FormalPara Key Points

Runners with varying levels of performance wear compression clothing to improve this performance and facilitate recovery, and the present systematic review summarizes relevant findings published to date.

The effect size values indicate that application of compression garments during running and/or recovery exerts no beneficial effect on racing performance (400-m sprint—marathon), various physiological parameters or the performance of strength-related tasks during recovery from running.

Small positive effects were observed with respect to the time to exhaustion (in connection with incremental or step tests), running economy, biomechanical variables, clearance of blood lactate, perceived exertion, maximal voluntary isometric contraction, and peak leg muscle power immediately after running, as well as markers of muscle damage and inflammation.

The effect size values for post-exercise pain, damage, and inflammation in muscles indicated large positive effects of compression.

1 Introduction

Runners at various levels of performance and specializing in different events (from 800 m to marathons) wear socks, sleeves, shorts, and/or tights with compression to improve their performance [1–3] and facilitate recovery [4, 5]. To date, the effects of compression clothing have been examined with a variety of running protocols, including short-term submaximal treadmill running [6–11], incremental treadmill tests to exhaustion [12–17], 10-km races [1], time to exhaustion at the pace used during a 10-km race [18], and 10-km submaximal running [14]. Few such investigations have involved real-life or simulated running competitions exceeding 1 hour in duration and conducted outdoors [2, 3, 19, 20], while others have applied compression garments only during recovery from running for 12 [21], 48 [4, 22], or 72 [5] h.

Recently, a number of publications concerning the influence of compression garments on running performance and perception of different intensities and durations of running have reported contradictory findings. In some cases, time-trial performance improved [23], but not in others [1, 3, 25]. Time to exhaustion was either unchanged [12, 18, 24], increased [16], or reduced [13]. In certain studies physiological parameters such as blood concentrations of lactate during [8, 15] and after [13] running and oxygen uptake [13, 18] were influenced to a considerable extent by compression clothing, whereas in others blood levels of lactate [1, 2] and oxygen uptake [12, 16, 26] were not altered.

Reviews of statistical findings in this field have summarized the multiple effects of compression clothing on exercise and recovery in various disciplines [27–30]. As pointed out earlier [29], assessment of effectiveness based on traditional deductive statistics may be prejudiced, since significance can be achieved either by increasing the number of participants and/or decreasing the variance of data comparing control and treatment conditions [31, 32]. Accordingly, as sometimes done [29, 33–35], calculation of effect sizes (ES; [36]) allows effectiveness to be compared and the practical relevance of the application of compression clothing assessed.

The aims of the present systematic review were as follows: (1) to review the available literature concerning compression garments and running, (2) to calculate the effect sizes associated with various markers related to performance and recovery; (3) to identify evidence-based application of compression in connection with distance running; and (4) to develop recommendations concerning the use of compression for distance runners.

2 Methods

2.1 Data Sources and Literature Searching

A comprehensive computerized search of the electronic databases PubMed, MEDLINE, SPORTDiscus, and Web of Science was performed during September of 2015 employing the following keywords: athlete, endurance, endurance running, blood flow, blood lactate, compression, compression clothing, compression garment, compression stockings, running, long distance running, exercise, fatigue, garments, heart rate, muscle damage, pain, swelling, oscillation, oxygenation, oxygen uptake, performance, perceived exertion, power, recovery, strength, stroke volume, textiles, thermoregulation, time to exhaustion, and time trial. In addition, the reference lists of the articles thus identified and from other relevant articles of which we were aware were examined for additional relevant titles.

2.2 Study Selection and Quality Assessment

Original research articles in peer-reviewed journals that investigated any kind of lower-limb compression garment (i.e., knee-high socks, sleeves, shorts, or tights) or whole-body compression garments during and/or after long distance running were included. These studies assessed physiological (VO2max (maximal oxygen uptake), VO2peak (peak oxygen uptake), submaximal VO2, blood lactate, post-exercise clearance of blood lactate, blood gases, cardiac parameters, inflammatory markers), biomechanical (ground contact time, step frequency, step length, swing time), psychological (rates of perceived exertion, perceived temperature, leg soreness), and/or performance parameters (running time, time to exhaustion, jump performance, maximal voluntary isometric contraction, peak leg muscle power). In the present analysis, we have only included data from investigations where (1) absolute values (means and measures of variability) were published or could be obtained from the authors and (2) both an experimental (compression) and a control group (non-compression) of runners at any level of performance (from untrained to elite) were included. Finally, only data concerning participants without any cardiovascular, metabolic, or musculoskeletal disorders were considered (Fig. 1).

