The effects of uphill training on the maximal velocity and performance of middle-distance runners: a randomized controlled trial

Abstract

This study investigated the effects of three uphill gradients; shallow (SHG ~ 2.5%), intermediate (IHG ~ 5.1%), and steeper (STHG ~ 7.6%) on maximal velocity (Vmax), 800 m time trial (TT), and strength endurance (SE) performance in middle-distance runners. Forty moderately trained athletes (aged 16–20) were randomly assigned to SHG, IHG, STHG, or control (CG) groups (n = 10 each). Over 8 weeks, participants completed gradient-specific training. Vmax (30 m sprint), TT (800 m), and SE (1-min burpees) were assessed pre- and post-intervention. Two-way RM-ANOVA revealed significant main effects on Vmax (F(3,36) = 2.87, p = .049, = 0.20), TT (F(3,36) = 7.60, p < .001, = 0.39), and SE (F(3,36) = 5.19, p = .004, = 0.30). Post-hoc analysis showed STHG significantly outperformed CG in Vmax (mΔ = 0.93, p = .040), TT (mΔ = -0.29, p = .001), and SE (mΔ = 3.90, p = .007). IHG also improved TT vs. CG (mΔ = − 0.25, p = .007). The finding revealed that, uphill training, particularly steeper gradients, significantly enhances speed, endurance, and running performances. Therefore, coaches and athletes should customize training programs using uphill while carefully matching the strengths and weaknesses of the athlete and the demands of the event.

Introduction

The key bio-motor skills that contribute to successful middle-distance running performance are often attributed to maximal velocity, speed endurance, and strength endurance¹. This is because middle distance running performance is dependent mainly on both anaerobic and aerobic energy systems^2,3. For example, a study reported that the performance of 800 m time trials was highly correlated with maximal aerobic speed and maximal anaerobic speed Støren, Helgerud⁴. This means that the runner operating at the lowest percentage of his or her maximal anaerobic speed would consume the least amount of anaerobic capacity per unit of time and endure the longest at a given supra-maximal effort in proportion to maximal aerobic speed. Additionally, performance in middle-distance running is influenced not only by aerobic and anaerobic capacity but also by strength endurance^5,6. The development of strength endurance has been shown to affect the velocity at VO2max^1,7. This is because strength endurance training enhances muscle power, allowing athletes to maintain higher speeds for longer durations. As a result, there is a direct increase in velocity at VO2max (vVO2max), as the muscles become more efficient at utilizing oxygen during high-intensity efforts^8,9.

Moreover, strength endurance training induces significant neuromuscular adaptations, improving coordination and muscle recruitment patterns⁹, such as: enhanced motor unit synchronization, increased firing rates, improved intramuscular and intermuscular coordination, and greater activation of both type I and type II muscle fibers. These adaptations allow for better motor control and the ability to sustain higher running velocities with reduced oxygen demand. Overall, integrating strength endurance into training regimens provides athletes with the capacity to maintain higher intensities for longer periods, directly improving middle-distance running performance.

Although research on uphill running dates back several decades, it has continued to gain attention as a form of high-intensity interval training (HIIT) aimed at enhancing competitive running performance^10–12. Coaches and athletes frequently incorporate uphill running to boost lower-body power output, increase running speed, and ultimately improve race performance. For instance, Burfoot¹³ reported that runners who engaged in a rigorous six-week program of challenging uphill running experienced noticeable gains in training distance, anaerobic ability, and muscle strength. Unlike other high-intensity resistance-to-movement exercises, uphill running can be seen as a much more sport-specific training strategy and may, therefore, prove more effective at improving running performance than other high-intensity resistance-to-movement exercises. The combined benefits of enhanced strength, speed, and cardiovascular fitness cause athletes who use uphill training to frequently achieve considerable gains in race time¹⁴. Another research also reported that, in comparison with running on flat ground, uphill running is characterized by a higher step frequency, more internal mechanical work, shorter swing/aerial phase duration, and a greater duty factor/contact time¹⁵. This is because, to increase the body’s potential energy during uphill running, lower limb muscles must perform a greater total mechanical effort than they do in level running.

Although, hill running training is a powerful method for targeting specific muscles and conditioning, leading to significant improvements in physical, physiological, and biomechanical aspects of running performance¹⁶, most studies in the field have focused on the acute effects of uphill running on various performance parameters^13,17–21, and others on comparing uphill running with those of downhill and level running^11,22–25. Therefore, there remains a notable gap in the literature regarding the long-term (chronic) effects of uphill running, particularly concerning the relative effectiveness of different uphill gradients on key performance outcomes such as; maximal velocity (Vmax), time trial performance (TT), and strength endurance (SE) in trained athletes. While uphill running is known to impose varying levels of intensity depending on the gradient, few studies have systematically compared these effects across distinct gradient levels. Hill gradients can be broadly categorized as shallow hill gradient (SHG 2–4%), intermediate hill gradient (IHG 4–7%), and steeper hill gradient (STHG > 7%)^26–28, . This classification provides a useful framework for examining how different intensities of uphill running influence athletic performance.

