Differences in adaptations to 1 year of aerobic endurance training: individual patterns of nonresponse

Abstract

Lacking responses to endurance training (ET) have been observed for several variables. However, detailed analyses of individuals' responses are scarce. To learn more about the variability of ET adaptations, patterns of response were analyzed for each subject in a 1-year ET study. Eighteen participants [42±5 years, body mass index: 24±3 kg/m², maximal oxygen uptake (VO_2max): 38±5 mL/min/kg] completed a 1-year jogging/walking program on 3 days/week, 45 min/session at 60% heart rate (HR) reserve. VO_2max, resting HR (rHR), exercise HR (eHR) and individual anaerobic threshold (IAT) were determined by treadmill and cycling ergometry respectively. Intraindividual coefficients of variation were extracted from the literature to distinguish random changes from training responses. Eight participants showed improvements in all variables. In 10 participants, one or two variables did not improve (VO_2max, rHR, eHR and IAT remained unchanged in four, four, three and one cases, respectively). At least one variable improved in each subject. Data indicate that ET adaptations might be detected in each individual using multiple variables of different adaptation levels and intensity domains. Nonresponse seems to occur frequently and might affect all variables. Further studies should investigate whether nonresponders improve with altered training. Furthermore, associations between patterns of nonresponse and health benefits from ET are worth considering.

Epidemiological studies indicate that a high endurance capacity is associated with a low risk of cardiovascular disease (Sandvik et al., 1993; Myers et al., 2002) and risk reduction might even be more strongly associated with fitness improvements than with physical activity alone (Williams, 2001; Myers et al., 2004; Sassen et al., 2009). Training-induced changes in endurance capacity vary considerably between individuals (Bouchard & Rankinen, 2001). Several previous studies have shown that a proportion of subjects demonstrated little or no improvements in maximal oxygen uptake (VO_2max), maximal work rate, submaximal exercise heart rate (eHR), submaximal respiratory exchange ratio or insulin sensitivity despite regular endurance training (ET) (Prud'homme et al., 1984; Kohrt et al., 1991; Bouchard et al., 1999; Skinner et al., 2001; Boulé et al., 2005; Hautala et al., 2006; Vollaard et al., 2009).

Nevertheless, systematic individual analyses of lacking training responses are scarce. Bouchard and Rankinen (2001) demonstrated in the HERITAGE Family Study that after 20 weeks of ET, nonresponders existed for different variables. However, it remained unclear whether these nonresponders were identical subjects for all variables. Vollaard et al. (2009) recently analyzed the magnitude of individual changes in eight variables after 6 weeks of ET. VO_2max low responders were not consistently low responders for other variables, but again, it was not apparent how many variables remained unchanged in each individual. To gain greater knowledge of the interindividual variability of training responses, it seems important to analyze individual patterns of nonresponse in more detail.

In some previous studies, a fixed proportion of subjects with the lowest training response were considered low responders, irrespective of the magnitude of their adaptations (Timmons et al., 2005; Vollaard et al., 2009). Other authors used the term nonresponder without defining it (Bouchard et al., 1999; Skinner et al., 2001). As small endurance changes within the day-to-day variability of a variable cannot be considered a worthwhile training-induced change, nonresponders should be defined as individuals who improve by not more than the biological variability of the respective variable.

The aim of the present study was to identify individual patterns of nonresponse within a tightly monitored 1-year ET study using common indicators of endurance capacity (Scharhag-Rosenberger et al., 2009). The ET prescription was designed to elicit both cardiovascular and metabolic adaptations and fell within the current recommendations for health-related recreational exercise [American College of Sports Medicine (ACSM), 1998]. VO_2max, resting heart rate (rHR), eHR and the individual anaerobic threshold (IAT) were chosen as indicators for endurance capacity. Different levels of adaptation like cardiocirculatory and metabolic changes as well as maximal, submaximal and resting parameters were thereby included. Negative changes and improvements by not more than the biological variability of the respective variable were defined as a nonresponse. This represents a novel approach to evaluate endurance changes on an individual basis.

Materials and methods

Subjects

Twenty-five healthy, middle-aged, untrained individuals were recruited for participation. The relatively small number of participants enabled complete control of compliance over the 1-year training period, which appeared essential for the present study. Participants gave written informed consent after the study had been approved by the local ethics committee. Eighteen subjects (7 males/11 females) completed the 1-year training program. Age, body mass index and pre-training VO_2max of these participants were 42±5 years (32–50 years), 24±3 kg/m² (19–28 kg/m²) and 38±5 mL/min/kg (26–47 mL/min/kg), respectively.

