The design of the run Clever randomized trial: running volume, −intensity and running-related injuries

In modern society, running has become a popular way to exercise and the proportion of adults who participate in recreational running has been growing since the 1970s [1]. In surveys of participation in leisure time physical activity, running is reported to be one of the most popular forms of exercise [2‐4] and a national report from the Danish Institute for Sports Studies, estimates that 31 % of the adult population in Denmark participate in recreational running [5]. Running is a cheap and easily accessible form of exercise, and the positive effects of running on health outcomes such as weight loss, cardio respiratory function and mortality are well known [6‐8]. Unfortunately, running is also associated with the risk of sustaining a running-related injury. A recently published systematic review and meta-analysis by Videbæk et al. report, that the incidence of running-related injuries per 1000 h of running varies between 2.5 and 33.0 and estimates a weighted injury incidence among recreational runners of 7.7 (95 % CI 6.9 – 8.7) [9]. The knee, lower leg and ankle/foot are the most common sites of injury and a recent study revealed the median time to recovery from injury to be 72 days [10‐13]. In addition, the proportion of injured runners receiving conservative treatment was 10.7 %, while 4.7 % underwent surgical treatment [10].

People engaged in recreational running or choosing running as a new and active lifestyle, should be offered evidence-based advice on running schedules with minimal injury risk. Running with a minimized risk of injury would aid the choice of an active lifestyle by decreasing discontinuation from running and, possibly, further reduce the proportion of people at risk of chronic diseases.

Training errors are recognized as a risk factor for running-related injuries and in order to develop running schedules with a minimized risk of injury, a better understanding of the different training variables’ influence on injury risk is needed [14, 15]. Trials investigating differences in injury risk in relation to the progression in running volume, the frequency and duration of running and participation in a preconditioning programme, have been conducted [16‐18]. The GRONORUN 1 and 2 trials both mentions the application of an subjective measure of running intensity as a limitation in the trials [19, 20]. The study by Pollock et al.[18] used maximal oxygen uptake as a relative measure of running intensity, comparing different durations of running and different weekly running frequencies. However, the sample consisted of male prison inmates between 20 and 35 years of age, which makes it difficult to generalize the results to recreational or novice runners. A review by Nielsen et al. concluded that there are moderate evidence suggesting that volume is associated with injury risk, but that this association possibly is influenced by the weekly progression in volume and the maturation of the runners [15]. The existing literature on running intensity and the development of injuries show conflicting results, possibly related to the subjective measure used in previous studies [15]. A possible association between the focus of a running schedule and the risk of sustaining specific injuries has also been hypothesized. Specifically, patellofemoral pain, illiotibial band syndrome and patellar tendinopathy were hypothesized to be distance-related, while achilles tendinopathy, gastrocnemius injury and plantar fasciitis were hypothesized to be pace-related [21]. Findings from both experimental and cohort studies support this hypothesized association. In an experimental study by Petersen et al. [22], an increase in running speed resulted in an increase in the peak plantar flexion moment, which was significantly higher than the increase in the peak knee extensor moment (p < 0.05). In another experimental study by Petersen et al. [23], a decreased running speed resulted in a cumulative load at the knee joint which was significantly higher, compared with the cumulative knee joint load at higher running speeds [22, 23]. An observational study by Nielsen et al. [24], with a 1-year follow-up and 847 novice runners included, investigated the progression in volume and found that greater progression in running volume increased the risk of the above mentioned distance-related injuries, but not the overall risk of injury. The aim of the Run Clever trial is therefore to conduct a training schedule intervention trial, comparing a running schedule which focus on running intensity with a running schedule which focus on running volume, with the purpose of extending the above mentioned studies, by investigating the difference in risk between the two training variables and the risk of sustaining specific injuries associated with a specific training variable. The following hypotheses are tested (H).

Furthermore, complementary risk factor analysis on BMI, age, gender, previous injury, running experience and general activity level will be performed. Of interest is the effect modification of the mentioned risk factors on the interventions’ influence on injury risk. Explorative analyzes will be performed on data collected using the Oslo Sports Trauma Research Center Questionnaire [25]. The focus of the explorative analyzes is the reported physical complaints progressing to time loss injuries and the relationship with the running performed.

The Run Clever trial is a randomized trial with a 24 - week follow-up. The follow-up period is divided into an 8 - week Preconditioning period and a 16 - week Intervention Training period. The randomization is performed after the 8 - week preconditioning period (Fig. 1). Participants are randomly allocated into two intervention groups, following different running schedules: Schedule Intensity (Sch-I) and Schedule Volume (Sch-V). The main outcome is injury and the physiotherapists diagnosing injured participants (diagnostic team) are blinded to group allocation.

