1

JEMonline

Post-Exercise Recovery Modalities: A Meta-Analysis

James W. Odom1, Frank B. Wyatt2, Frank J. Spaniol1

1Texas A&M University – Corpus Christi, Corpus Christi, TX, USA, 2Midwestern State University, Wichita Falls, TX, USA

ABSTRACT

Odom JW, Wyatt FB, Spaniol FJ. Post-Exercise Recovery Modalities: A Meta-Analysis. JEMonline2016;1(5):1-20. The aim of this study was to examine the effects of various post-exercise recovery modalities. A meta-analytic review was undertaken, assessing the effects of active recovery (ACT), cold water immersion (CWI), contrast water therapy (CWT), and compression garments (CG) on various psychological (PSY), biological (BIO), and performance (PERF) variables.A thorough search of literature databases yielded 12 studies for analysis. All examined modalities appear to focus on enhanced blood circulation, enhancing metabolite clearance, edema reduction, and enhancing nutrient uptake. ACT, CWI, CWT, and CG all show beneficial effects on PSY recovery. However, their effects on PERF are ambiguous at best, with certain modalities affecting specific PERF variables. ACT seems most appropriated when successive bouts of exercise are required within 30 min of each other. PERF effects of CWI and CWT may positively affect repeated sprint ability (RSA), with other indicators showing differing results. For CG, there is evidence of a beneficial effect on change of direction. Evidence pertaining to the effects of BIO recovery is lacking in this analysis.Further research is required to better define post-exercise recovery and determine the effects of these recovery enhancing modalities.

Key Words: Recovery, Cold Water Immersion, Compression Garments, Athletes

1

INTRODUCTION

Athletic competition and associated training sessions result in various fatiguing mechanisms, which may lead to subsequent and temporary decrements in performance as well as reduced adaptation responses. This has led to the development of various post-exercise modalities aimed at attenuating the losses in performance, including: cold water immersion (CWI); contrast water therapy (CWT); compression garments (CG); and active recovery sessions (ACT). Whileother methods exist, the aforementioned modalities have been thoroughly studied and reviewed (8,13,18,28,35,36) with a greater level of affordability and ease of implementation. For this reason, these methods have been chosen for meta-analysis.

CWI is a form of hydrotherapy in which the athlete is submerged up to the waist, chest, or neck in water between 12°C to 20°C for periods of 5 to 15 min(11,12,19,20,23,28,37). This has been widely implemented by exercise professionals for its effects on local and systematic blood flow, nerve conduction, and thus lowered perception of pain, enhanced metabolite clearance, and accelerated return to baseline performance (28,35). Additional benefit is derived from the change in hydrostatic pressure, increasing cardiac output (Q), oxygenation, and reducing edema (28). CWT is a similar modality, alternating between hot (36°C to 40°C) and cold (12°C to 20°C) baths or showers for 5 to 7 rotations, lasting between 1 to 2 min each. Various protocols utilize different ratios, usually 1:1 or 2:1 hot to cold (9,11,12,15,26). With previously mentioned physiological mechanisms, the fluctuation in temperature induces vasodilation/vasoconstriction of the blood vessels, known as a “vascular pump” that further enhances systematic blood flow (35,36). Accelerated recovery from CG may be resultant of increased external pressure gradient, reduced mechanical strain, and stabilization of muscle (13,18). These mechanisms lessen edema through decreased interstitial space, enhanced blood flow, and improved metabolite flux (13,18). These garments tend to increase distally in pressure and can be worn as full-length, lower body tights or as stockings for a period of 24 to 48 hrs (6,10,20). ACT is the most widely used recovery protocol, applying different modes of light, aerobic activity to increase blood flow and clear waste products (16,32-34).

In the literature, numerous variables have been examined to assess the level of fatigue. The return of these measures to baseline is considered indicative of recovery. These variables can be placed into three categories: psychological (PSY), biological (BIO), and performance (PERF). BIO variables include various biomarkers, such as: creatine kinase (CK); lactate dehydrogenase (LDH); and blood lactate (BLa). Additional BIO measures may include oxygen consumption (VO2) and heart rate (HR). PSY variables involve subjective measures through the use of Likert scales or questionnaires, such as: profile of mood states (POMS); perceived fatigue (PF); and delayed onset muscle soreness (DOMS), which is perhaps the most examined indicator of fatigue or recovery in this study (6,9,11,20,25).The translation of these changes into PERF has also been investigated, focusing on tests of jumping, sprinting, agility, and sport-specific ability.

