Introduction

Road cycling is widely recognized for its physical and mental health benefits, including improvements in metabolic and cardiovascular function, reduced risk of falls, and decreased symptoms of depression, particularly among older adults and recreational populations1,2,3,4. Given these benefits, many individuals engage in road cycling regularly, often spending prolonged periods on the bike due to the repetitive and endurance-based nature of the activity5.

This extended time in a fixed posture makes the rider’s position on the bike a critical factor for both performance and well-being. Previous studies have shown that cycling posture and bike setup can significantly influence muscle activation patterns, perceived comfort, and the risk of overuse injuries6,7,8,9. These findings underscore the importance of optimizing bicycle configuration to suit each cyclist’s individual characteristics and needs10particularly in the context of recreational cyclists.

The ergonomic bicycle adjustment process (also known as “bike fitting”) aims to improve riding comfort and reduce injury risk11 by tailoring bicycle components to the cyclist’s needs, goals12and body position on the bicycle10. Among the different bicycle configuration parameters, the saddle height has been one of the most studied13,14. Manipulation of it can generate changes in oxygen consumption15energy cost of pedaling16,17 and even increase the risk of overuse injuries18 as alterations in the range of motion and mechanical work of the lower limbs occur13,19.

Improper bike setup, including saddle height, has been shown to affect cyclists’ comfort and perceived exertion, with lower saddle heights often linked to increased discomfort, higher exertion, and greater reports of fatigue and pain20,21,22. These findings underscore the relevance of saddle positioning in optimizing rider experience and performance. However, while saddle height has been extensively studied, the specific role of saddle setback remains less explored, despite its potential biomechanical implications.

Saddle setback can significantly influence cycling performance and biomechanics, as it affects variables such as pedal force23joint loading, and power output7. While some studies suggest that a greater setback can improve force effectiveness and peak power7others highlight increased muscular24 and joint stress8particularly with excessive rearward positioning. Despite its relevance, the specific impact of saddle setback on lower limb kinematics remains underexplored, especially in mixed-sex samples14highlighting the need for further investigation.

Moreover, kinematic patterns in cycling are influenced by intensity, with higher training loads, commonly measured by Rate of Perceived Exertion (RPE) and power output (W)25,26leading to compensatory changes in the lower limbs27. These include increased ankle dorsiflexion and greater knee and hip extension28,29,30,31alongside more flexed positions in the elbows and spinal segments27. As intensity rises, joint angular ranges expand, particularly at the ankle and knee32while nearing the second ventilatory threshold prompts a shift in joint contribution: the ankle reduces its role as the knee and hip increase their flexion-extension movements33. Despite these changes, saddle pressure often decreases at higher intensities, likely due to altered posture and load distribution34. These adaptations highlight the complex interplay between intensity, joint mechanics, and comfort in cycling performance.

Given the limited research analyzing the effects of different saddle setback positions, particularly considering both male and female cyclists, pedaling intensity, and comfort; this study aimed to examine how saddle setback and pedaling intensity influence comfort, perceived exertion, and lower-limb angular kinematics in recreational male and female cyclists. We hypothesized that a backward saddle setback would reduce comfort levels, increase the perception of exertion and the extension levels of the lower limbs. We also hypothesized that higher intensities would reduce the flexion levels of the hip, knee and ankle and increase the perception of exertion.

Materials and methods

Participants

The study involved a sample of 34 recreational road cyclists (20 men and 14 women; Table 1). All participants met the study inclusion criteria: they were recreational cyclists aged between 18 and 50 years; they had not suffered serious lower limb injuries in the last 6 months, nor had any discomfort at the time of the study; they were not taking any medication or stimulants prior to or during the study. All participants were duly informed about the nature and purpose of the study and provided their written informed consent before participating. The study protocol was reviewed and approved by the Ethics Committee of the University of Valencia (No. 1954944) and followed the ethical principles established in the Declaration of Helsinki.

