusion criteria of LIFE-P have been reported (34, 49). In brief, subjects were eligible if they were between 70–89 yr of age, were sedentary, had an SPPB score ≤ 9, and were able to walk 400 meters within 15 min. A total of 424 participants were randomized into the PA and SA arms at four sites (Cooper Institute, Stanford University, University of Pittsburgh, and Wake Forest University) and followed for at least 12 mo. All participants provided written informed consent, and the study protocol was approved by the Institutional Review Boards of all participating institutions. The present study includes data only from Caucasian participants because the number of minority participants was insufficient to perform the race-specific analyses necessary to control for population stratification due to racial/ethnic differences in I/D variant frequencies (37, 41) and the association between genotype and serum ACE levels (6).

Medical screening.

To ensure the safety of the exercise program for potential participants, a thorough medical screening was performed (29). These screening procedures included the completion of a medical history questionnaire, an electrocardiogram, a physical exam performed by a study physician, and completion of the Community Healthy Activities Model Program for Seniors physical activity questionnaire (54). Furthermore, participants were asked to bring to this screening visit all prescription and nonprescription medications they had taken in the previous 2 wk. Drug names and whether the medication was prescribed were recorded. Medications were later coded to reflect their function (e.g., antihypertensive) and drug class (e.g., ACE inhibitor, diuretic, β-blocker, etc.).

SA intervention.

The SA intervention (serving as an attention control group) was designed to provide health education and maintain contact with participants. Those randomly assigned to SA attended weekly group presentations for the first 26 wk and then monthly until the end of the trial. Presentations were given on health topics that were relevant to older adults such as nutrition, medication use, foot care, and preventive medicine. All SA participants received basic information about physical activity participation, and each class was concluded with upper extremity stretching. Regular telephone contact was made to encourage participation.

Physical activity intervention.

Participants randomized to the PA intervention performed walking, strength, flexibility, and balance training in both center- and home-based settings. In the adoption phase (weeks 1–8), three supervised center-based physical activity sessions per week were conducted. These sessions were 40–60 min in length and used to initiate the walking program and to introduce participants to the strength, stretching, and balance exercises in a safe and effective manner. The strength exercises included standing chair squats, toe stands, leg curl, knee extensions, and side hip raises with ankle weights. The balance exercises involved a series of dual and single leg standing movements.

The intensity of training was gradually increased over the first 2–3 wk. Moderate intensity exercise was promoted and assessed by the Borg scale (7), a numerical scale indicating a rating of perceived exertion from minimal (6) to maximal (20). Participants were asked to walk at a target intensity of 13 (somewhat hard) and perform strength training at an intensity of 15–16 (hard). In the transition phase (weeks 9–24), the number of center-based sessions was reduced to two times per week and home-based activities were increased. In the maintenance phase (week 25 to the end), participants were encouraged to perform home-based physical activity a minimum of 5 days per week, and one weekly center-based session was offered. The maintenance phase was continued until the final closeout assessment visits.

Measures of physical function.

The primary measures of physical function for this analysis were usual-paced gait speed during a 400 m test and score on the SPPB. During the 400 m walk test, participants were asked to walk 10 laps of a 40 meter course at their usual pace. Participants were allowed to stop and rest if necessary, but without sitting. If a participant did not complete the course, the time and distance completed were recorded and gait speed was determined accordingly. The SPPB is a well-validated battery of tasks designed to examine lower-extremity function in older adults. The test is based on timed measures of standing balance, short duration walking (4 m), and ability to rise from a chair. Each task was scored on a 0–4 scale depending on the ability and time needed to complete each task. A summary score (range, 0–12) was subsequently calculated by summing the three scores (25).

DNA collection and genotyping.