As a result, longitudinal studies assessing the chronic effects of uphill training interventions and comparing the relative effects of various uphill gradients on athletes with some training experience are crucial for refining and optimizing training strategies. Therefore, this study aimed to investigate the chronic effects of uphill training on Vmax, and performance (TT, and SE) and to determine the most effective hill gradient in improving those performance measures in middle-distance runners. It was hypothesized that 8 weeks of uphill training can have a positive impact on maximal velocity (Vmax) and performance (TT and SE) of middle-distance runners. Furthermore, it was expected that higher gradients would yield greater improvements in Vmax, and performance (TT and SE) compared to lower gradients.

Materials and methods

Participants

A total of forty moderately trained middle-distance runners (24 males and 16 females), with age (M-18.54 ± 1.02, F-17.44 ± 1.09 years [mean ± SD], depicted in Table 1), with 2.11 ± 0.74 years of structured training experience took part in the study. Written informed consent was obtained before the intervention, and the study was approved by the Ethics Committee of Bahir Dar University, Bahir Dar, Ethiopia (Protocol no. IRERC 05/2024). All procedures were performed following the Declaration of Helsinki on the use of human subjects. All participants gave their written consent to participate after having received information about the study.

Table 1.

General characteristics of participants included in the study (means ± SD).

Variable	Male	Female
Age (years)	18.54 ± 1.02	17.44 ± 1.09
Height (m)	1.71 ± 0.04	1.68 ± 0.05
Body mass (kg)	54.67 ± 4.04	54.00 ± 3.44
Body mass index (kg/m2)	18.64 ± 1.17	19.14 ± 1.31
Vmax (ms− 1)	7.67 ± 0.73	6.80 ± 0.48

The inclusion criteria were all healthy youth middle-distance runners in Bahir Dar University’s sport academy with no muscular, neurological, or tendon injuries and no drug consumption. Prior to inclusion, medical history questionnaire were used to screen participants: Each participant completed a standardized medical history form to report any current or past muscular, neurological, or tendon injuries, and to disclose medication or drug use. The exclusion criteria were athletes with less than 6 months of training experience, athletes with lower body injury in the previous 3 months, and those taking any medication.

Study design

The study adopted a pre-post parallel group experimental design, with measurements conducted before and after the 8-week intervention period. The participants were randomly assigned to one of the three training groups or the control group. Each group included 10 participants (steeper hill group, 7.6%; intermediate hill group, 5.1%; shallow hill group, 2.5%; and control group). The reason behind including specific gradient levels is because different hill gradients provide varying levels of intensity based on the given categories^26,27, which helps in examining how each intensity level affects performance measures in our case maximal velocity, 800 m time trials, and strength endurance performance. This variation allows researchers to determine which gradient offers the optimal balance between training load and performance improvement.

The study was registered with pactr.samrc.ac.za (Registration Number NCT01234567) on 29/11/2024. The study follows CONSORT guidelines to ensure the quality and transparency of reporting²⁹ (Fig. 1). A priori power analysis with F-test family was conducted using G*Power version 3.1.9.4 to determine the required sample size³⁰. Based on the result of previous study²², all the responses within each subject group are assumed to be normally distributed within group standard deviation for each measurable variable and to detect 0.62 effect size between the pre and post experimental group with a statistical power (1–β) 80% at a significant level (α) 0.05, the estimated sample size we need was 36 participants. By providing an additional allowance of 10% sample recruitment, due to possible dropout, the required sample size has been increased to 40.

Fig. 1.

Flow diagram through the phases of a parallel randomized trial of four groups (enrolment, allocation, intervention, follow-up and data analysis).

To avoid possible bias, trained personnel were used to generate random allocation, enrolled participants, and assigned participants to interventions. Due to having small sample size and its easiness to execute, simple random, lottery method was used to assign participants to the four study groups (each with ten participants): steeper hill group (STHG), intermediate hill group (IHG), shallow hill group (SHG), and control group (CG) at 1:1:1:1 ratio as indicted in (Fig. 1). Stratified randomization by gender was implemented to ensure equal distribution of male and female participants across the four groups. Allocation concealment was ensured by using sealed envelopes. To minimize bias, blinding was employed where feasible. The outcome assessors were blinded to the group assignments.

Procedures

The study was conducted using a pre-test and post-test design. At baseline, participants underwent testing on two separate days to establish initial measurements. Following the intervention after 48 h, participants were tested again on two separate days to assess the effects of the experimental conditions on Vmax and performance (TT and SE) athletes. Day one consisted of a 30-m Vmax and one-minute burpee strength endurance test with at least 20 min of rest between tests, and day two consisted of 800-m running performance test at the university’s outdoor track. All performance assessments were conducted under consistent environmental conditions, outdoors on a dry, flat surface between 8:00 am and 10:00 am to minimize the influence of circadian variation on physical performance. Weather conditions (temperature and wind) were monitored to ensure consistency across testing days. Participants were instructed to maintain their usual sleep and nutrition routines prior to testing days. On the first day, the Vmax was tested after a standard warm-up period of 10–15 min followed by dynamic stretching³¹. A standard flying 30’s consisting of a 30 m acceleration zone and a 30 m maximal velocity zone on a 400 m oval outdoor track at the university was used to test Vmax, as in previous studies^32,33. The Flying 30 m Test was selected as a measure of maximal sprint velocity (Vmax) due to its ability to isolate and assess an athlete’s top-speed phase, which is critical in las laps of middle distance running. Unlike standing-start sprints, the Flying 30 m Test includes a run-up phase (typically 20–30 m), in our case 30 m that allows athletes to accelerate before entering the timed zone. This design ensures that the measurement captures the athlete’s peak sprinting velocity, rather than acceleration capacity. As such, it provides a more accurate and sport-relevant assessment of Vmax^34,35.