Table 1 Anthropometric characteristics, estimated exercise intensity, and training habits of the study participants, presented as mean (M) ± standard deviation (SD).
Full size table

Procedures

Participants were measured and evaluated in a 1-day session under similar environmental and laboratory conditions (382 m altitude, 30% relative humidity, 16–18 ºC) and after a period of 48 h without any high-intensity training session or stimulants.

First, cyclists’ anthropometrics and bicycle characteristics were measured. Second, the researchers adjusted the indoor bike to the own bicycle geometry of each cyclist (Table 2). Subjects performed the same protocol under 3 different setback positions and 2 pedaling intensities in a randomized order. We used C + + software specifically developed (Visual Studio Code, version 1.83) to ensure balanced randomization and avoid selection or accidental bias. Lower-limb 3D kinematics were measured while participants pedaled and comfort and RPE were recorded after each intensity.

Table 2 Bicycle configuration and setup parameters used by the participants, presented as mean (M) ± standard deviation (SD).
Full size table

Once the measurements were taken and the bike was adjusted, the cyclists performed a 10-minute warm-up at a free intensity6. They performed 6 sets of 5 min of cycling on the same indoor bicycle (Wattbike Pro, Wattbike, Nottingham, England), being instructed to maintain a constant cadence of 90 rpm throughout each trial, with real-time biofeedback provided by the ergometer to support cadence control. Participants were asked to keep hands on the levers. A rest of 2 min between intensities and 5 min between saddle positions was defined6.

The intensity (watts normalized to body weight) was calculated following Valenzuela et al.,35 recommendation for recreational cyclists. Concretely, for men, intensity 1, watts corresponding to the first ventilatory threshold (VT1) was fixed at 2.4 W·kg− 1 with an allowed range of 2.3 to 2.5 W·kg− 1; intensity 2, corresponding to the second ventilatory threshold (VT2) was fixed at 3 W·kg− 1 with an allowed range of 2.9 to 3.1 W·kg− 1; whereas for women, intensity 1 (VT1) was fixed at 2.1 W·kg− 1 [2.0-2.2 range], and intensity 2 (VT2) in 2.6 W·kg− 1 [2.5–2.7 range]35.

Anthropometric measurements and bicycle characteristics

Anthropometric measurements were taken with a wall-mounted height rod (Seca 216®, Hamburg, Germany) and a bioimpedance scale (Tanita MC 780-P, Tokyo, Japan). Inseam length was measured as the distance from the pubis to the floor with the cyclist barefoot. To determine the foot length, the distance between the fifth metatarsophalangeal joint and the calcaneus was measured14. According to international guidelines for anthropometry, all anthropometric measurements were performed by the same researcher as in a previous study36. For the bicycle characteristics an anthropometric tape, plumb bob and set square were used14. Finally, saddle height was measured according to the recent recommendations of an international consensus37as the distance from the highest point of contact on the saddle to the pedal spindle. Since participants used different saddle models, the contact point was standardized by defining it as the center of the saddle. This measurement was taken as the shortest distance from the center of the saddle to the center of the bottom bracket, and the crank arm length was then added to calculate the true saddle height.

Since the crank and saddle lengths of the participants’ bicycles were not the same, the following adjustments were made to unify the saddle setback (Table 2). The saddle setback preferred (Preferred) in this study was calculated as the horizontal distance from the center of the saddle to the center of the pedal axle, accounting for crank length. This method aligns with recent international recommendations for standardized measurement protocols in cycling biomechanics, as outlined in the consensus statement by Priego-Quesada et al.37. While UCI saddle setback (preferred, forward and backward) is traditionally defined as the horizontal distance from the saddle nose to the bottom bracket -consistent with UCI regulation 1.3.013, which requires the saddle peak to be at least 5 cm behind the vertical plane of the bottom bracket spindle - this standard definition may be limited by variations in saddle geometry. Alternative approaches have been proposed in the literature, such as measuring the distance between the center of pressure on the saddle and the bottom bracket9or from the bottom bracket to the saddle clamp38. Our approach, based on the consensus guidelines, provides a practical and reproducible method that also incorporates crank length, offering a more complete representation of the cyclist’s effective position. The other setbacks, saddle setback forward (Forward, -10% from Preferred) and saddle setback backward (Backward, + 10% from Preferred) were calculated from the saddle setback preferred. Vertical difference between handlebar and saddle height and the diagonal distance between handlebar and the saddle was measured following the consensus statement by Priego-Quesada et al.37.