Fig. 1
figure 1

Pathway of identified and subsequent excluded or reviewed articles. PEDro physiotherapy evidence database

Full size image

Each study meeting our inclusion criteria was also evaluated by two independent reviewers according to the Physiotherapy Evidence Database (PEDro) Scale [37], where a “yes” answer adds 1 point, and “no” 0 points, and the maximal score is 10 points. This approach has been applied previously in connection with systematic reviews to assess methodological quality [38–40].

2.3 Statistical Analyses

To compare and quantify each parameter of performance and recovery, the ES (Hedges’ g) and associated 95 % confidence interval were calculated as proposed by Glass [41]. Hedges’ g was computed as the difference between the means of experimental (compression) and control (no compression) values divided by the average standard deviation for the population concerned [41]. To optimize the calculation of ES and estimate the standard deviation for Hedges’ g, the standard deviations of the experimental and control groups at baseline were pooled [36]. In accordance with standard practice, the ES values obtained were then defined as trivial (<0.10), small (0.10–0.30), moderate (0.30–0.50), or large (>0.50) [32]. All statistical analyses were carried out in version 11.5.1.0 of the MedCalc software (MedCalc, Mariakerke, Belgium).

3 Results

3.1 Characteristics of the Studies Analyzed

Of the 643 studies initially identified, 32 (published between 1987 and 2015) were examined in detail (Fig. 1). Their average PEDro score was 6.5 (range 5–9). The participants and compression clothing, parameters measured, and protocols in each study are summarized in Table 1.

Table 1 Summary of studies included in a systematic review of investigations into the effect of compression clothing on long-distance running performance and subsequent recovery
Full size table

Thirty-two of these studies involved performance of different running protocols by a total of 494 participants (458 men and approximately 36 women (in one case, the number of women was not reported [1])). Twenty-four included only male participants, one only women, and the remaining seven included both. The mean sample size was 15.0 ± 7.7 (mean ± SD; range 6–36) and age 29.0 ± 7.2 (19–48) years.

The compression garments employed included knee-high socks (n = 17), tights (n = 10), knee-high calf sleeves (n = 4), shorts (n = 3), and tights and a long-sleeve shirt providing whole-body compression (n = 1). Sixteen studies included elite or well-trained subjects, 13 recreational athletes or participants competing at a regional level, and three involved untrained participants. In 26 of these investigations graduated compression, with pressure decreasing in the distal to proximal direction, was applied. Moreover, 27 provided information concerning the level of pressure exerted (7–40 mmHg), six included no such information, and five referred to information from the manufacturer on this matter (Table 1).

The various ESs relating to the effects of applying compression clothing during running and recovery are illustrated in Fig. 2.

Fig. 2
figure 2

Hedges’ g effect sizes (square) and associated 95 % confidence interval (lines) of the application of compression clothing according to various markers of performance and recovery during and after running

Full size image

3.2 Analysis of Performance

Altogether, the findings indicate that compression clothing has little or no positive effect (mean g = 0.03 ± 0.15; range −0.23 to 0.23 [1–3, 19, 20, 23, 25, 42]) on running performance (Table 1), as reflected in the times for a (half) marathon, 15-km trail running, 5- and 10-km runs and 400-m sprint. Of the 11 studies in which the time to exhaustion in incremental or step tests or runs until exhaustion were examined, seven reported small mean effects of compression garments on variables related to performance (mean g = 0.27 ± 0.33; range 0.01–0.96 [4, 13, 15–17, 26, 43]). Three studies documented a moderate-to-large effect of compression garments (g = 0.39 [16]; g = 0.41 [43]; g = 0.96 [4]), whereas four found small negative effects on the time to exhaustion (mean g −0.22 ± 0.11; range −0.31 to −0.07 [12, 15, 18, 24]).

3.3 Running Economy

In the four investigations that evaluated the influence of calf compression sleeves [6], compression socks [18], compression tights [15], or three different compression garments (socks, tights, whole-body compression) [26] on the running economy of well-trained and highly trained runners, a small positive effect was observed (mean g = 0.21 ± 0.38; range 0.00–0.88).

3.4 Biomechanical Parameters

Compression sleeves [6] and stockings [18] exerted a small positive effect (mean g = 0.21 ± 0.38; range −0.33 to 0.72 [6, 18]) on biomechanical parameters (i.e., ground contact time, step frequency, step length, and swing time).