Participants were required to wear their regular running shoes both in the pre- and post-test assessments to maintain ecological validity and ensure consistency. No specific footwear was mandated or restricted. The tests were conducted using a high-speed camera Canon SX70 HS capable of recording in 1080p resolution at a sampling rate of 240 Hz, was utilized to capture a 30-meter sprint time. The camera was positioned midway on a tripod at a height of 1.5 m, located 7.5 m from the side of the runway to clearly capture both the start and finish lines. Performance times were determined through frame-by-frame video analysis. This method allowed for consistent and reliable measurement across all participants and testing sessions. The 30-meter sprint test was performed using a flying start. Participants had a 30-meter acceleration zone prior to the 30-meter timed segment. Timing began when the athlete’s torso crossed the start of the 30-meter flying zone, not from a stationary start. Three trials with 6–8 min recovery times between repetitions were allowed, and the best time was recorded to the nearest two decimal places. The timing starts when the athlete’s torso passes through the start of the 30 m cone and finishes at the 60 m cone marker^36,37.

After a recovery period of more than 20 min, a one-minute burpee test was conducted to measure strength endurance. The Burpee Test, has been examined in several studies for its validity and reliability as a measure of whole-body muscular strength endurance^38,39. The rationale is that while some tests concentrate on a specific area of the body or set of muscles, burpee tests are a total body workout in which the participant attempts to perform the maximum number of burpees in the allotted time⁴⁰. Athletes performed squatting by down from the standing posture and placing their hands in front of their feet. The legs are then pushed back into a push-up stance with a straight line from the shoulders to the heels while the body weight is transferred to the hands. Returning to a squatting position after the legs are pulled back will allow them to stand up straight again. Each repetition must meet these criteria to be counted as a valid burpee. The maximal possible number of burpees in one minute was recorded and failing to do one of the criteria in each repetition was not counted (incomplete push-up position: If the participant’s body does not form a straight line from shoulders to heels (e.g., hips are sagging or too high), Feet Placement: If the participant fails to thrust their legs back fully into the push-up position or does not bring them back fully to the squat position, Jump and Reach: If the participant does not perform a full jump or does not raise their hands above their head, Chest to ground: If the chest does not touch the ground during the push-up phase, Improper form: Any lack of proper form that compromises the integrity of the movement (e.g., not squatting fully, not extending the body properly)). Three trials with 8–10 min recovery times between trials were allowed, and the best performance was recorded.

On the second day, running performance was tested through an 800 m time trial after a warm-up period of 10–15 min followed by dynamic stretching. 800 m run test can be considered to test the middle-distance running performance on a 400 m oval outdoor track⁴¹. The total time taken to run 800 m was recorded. All testing was performed at a similar time of the day for each subject.

Training protocol

Before starting the investigation, all the participants regularly engaged in low to high intensity running four times a week, in average 21.748 miles per week. However, during the intervention, the researchers introduced one additional hill training program to their routine and substituted their strength training with uphill training on the motorized treadmill Cybex 530T pro plus USA. Heart rate (HR) was measured using the Polar H10 heart rate monitor, which has been validated against electrocardiogram (ECG) standards in previous research^42,43. During the training sessions, HR was continuously monitored, and athletes were instructed to maintain their HR within the 85–100% of HRmax range. HRmax was estimated using the formula 220 – age, and real-time feedback was provided to ensure athletes remained within the target zone. Any deviations were corrected immediately by adjusting running speed. The intervention groups performed 2 uphill training sessions per week over an 8-week period while maintaining their normal running training outside of the weekly uphill training rest sessions, whereas the control group continued their normal training programs. The steeper uphill training sessions consisted of completing 3–4 sets of 6–10 bouts for 30–90 s on a treadmill set to a 7.6% grade while running at 85–100% HR max with 4–6 min rest duration between sets and 2–4 min between repetitions. the intermediate uphill training sessions consisted of completing 2 sets of 6–10 bouts for 1–2 min on a treadmill set to a 5.1% grade while running at 85–100% HR max with 4–6 min rest duration between sets, and 2–4 min between repetitions, and the shallow uphill training sessions consisted of completing 6–10 bouts for 2–3 min on a treadmill set to a 2.5% grade while running at 85–100% HR max with 2–4 min rest duration between repetitions as presented in Table 2. Participants in the control group continued their normal weekly training programs. Training adherence was verified by maintaining detailed attendance records for each training session. Participants were required to attend at least 85% of the sessions to be included in the final analysis. Attendance was recorded by supervising staff at each session. Exercise intensity was monitored using heart rate (HR) monitors worn by all participants during training. HR data were recorded in real time to ensure participants maintained the target intensity range of 85–100% of their age-predicted maximum heart rate (HRmax). Any deviations were addressed immediately by adjusting pace to maintain the prescribed intensity.