DNA was isolated from whole blood, and the concentration and purity (Abs260/Abs280) were determined by a UV spectrophotometer. The ACE I/D polymorphism was determined using polymerase chain reaction (PCR) amplification with subsequent visualization of PCR products on 2% agarose gels by electrophoresis. The sequences of the sense and anti-sense primers were 5′-CTGGAGACCACTCCCATCCTTTCT-3′ and 5′-GATGTGGCCATCACATTCGTCAGAT-3′, respectively. The insertion allele (I) is detected as a 490-base pair band, and the deletion allele (D) is visualized as a 190-base pair band. The PCR products were visualized independently by two laboratory technicians, and genotypes that were not scored identically were reanalyzed. Because the D allele in heterozygous samples is preferentially amplified (52), all identified DD samples were reamplified using a primer pair that recognizes an insertion-specific sequence (5′-TGGGACCACAGCGCCCGCCACTAC-3′, 5′-TCGCCAGCCCTCCCATGCCCATAA-3′). The reaction yields a 335-base pair band only in the presence of an I allele and no product in DD homozygotes. This procedure identifies ∼1% of true ID genotypes that are misclassified as DD with the initial primers.

Statistical analysis.

Baseline characteristics across genotypes were compared by analysis of variance (ANOVA) for continuous variables and chi-square tests for discrete variables. An a priori alpha level of 0.05 was established to determine statistical significance. When continuous values were nonnormally distributed, the Kruskal-Wallis test was used for comparisons. To compare baseline characteristics of randomized groups within each genotype, t-tests for independent samples (continuous variables) and chi-square tests (discrete variables) were used. The Mann-Whitney test was used when continuous data were nonnormally distributed. Statistical evaluation of baseline characteristics was conducted with SPSS Statistics version 21 (IBM, Armonk, NY).

The central outcomes of interest for this study were the genotype*treatment interaction for the change in gait speed (primary) and SPPB (secondary) at the 12 mo study visit. Genotype association testing was conducted with JMP Genomics version 6.0 (SAS, Cary, NC). Hardy-Weinberg equilibrium was checked using the chi-square test. The genotype*treatment interaction for the change in each outcome (gait speed, SPPB) at 12 mo was evaluated by linear regression according to 2-degree of freedom general association genetic model. To adjust for potentially confounding factors, covariates included in the model were clinic site, body mass index, age, sex, baseline level of the outcome, the use of either angiotensin receptor blockers (ARBs) or ACE inhibitors, and comorbidity. As published previously (13), comorbidity was calculated as a composite score based on the presence/absence of 10 prevalent comorbidities: hypertension, heart attack, heart failure, stroke, cancer, diabetes, broken hip, arthritis, liver disease, and lung disease. Because of the potential for variability due to the inclusion of both normotensive and hypertensive individuals, an exploratory analysis was also performed to evaluate responses only among persons with hypertension.

Following association analyses, an a posteriori analysis was conducted to evaluate the estimated overall effect of the exercise intervention with the II genotype to those with at least one copy of the D allele (ID/DD). These data were calculated for each genotype as the change in each participant in the PA group minus mean change in the SA group. To examine the clinical significance of the genotype*treatment interaction, the proportion of participants who obtained clinically significant benefits at 12 mo was computed between those with the II genotype and to those with ID/DD for 400 m gait speed and SPPB score. Participants were coded as displaying clinically significant improvement, no clinically significant change, or clinically significant decline in each measure based on previously established cut-points (SPPB: ± 1 point; gait speed: ± 0.05 m/s) (33, 47) The chi-square test was used to determine for each genotype group (II, ID/DD) if the proportion of clinically significant changes differed between intervention arms and Bonferroni-corrected z-tests were used to determine which change classification(s) differed.

Finally, linear regression models were created to evaluate, relative to traditional clinical predictors, the utility of the ACE I/D polymorphism in predicting changes in gait speed and SPPB performance among participants in the PA intervention. Models were created for each performance outcome by the stepwise procedure, which computes a series of F tests to determine the optimal model. Variables were entered into the model at P ≤ 0.05 and removed from the model at P ≥ 0.10. Clinical predictors added to the model were age, sex, body mass index (BMI), comorbidity, intervention attendance, baseline level of function (i.e., gait speed or SPPB score), and single nucleotide polymorphism (SNP) recoded according to a recessive model where II = 0 and ID/DD = 1.