Training

All participants performed 50 weeks of jogging or walking, 3 days/week, 45 min/session with a constant heart rate (HR) prescription of either 60% HR reserve or HR at the lactate threshold (LT) according to Hagberg and Coyle (1983). To ensure fitness changes, the higher HR prescription from both formulas was chosen. Training HR was determined using these two methods to ensure a stimulus to both the cardiocirculatory and the metabolic system. Indeed, the two training prescriptions differed by 2±11 min⁻¹ only. Training intensity corresponded to 62±9% pre-training VO_2max. The training program was designed to be within the recommendations of the ACSM “for developing and maintaining cardiorespiratory and muscular fitness in healthy adults” (ACSM, 1998). All training sessions were recorded by means of telemetric systems (F6, Polar Electro, Kempele, Finland), and the training frequency, duration and average HR of each session in each subject were analyzed to evaluate compliance.

Exercise testing

Following an initial habituation test on a treadmill, maximal treadmill and cycling ergometer tests were conducted before and after the training period. All treadmill tests were performed on an ELG70 (Woodway, Weil am Rhein, Germany). The protocol started with five submaximal exercise stages, individually beginning at 4, 5 or 6 km/h (according to the results of the habituation test) and increasing by 1 km/h every 3 min. After these five stages, the tests continued with rampwise increments by 0.8 km/h every minute until exhaustion. The protocol allowed assessment of submaximal exercise HR and VO_2max in the same test (Meyer et al., 2006, 2007). Resting HR was measured after 10 min of rest in a supine position before the test. Submaximal exercise HR was calculated as the average HR at the end of the first five stages of the treadmill test. VO_2max was determined by means of a metabolic device (MetaMax II, Cortex, Leipzig, Germany) as the highest 30-s average during the rampwise part of the test. It was only analyzed if at least one of the following criteria was fulfilled: (i) leveling off, defined as an increase in oxygen uptake during the last 60 s of the test of <100 mL/min, (ii) maximal HR (HR_max)>(220−age)−10%, (iii) maximal blood lactate concentration (La_max)>8 mmol/L and (iv) maximal respiratory exchange ratio (RER_max)>1.1 (Howley et al., 1995; Midgley et al., 2007). One subject failed to meet these criteria and was excluded from analysis of VO_2max.

For the cycle ergometer tests, an Ergofit261 ergometer (Ergo-Fit, Pirmasens, Germany) was used. The tests started at 25 or 50 W and the workload was increased by 25 or 50 W every 3 min until exhaustion. From the lactate curve attained during the cycle ergometer tests, the IAT was determined according to Stegmann et al. (1981). Determination was impossible in three subjects who showed rather linear lactate increases in the pre-training test presumably due to low fitness levels. All the test protocols remained unchanged between pre- and post-testing for a given subject.

Analyses of individual changes

As a measure of day-to-day variability, within-subject coefficients of variation (CVs) of the four variables were extracted from the literature. They were 7.5% for rHR (Stanforth et al., 2000), 2.7% for eHR (Bagger et al., 2003) and 5.6% for VO_2max (Katch et al., 1982). As no CV for the IAT according to Stegmann et al. (1981) has been published, an average CV of six other LT concepts of 1.9% was used (Pfitzinger & Freedson, 1998). Individual pre–post differences of more than one CV were considered as improvements, whereas negative changes and smaller improvements than one CV were defined as a nonresponse.

Statistics

Outcomes of the entire group are given as means±standard deviations (SD). All data were tested for normal distribution using the Shapiro–Wilks W-test. As this was present throughout most dependent variables, t-tests for repeated measures were chosen to compare pre- and post-training values. In general, training responses can be influenced by age, gender, baseline level and compliance (ACSM, 1998; Bouchard et al., 1999; Spina, 1999). To evaluate whether these factors were associated with training responses in the present study, multiple linear regression analyses were conducted (Statistica 6.0 software for windows). For each of the four endurance parameters, a model was created with age, gender, the parameter's baseline level and compliance (average number of training sessions per week) as independent variables and the pre–post change of the parameter (Δ) as the dependent variable. A P-level <0.05 was considered as significant.

