Determinants of Uphill Running

With increased participation in trail-running events in the last decade, new research has emerged attempting to explain performance outcomes. Determinants of performance may include physiologic variables we can measure in the lab or other descriptive information such as training history or anthropometric measures. With this information, sport scientists create 'profiles' on individuals within a group of athletes and statistically determine which variables best explain performance outcomes. In traditional marathon running, performance is described by the combination of:

1.     VO2max - maximal aerobic capacity

2.     Lactate Threshold - % of VO2max where Lactate begins to rise disproportionately, and steady state conditions cannot be achieved.

3.     Running economy (RE) - the energy cost for a specific pace. Expressed as mL(O2) per kg-1per kilometer-1. Or with an associated velocity, RE may be represented as velocity at a specific VO2 (i.e., vVO2max).  

4.     Durability - the maintenance of physiologic variables over time. 

This type of research is important to coaches and athletes as it influences training design and periodization. Training design is the strategic planning of training variables including volume (duration), intensity, frequency, and modality (type-of training) used to describe a specific workout or workouts within a training cycle. Whereas periodization is the planned variation of these variables over time to optimally develop the physiological capacities needed at the time of a race or event.  For example, in some cycling-disciplines there is an association between an athlete's 'threshold-power' (often referred to as: functional threshold power or critical power) and race success. This metric represents an output you can sustain for a relatively long time (30-60 mins) but is the upper end of what can be sustainable. Cyclists therefore plan their training design and periodization to elicit the highest threshold power possible on race day.

In a 2020 study by Coates et al., researchers investigated determinants of trail-running performance in a group of runners competing on a 20-km loop with 620 m of vertical for either 50 km, 80 km, or 160 km. Prior to competition, each participant completed traditional laboratory-based tests designed to measure VO2max, ventilatory thresholds, and running economy.

For the incremental running tests, they used 3-minute stages at 1% incline and increased speed by .8 kph (0.5 mph) from what was estimated to be their half marathon pace. Following 3 of these submaximal stages, participants were given a choice of increasing speed (+.8kph) or incline (+1%) every minute until volitional fatigue. Peak velocity was recorded as the highest velocity achieved or was calculated as peak velocity = max velocity (% grade x 0.2 kph). Running economy was calculated using VO2 data from the stage prior to the first ventilatory threshold.

The study revealed that VO2max and peak velocity achieved during the incremental test were related to 50 km performance while running economy was not correlated with performance.

In the 80-km race, peak velocity achieved during the incremental test was the only variable related to performance. Finally, in the 160 km distance, no physiological variables were associated with performance. Researchers concluded that running economy is an inadequate predictor of trail and ultra-running performance, despite its importance in road and track running.

Previously, Millet et al. (2012) proposed that ultra-runners prioritize musculoskeletal adaptations needed to limit tissue damage from repeated impact during long runs in mountainous terrain. This emphasis on factors like increased muscle mass creates a tradeoff which then sacrifices running economy. The highest running economy values have been observed in East African distance runners where there is a tendency for athletes to have long Achilles’ tendons, low body fat, long stride length, high percentage of type I fibers, and long legs relative to body mass.

Considering the vast biomechanical differences between trail and road running, one must question whether traditional measures of running economy are appropriate for trail runners?

For instance, researchers compared running economy on a flat surface at half marathon pace to performance on a trail running course with 620 m per 20 km. For the 80 km course, there would be 2480 m (8136 ft) of climbing, which isn't the most impressive vertical profile relative to other races, but it is still not flat. Theoretically, running uphill at a lower energy cost seems advantageous for trail running and might even be the difference between a competitor having to walk vs. comfortably run.

As an applied sports scientist & coach who has tested hundreds of trail runners, I became interested in trying to make sense out of this conclusion. Since the release of this paper, a heated debate has broken out on whether running economy and the training associated with improving it should even be considered for trail runners.