Table 2.

8-week uphill training protocols for groups (STHG, IHG, SHG, and CG).

	SHG (n = 10)	IHG (n = 10)	STHG (n = 10)	CG (n = 10)
Gradient	2.5%	5.1%	7.6%	NA
HRmax	85–100%	85–100%	85–100%	NA
Recovery b/n set	4–6 min	4–6 min	4–6 min	NA
Recovery b/n rep.	2–4 min	2–4 min	2–4 min	NA
Progression
Week 1	6 × 2 min	2 × 6 × 60 s	4 × 6 × 30 s	NA
Week 2	8 × 2 min	2 × 8 × 60 s	4 × 8 × 30 s	NA
Week 3	10 × 2 min	2 × 10 × 60 s	4 × 10 × 30 s	NA
Week 4	6 × 3 min	2 × 6 × 90 s	4 × 6 × 45 s	NA
Week 5	8 × 3 min	2 × 8 × 90 s	4 × 8 × 45 s	NA
Week 6	10 × 3 min	2 × 10 × 90 s	3 × 10 × 60 s	NA
Week 7	6 × 4 min	2 × 6 × 2 min	3 × 5 × 90 s	NA
Week 8	8 × 4 min	2 × 8 × 2 min	3 × 7 × 90 s	NA

Adapted from a previous study¹⁶.

NA not applicable, HRmax maximal heart rate, vVO2max velocity at maximal oxygen consumption, STHG steeper hill group, IHG intermediate hill group, SHG shallow hill group, CG control group.

Statistical analysis

Statistical analysis was performed via IBM SPSS Statistics version 27 (IBM Corporation), which is widely applied in behavioral, health, and exercise science research for conducting repeated-measures ANOVA and calculating effect sizes such as partial eta squared (ηp²), which is commonly reported in these fields. The authors determined the sample size with a priori statistical power analysis with G-power 3.1 with a reasonable power of > 80%. Descriptive statistics of each outcome variable are presented as the mean and standard deviation. Two-way mixed (time [pre-post] X group) repeated-measures analysis of variance (ANOVA) was used to establish whether there were any significant differences between the pre-training and post-training tests, the training groups, and any interaction effects for each variable. In case of significant main effects or interactions, post hoc Tukey’s HSD test in correction to adjust p-values for all pairwise comparisons. This approach helps the researchers in control of confounding variables by incorporating both within-subjects and between-subjects factors, ensuring that observed effects are the result of experimental interventions rather than external factors⁴⁴. In addition, it also allows to look into the interaction effects of within-subjects and between-subjects elements. This is critical for understanding how different conditions or treatments interact over time or across groups.

Data normality was verified with the Shapiro‒Wilk W test, and homogeneity of variance was tested via Levene’s test. All the dependent variables (V̇max, TT, and SE) were assessed at a significance level of p < .05 for all statistical analyses. Partial eta squared was used to measure the effect size which indicates, the proportion of the total variance in the dependent variable that is associated with a specific factor, after accounting for (or partialling out) the variance explained by other factors and interactions in the model⁴⁵. Also, effect sizes using partial Etha square defined as either small; 0.01, medium; 0.06, large; ≥ 0.14⁴⁶, the magnitude of the mean ± SD of the mean difference (MΔ), 95% confidence interval (95% CI), and p-values were used for data interpretation⁴⁷.

Results

Table 3 summarizes the descriptive data at baseline and the adjusted absolute changes in Vmax, TT, and SE performance during the study period. The average ages of the participants in the respective groups were STHG = 18.5 ± 0.97 years, IHG = 18.6 ± 1.07 years, SHG = 18.2 ± 0.63 years, and CG = 17.1 ± 1.37 years and the average heights were 1.71 ± 0.04 m, 1.68 ± 0.05 m, 1.71 ± 0.04 m, and 1.68 ± 0.05 m respectively. In addition, the average weight of the participants was STHG = 54.70 ± 3.83 kg, IHG = 54.10 ± 4.01 kg, SHG = 54.70 ± 3.83 kg, and CG = 54.10 ± 4.01 kg, and the body mass index were 18.66 ± 1.04, 18.99 ± 1.50, 18.72 ± 0.98, and 18.99 ± 1.50 respectively.

Table 3.

Mean ± SD of demographic and outcome variables for participants at baseline and follow-up (n = 40).