3.5 Physiological Parameters During Running

Whereas maximal oxygen uptake was not affected in most cases (mean g = 0.05 ± 0.35; range −0.28 to 1.16 [9, 12, 13, 15–18, 26]), one study found a large positive effect on this parameter in well-trained runners performing a test to exhaustion (best g = 1.16 [18]). In the case of parameters related to endurance, such as submaximal oxygen uptake (mean g = −0.01 ± 0.10; range −0.21 to 0.16 [6, 8, 15, 26, 44]), compression clothing had no effects. In most cases there were no effects on blood lactate concentrations either (mean g = −0.04 ± 0.33; range −0.96 to 0.54 [1–3, 8–10, 12, 15, 16, 18, 26, 44]), although two studies reported moderate positive effects on this parameter (mean g = 0.49 ± 0.33; range 0.46–0.54 [8, 15]).

Neither blood saturation and partial pressure of oxygen (mean g = −0.05 ± 0.20; range −0.37 to 0.31 [9, 12, 26] nor cardiac parameters, including heart rate, cardiac output, cardiac index, and stroke volume, were influenced to any great extent by the compression garments (mean g = 0.08 ± 0.37; range −0.48 to 1.77 [1, 3, 8–10, 12, 13, 15, 16, 18, 19, 21, 24, 42, 44]), although two studies did observe moderate-to-large positive effects on maximal heart rate (g = 1.77 [18]) and heart rate during submaximal running (g = 0.53 [10]).

3.6 Body and Perceived Temperature

Body core temperature during and after running was affected to small and moderate extents by compression clothing in two studies [20, 45], with no effect on body temperature (p = 0.00) (mean g = 0.24 ± 0.21; range 0.00–0.41 [20, 45]). Although in one case the perceived temperature during running with compression was not altered (g = 0.00 [25]), Venckūnas and colleagues [42] detected a moderate negative effect on perceived body temperature during 30 min of submaximal running (g = −0.32).

3.7 Psychological Variables While Running

The overall effect of compression clothing on perceived exertion during running was small but positive (mean g = 0.28 ± 0.38; range −0.31 to 1.21 [1–3, 7, 13, 18–20, 23–26, 42, 44–46]), although certain investigations showed moderate-to-large positive effects [2, 23, 26, 45] and others small-to-moderate negative [19, 20] or no effects [7, 13, 18, 24, 25, 42].

3.8 Responses After Running

3.8.1 Recovery of Parameters Related to Performance

The present analysis revealed trivial, small, moderate, and large positive, as well as small and large negative effects on recovery in strength-related tasks such as jumping after running (mean g = 0.02 ± 0.33; range −0.64–0.53 [1–3, 7, 19, 44]). Single and repeated jumping performance (counter movement jump) following different running protocols was associated with small to large negative effects [1], as well as small [7, 20], moderate [19], and large [3] positive effects.

Maximal voluntary isometric contraction and peak leg muscle power immediately after running were somewhat greater with than without compression garments (mean g = 0.19 ± 0.22; range −0.05 to 0.53 [2, 3, 5, 19–21]). Leg strength following running protocols as reflected in maximal voluntary (isometric) contractions and isometric knee extensor torque, showed trivial negative effects [3, 21], as well as trivial [2], small [19, 20], moderate [5], and large [3] positive effects.

3.8.2 Clearance of Blood Lactate

Clearance of blood lactate following running exhibited a small positive effect (mean g = 0.29 ± 0.32; range −0.02 to 0.62 [10, 13]).

3.8.3 Markers of Muscle Damage

On average, compression clothing exerted a small effect on post-exercise levels of creatine kinase, a marker for muscle-damage (mean g = 0.11 ± 0.47; range −0.42 to 0.73 [2, 5, 19, 20]. In three studies this effect was moderate-to-large and positive (range 0.35–0.73 [2, 5, 20] and in another small-to-moderate negative at three different time-points (1, 24, and 48 h) after exercise [19]. For other markers of muscle damage, such as serum levels of myoglobin, interleukin 6, and C-reactive protein, both trivial negative [19] as well as trivial, moderate, and large positive effects were detected [2, 5, 19, 20]. The overall average ES for different inflammatory markers was small and positive (mean g = 0.24 ± 0.44; range −0.09 to 1.14 [2, 5, 19, 20]).

3.8.4 Perceived Muscle Soreness

Compression exerted a large positive effect on post-exercise leg soreness and delay in the onset of muscle fatigue (mean g = 0.67 ± 1.06; range −0.44 to 3.80 [2, 5, 14, 19, 20, 22, 23, 26, 47]).

4 Discussion

Compression clothing led to trivial [1, 3, 19, 20, 25, 42] and small [1, 2, 20, 23] ES values in connection with running performance (400 m to 42,195 km). Comparable ES values for improving time to exhaustion were obtained in eight studies [12, 13, 15, 18, 24, 43], whereas three others reported moderate-to-large values [4, 16, 43]. This degree of improvement is in line with that reported by Born and colleges [29], as well as with the influence of compression garments on performance in other disciplines such as cycling and repeated sprinting.