An important difference between my data set for trail runners is that we use an incremental test that increases treadmill incline during each stage while these researchers are increasing by speed. With my protocol, I typically increase incline 1% every 3-4 minutes and only add speed in small increments if the test is progressing too slowly. The starting speed is set at an intensity that is below their first ventilatory threshold where they can easily hold a conversation. The benefit of using this protocol is that trail runners are more familiar with running on incline at relatively lower velocities on a treadmill. Whereas the drawback of this protocol is that we cannot include a traditional measurement of running economy due to the incremental increase in treadmill incline. Instead, to quantify the treadmill intensity, I use a vertical calculation of 'vertical feet per hour' (VF/hr).

 Equation 1: VF/hr = (Incline (%) / 100) * (speed (mph) / 1) * (5280/1)

Using a metric describing uphill velocity, we can investigate to see if there are differences in the energy cost of uphill running economy within my data set and how this corresponds with uphill performance. In addition, considering how peak velocity achieved during incremental tests predicted performance best in studies by Coates et al. (2020) and Pastor et al. (2022), I became interested in understanding determinants of 'peak vertical velocity.'

Information on the dataset

The dataset I’m using includes 51 runners who completed a test format in 2023 that measured both VO2max and Blood Lactate using a vertical step-test protocol. This includes both males and females with ages ranging between 16-76. All tests were completed with a VO2Master portable metabolic analyzer and a Lactate Plus analyzer. Tests were performed in Bozeman, MT at 1470 m (4,820 ft).

Variables Used in Analysis

·      VO2max was determined from each test as the highest 30-s average VO2 value measured.

·      VO2 at LT1 & LT2 values were determined from the final averaged minute of VO2 at the testing stage at which LT1 and LT2 were determined.

·      Vertical velocity was calculated using equation 1.

·      Vertical RE was calculated as the mL of VO2 per kg-1 per 300 vertical meters-1.

·      LTmin was the lowest measured lactate value.

·      LTmax was the highest measured lactate value.

Table 1: Characteristics of Physiologic Profiles

Vertical Running Economy 

Instead of using the traditional unit of measure of running economy expressed at mL(O2) per kg-1 per km-1, I wanted to use a vertical metric. To my knowledge, this is a new concept so there is obviously room for further investigation. I wanted to use a vertical metric that had value to runners, so I chose 300 m (~1000 ft) for vertical running economy since it is common for trail runners to pace themselves at this scale. Due to a much larger metabolic effort required to run 300 vertical meters (~1000 ft) than 1 flat kilometer, the running economy numbers vary greatly from running economy measures using a kilometer (see table 2 & table 3). In the future, I may consider using a shorter vertical metric like 100 m or 50 m to match closer to flat running economy normative values.

Table 2: Normative Data for Running Economy Values at 1% incline.

To show how running intensity impacts vertical running economy, I've included normative values from my data set at LT1, LT2, & VO2max.

Three running economy curves relating uphill running velocity to VO2 were established. Running economy values from the two most economical subjects (lowest O2 cost for a given uphill velocity) were averaged, and a linear regression equation between uphill running velocity and VO2 was calculated. The same was done for two of the least economical runners and two runners with average economy. Individuals were matched across the 3 groups to having similar VO2 values at LT1, LT2, and VO2max to demonstrate variance in economy.

Figure 3: High, average, and low vertical running economy curves

Discussion

The first key take-away was that there is a wide variance in the energy cost required to run uphill. 3 vertical running economy categories (high, average, & low) were created to evaluate physiologic profiles across the dataset. Normative values of vertical running economy were compiled across the data set at three intensities - LT1, LT2, & VO2max.

Secondly, prediction models of uphill running velocity were created to identify differences in individual profiles. These models are intended to be used in an applied setting to explain outliers in treadmill performance based on the magnitude of the input variables. An individual can then individualize their training to target the variables that are limiting their uphill performance.

Does uphill running economy matter?