To ensure consistency across all testing conditions, the saddle height, the vertical difference between the handlebar and saddle height, and the diagonal distance between the handlebar and saddle were kept constant throughout the three saddle setback configurations (preferred, forward, and backward). This was achieved through an iterative adjustment process in which the new setback was first applied, followed by a verification of the saddle height, which was corrected if necessary. The saddle setback was then rechecked, and final adjustments were made to the diagonal distance and the handlebar height relative to the saddle. This process ensured that the participants’ biomechanical posture remained unchanged, isolating the effect of saddle setback on the outcomes of interest.

3D procedures and data analysis

Lower limb kinematics were recorded using an infrared optoelectrical camera system (Optitrack™ V120 Trio, NaturalPoint, Inc., Oregon USA) controlled by the Motive software (Motive V2.3.1., NaturalPoint, Inc., Oregon USA). The system was calibrated according to the manufacturer’s guidelines. Following and adapting the protocol from previous studies6,14seven 10 mm diameter reflective markers were attached to specific anatomical landmarks (anterosuperior iliac spine, posterosuperior iliac spine, greater trochanter, lateral condyle, lateral malleolus and fifth metatarsal) and to the center of the pedal to record angular kinematics. Angle convention was established as Fig. 1 shows. Kinematic measurements were recorded by two bouts of 15 s at minutes 2 and 4 of pedaling for every trial, ensuring more than 10 pedaling cycles at a sampling frequency of 120 Hz. The right leg was analyzed, as previous studies have shown that there are no significant kinematic asymmetries between the left and right legs in cyclists of different performance levels39.

Prior to the kinematic recording during pedaling, a calibration of the bottom bracket centre was performed. A reflective marker was placed at the centre of the bottom bracket and its position was recorded for 10 s. After applying the appropriate filtering, this position was used as a reference for calculating the crank position throughout the pedaling cycle, in combination with the marker placed at the bottom bracket centre. To determine the pedal position, a 2D analysis (X and Y axes) of the crank vector was implemented in a MATLAB routine, and circular statistics were used to calculate the crank position throughout the pedaling cycle, with 0° corresponding to the top dead centre of the pedal stroke.

Additionally, a static posture recording was performed, as recommended by the consensus study37. Before the dynamic recording, participants stood in an upright posture with their arms crossed and aligned with the plane of the cycle ergometer. The joint angles in this standing position were recorded. To normalize the dynamic kinematic data, the angles from the static calibration were subtracted from the raw values recorded during pedaling. The formula used was:

$$Normalized{\text{ }}jo\operatorname{int} {\text{ }}angle\,=\,raw{\text{ }}jo\operatorname{int} {\text{ }}angle--angle{\text{ }}at{\text{ }}s\tan ding{\text{ }}posture.$$
Fig. 1
figure 1

Marker set and angle convention employed.

Full size image

A Butterworth fourth order filter with a cut-off frequency of 8 Hz was applied to smooth the data. The cut-off frequency of 8 Hz was established based on findings from previous studies40which demonstrated that using lower cut-off frequencies, such as 6 Hz, reduced the precision of kinematic data. Then, a customized MATLAB routine (MATLAB R2022b, MathWorks, MA) was employed to extract the parameters. The accuracy of the 3D reconstruction, of the angular values and the test-retest reliability was previously calculated in the study of Encarnación-Martínez et al.14.

Perceived exertion and comfort protocols

Perceived exertion was recorded using the Borg scale of 6–20 RPE. It includes a score from 6 perceiving “no exertion at all” to 20 perceiving a “maximal exertion” (Borg, 1982). Overall comfort was recorded using the 15 cm visual analogue comfort perception scale (VAS)41 adapted for this study. This item is rated from left side, labeled as “not comfortable at all” (0 comfort points), to the right side, labeled as “best comfort imaginable” (15 comfort points). These data were recorded after each intensity and saddle setback condition.