The mean age of the study sample was 77.2 ± (SD = 4.3) yr; 65.7% were women, 82.0% had hypertension, and 44.5% were taking either an ACE inhibitor or ARB. On average, participants were also overweight (BMI: 29.7 ± 5.3) with elevated systolic blood pressure (131.7 ± 17.8 mmHg). A total of 55 participants (19.4%) possessed the II genotype, 120 (42.4%) possessed the ID genotype, and 108 (38.2%) possessed the DD genotype. The ACE I/D genotype distributions for each randomized group were in Hardy-Weinberg equilibrium (P > 0.05). Participant demographic characteristics are listed in Table 1 by genotype; no significant differences in demographic characteristics were observed between genotype strata. Likewise, demographic characteristics were largely similar between PA and SA intervention groups within each genotype strata (Table 2).

View this table:
Table 1.

Baseline characteristics of Caucasians in LIFE-P according to ACE I/D genotype

View this table:
Table 2.

Baseline characteristics of Caucasians in LIFE-P according to genotype and randomized group

Self-reported physical activity and exercise adherence.

At baseline, the overall amount of self-reported moderate-intensity activity completed per week [median: 30 min/wk (25th–75th %: 0–135 min/wk)] did not significantly differ between genotypes (P = 0.624) or intervention groups (P = 0.728). After the 12 mo intervention, activity levels remained similar across genotypes (P = 0.457), but a significant difference in moderate-intensity activity was observed between intervention groups [PA: 135 (0–270); SA: 60 (0–188) min/wk; P = 0.008]. Among those in the PA intervention, attendance to the intervention sessions did not differ between genotypes (P = 0.469). Over the course of the study, median PA session attendance for participants within this sample was 74% (25th–75th percentiles, 48–84) for participants with the II genotype, 73% (53–83) for those with the ID genotype, and 68% (45–86) for those with the DD genotype.

Physical function.

At baseline, functional performance of this study sample was similar to that of the full LIFE-P cohort (34). Mean gait speed of the study sample was 0.86 ± 0.18 m/s and did not significantly differ between genotypes (P = 0.096) or intervention groups (P = 0.553). Baseline SPPB scores (7.6 ± 1.4 points) were similar between intervention groups (P = 0.687), but a significant difference was observed among genotypes [P = 0.029 (II: 7.1 ± 1.6; ID: 7.7 ± 1.4; DD: 7.5 ± 1.5)]. Importantly, however, baseline SPPB scores were similar between intervention groups within each genotype strata (all Ps > 0.500).

After 12 mo, gait speed among participants in the SA intervention had significantly declined (unadjusted change = −0.05 ± 0.12 m/s), while gait speed of those in the PA intervention remained stable (unadjusted change = 0.00 ± 0.14 m/s; P = 0.005 between groups). Meanwhile, unadjusted changes in SPPB were 1.0 ± 2.0 points for the PA intervention and 0.4 ± 1.9 points in the SA intervention (P = 0.037 between groups). The distribution of gait speed and SPPB changes is shown in Fig. 1. These figures highlight the heterogeneity of responses to each of the interventions. Among those in the PA intervention, ∼29% of participants improved gait speed by at least 0.1 m/s, while ∼33% experienced a decline of 0.1 m/s or more. Meanwhile, ∼37% of PA participants improved their SPPB score by at least 2 points, while another 37% experienced either no change or a decline in SPPB score.

Fig. 1.

Distribution of changes in performance. Change in gait speed (top) and short physical performance battery (SPPB) performance (bottom) among participants in the physical activity (left) and health education (right) intervention groups. Data for gait speed indicate individual responses to the interventions, while SPPB data reflect the proportion of individuals with a respective change in SPPB score following treatment.