Variables	STHG (n = 10)		IHG (n = 10)		SHG (n = 10)		CG (n = 10)
	Baseline (Mean ± SD)	Follow-Up (Mean ± SD)	Baseline (Mean ± SD)	Follow-Up (Mean ± SD)	Baseline (Mean ± SD)	Follow-Up (Mean ± SD)	Baseline (Mean ± SD)	Follow-Up (Mean ± SD)
Age (years)	18.5 ± 0.97		18.6 ± 1.07		18.2 ± 0.63		17.1 ± 1.37
Height (m)	1.71 ± 0.04		1.68 ± 0.05		1.71 ± 0.04		1.68 ± 0.05
Weight (kg)	54.70 ± 3.83		54.10 ± 4.01		54.70 ± 3.83		54.10 ± 4.01
BMI (kg/m2)	18.66 ± 1.04		18.99 ± 1.50		18.72 ± 0.98		18.99 ± 1.50
Vmax (m/s− 1)	7.46 ± 0.75	8.74 ± 0.55	7.34 ± 1.06	7.93 ± 0.79	7.31 ± 0.50	7.65 ± 0.55	7.17 ± 0.76	7.17 ± 0.76
TT (min)	2.11 ± 0.08	1.53 ± 0.18	2.07 ± 0.19	1.64 ± 0.25	2.07 ± 0.18	2.01 ± 0.24	2.11 ± 0.06	2.12 ± 0.06
SE (burpee/min)	20.30 ± 2.40	25.80 ± 2.65	20.60 ± 2.63	23.40 ± 3.40	19.70 ± 2.49	20.40 ± 2.54	19.40 ± 2.36	18.90 ± 1.59

BMI body mass index, Vmax maximal running velocity, TT time trial performance, SE strength endurance, SD standard deviation, STHG steeper hill group, IHG intermediate hill group, SHG Shallow hill group, CG control group.

The statistical analysis of all of the pre-training assessments revealed no statistically significant differences between groups in all variables. This demonstrates that groups that were similar at baseline were created by the randomization process, which provides a solid foundation for comparing the three uphill groups to the control group.

A two-way repeated measure ANOVA revealed a significant time effect on Vmax (F (1, 36) = 173.68, p < .001, = 0.82), indicating that Vmax significantly changed over the training period, as indicated in Table 4; Fig. 2. Average Vmax were significantly higher on post-test result (M = 7.87 ± 0.86 m·s⁻¹) than pre-test result (M = 7.32 ± 0.77 m·s⁻¹). Additionally, there is a significant main effect of hill gradient level on Vmax (F (3, 36) = 2.87, p = .049, = 0.20), suggesting differences in Vmax between the different hill gradients as indicated in Table 5. Vmax were significantly higher on STHG (M = 8.74 ± 0.55 m·s⁻¹), than IHG (M = 7.93 ± 0.79 m·s⁻¹), SHG (M = 7.64 ± 0.54 m·s^− 1), and CG (M = 7.16 ± 0.76). There was also a significant interaction effect between time and hill gradient level (F (3, 36) = 42.67, p < .001, = 0.78), indicating that the effect of training time on Vmax varied depending on the hill gradient.

Table 4.

Summary of statistical results for the effects of time and gradient interaction on vmax (maximum velocity in m/s), TT (800 m time trial performance in min), and SE (strength endurance in burpees/min).

Variables	Within subject effects			Pairwise comparison					Interaction effect Time * gradient
	F	Sig.		Mean difference	St. error	Sigb	95% CI		F	Sig.
							Lower bound	Upper bound
Vmax (m/s− 1)	173.68	0.001	0.82	0.55	0.04	0.001	0.63	0.46	42.67	0.001	0.78
TT (min)	155.02	0.001	0.81	0.27	0.02	0.001	0.22	0.31	44.01	0.001	0.78
SE (burpee/min)	107.12	0.001	0.75	2.12	0.20	0.001	1.71	2.54	41.05	0.001	0.77

The table presents within-subject effects (F-values, significance levels [Sig.], and partial Eta squared [η_p²]), mean differences, standard errors, significance, and 95% confidence intervals [CI]), and interaction effects. All effects were statistically significant (p < .001), indicating strong influences of both time and gradient interaction on performance variables.

Fig. 2.

Within and between group effect of hill training on maximal velocity across various gradients.

Table 5.

Summary of statistical results for the effects of gradient on vmax (maximum velocity in m/s), TT (800 m time trial performance in min), and SE (strength endurance in burpees/min).

Variables	Between subject effects			Pairwise comparison
	F	Sig.		Treatment groups	Mean difference	St. error	Sigb	95% CI
								Lower bound	Upper bound
Vmax (m/s− 1)	2.87	0.049	0.20	STHG – IHG	0.47	0.32	0.489	− 0.41	1.34
				STHG – SHG	0.62	0.32	0.236	− 0.25	1.50
				STHG – CG	0.93*	0.32	0.032	0.06	1.81
				IHG – SHG	0.15	0.32	0.963	− 0.72	1.03
				IHG – CG	0.46	0.32	0.487	− 0.41	1.34
				SHG - CG	0.31	0.32	0.776	− 0.56	1.18
TT (min)	7.60	0.001	0.39	STHG – IHG	− 0.04	0.072	0.948	− 0.23	0.15
				STHG – SHG	− 0.22*	0.072	0.023	− 0.41	− 0.02
				STHG – CG	− 0.29*	0.072	0.001	− 0.48	− 1.00
				IHG – SHG	− 0.18	0.072	0.083	− 0.37	0.02
				IHG – CG	− 0.25*	0.072	0.006	− 0.45	− 0.06
				SHG - CG	− 0.08	0.072	0.716	− 0.27	0.12
SE (burpee/min)	5.19	0.004	0.30	STHG – IHG	1.05	1.10	0.778	-1.92	4.02
				STHG – SHG	3.00*	1.10	0.048	0.02	5.97
				STHG – CG	3.90*	1.10	0.006	0.92	6.87
				IHG – SHG	1.95	1.10	0.307	− 1.03	4.93
				IHG – CG	2.85	1.10	0.065	− 0.13	5.83
				SHG - CG	0.90	1.10	0.847	− 2.08	3.88