Statistical analysis

Statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS) software (Version 25, IBM Corp., Armonk, NY). As the variables followed a normal distribution (p > 0.05; Shapiro-Wilk test), mean and standard deviation were used to present the data. A three-way mixed-design analysis of variance (ANOVA) was performed, with saddle setback (Preferred, Forward, Backward) and intensity (VT1, VT2) as within-subjects factors, and sex as a between-subjects factor, to analyze their effects on each of the dependent variables (perceived exertion, comfort, and angular kinematics).

The significance levels in the pairwise comparison were adjusted using the Bonferroni correction. Sphericity was also checked with the Mauchly test, and in those cases where this assumption was not met, the Huynh-Feldt adjustment was used. Finally, effect size (ES) was calculated using partial eta squared (η²ₚ) that were classified as small (0.01–0.059), moderate (0.060–0.140), and large (> 0.140)42. The significance level was set at p < 0.05.

The relationship between the saddle setback and the independent variables was analyzed by means of a stepwise regression analysis. For that purpose, only the setbacks that showed great comfort and a knee extension angle between 30 and 40º were measured. Then, a Bland–Altman plot was done to check the agreement between real and estimated saddle setback. Differences between real and estimated saddle setback were plotted against the mean results43. Intraclass correlation coefficient (ICC) was performed to check reliability between real and estimated saddle setback. ICCs were interpreted as: excellent (0.75–1), modest (0.4–0.74), or poor (0–0.39)44.

‘Equity, diversity, and inclusion statement’

All the recreational cyclists included in the study were Caucasians.

Results

Results on perceived exertion and comfort

Table 3 shows that the perceived exertion values reported by cyclists were significantly higher at VT2 compared to VT1 (p < 0.05, η²ₚ = 0.728; mean difference = − 2.636; 95% CI = − 3.216 to − 2.055). No significant differences were observed as a function of saddle setback or sex (p > 0.05). Regarding comfort, overall ratings were lower in the Backward position than in the Preferred and Forward conditions (p < 0.05, η²p= 0.593), with no differences across pedaling intensities. However, at the Preferred setback, men reported greater comfort than women (p = 0.038, η²p= 0.105).

Table 3 Descriptive statistics of the variables RPE and overall comfort, presented as mean (M) ± standard deviation (SD).
Full size table

Results on angular kinematics

Three-way repeated measures ANOVA showed that there were significant differences relating saddle setback, intensity, and sex (Tables 4 and 5).

Results obtained as a function of saddle setback

Changes in saddle setbacks affect most of the variables analyzed in angular kinematics.

Figure 2 shows that maximal hip flexion, hip extension, and ankle plantarflexion were significantly greater in the Backward condition compared to both the Preferred and Forward conditions (p < 0.05). Similar differences were observed for the vertical displacement range of the hip and the knee range of motion (p < 0.05). Additionally, both minimum and maximum knee flexion angles were significantly lower (more extended limb) in the Backward condition (p < 0.05), as were the crank positions at maximum hip flexion, maximum hip extension, and minimum knee flexion (p < 0.05 for all comparisons). Internal knee displacement and medial-lateral knee displacement range were also significantly greater in the Backward condition than in the Preferred condition (p = 0.015 and p = 0.041, respectively). In general, the Backward condition increased joint ranges and knee displacement.

Table 4 Descriptive statistics of angular kinematic variables in women (n = 14), presented as mean (M) ± standard deviation (SD).
Full size table
Table 5 Descriptive statistics of the angular kinematics in men (n = 20), presented as mean (M) ± standard deviation (SD).
Full size table

Results obtained as a function of pedaling intensity

Post-hoc comparisons revealed that maximal hip flexion, hip extension, and mean medial knee displacement were significantly reduced at VT2 compared to VT1 (p < 0.05). In contrast, ankle dorsiflexion was significantly greater at VT2 than at VT1 (p < 0.05).