The genotype*treatment interaction was significant for both 400 m gait speed (P = 0.002) and SPPB (P = 0.020). Changes in function are displayed in Fig. 2 by genotype and randomized group. These data reflect changes from adjusted means of 0.87 m/s and 7.5 points for gait speed and SPPB, respectively. These data indicate that, relative to the SA intervention, exercise improved performance on both tests among persons with either the ID or DD genotype. However, the change in performance among those with the II genotype who exercised was actually poorer than those who received health education. In contrast, exercise improved gait speed by 0.06 m/s and SPPB score >0.7 points among those with ID/DD genotypes (Fig. 3). Among only hypertensive individuals, the data were similar with the genotype*treatment interaction reaching statistical significance for gait speed (P = 0.008) but not for SPPB (P = 0.071).

Fig. 2.

Change in function by genotype and treatment. Change in 400 m usual-paced gait speed and SPPB score among Caucasian participants by ACE I/D genotype and randomized treatment group. Data adjusted for clinic site, body mass index, age, sex, baseline level of the outcome, use of either angiotensin receptor blockers (ARBs) or ACE inhibitors and comorbidity. Columns indicate adjusted mean, bars indicate SE. Numbers within columns indicate sample size.

Fig. 3.

Influence of exercise on function by genotype. Effects of the LIFE-P exercise program on 400 m usual-paced gait speed and SPPB score by ACE I/D genotype among Caucasian participants. Data expressed as change among participants in the exercise group minus mean change in the control group from baseline to 12 mo. Columns indicate adjusted mean, bars indicate SE.

Measures of clinical meaningfulness.

With respect to promoting a clinically significant improvement in gait speed, a significant difference existed between intervention arms for ID/DD carriers (P = 0.018) but not II carriers (P = 0.930). Among ID/DD carriers, 29.9% of those in the PA group experienced a clinically significantly improvement in gait speed (0.05 m/s) compared with only 13.7% in the health education group (Fig. 4). In contrast, the proportion of II carriers who experienced a clinically significant improvement in gait speed was actually lower among those who exercised (18.5%) than those who received health education (20.0%). These differences in clinically significant outcomes between genotype groups were even starker for the change in SPPB performance. Again, a significant difference existed between intervention arms for ID/DD carriers (P = 0.015) but not II carriers (P = 0.275). For ID/DD carriers, 68.2% of those who exercised made clinically significant improvements in SPPB performance compared with 51.5% who received health education (Fig. 5). Among II carriers, however, these outcomes were nearly reversed as the proportion of exercisers experiencing clinically significant improvements in SPPB performance (42.9%) was actually 21.1% lower than that of the health education group (64.0%). Moreover, the percentage of II exercisers who obtained a clinically significant improvement in SPPB performance was actually lower than that of the ID/DD individuals who received health education (42.9% vs. 51.5%). Regarding the prevention of clinically significant declines in SPPB performance, exercise prevented these declines among 15.6% of participants with the ID/DD genotypes. For II individuals, however, participants in the exercise intervention again performed poorly compared with those receiving health education as the proportion of clinically significant declines was higher among exercisers (39.3% vs. 28.0%).

Fig. 4.

Clinically significant changes in gait speed. Proportion of study participants obtaining clinically significant benefits in usual-paced gait speed following 12 mo of exercise, as measured during the 400 m walk test. Data are dichotomized by persons with (ID/DD) and without (II) at least 1 copy of the D allele and depict proportional changes in both the exercise [physical activity (PA)] and health education [successful aging (SA)] intervention groups. The chosen clinically significant change was 0.05 m/s as established in the literature (33, 47). *P < 0.05 between intervention arms.

Fig. 5.

Clinically significant changes in SPPB score. Proportion of study participants obtaining clinically significant benefits in lower extremity function following 12 mo of exercise, based on changes in SPPB score. Data are dichotomized by persons with (ID/DD) and without (II) at least 1 copy of the D allele and depict proportional changes in both the exercise (PA) and health education (SA) intervention groups. The chosen clinically significant change was 1 point as established in the literature (33, 47). *P < 0.05 between intervention arms.

Regression modeling.

