Key factors influencing cycling performance and overall race time in the Ironman 70.3 for amateur athletes

Triathlon is a multidisciplinary sport that consists of swimming, cycling, and running in that order [1]. The distance covered by each discipline can range from very short to very long. Sprint distance (0.75 km swimming, 20 km cycling, and 5 km running), Olympic distance (1.5 km swimming, 40 km cycling, and 10 km running), Half Ironman (IM 70.3; 1.9 km swimming, 90 km cycling, and 21 km running), and Ironman total distance (IM 140.6; 3.8 km swimming, 180 km cycling, and 42 km running) are the most common triathlon events [2].

Researchers and coaches have recently been interested in the best discipline to predict triathlon performance [2, 3]. Given that the distance covered by swimming, cycling, and running varies significantly between triathlon race distances, the discipline that best predicts the overall performance varies between Sprint, Olympic, Half Ironman, or Ironman race distances. Swimming is the most predictive discipline among professional athletes in Sprint- and Olympic-distance triathlons, whereas cycling is the most predictive in IM 70.3 [2]. Cycling and running were the most predictive in IM 70.3 among amateurs [4]. Considering professional athletes (those who completed the Ironman total distance in below 8 h) cycling split time better predicted overall race time in the Ironman total distance [3].

Given the importance of cycling for overall performance in triathlon races, coaches and athletes must identify the best predictors for cycling performance. IM 70.3 requires athletes to cycle 90 km; thus, unlike shorter triathlon distances that require more strength and power, long distances require more aerobic fitness and strategy [5]. Indeed, long -distance events have traditionally been associated with body composition, experience time, and physiological variables, such as maximal oxygen uptake (\(\dot{\text{V}}\)O₂max) and percentage of \(\dot{\text{V}}\)O₂max, which can be sustained for an extended period and movement economy [6, 7].

The \(\dot{\text{V}}\)O₂max reflects the maximum rate at which oxygen can be taken up and used by the body during strenuous exercise involving large muscle mass [8]. It can be measured during a maximum incremental effort test, with higher \(\dot{\text{V}}\)O₂max values, indicating better cardiorespiratory fitness [9]. However, because the intensity associated with \(\dot{\text{V}}\)O₂max cannot be sustained for more than a few minutes during a long-distance event, a percentage of \(\dot{\text{V}}\)O₂max is maintained throughout the race. In this context, the \(\dot{\text{V}}\)O₂ sustained during a long event is determined by both the \(\dot{\text{V}}\)O₂max and the percentage of it that can be sustained [10]. These characteristics are typically measured during a maximum incremental test using lactate or ventilatory thresholds [10]. In cycling, another test called functional power threshold (FTP) has also been proposed to determine the highest mean power output that can be sustained over extended periods, providing a more feasible and cost-effective alternative to the more labor-intensive and expensive lactate or ventilatory threshold tests. FTP was originally defined as the highest power output that can be maintained in a quasi-steady state without fatigue for approximately 60 min [7]. However, a 20-min time trial (TT) test has been suggested as a more practical alternative, with FTP estimated as 95% of the average power produced during the 20-min TT [7]. Additionally, a specific warm-up protocol for FTP testing consists of a 45-min session with periods of low-intensity cycling, fast accelerations, and a 5-min time trial[7]. Despite the standard warm-up being relatively long, a shorter protocol consisting of 10 min of cycling at a light to moderate intensity has been shown not to affect FTP results [11]. Regarding the validity of FTP in predicting maximal lactate steady state (MLSS), previous research has demonstrated a nearly perfect correlation (r = 0.91) between FTP, calculated as 95% of the mean power sustained during the 20-min TT, and MLSS among trained and well-trained athletes [12]. Although FTP test is considered a reliable test and that there is a good association between FTP and MLSS and other physiological markers, there are a large limit of agreement between these measurements suggesting these parameters cannot be used interchangeably [13].

Finally, body composition, particularly low fat mass, has been associated with overall performance in the Ironman distance (r = 0.41, p < 0.001) [14, 15]. It can be measured using various methods, such as magnetic resonance imaging or dual-energy X-ray absorptiometry (DXA), which are particularly interesting methods as they produces reliable data [16].