It is important to note that in the study by Coates et al. (2020), researchers compared physiologic testing variables to actual performance times in a race. Whereas in this analysis, we are looking for relationships between physiologic variables and treadmill performance.[TC1]  The fundamental issue with available research is that sport scientists are using a measurement of running economy that is not task specific to the demands of trail running. This then leads to misinterpretations that leave out the nuances of what contributes to successful performance.

One of the main misinterpretations by 'evidence-based' coaches & practitioners is that "high VO2 = good, low VO2 = bad." VO2 is simply a measure of how much oxygen one is consuming and the associated energy expenditure. It certainly helps to have a high VO2max, but this does not always correspond to increased speed. For example, an individual may travel up to high altitude and experience benefits in increased hemoglobin mass, ventilatory drive, and capillary density; all leading to higher oxygen transport and therefore increased VO2. However, this does not guarantee they will run faster, especially if they are not training to run faster.

In figure 1, there is a noticeably large variance in uphill running economy across the 3-categories. When comparing the high and low categories, there is approximately a 1000 vertical feet/hour difference at the same VO2 across all intensities. This discrepancy suggests further research is needed to understand uphill running economy in the context of trail running. Discrepancies in exercise economy exist in other endurance sports such as Nordic skiing (Losnegard, 2014) and running (Daniels, 1992) even in highly trained athletes. This means that athletes competing against each other can have diverse combinations of physiologic profiles yet still perform at the similar levels.

Future research should therefore attempt to understand which factors (training history, physiologic variables, etc.) contribute to the diversity in physiologic profiles. Furthermore, are there considerations coaches should make based on the profile of an athlete? [TC2] For instance, in marathon running, the most optimal profile for an athlete is having a high VO2max, high running economy, and a high lactate threshold pace (LT2% of VO2max). However, this profile exists only in a handful of world class runners. So, for athletes with deficits in their profile, are there specific training interventions that lead to the optimal training responsiveness?

In the context of improving running economy, various training interventions been researched such as strength training, plyometrics, short-duration intervals (i.e. strides, fartlek), and increasing weekly mileage. In a recent meta-analysis by Llanos-Lago et al (2024), it was reported that improvements in running economy associated with strength training depend on the type used (i.e., heavy loads, plyometrics, isometrics, or submaximal) and the VO2max of the athlete. It appears that strength training with heavy loads (> 80% of one repetition maximum) can improve running economy at high speeds (> 12 km/h) whereas plyometric training can improve running economy at speeds less than 12 km/h. The combination of two of more strength training methods may induce even greater improvements in running economy than isolated methods. Athletes possessing a high VO2max experience the greatest benefit in running economy when using strength training.

Considering the trade-off between VO2max and running economy, this research can help provide some guidance in navigating the nuances of this topic. For obvious reasons, having a high VO2max is advantageous for endurance performance, however, athletes should always consider strategies to reduce energy cost based on the demands of their sport. For instance, if an ultra-runner plans to walk most of their race, running economy may not be the highest priority. If a runner has enough of an aerobic capacity (i.e., VO2max, VO2 at LT1, etc.) to comfortably run their event and they experience low running economy, then inventions to improve running economy should be included in their training.

Vertical velocity at VO2max

The variables that contributed most to peak vertical velocity in this model were VO2max, vertical running economy at VO2max, VO2 at LT1, and LT1% of VO2max. Based on available research, VO2max appears to be of greater importance the shorter the duration. This is due to higher fractional utilization of VO2max at race pace.

For instance, VO2max likely matters most in 'Vertical-Kilometer (VK)'-style race, which is an uphill mountain running race with an elevation gain of 1,000 meters is less than 5-km. Fornasiero et al (2022) evaluated the relationship between VO2max and VK performance in national and international level runners using cardiopulmonary test data from 64 mountain runners. Using predictive models, they estimate that at an international level, runners require a VO2max of 86 and 74 mL/kg/min to win a race for males and females, respectively. Whereas at the national level, runners require a VO2max of 83 and 66.8 mL/kg/min to win a race for males and females, respectively.