The table presents between-subject effects (F-values, significance levels [Sig.], and partial Eta squared [η_p²]), mean differences, standard errors, significance, and 95% confidence intervals [CI]). Effects were statistically significant when (p < .05), indicating strong influences of gradient on performance variables.

Vmax maximal running velocity, TT time trial performance, SE strength endurance, CI confidence interval, STHG steeper hill group, IHG intermediate hill group, SHG shallow hill group, CG control group.

* Significantly different at (P < .05) as determined by two-way repeated measure ANOVA, and post-hoc Tukey’s HSD test.

The post hoc Tukey’s Honestly Significant Difference (HSD) indicated that Vmax was significantly higher in the STHG compared to the CG (MΔ = 0.93, 95% CI [0.06–1.81], p = .032). No statistically significant differences were detected in all other possible pairs of comparisons (p > .05). The overall findings are presented in Table 5.

The analysis for time trial performance revealed a statistically significant time effect of (F (1, 36) = 155.02, p < .001, = 0.81), indicating that TT performance significantly changed over the training period. Average time trial performance was significantly lower in the post-test result (M = 1.82 ± 0.31 min) than in the pre-test result (M = 2.09 ± 0.14 min) (see Fig. 3). Additionally, there is a significant main effect of hill gradient level in TT (F (3, 36) = 7.60, p < .001, = 0.39), suggesting differences in TT between the different hill gradients. TT was significantly lower in STHG (M = 1.53 ± 0.18 min) than in IHG (M = 1.64 ± 0.25 min¹), SHG (M = 2.0 ± 0.24 min), and CG (M = 2.11 ± 0.66 min). There was also a significant interaction effect between time and hill gradient level (F (3, 36) = 44.01, p < .001, = 0.78), indicating that the effect of training time on TT varied depending on the hill gradient. The post hoc Tukey’s HSD indicated that TT was significantly lower in the STHG compared to the SHG and the CG (MΔ = − 0.22, 95% CI [− 0.41 − 0.02], p = .023), and (MΔ = − 0.29, 95% CI [− 0.48–1.00], p = .001) respectively. It was also observed a lower TT in IHG compared to the CG (MΔ = − 0.25, 95% CI [− 0.45 – (− 0.06)], p = .006). No statistically significant differences were observed in all other possible pairs of comparisons (p > .05).

Fig. 3.

Within and between group effect of hill training on 800 m TT performance across various gradients.

In addition, the analysis for SE showed a significant time effect of (F (1, 36) = 107.12, p < .001, = 0.75), indicating that SE significantly changed over the training period depicted in Fig. 4. The average SE was significantly higher on the post-test result (M = 22.12 ± 3.70 burpee) than the pre-test result (M = 20.0 ± 2.43 burpee). Additionally, there is a significant main effect of hill gradient level on SE (F (3, 36) = 5.19, p = .004, = 0.30), suggesting differences in SE between the different hill gradients. SE were significantly higher on STHG (M = 25.80 ± 2.66 burpee) than IHG (M = 23.40 ± 3.40 burpee), SHG (M = 20.40 ± 2.55 burpee), and CG (M = 18.9 ± 1.60 burpee). There was also a significant interaction effect between time and hill gradient level (F (3, 36) = 41.05, p < .001, = 0.77), indicating that the effect of training time on SE varied depending on the hill gradient. The post hoc Tukey’s Honestly Significant Difference (HSD) indicated that SE was significantly higher in the STHG compared to the SHG and the CG (MΔ = 3.0, 95% CI [0.02–5.97], p = .048), and (MΔ = 3.9, 95% CI [0.92–6.87], p = .006) respectively. No statistically significant differences were detected in all other possible pairs of comparisons (p > .05).

Fig. 4.

Within and between group effect of hill training on SE across various gradients.

Discussion

The purpose of this study was to examine the chronic effects of uphill training on Vmax, and performance (TT and SE). Additionally, the study aimed to determine the most effective uphill gradient for enhancing these performance measures in young, middle-distance runners. The major findings of the present investigation suggest that 8 weeks of high intensity uphill training can significantly improve the maximal velocity, 800 m time trial performance and strength endurance performance, as shown in Table 5. To the authors’ knowledge, this is among the few studies covering the relative effectiveness of different uphill gradients on Vmax, TT performance, and SE performance. Therefore, the present study demonstrated that steeper uphill training ≥ 7.6% can only significantly improve Vmax than the intermediate and the shallow hill groups.