Given the importance to coaches and athletes of identifying the best predictors of cycling performance during an IM 70.3 event and planning their training section, the purpose of this study was to assess \(\dot{\text{V}}\)O₂max and percentage of \(\dot{\text{V}}\)O₂max that is reached at ventilatory thresholds, economy of movement, and FTP. Furthermore, we investigated whether these variables are related to the overall race time and cycling time of an IM 70.3 race and created an equation to predict performance on it. The study also sought to validate the significance of body composition to physiological variables associated with performance. The present study hypothesizes that because FTP is a particular measurement for cycling, it can predict performance in this part of the triathlon. However, because \(\dot{\text{V}}\)O₂max is a more comprehensive measure of physical fitness level, it would strongly correlate with overall race performance. Furthermore, we hypothesized that lean mass and fat mass % would be equally associated with performance-related physiological variables.

Male athletes had higher total body mass (kg) (p = 0.004, d = − 2.16), height (cm) (p = 0.039, d = − 1.37), absolute lean mass (kg) (p < 0.001, d = − 4.03), as well as a lower fat mass percentage (%) (p = 0.049, d = 1.29) compared to female athletes (Table 1).

The male athletes had significantly higher PO_peak(W) values than female athletes (p = 0.001, d = − 2.52). In terms of physiological variables, there were no significant differences between male and female athletes in the percentage of \(\dot{\text{V}}\)O₂max at VT (p = 0.541, d = − 0.37) or RCP (p = 0.657, d = 0.26) and \(\dot{\text{V}}\)O₂max (mL/kg/min) (p = 0.103, d = − 1.04). Swimming, cycling, and running hours per week did not differ significantly between sexes (p = 0.862, d = − 0.10; p = 0.896, d = 0.08; and p = 0.111, d = − 1.01, respectively). Furthermore, there was no significant difference in triathlon experience time between sexes (p = 0.118, d = 0.99) (Table 1).

Male and female athletes had the same cycling split time and overall race time (p = 0.107, d = 1.02, and p = 0.197, d = 0.80, respectively) (Table 1).

Cycling performance was significantly correlated to \(\dot{\text{V}}\)O₂max (r = − 0.764, p = 0.004), power at VT (r = − 0.698, p = 0.012), power output at RCP (r = − 0.646, p = 0.023), lean mass (r = − 0.639, p = 0.025), fat mass % (r = 0.699, p < = 0.011), and FTP (r = − 0.799, p = 0.002) (Fig. 1).

Overall race time had presented a significant relationship with \(\dot{\text{V}}\)O₂max (r = − 0.817, p = 0.001), power output at GET(r = − 0.692, p = 0.013), power at RCP (r = − 0.606, p = 0.037),, lean mass (r = − 0.504, p = 0.095), fat mass % (r = 0.818, p = 0.001), and FTP (r = − 0.774, p = 0.003) (Fig. 2).

Multiple linear regression models were fitted to determine which variables influenced cycling performance and overall race time (Table 2). Because cycling split time and overall race time did not differ significantly between male and female participants, sex was not considered in the regression models. Given the multicollinearity effect of the physiological and anthropometric variables studied, only one of these variables (cycling and overall race time) comprised both prediction models.

Functional threshold power (β = − 0.799, t = − 4.196, p = 0.002) can predict 63.8% of the cycling split time in an Ironman 70.3 according to the multiple regression model (Table 2). \(\dot{\text{V}}\)O₂max (β = − 0.817, t = − 4.477, p = 0.001) can predict 66.7% of the overall race time in IM 70.3 (Table 2).

When the relationship between lean mass and FTP or lean mass and \(\dot{\text{V}}\)O₂max was examined, the results revealed a significant relationship (r = 0.882, p < 0.001, and r = 0.673, p = 0.016, respectively). Finally, the results also showed a significant relationship between fat mass (%) and FTP or fat mass (%) and \(\dot{\text{V}}\)O₂max (r = − 0.882, p < 0.001, and r = − 0.875, p < 0.001, respectively).