Vertical velocity at LT1

During ultra-running events, the highest intensity an athlete may compete at is likely close to their 1st Lactate Threshold (LT1). This threshold is identified at the intensity where lactate first begins to rise due to release of epinephrine, which facilitates the recruitment of fast-twitch muscle fibers and utilization of glycogen. By measuring both VO2 and lactate simultaneously during a test, we can then assign a metabolic rate (VO2) where lactate begins to rise. This metric, VO2 at LT1, therefore represents the upper end of a metabolic rate that can be sustained for a long period of time.

Using Figure 5, we can start to see a relationship between VO2 at LT1 and VF/hr at LT1.

With VO2 at LT1 only explaining 40% of the variance in VF/hr at this intensity (R-squared = .4), I decided to look deeper into the results. One trend I have noticed with this style of testing is that there is a high degree of the variability in the absolute level of lactate when exercising at low intensities. To begin documenting this, I included the minimum lactate (LTmin) measured during each test. Interestingly, in our dataset this value ranged from .6 - 3.5 mmol/L and averaged at 1.5 mmol/L.

To visualize the relationship between lactate and uphill speed, I graphed LTmin vs. VF/hr at LT1 (see figure 4) and fit a trend line to the data. Interestingly, those individuals with LTmin values above 2 mmol/L never reached a VF/hr pace above 1000 ft/hr. Anecdotally, I can add that most of the athletes with elevated lactate levels engage in workout regimens that put a large-emphasis on high intensity calling for a need for training that reduces lactate levels at low-intensities.

Next, I completed regression analysis of VF/hr at LT1 using VO2 at LT1, LTmin, vertical running economy at LT1, and VO2max. Together, these variables were able to explain 81% of the variance in the dataset with VO2 at LT1, LTmin, and vertical running economy at LT1 all significantly associated with vertical speed at LT1, whereas VO2max was not correlated.

References:

 Coates, Alexandra & Berard, Jordan & King, Trevor & Burr, Jamie. (2020). Physiological Determinants of Ultramarathon Trail Running Performance. 10.31236/osf.io/y2kdx.

Pastor FS, Besson T, Varesco G, Parent A, Fanget M, Koral J, Foschia C, Rupp T, Rimaud D, Féasson L, Millet GY. Performance Determinants in Trail-Running Races of Different Distances. Int J Sports Physiol Perform. 2022 Jun 1;17(6):844-851. doi: 10.1123/ijspp.2021-0362. Epub 2022 Feb 25. PMID: 35213820.

G. Y. Millet, M. D. Hoffman, and J. B. Morin.

Sacrificing economy to improve running performance—a reality in the ultramarathon?

Journal of Applied Physiology 2012 113:3, 507-509

 Fornasiero A, Savoldelli A, Zignoli A, Callovini A, Decet M, Bortolan L, Schena F, Pellegrini B. Eager to set a record in a vertical race? Test your VO2max first! J Sports Sci. 2022 Nov;40(22):2544-2551. doi: 10.1080/02640414.2023.2172801. Epub 2023 Feb 1. PMID: 36725692.

 Losnegard T, Schäfer D, Hallén J. Exercise economy in skiing and running. Front Physiol. 2014 Jan 24;5:5. doi: 10.3389/fphys.2014.00005. PMID: 24478718; PMCID: PMC3900875.

 Daniels J, Daniels N. Running economy of elite male and elite female runners. Med Sci Sports Exerc. 1992 Apr;24(4):483-9. PMID: 1560747.

 Llanos-Lagos C, Ramirez-Campillo R, Moran J, Sáez de Villarreal E. Effect of Strength Training Programs in Middle- and Long-Distance Runners' Economy at Different Running Speeds: A Systematic Review with Meta-analysis. Sports Med. 2024 Jan 2. doi: 10.1007/s40279-023-01978-y. Epub ahead of print. PMID: 38165636.