Although studies on the chronic effects of uphill training at steep slopes on the maximal velocity and performances of runners are scarce, research on intermediate combined uphill and downhill sprint training has shown a significant effect on running performance. Combined uphill-downhill training on an intermediate slope of 3° improved maximal velocity by 4.8% in experienced sprinters²⁵ and 3.5% in physical education students⁴⁸. Bissas, Paradisis²² also reported a 3.7% improvement from a similar design after 6 weeks of training on sport science and physical education students. This is contradictory to the present study, which doesn’t show a significant change. This may be due to the combined effect of uphill with downhill training because training on both uphill and downhill gradients may ensure a balanced development of both concentric and eccentric muscle action and development. The resistance encountered during uphill running demands greater force production from the lower limb muscles, particularly the quadriceps, hamstrings, and calves⁴⁹. This heightened muscular effort promotes hypertrophy and neuromuscular adaptations, enhancing the athlete’s capacity to generate force rapidly⁵⁰. In contrast, downhill training enhances stride frequency and neuromuscular coordination, both of which are essential components of maximal running speed²⁵.

Additionally, uphill running alters stride length and frequency, thereby refining sprinting mechanics and contributing to more effective acceleration and overall speed performance^51,52. However, the reason why the shallow hill might not produce significant change is due to insufficient resistance to significantly overloading the muscles required for sprinting.

This study also showed that steeper and intermediate uphill training can improve 800 m time trial performance, while the shallow hill training has showed no statistically significant change. The result is supported by previous studies^11,16, while no specific studies support the intermediate hill training result. Barnes, Hopkins¹⁶ demonstrated that uphill training on a 10 − 15% gradient with 100 − 110% vVO2max improved 5 km time trial performance by 2%. Similarly, Ferley, Osborn¹¹ also reported that a six-week interval uphill training performed on a 10% gradient improved running performance significantly by 2%. This demonstrated that uphill training on intermediate to steeper uphill gradients with some type of high-intensity training can improve 800 m time-trial performance by enhancing both anaerobic and aerobic capacities. The anaerobic glycolytic system is highly engaged during uphill running, improving the athlete’s ability to tolerate and clear lactate⁵³. This increased lactate threshold allows athletes to maintain higher intensities for longer periods. Additionally, the aerobic benefits of uphill running, such as improved VO2 max and cardiovascular efficiency, contribute to better overall endurance^13,54. Consequently, athletes experience faster 800 m times, as they can sustain higher speeds and resist fatigue more effectively. Uphill running can be a more sport-specific training strategy than other high-intensity resistance-to-movement workouts such as plyometric and weightlifting. Research indicated that enhancing the amount of time spent at or near V̇O2max, or the amount of work done at a high intensity, through training such as uphill training is essential because the ability to sustain near maximal efforts in running is highly correlated with running performance in races ranging from 800 m to 10 km^13,55.

The performance in terms of strength endurance also improved across intermediate and steep hill gradients in the present study, but no statistically significant change was found in the shallow hill group. A previous study confirmed that treadmill running at a 10% gradient with 90% maximal aerobic speed can significantly improve muscle endurance by 21.2%⁵⁶. In addition, uphill training performed at 3^o has been shown to increase maximal isometric force and rate of force production in leg muscles, with improvements of 7.1% and approximately 25%, respectively⁵⁰. This indicates that uphill running necessitates greater net mechanical work compared to level or downhill running, leading to enhanced muscular endurance¹⁰. The continuous effort required to overcome gravity during uphill runs promotes muscular endurance, allowing athletes to perform at a high intensity for extended durations. This type of training also improves running economy, as athletes learn to optimize their stride mechanics and energy expenditure⁵⁷. As a result, they can maintain a faster pace with less effort over long distances.

The classification of hill gradients into steeper, intermediate, and shallow categories is critical for customizing training programs. The effect size resulted from the time, hill gradient level, and the interaction between the two were presented in Tables 4 and 5). Relative to intermediate hill training, steeper uphill training significantly enhances performance metrics by optimizing biomechanical efficiency, energy expenditure, and kinematic adaptations. This is because a steeper incline leads to distinct kinematic features, such as increased joint angles and higher ranges of motion during sprinting, which are crucial for acceleration⁴⁹, and steep gradients exhibit faster stride frequencies and shorter foot-ground contact times compared to lower hills, indicating a more efficient running gait⁵⁸. Although steeper and intermediate hill gradients resulted notable performance benefits, training on shallow inclines does not appear to elicit significant improvements across key performance parameters. This discrepancy may be attributed to differences in muscle recruitment patterns, metabolic demands, and perceptual responses. Steep and intermediate gradients are known to enhance activation of major lower-limb muscles particularly the gluteus maximus, hamstrings, and quadriceps more effectively than shallow gradients, which may not sufficiently challenge these muscle groups to induce meaningful adaptations^59,60.