Run Clever is the first randomized trial to investigate how a running schedule focused on running intensity compared with a running schedule focused on running volume influences the overall risk of injury and the type of injuries sustained.

The population of interest is rapidly growing and should be prioritized, as it possibly is the largest group of recreational active adults worldwide [1, 41]. Recreational runners as a group are characterized by a large heterogeneity, especially concerning running experience and training habits. Consensus regarding the definition of a recreational runner is not established and the term is used inconsistently in relation to a wide variety of running populations. In three studies all describing the included runners as recreational, running experience in years ranged from 5.0 to 8.6 and the weekly volume in kilometers ranged from 28 to 38. However, the weekly running frequency only ranged from 3.0 to 3.8 [35, 42, 43]. The present trial defines the population as recreational runners, using the inclusion criteria of an average weekly frequency between 1 and 3 running sessions the past 6 months. The choice of weekly frequency was primarily based on a descriptive study by the Danish Institute for Sports Studies, which showed that the largest proportion of recreational runners in Denmark averaged 1 – 3 weekly running sessions [44].

The ideal percentage of weekly progression is unclear and a trial by Buist et al. comparing a weekly progression of 10.5 and 23.7 % found no difference in risk of RRI. However, a weekly progression of 30 % has been found to increase the risk of specific RRI [19, 24]. In an attempt to allow for adaptation to the progression in running, the research team developed the 4-week cycle, based on a theoretical step loading approach [29].

The follow-up Run Clever applies, aims at increasing the homogeneity of the included participants, in relation to the running performed most recently, prior to allocation to the intervention. Other running-related factors, such as running experience and previous running-related injury, are non-modifiable parts of the heterogeneity characterizing recreational runners, and can only be sought controlled for through the baseline data collection and complementary risk factor analysis.

The superiority power calculation based on a minimum risk difference between groups of 5 %, and taking a hypothesized loss to follow-up of 15 % into account, revealed a required sample size of 713 participants to reach a power of 80 %. A predetermined inclusion period of 6 months has been determined. The challenge of including 713 participants over a 6 month period is acknowledged. However, the estimated 31 % of the adult population in Denmark who participates in recreational running, provides an approximate background population of 904.685 persons, from which recruitment of participants can take place (Based on numbers from Statistics Denmark) [45]. Information regarding the possibility of participation in the trial is announced through the two largest Danish recreational running organizations. Possible participants can easily access and fill out the inclusion/exclusion questionnaire and indicate their interest in participation. The baseline information is collected through the online training diary, making data collection less time consuming for both participants and investigators. This gives us reason to assume that such a large scale inclusion is possible. However, a sample size of 357 participants, based on a minimum risk difference between groups of 1 %, with a power of 80 % and taking the hypothesized loss to follow-up of 15 % into account, is the minimal acceptable number of included participants and also the minimal acceptable clinical relevant difference. At the end of the 6 month inclusion period, no additional inclusions will be made.

The use of GPS to collect data on running exposure is not new and the applicability of this method has been investigated both in runners and in field-based sports [46‐50]. In the current trial, GPS is used to quantify volume (kilometers) and intensity (minutes per kilometer). Nielsen et al. [47] investigated the feasibility of using GPS to quantify training volume in kilometers. The study concluded that the use of GPS caused no clinically relevant measurement errors. Townshend et al. [46] investigated the accuracy of GPS in the measurement of running speed. The study concluded that the use of GPS can provide accurate data on running speed. Further, unpublished data found a Root Mean Square (RMS) of the average pace in minutes per kilometer < 5 s per kilometer, when collected over approx. 300 m. If the distance was increased to 1000 m, the RMS of the average pace in minutes per kilometer was 3 – 5 s per kilometer [51]. Based on this information, it seems feasible to use GPS to quantify running exposure data on running volume (kilometers) and running intensity (minutes per kilometers). Running intensity will, however, not be averaged using distances < 500 m in the current trial.

A modification of the fourth and last question in the OSTRC questionnaire [25], used to collect data on the secondary outcome is applied. The original severity score ranging from 0 to 100 is not changed. In order to accomplish this, the added fifth answering possibility is given the same response value as the fourth answering possibility. The modified version was employed based on experiences from an ongoing study in our lab on handball injuries, providing previously injured respondents with a possibility of reporting “no participation due to pain”.