A multiple linear regression model was also used to determine which variables influenced FTP and \(\dot{\text{V}}\)O₂max. Table 3 shows the results of the statistical model’s analyses. FTP (r² = 0.777) can be predicted by LM (β = 0.882, t = 5.909, p < 0.001), and \(\dot{\text{V}}\)O₂max (r² = 0.766) can be predicted by fat mass percentage (β = − 0.875, t = − 5.718, p < 0.001) (Table 3).

The level of association between physiological and anthropometric variables and triathlon experience in cycling split time and overall race time in an Ironman 70.3 event was confirmed in this study. Furthermore, regression models were created to predict cycling split time and overall race time performance. The main findings of the present study were as follows: (a) FTP, power output at GET, power output at RCP (W), \(\dot{\text{V}}\)O₂max, fat mass (%), and lean mass (kg) presented significant association with cycling and overall race time (coefficient of correlation ranges between 0.606 and 0.818),, (b) FTP is the best variable associated with cycling split time and it can predict; (c) \(\dot{\text{V}}\)O₂max is the best variable associated with overall race time; (d) lean mass can predict 77.7% of FTP; and (f) low fat mass (%) can predict 76.6% of \(\dot{\text{V}}\)O₂max.

Despite the high association level between all the outcomes and cycling or overall performance, FTP is the better independent variable associated with cycling split, whereas \(\dot{\text{V}}\)O₂max is the best variable related to overall race time. However, only FTP can be predicted by lean mass and fat mass, while \(\dot{\text{V}}\)O₂max can be predicted only by fat mass percentage. The observed phenomenon may be attributable to the pivotal role of muscle mass in generating force and, consequently, power during cycling. Conversely, \(\dot{\text{V}}\)O₂max is predominantly constrained by cardiovascular factors rather than muscular factors. Consequently, relative \(\dot{\text{V}}\)O₂max (i.e. ml/kg/min) is significantly affected by fat mass, given that adipose tissue is not involved in muscle contraction and then does not contribute much to the increase in consume oxygen.

In light of the significant correlation between the FTP test and cycling performance outcomes, a critical evaluation of the protocol employed, and the test's reliability is warranted. Regarding the reliability, the FTP values assessed by a 20 min test have been reported as both reliable and repeatable in several previous researchers [27‐29]. Considering the test prtocol, the warm up originally prescribed by Allen and Coggan is 45 min, including 20 min at self-selected low intensity, 3 × 1 min of fast pedalling accelerations, 5 min at self-selected low intensity, 5 min at maximal effort and 10 min at self-selected low intensity. Despite Allen and Coggan justify this long warm-up protocol and the high intensity as necessary to “open up the legs for the rest of the effort” and to “capture your ability to produce watts”, it has been reported that the blood lactate data were too high between the end of the warm-up and the commencement of the 20-min time-trial (6.5 ± 2.9 mmol/L) [30]. Considering that the recommendation suggests a shorter duration of ~ 15–20 min and [La] < 3 mmol/L before the endurance performance [31, 32], the original warm-up procedure is not in accordance with it, which may affect negatively the participants performance. In a recent review study, Mackey and Horner [13] noted that out of 15 studies analyzed, only 5 followed the prescribed warm-up protocol as suggested by Allen and Coggan [7].

Considering the association between FTP and other candidate markers of the heavy-to-severe intensity, such as RCP and maximal lactate steady state (MLSS), should be discussed. Sitko et al. [23] and Mackey et al. [13] conducted studies aiming to verify the association level between FTP and the RCP. The authors showed very primarily to near-perfect correlations between power output at RCP and FTP. In the same direction, Barranco-Gil et al. [11] also found high correlation between FTP and RCP, even when using a warm-up protocol similar to the one used in the present study (10 min at a self-selected intensity). Despite the high correlation between tests, FTP corresponds to a significantly lower power output than RCP [11]. The elevated power output associated with the RCP in relation to the FTP was anticipated given the protocol-dependent nature of the RCP measurement [33]. In a previous study comparing incremental tests characterized by different ramp slopes, the power output associated with RCP rose from 5 W/min protocol to 10 W/min, 15 W/min, 20 W/min, 25 W/min and 30 W/min [34]. This phenomenon is attributable to the fact that \(\dot{\text{V}}\)O₂ increases in response to augmented workload, and the extent to which the workload exceeds \(\dot{\text{V}}\)O₂ is contingent upon how rapidly the work rate increases [35]. That is, the greater the augmented workload, the greater the dissociation between \(\dot{\text{V}}\)O₂ and power out at RCP. The authors showed that only at 5 W/min protocol the power output at RCP was similar to power output at maximal metabolic steady state [34]. As in the present study a 15 W/min protocol was applied, higher power output at RCP than at FTP were already expected. Therefore, despite the association between power output at FTP and RCP, it has already been shown large limits of agreement between both, implying that caution should be exercised when using both variables interchangeably [13, 23, 30].