Results

Compliance and maximal effort

The subjects trained on 2.8±0.2 days/week for 48±2 min on average with an HR of 1±1 min⁻¹ above their prescription. Maximal effort during the treadmill tests was not significantly different between pre- and post-training when HR_max, La_max and the number of participants showing a leveling off were considered (HR_max: P=0.17, La_max: P=0.74, leveling off: N=8 both pre- and post training). However, RER_max was significantly lower after the training program (1.11±0.05 vs 1.05±0.03, P<0.001, N=17).

Endurance changes

Training-induced changes of the four indicators of endurance capacity are presented in Table 1 as means, SD and ranges. Table 2 shows the individual patterns of training responses. Nonresponders were found for each variable (horizontal view of Table 2). If Table 2 is read in columns, it shows that eight subjects (44.5%) improved in all four variables. In eight other subjects (44.5%), one variable did not change and in two subjects (11%), two variables did not change. Each subject improved in at least one indicator of endurance capacity.

Table 1.  Mean endurance changes

	Mean ± SD	Minimum–maximum	P
VO2max, maximal oxygen uptake (N=17); rHR, resting heart rate (N=18); eHR, submaximal exercise heart rate (N=18); IAT, individual anaerobic threshold (N=15); SD, standard deviation.
VO2max (L/min)	+0.36 ± 0.32	−0.38 to +0.87	<0.001
rHR (min^−1)	−9 ± 6	−24 to +1	<0.001
eHR (min^−1)	−11 ± 7	−22 to +2	<0.001
IAT (W)	+16 ± 9	−1 to +35	<0.001

Table 2.  Individual endurance changes

Influences on the training effects

Multiple linear regression analyses revealed that the potential influencing factors age, gender, baseline level and compliance were not significantly associated with the changes of the four endurance variables. The statistical results are presented in Table 3.

Table 3.  Potential influences on the training effects

Outcomes of multiple linear regression analyses. VO2max, maximal oxygen uptake; rHR, resting heart rate; eHR, submaximal exercise heart rate; IAT, individual anaerobic threshold.
Age, gender, compliance, baseline level and ΔVO2max	P=0.70	R=0.40	N=17
Age, gender, compliance, baseline level and ΔrHR	P=0.51	R=0.46	N=18
Age, gender, compliance, baseline level and ΔeHR	P=0.21	R=0.59	N=18
Age, gender, compliance, baseline level and ΔIAT	P=0.87	R=0.33	N=15

Discussion

In a population of recreational athletes between 32 and 50 years of age, 1 year of ET within the ACSM recommendations (ACSM, 1998) elicited significant mean changes in the four observed indicators of endurance capacity. However, interindividual variability of the training response was high as demonstrated by wide ranges of training-induced changes and a proportion of nonresponders for each variable. Individual patterns of nonresponse showed that in slightly more than half of the subjects, one or two variables remained unchanged. However, each subject responded in at least one variable. Nonresponse occurred in all variables. The pattern of lacking changes was not influenced by age, gender, compliance or fitness level at the onset of the study.

Ranges of the training adaptations

The present data confirm findings by Bouchard and Rankinen (2001), Vollaard et al. (2009) and others (Prud'homme et al., 1984; Kohrt et al., 1991; Hautala et al., 2006) that training-induced changes in common endurance variables differ considerably between individuals. VO_2max changes in the present study ranged from slight decreases to considerable increases. Similarly, previous investigations revealed VO_2max changes between −5% and +56% after 2 weeks to 1 year of training (Prud'homme et al., 1984; Kohrt et al., 1991; Bouchard & Rankinen, 2001; Skinner et al., 2001; Hautala et al., 2006; Vollaard et al., 2009). However, only mean changes are reported in the vast majority of ET studies. To avoid misleading information, at least ranges of the endurance changes (if not individual responses) should be reported in the future.

Nonresponse in the four examined indicators of endurance capacity

Nonresponders were observed in each of the four examined endurance variables. Their proportion was about the same for VO_2max, rHR and eHR. VO_2max generally depends on the subjects' maximal effort during the exercise test and, therefore, indicators of maximal effort have to be considered (Meyer et al., 2005). In the present study, four such indicators were assessed: RER_max was lower post-training than pre-training, whereas HR_max, La_max and the number of participants showing a leveling off did not change. The lower RER_max might indicate reduced effort and thereby explain lacking changes in VO_2max. The frequency of VO_2max nonresponders might thus have been slightly overestimated. The only variables that remained unchanged in the same individuals were resting and submaximal exercise HR, which is probably due to their physiological similarity. Interestingly, only one nonresponder was observed for the IAT. As the IAT has the lowest intraindividual CV of 1.9%, one might assume that improvements are comparably easy to reach. However, all participants except for the nonresponder demonstrated increases ≥5.6% (which corresponds to the CV of the VO_2max used in this study). Thus, the low proportion of IAT nonresponders was not due to methodological inconsistency between dependent variables. However, as the IAT could not be determined in three subjects, these results are based on a slightly smaller sample.