At higher intensity (VT2), hip motion decreased, and ankle dorsiflexion increased, reflecting changes in joint kinematics.

Fig. 2
figure 2

Hip, knee, and ankle differences between saddle setbacks. Full sample of recreational cyclists (n = 34). *Statistical differences (p < 0.05).

Full size image

Results obtained as a function of sex

There was a trend towards statistical significance in ankle plantar flexion, with women obtaining higher values than men (p = 0.052). Thus, the crank position in ankle plantar flexion was achieved before in the pedaling cycle in women compared to men (p = 0.045, η²ₚ = 0.127).

Pedaling cadence showed an interaction between the variables intensity and sex where women had a significantly lower cadence than men at VT1 intensity (p = 0.03; η²ₚ = 0.147). Hip range was also significantly higher for women at VT1 intensity compared to men (p = 0.008, η²ₚ = 0.203) and pelvic tilt angle showed an interaction between the variables saddle setback, intensity and sex where women pedaled more inclined than men at Backward condition at VT1 intensity (p = 0.029, η²ₚ = 0.140). In contrast, men had a greater mean external knee displacement at Preferred condition at both intensities (VT1 and VT2) compared to women (p = 0.036, η²ₚ = 0.097; p = 0.036, η²ₚ = 0.087, respectively).

Although some sex-related kinematic differences exist, they are generally small and do not significantly affect the overall movement patterns observed during cycling.

Stepwise regression analysis

The multiple linear regression analysis showed the existence of a relationship between the variables that is expressed by the equation.

$$Y = ~0.421h + 0.278SH - 0.509IL + 0.068FL - 21.960$$

Where Y is the saddle setback, h is the height of the cyclist in cm, SH the height of the saddle in cm (including crank length), IL is the inseam length in cm, and FL is the length of the foot in cm (distance from the first metatarsal to the calcaneus). All variables in the model had a significant association with saddle setback. The coefficient of determination was R2 = 0.840 and the root mean square error (RMSE) was 0.962, indicating that near of the 84% of variance in the saddle setback was accounted by the predictor variables.

The Bland–Altman plot for the real and the estimated setback is in Fig. 3. The mean bias of the differences between the real and estimated setback is 0.220 cm, and the ICC was 0.95.

Fig. 3
figure 3

Bland-Altman plot for real and estimated saddle setback.

Full size image

Discussion

As we hypothesized, the results of this study confirmed that a backward saddle setback reduced comfort levels and increased lower limbs extension levels. Furthermore, a high intensity increased perceived exertion and reduced lower limbs flexion levels. The results are discussed in more detail below.

Perceived exertion and comfort

Regarding the influence of saddle setback and sex on perceived exertion, the results indicated that there were no statistically significant differences (Table 3), thus agreeing with other studies20,45. The results concerning pedaling intensity showed a greater perception of exertion at higher intensity, consistent with other studies46 where perception of exertion also increases progressively and linearly with increasing intensity and time.

Overall comfort was reduced when cyclists pedaled at backward saddle setback (Table 3). These results are consistent with previous studies47 where discomfort is also increased when cyclists pedal at high saddle height and backward saddle setback compared to a neutral position. It has been shown that there is a relation between the adjustment of body position on the bicycle and comfort11. In the study by Kraus et al.48 cyclists reacted to changes in saddle setback by adjusting their position on the saddle to try to maintain the same knee over pedal spindle position and the same center of pressure. Other studies reported less comfort and more fatigue when using a low saddle height compared to a preferred saddle height or other saddle heights20,22. This finding is in line with our study where men reported greater overall comfort in Preferred saddle position than women. The differences in comfort between men and women may be also explained by the sex differences in anatomical structure such as pelvic structure and foot length14. It should be also considered that most of the studies on bike fitting (specifically saddle height adjustment) have been carried out on male cyclists and therefore the methods traditionally used do not fit well in female cyclists6,14,36,49.

Angular kinematics

The body position on the bicycle can be modified by factors such as intensity27,50cadence28fatigue state51cyclists’ perceived comfort22discipline52gender14 and training level36 and therefore produce changes in joint kinematics and performance.