For gait speed, regression modeling indicated that that ACE I/D genotype was a significant predictor of functional changes relative to traditional clinical factors (Tables 3 and 4). For gait speed, the strongest predictor of response was comorbidity burden, which explained roughly 7% of the variance in response. Meanwhile, adherence to the intervention, genotype, and baseline gait speed explained ∼3, 4, and 5% of the variance, respectively. BMI, age, and sex did not contribute significantly to the final model. For change in SPPB performance, baseline function and age each explained ∼5% of the variance, while genotype again explained ∼4% of the variance. BMI and adherence each contributed to ∼3% of the variance, while sex and comorbidity were not significant to the model.

View this table:
Table 3.

Results of the stepwise regression model for the gait speed response to the exercise intervention

View this table:
Table 4.

Results of the stepwise regression model for the SPPB response to the exercise intervention


Exercise training is an important behavioral intervention that has numerous beneficial health effects. However, as reviewed by members of our group and others (11, 14, 45), there is large interindividual variation in the magnitude of change in performance following training. Extensive evidence has now demonstrated that genetic variability is a key contributor to this heterogeneity of exercise responsiveness (8, 51). The ACE I/D polymorphism is the most studied genetic variant in terms of its contribution to changes in physical performance following exercise training. Biologically, this variant is responsible for a significant proportion of interindividual variance in serum ACE levels and accordingly is a critical regulator of blood pressure (50). Moreover, this variant not only modulates systemic ACE activity but also influences local metabolic function within tissue-specific RAS such as that within skeletal muscle (18, 28).

While many studies have investigated the association of ACE I/D genotypes with changes in performance among younger individuals, few have evaluated the influence of this polymorphism on changes in physical performance of older adults following training. Moreover, the results of this select number of studies have been mixed. Accordingly, the objective of this study was to evaluate the influence of the ACE I/D genotype on changes in physical function among Caucasian older adults following participation in a 12 mo randomized clinical trial comparing center-based, multimodal exercise to health education for the improvement of physical function among older adults. The findings here suggest that variability in ACE I/D genotype may modulate posttraining changes in physical function among this population. To our knowledge, this is the first study to report findings from a multisite RCT to directly implicate the ACE I/D polymorphism as a source of variability in seniors' functional responses to exercise. More specifically, the present data indicate that Caucasian seniors who inherited the II genotype displayed an impaired responsiveness to training compared with persons with the ID or DD genotypes and that the genetic influence is similar in magnitude to that of clinical indices such as age and comorbidity burden.

To our knowledge, the first data related to older adults regarding a potential interaction between physical activity and the ACE I/D polymorphism were epidemiologic findings from the Health, Aging, and Body Composition Study. In this study, Kritchevsky et al. reported that the II genotype was associated with an increased risk of incident mobility limitation among physically active, but not sedentary, older adults (32). Subsequent training studies appeared to indicate, as in younger individuals, a genotype effect that was dependent on the mode of exercise utilized. For instance, Giaccaglia et al. (21) reported that, while responses were robust among DD carriers, older adults with the II genotype did not experience significant increases in knee extensor strength following 18 mo of resistance training. In contrast, however, Defoor et al. (17) reported that improvements in aerobic power were greatest in II carriers following 3 mo of aerobic training.

LIFE-P differed from each of the training studies as it utilized a multimodal intervention intended to enhance physical function across a variety of physiological dimensions (i.e., aerobic power, strength, balance, flexibility). Based on the prior literature referenced above, the present results are surprising given that walking was the primary mode of exercise within the combination intervention. It is possible that the selection of functional outcomes is responsible for the findings favoring carriers of the D allele. The SPPB encompasses short-duration tests for which performance depends minimally on aerobic capacity. Additionally, it is possible that results may have differed had a maximal gait speed been utilized during the walk test rather than usual-paced gait speed.