From a physiological point of view and depending on the duration and environmental conditions, performance during running is determined mainly (although not exclusively) by the athlete’s peak oxygen uptake, fractional utilization of VO2peak, velocity at the lactate threshold, running economy (including biomechanical factors [48], and heat exchange processes [49]). Other physiological mechanisms of relevance in this context include enhanced hemodynamics, i.e., elevated venous return [50, 51], arterial inflow [52], and lymphatic outflow [53]. Since our statistical analysis revealed that the ES between compression and non-compression running for peak oxygen uptake, oxygen uptake, and blood lactate was trivial, a runner will most probably not benefit from compression in these respects.

However, the small positive ES for body temperature, running economy, and biomechanical variables during running indicate a potential (small) benefit of compression with regard to running economy (due potentially to altered biomechanics). This is consistent with previous reports [28, 29] that compression clothing improves neuro-mechanical parameters, including lower presynaptic inhibition [54, 55] and coordinative function [56], as well as recruitment of fewer muscle fibers [45, 57].

In contrast to performance and physio-biomechanics, compression appears to exert positive effects on the psychological parameters examined, i.e., a small positive effect on perceived exertion [1–3, 7, 13, 18–20, 23–26, 42, 44–46], as well as a large positive effect regarding leg soreness [2, 5, 14, 19, 20, 22, 23, 26, 47] both during running and recovery. As described previously [29], this psychological improvement may be a result of (1) attenuated oscillatory displacement of the leg muscles [57, 58], (2) a reduction in the number of muscle fibers recruited [59], (3) less structural damage to muscles [47, 60], and/or (4) improved lymphatic outflow leading to less muscle swelling and, thereby, greater comfort [53]. Since it is difficult to design an appropriate placebo condition for wearing compression garments, the possibility that the improvement in psychological parameters is influenced by more positive perceptions and the participants’ intuitions concerning the results to be expected cannot be excluded.

The mean serum levels of creatine kinase were reduced in all of the studies (small ES), indicating that compression may help diminish structural damage to muscle and facilitate the clearance of metabolites through improved lymphatic outflow. The analytical review by Hill and colleagues [30] found higher (moderate) ES values for post-exercise levels of creatine kinase, but their investigation involved vertical jumping, repeated sprinting, and resistance training, rather than running. These effects have been attributed to attenuated release of creatine kinase into the blood and improved clearance of metabolites [61], as well as better muscle repair [62]. Furthermore, improvement of the pump function of muscles by compression clothing (described in detail by Born et al. 2013 [29]) may enhance circulation and thereby removal of creatine kinase. In this connection it is important to note that we also calculated increases (small ES) in the temperatures of both muscles and the whole body [20, 45] with compression. Since biochemical processes are controlled by temperature, these changes may also contribute to the differences in physiological and psychological variables.

The various hemodynamic (venous return and arterial inflow) and neural mechanisms and mechanical properties by which compression enhances performance and recovery have been described in detail previously [28, 29]. Since the methods (apparatus, study design, intensity and duration of running, range of compression) employed in the different studies examined here varied, we have refrained from meta-analysis. However, unlike Born [29], we have focused on running only and the reasonably large number of studies (n = 32) involving 494 participants analyzed here provides an adequate overview (Fig. 2) of the responses of various parameters of importance for running performance and recovery to compression.

5 Conclusions

On the basis of the mean ES-values for variables related to performance and recovery calculated from 32 studies, we conclude that compression exerts a trivial mean effect on running performance (times for a (half) marathon, 15-km trail running, 5- and 10-km runs and 400-m sprint), maximal and submaximal oxygen uptake, blood lactate concentrations, blood gas kinetics, cardiac parameters (including heart rate, cardiac output, cardiac index, and stroke volume), and body and perceived temperature, or on the performance of strength-related tasks after running. Small positive effects were observed for the time to exhaustion (in incremental or step tests), running economy (including the biomechanical variables ground contact time, step frequency, step length, and swing time), clearance of blood lactate, perceived exertion, maximal voluntary isometric contraction and peak leg muscle power immediately after running, and markers of muscle damage and inflammation. Body core temperature was calculated to be moderately affected by compression. The ES values for post-exercise leg soreness and the delay in onset of muscle fatigue indicated large positive effects of compression.

Apparently, by wearing compression garments runners might slightly improve variables related to endurance performance (i.e., time to exhaustion), due to improvements in running economy, biomechanical variables, perception, and muscle temperature. They should also benefit from reduced muscle pain, damage, and inflammation during recovery.