Despite the concerns about FTP, the current findings show that it is a very good predictor variable of cycling split time. These findings have important practical implications because FTP measurement is more accessible than other physiological measures assessed in a physiology laboratory because it does not require biological sampling, can be performed by home trainers or virtual exercise platforms, and is less expensive.

Regarding overall race time, \(\dot{\text{V}}\)O₂max is the best variable associated with it; additionally, \(\dot{\text{V}}\)O₂max was able to explain 66.7% of the race performance. These data are not surprising given that high \(\dot{\text{V}}\)O₂max values have been extensively demonstrated to be required for good performance in long-term races [10].

Given the importance of FTP and \(\dot{\text{V}}\)O₂max in predicting cycling split time and overall race time, the present study investigated the role of body composition in these two physiological variables.

Even though both fat mass (%) and lean mass (kg) were associated with \(\dot{\text{V}}\)O₂max (ml/kg/min), fat mass (%) had a stronger association with \(\dot{\text{V}}\)O₂max than lean mass (kg). Furthermore, fat mass predicts 76.6% of \(\dot{\text{V}}\)O₂max (ml/kg/min). These findings show that low levels of fat mass are more important for \(\dot{\text{V}}\)O₂max than muscle mass. Even though muscle mass consumes oxygen, during exercises, the \(\dot{\text{V}}\)O₂max is not limited by the ability of the muscles to take up oxygen from the blood and consume it, but it is limited by the ability of the cardiorespiratory system to deliver oxygen to the exercising muscles [10], which may explain the not very high association between muscle mass and \(\dot{\text{V}}\)O₂max. On the other hand, the fat mass percentage showed a high and negative correlation level with VO₂max. In the present study, VO₂max relative to body mass was studied, therefore, for the same absolute \(\dot{\text{V}}\)O₂max value, the greater the body mass the lower relative to body mass \(\dot{\text{V}}\)O₂max is. In this context, a significant and inverse relationship between fat mass percentage and \(\dot{\text{V}}\)O₂max was expected, since this is a mass that does not contribute to the \(\dot{\text{V}}\)O₂, therefore when the absolute \(\dot{\text{V}}\)O₂ value is divided by a greater body mass, the result will be lower.

Functional threshold power, on the other hand, showed the same level of association with lean mass (r = 0.822, p < 0.001) or fat mass (r = 0.822, p < 0.001), indicating that higher lean mass and lower fat mass percentage are important for FTP. Furthermore, this variable explained 77.7% of the FTP values, indicating that this is a very good prediction model. Unlike \(\dot{\text{V}}\)O₂max, which is limited by the cardiorespiratory system, FTP is a measure of aerobic endurance that is linked primarily to adaptations in muscles mass [36]. Therefore, it is already expected that the muscle mass is more important for FTP than for \(\dot{\text{V}}\)O₂max.

The fact that FTP is strongly related to cycling split time and that low fat mass % and high lean mass can explain a large portion of FTP variation emphasizes the importance of body composition for cycling performance during an Ironman 70.3. Similarly, the fact that a low-fat mass % is also strongly associated with \(\dot{\text{V}}\)O₂max supports the importance of body composition in an Ironman 70.3 race. As a result, the authors argue that coaches and athletes who are aware of this information can set goals for aerobic training and body composition.

Ventilatory thresholds, \(\dot{\text{V}}\)O₂max, FTP, and body composition were strongly associated with cycling and overall race time in an Ironman 70.3. \(\dot{\text{V}}\)O₂max is the best predictor of overall race time and FTP of the cycling split time. Furthermore, high lean mass is strong predictor of FTP, whereas low fat mass is a strong predictor of \(\dot{\text{V}}\)O₂max.