There are several potential consequences of these findings. First, a person's I/D genotype could be utilized to identify individuals likely to display a less robust response to traditional training and indicate the need for the development of an alternative training strategy. This personalized strategy may include the use of adjuvant therapies or adjustments in the frequency, volume, mode, and/or intensity of training. Conceptually, it might be possible that testing for ACE I/D genotype and other variants with replicated findings could be implemented similarly to pharmacogenetic implementation programs designed to prevent adverse responses to drug therapy (53). However, additional investigation is needed before such an implementation program could become a reality. For instance, genetic associations may be identified because the identified SNP is in linkage disequilibrium with another which is actually the functional polymorphism. Accordingly, replication of genetic findings is a critical step in determining the true value of the association (23, 27). This step is particularly important here given the inconsistency in the literature as well as the apparent influence of exercise modality and/or performance outcome on the association between genotype and change in performance. A potential strength of this study is the availability of the primary LIFE trial, which includes a larger sample size (n = 1,635) and utilizes the same study interventions (20, 36). This availability of the LIFE trial may provide a unique opportunity to replicate this association.

Additional information is also needed regarding the interactions between ACE I/D genotype and RAS-modulating drugs in combination with physical exercise. For instance, members of our group previously reported that ACE inhibitor use by seniors is associated with improved functional responses to exercise (13). However, the present study indicates that those with the D allele, associated with higher concentrations of ACE, demonstrated the most robust improvements in function. Investigations are needed that will evaluate the pharmacogenetic influences of ACE activity in combination with exercise. Though we controlled for ACE/ARB use in our analyses, the study sample size is insufficient to conduct a pharmacogenetic analysis to evaluate drug*genotype interactions among those who received physical exercise. Though speculative, the findings of this study may be at least partially attributable to undocumented issues of hypotension in II carriers treated with ACE inhibitors or ARBs. Independently, exercise (30) and antihypertensive medications (5) each pose risks for older adults related to the development of hypotension, including syncope or falls. It is possible that adverse reactions to the combination of exercise and antihypertensive drugs may be magnified among II carriers due to lower basal ACE activity. Future larger-scale studies are needed to investigate this hypothesis.

The strengths of the present study include the clinically relevant population and study outcomes, the multisite RCT design, and the relatively long period of intervention and follow-up. However, like any study, the present investigation is not without limitations. For instance, though the study had a relatively large sample size, it was insufficient to conduct analyses for races other than Caucasians. Secondly, the addition of serum ACE levels would have strengthened this study's findings. Thirdly, the multimodal nature of the intervention could be considered a weakness as it impedes the interpretation of the effects of individual exercise components (aerobic, resistance, etc.). Moreover, the inclusion of home-based exercise could have affected the overall dose of walking achieved across genotype strata. However, multimodal intervention that includes home-based activity follows clinical guidelines and therefore may be more relevant and/or generalizable than regimens that include only a single modality or are solely center based. Additional measures of exercise adherence or intensity may have aided in the interpretation of study findings. Finally, the inclusion of only one SNP is a limitation in identifying the full impact of genetic polymorphisms on seniors' responses to exercise. Larger-scale studies are needed to identify additional polymorphisms that contribute to response heterogeneity. Conduct of such studies will be a critical step in developing personalized guidelines for exercise therapy.

Despite these limitations, this study was the first to our knowledge to report findings from a multisite RCT to indicate that ACE I/D genotype modulates older adults' changes in physical function following engagement in long-term exercise training. Future studies are needed to further evaluate this association in nonwhite racial/ethnic groups and to identify other variants associated with changes in physical function. Moreover, larger-scale studies are needed to evaluate interactions between ACE I/D genotype and antihypertensive medications in conjunction with exercise training.

In conclusion, the present study demonstrated that ACE I/D genotype influences older adults' functional responses to regular exercise as changes in physical function were impaired among II carriers compared with ID/DD carriers. Accordingly, adjuvant or alternative therapies may be needed to improve physical function among older adults with the II genotype. These findings add to prior literature that has indicated genetic contributions to heterogeneity in exercise responsiveness. Future studies are needed to provide further data regarding the influence of other polymorphisms as well as the concomitant influence of other environmental factors (e.g., medication use) to develop the optimal exercise prescription for these individuals. Such studies will have an important impact on the development of tailored interventions for the prevention of physical disability among older adults.