The present study showed an increase in lower limb extension when cyclists pedaled at the backward saddle setback (Fig. 2). These findings are in line with another study16 which found that a higher saddle position increases the hip and knee extension and the ankle plantarflexion and decreases the flexion and the range of motion of these joints. More specifically, the results showed a significant increase in hip extension, knee range, internal mean knee displacement and medial-lateral displacement range and, conversely, a decrease in maximum and minimum knee flexion at backward saddle setback (Tables 4 and 5). The results of the study by Ménard et al.8 showed that the exacerbation of these kinematic changes could increase the risk of iliotibial band syndrome by increasing the compression force of the lateral femoral band-epicondyle. The tibiofemoral mean and peak compression force were also higher in the backward saddle setback than in the forward saddle setback in the study by Menard et al.9. Changes in saddle setback have also been shown to influence the amplitude and the pattern of muscle forces. A backward saddle position (-5 cm) led to a greater peak of plantar flexors and hamstrings forces compared with a forward position (+ 5 cm)24.

The results also showed a decrease in maximum hip flexion, maximum hip extension and internal mean knee displacement at higher intensities (Tables 4 and 5) in agreement with other authors such as Pouliquen et al.53 and Ferrer-Roca et al.54. Ankle dorsiflexion was increased at VT2, consistent with previous findings46. These results differ from other studies such as Holliday et al.27 or Ferrer-Rocal et al.54 which concluded that ankle dorsiflexion and ankle range of motion increased at higher intensities. Moreover, the study by Lanferdini & Vaz55 postulates that fatigue causes joint changes in the lower limbs, greater range of motion in the ankle and lower contribution to the total force moment. A change in saddle setback could also lead to joint adjustments since even a slight change in the bicycle’s geometry could cause modifications in the lower limb’s kinematics, which can lead to an increase or decrease in range of motion16. These kinematic differences can also be affected by the interaction of the effects of pedaling technique, fatigue and workload28.

Hopker et al.56 found no significant difference in pedaling cadence between men and women, whereas the results of this study showed that women had a lower pedaling cadence than men at lower intensities. The greater pelvic angle of women could be the factor that explains why they pedaled more inclined than men at a backward saddle setback at lower intensities14.

One important aspect to consider is the limited number of studies analyzing the pedaling biomechanics of female cyclists14. Our study contributes to filling this gap by providing reference kinematic data for female amateur cyclists, which can be compared to that of male cyclists. This new dataset, presented in the updated table, offers a crucial reference point for future systematic reviews and can help to better understand the potential differences and similarities between sexes in cycling biomechanics. By including female participants in our analysis, we hope to encourage further research in this area and contribute to a more inclusive understanding of cycling performance.

One limitation of this study is that the crank length on the Wattbike Pro ergometer could not be modified, potentially introducing minor biomechanical variations. While 38.2% of participants typically used 170 mm cranks (matching the ergometer), others used slightly different lengths. Additionally, although cadence was fixed and biofeedback was provided, participants deviated slightly from the target cadence. However, no meaningful differences in cadence were observed across the experimental conditions. However, to minimize this impact and approximate the cyclist’s real-world setup, we adjusted both saddle setback and saddle height based on crank length, aiming to replicate their habitual riding posture as accurately as possible. These procedures followed the recommendations proposed by the recently published international consensus on bicycle setup and cycling kinematics37.

Conclusions

The main findings of this study, conducted on recreational road cyclists, indicate that the backward saddle position reduced overall comfort. Additionally, the Backward condition showed a general trend toward greater extension at the knee and ankle joints, and increased flexion at the hip, reflecting an adaptation to the more rearward saddle placement. Additionally, high intensity pedaling led to increased perceived exertion and a decrease in lower limb flexion. It is important to note that these results reflect the specific biomechanical responses of a recreational road cycling population, whose characteristics may differ from those of elite cyclists. Furthermore, a new regression equation was developed, which may assist in performing an initial adjustment of saddle setback based on individual anthropometric parameters.