Individual patterns of nonresponse

The present study, for the first time, evaluated individual patterns of nonresponse to ET. From a physiological point of view, the observed heterogeneous adaptations are not surprising: ET potentially affects multiple organs of the human body (Wilmore & Costill, 2004) and presumably elicits measurable adaptations in some of them in each individual. The present data indicate that a combination of primarily cardiocirculatory and metabolic parameters as well as parameters measured at rest, during submaximal and maximal exercise, might enable detection of ET adaptations in each individual.

Vollaard et al. (2009) reported recently that low responders for VO_2max were not consistently low responders for other variables. In the present study, individual patterns of response showed that nonresponders for one variable usually respond in other variables. As a possible reason for missing training adaptations in some subjects and some variables, Vollaard et al. (2009) discuss that their training stimulus as a fixed percentage of VO_2max might have led to inhomogeneous metabolic responses during endurance exercise, as also described by others (Meyer et al., 1999; Scharhag-Rosenberger et al., 2010). In the present study, however, the training stimulus was designed to relevantly affect both the cardiocirculatory and the metabolic system. Nevertheless, nonresponders were observed for all variables regardless of their cardiocirculatory or metabolic background.

The training program in the present study consisted of prolonged moderate-intensity training sessions as recommended for beginners in recreational physical activity (ACSM, 1998). It remains unclear whether missing endurance changes in some subjects and some variables are caused by the training regimen or are inherent to the subject. The literature suggests that the reasons for nonresponse may be genetic. Familial aggregation of VO_2max changes has been observed in the HERITAGE Family Study (Bouchard et al., 1999) and differences between high and low responders in the activation of angiopoietin 1 and EGF-like domain 2 gene have been found in the human muscle (Timmons et al., 2005). Recently, Timmons et al. (2010) identified 11 single nucleotide polymorphisms that together explained 23% of the interindividual variance in training-induced VO_2max changes, corresponding to about 50% of the estimated genetic variance for VO_2max. They analyzed three groups of subjects who conducted ET programs of different intensities, frequencies and durations. In each group, about 20% of the subjects demonstrated VO_2max changes <5%. One of the training programs consisted of prolonged and interval-based low- to high-intensity training sessions and therewith included all kinds of ET regimens. This finding argues against a systematic impact of the training regimen on nonresponse, but nevertheless it seems worthwhile investigating whether nonresponders can improve the respective variable through customized alterations of the training program.

Regarding the strong relationship between endurance capacity and cardiovascular or metabolic health (Myers et al., 2002), the question arises as to whether a certain pattern of training response is necessary to gain health benefits from ET. Current knowledge indicates that among parameters of endurance capacity, VO_2max is the most important indicator of mortality (Wilmore et al., 2001; Sassen et al., 2009). However, so far, it has not been investigated whether VO_2max nonresponders who demonstrate changes in other endurance variables actually gain less health benefits from ET than VO_2max responders. Nonresponder analyses should also be extended to other health-related variables. Resting blood pressure, total cholesterol, high- and low-density lipoprotein cholesterol, C-reactive protein and uric acid did not change significantly in the present study and were thus not analyzed.

Perspectives

Individual patterns of nonresponse to ET are of considerable interest and were evaluated for the first time in the present study. Nonresponders were defined as subjects who improved by not more than the biological variability of the respective indicator of endurance capacity. The data confirm that ET adaptations vary considerably between individuals as demonstrated before in a couple of studies (Prud'homme et al., 1984; Kohrt et al., 1991; Bouchard & Rankinen, 2001; Vollaard et al., 2009). Individual patterns of nonresponse showed that in more than half of the subjects, some of the observed endurance variables did not improve. Nonresponders were observed for each of the four indicators of endurance capacity and their frequency was the highest for VO_2max and the lowest for the IAT. However, at least one variable responded in each subject, indicating that a detection of endurance changes might be possible in each individual using a combination of variables of different adaptation levels and intensity domains. It seems important to investigate whether nonresponders can improve the respective variable through customized alterations of the training regimen. Furthermore, associations between patterns of nonresponse and health benefits from ET are worth considering.