This work was supported by the following grants from the National Institutes of Health/National Institute on Aging: Lifestyle Interventions and Independence for Elders Pilot (U01AG-22376), Claude D. Pepper Older Americans Independence Centers (University of Florida, P30AG-028740; Wake Forest, P30AG-21332; Pittsburgh, P30AG-024827).


No conflicts of interest, financial or otherwise, are declared by the author(s).


Author contributions: T.W.B., C.S.C., B.J.N., and M.P. conception and design of research; T.W.B. and F.-C.H. analyzed data; T.W.B., F.-C.H., T.E.B., C.S.C., T.S.C., J.A.D., B.H.G., M.M.M., B.J.N., V.Y., J.A.J., and M.P. interpreted results of experiments; T.W.B. prepared figures; T.W.B. drafted manuscript; T.W.B., F.-C.H., T.E.B., C.S.C., T.S.C., J.A.D., B.H.G., M.M.M., B.J.N., V.Y., J.A.J., and M.P. edited and revised manuscript; T.W.B., F.-C.H., T.E.B., C.S.C., T.S.C., J.A.D., B.H.G., M.M.M., B.J.N., V.Y., J.A.J., and M.P. approved final version of manuscript; T.E.B., T.S.C., B.H.G., B.J.N., and M.P. performed experiments.


LIFE-P Investigators are at Cooper Institute, Dallas, TX: Steven N. Blair (field center principal investigator), Timothy Church (field center coprincipal investigator), Jamile A. Ashmore, Judy Dubreuil, Georita Frierson, Alexander N. Jordan, Gina Morss, Ruben Q. Rodarte, Jason M. Wallace; National Institute on Aging: Jack M. Guralnik (coprincipal investigator of the study), Evan C. Hadley, Sergei Romashkan; Stanford University, Palo Alto, CA: Abby C. King (field center principal investigator), William L. Haskell (field center coprincipal investigator), Leslie A. Pruitt, Kari Abbott-Pilolla, Karen Bolen, Stephen Fortmann, Ami Laws, Carolyn Prosak, Kristin Wallace; Tufts University: Roger Fielding, Miriam Nelson; University of California, Los Angeles, Los Angeles, CA: Robert M. Kaplan; VA San Diego Healthcare System and University of California, San Diego, San Diego, CA: Erik J. Groessl; University of Florida, Gainesville, FL: Marco Pahor (principal investigator of the study), Michael Perri, Connie Caudle, Lauren Crump, Sarah Hayden, Latonia Holmes, Cinzia Maraldi, Crystal Quirin; University of Pittsburgh, Pittsburgh, PA: Anne B. Newman (field center principal investigator), Stephanie Studenski (field center coprincipal investigator), Bret H. Goodpaster, Nancy W. Glynn, Erin K. Aiken, Steve Anthony, Judith Kadosh, Piera Kost, Mark Newman, Christopher A. Taylor, Pam Vincent; Wake Forest University, Winston-Salem, NC: Stephen B. Kritchevsky (field center principal investigator), Peter Brubaker, Jamehl Demons, Curt Furberg, Jeffrey A. Katula, Anthony Marsh, Barbara J. Nicklas, Jeff D. Williamson, Rose Fries, Kimberly Kennedy, Karin M. Murphy, Shruti Nagaria, Katie Wickley-Krupel; Data Management, Analysis and Quality Control Center (DMAQC): Michael E. Miller (DMAQC principal investigator), Mark Espeland (DMAQC coprincipal investigator), Fang-Chi Hsu, Walter J. Rejeski, Don P. Babcock, Jr., Lorraine Costanza, Lea N. Harvin, Lisa Kaltenbach, Wei Lang, Wesley A. Roberson, Julia Rushing, Scott Rushing, Michael P. Walkup; Yale University: Thomas M. Gill.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
View Abstract