Additionally, compared to intermediate inclines, shallow hill gradients may not sufficiently challenge the body’s physiological systems to trigger meaningful adaptations, as they place lower demands on both aerobic and anaerobic energy systems⁶¹. From a biomechanical standpoint, intermediate gradients can promote more effective force application and stride mechanics, than shallow gradients in improving running economy. Furthermore, the perceived exertion at intermediate gradients may strike an optimal balance challenging enough to drive adaptation, but not so intense as to impair training quality or recovery. In contrast, shallow gradients may not provide a sufficient training stimulus, resulting insignificant performance improvements.

Although, treadmill running can replicate certain physiological responses and shows strong correlations with outdoor performance^62–64, its ecological validity remains a subject of debate. Forinstance, a study confirmed that treadmill running results in longer completion times and higher perceived exertion, along with elevated blood lactate levels compared to outdoor running, suggesting discrepancies in effort and physiological responses⁶³. Additionally, treadmill conditions may restrict natural running mechanics, particularly affecting acceleration and deceleration patterns that are crucial in outdoor races⁶⁵. Similarly, a study highlighted that treadmill running may not fully replicate outdoor race conditions due to differences in pacing strategy, heart rate responses, and environmental factors, suggesting that performance on a treadmill is generally slower and less motivating than on a track⁶⁶. This finding suggests that treadmill running offer a controlled setting ideal for monitoring physiological responses and training adaptations enhances experimental reliability and repeatability, it may not fully capture the complexities and demands of outdoor race conditions including environmental variability such as terrain variablity, wind resistance, humidity, temprature, visual cues and physiological conditions experienced during outdoor races limits their ecological validity. These differences can influence running biomechanics, energy expenditure, and psychological responses.

Limitations of the study

While this study constitutes one of the few studies in the literature, comparing the chronic effects of three different uphill gradients on various performance measures, it has methodological limitations that need to be acknowledged and addressed in the future. A key limitation of this study is the small sample size, which may reduce statistical power and limit the generalizability of the findings. Additionally, due to the small number of athletes available in the field, the study included a range of training experiences (6 months to 4 years). This performance heterogeneity among participants such as differences in training background and fitness level may introduce variability that affect the interpretation of results^67,68. Potential sex differences also, as male and female can exhibit distinct physiological responses to training, including differences in oxygen-carrying capacity, body composition, and metabolic efficiency⁶⁹. These may all affect the result of the study, as athletes at different stages of their training journey and due to gender may respond differently to the same training regimen. This diversity of the study population may be considered a limitation. Additionally, a priori power analysis was conducted using G*Power to determine the target sample size, limitations exist in accurately predicting the number of participants required, particularly due to uncertainty in estimating true effect sizes. Therefore, while the study was powered to detect moderate to large effects, caution is advised in interpreting smaller between-group differences. Future studies should consider larger sample sizes, sex-balanced cohorts and stratify participants by performance level to better account for these variables or multi-site recruitment to improve power and generalizability. However, the intervention study on athletes having some training experience while using well-established controlled training protocol can be viewed as an advantage in the current study and for the future development of efficacious training programs.

Conclusions

The present study investigated the effects of different uphill training gradients on Vmax, 800 m TT, and SE performances. The findings demonstrated that steeper uphill training at 7.6% gradient enhances all performance measures, making it the most effective gradient tested. However, intermediate uphill training at 5.1% also showed significant improvement in 800 m time trial performance. These results highlight the importance of gradient-specific training in optimizing performance and contribute to the limited literature on the chronic adaptations to uphill running. Moreover, the study underscores the role of uphill training in promoting neuromechanical efficiency, muscle adaptation, and training specificity in endurance and speed development.

Practical application

Our findings support that incorporating uphill running in the training programs of middle-distance runners helps to improve key performance parameters. Different uphill gradients appear to induce distinct bio-motor adaptations, suggesting that coaches and athletes should customize training programs using uphill training while carefully matching the strengths and weaknesses of the athlete and the underlying demands of the specific event. For example, steeper uphill (≥ 7.6%) may be particularly applicable in middle-distance running, where high power output, anaerobic capacity, and muscle adaptation are critical for achieving high velocities over moderate distances⁷⁰. In contrast, intermediate hills (~ 5.1%) may be more suitable for longer distances such as 3000 m, requiring relatively less power output, efficient use of energy through improved running mechanics and running economy⁷¹. Future research should aim to explore whether modifications in training variables such as frequency, duration, volume, and periodization can replicate or enhance the performance benefits of uphill training. Moreover, it is important to investigate whether these benefits are true in well-trained or elite athletes, who may already exhibit high levels of neuromuscular efficiency and physiological adaptation.

Author contributions

Y.A, T.T, and Z.B conceived and designed the study. Y.A and Z.B conducted the experiments and collected the data. Y.A, T.T, and ZB analyzed the data. Y.A, T.T, and Z.B wrote the manuscript. All the authors read and approved the manuscript.

Funding

The present study was financially supported by Bahir Dar University and Bahir Dar University sport Academy PhD Program.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

The studies involving human participants were reviewed and approved by the Bahir Dar University Ethical Review Committee (reference number IRERC 05/2024). All the participants were informed about the possible adverse effects before the intervention and provided their written informed consent to participate in the study.