While there is a considerable amount of literature pertaining to this subject, the results are largely equivocal (13,18,28,35,36). Further exacerbating this issue is the variability in research protocols and recovery measures, making summarization and consensus problematic. Clarifying the effects of these modalities will enhance the athlete’s ability to reduce the deleterious effects of repetitive competition and training stress, while assessment of measures and methods will advance the research field. Therefore, a meta-analytic review has been undertaken to investigate: (a) the efficacy of post-exercise recovery modalities; and (b) the proper measures and methods undertaken in the related research.

METHODS

Search of Literature

A meta-analytic review was conducted thatinvolved a thorough search of electronic databases (Pub Med, SPORTDiscus, Google Scholar), cross references from other studies, and manual search. Keywords used in the search included: “recovery modalities”, “cold water immersion”, “contrast water therapy”, “ice baths”, “contrast baths”, “contrast showers”, “compression garments”, “active recovery”, and “athletes”.Following the collection of literature, 21 studies were selected to be coded. Upon review, 9 studies were excluded due to insufficient data or the use of modalities during exercise, as this was viewed as an aid in performance rather than post-exercise recovery. However, these findings will be discussed. Consequently, this yielded 12 studies for final analysis.

Inclusion Criteria

Inclusion criteria included: (a) healthy junior and adult athletes as participants; and (b) the use of hydrotherapy, CG, and/or ACT post-exercise as part of the individual methodology. Studies involving injured athletes were excluded, as was the case with studies that involved interventions of sleep, nutrition, supplementation, pharmacological agents, and animals. These criteria were developed to increase the external validity of the meta-analysis and simplify the statistical process.

Statistical Analysis

Upon collection, means and standard deviations of the experimental and passive control groups (PAS) were extracted to calculate effect size (ES) using Cohen’s d. The magnitude of change from pre-test to all post-test time points was determined for all groups. When summarizing results, differences between PAS and experimental ES’s were calculated and given an inference level (0.00-0.20: trivial; 0.20-0.60: small; 0.60-1.20: moderate; 1.20-2.00: large; 2.00-4.00: very large (15).

RESULTS

In total, 12 studies were included for meta-analysis, investigating the effects of 4 recovery modalities on 223 subjects. Further, 393 total ES’s were calculated (PSY: 130, BIO: 101, PERF: 162), examining magnitude of change from baseline for all variables in all groups. Additional ES’s were calculated between-group in the cases where pre-test values were not provided in the original research. Notable differences in ES between PAS and experimental groups were reported with inference levels assigned (Tables 1-4).

Active Recovery

ES’s for variables in all studies involving ACT are displayed in Table 1. Few PSY variables were examined, with several small to moderate differences in ES reported. BIO variables appear more effected by ACT, with greater moderate responses. However, differences in ES for PERF variables were mostly trivial with very few small changes.

Table 1. Summary of Results: ACT

Study/Modality / N / PSY (d) / BIO (d) / PERF (d)
Losnegard et al.: ACT / 10 / none / BLa: PASpost (1.08)
ACTpost (0.33)**
VO2: PASpost (0.35)
ACTpost (0.82)*
VO2peak: PASpost (-0.02)
ACTpost (0.46)*
HRpeak: PASpost (0.00)
ACTpost (0.22)*
Vepeak & O2 deficit: all trivial / 800m TT: PASpost (0.43)
ACTpost (0.10)*
Suzuki et al. :ACT / 15 / POMSanger: PASimm (0.26)
ACTimm (1.26)**
POMSfatigue: PASimm (0.85)
ACTimm (1.31)*
POMStension: PAS48 (-0.54)
ACT48 (-1.04)* / CK: PASimm (0.17)
PAS24 (0.52)
PAS48 (-0.14)
ACTimm (0.93)**
ACT24 (2.76)££
ACT48 (0.74)**
LDH: PASimm (1.04)
PAS24 (0.14)
PAS48 (-0.59)
ACTimm (2.52)£
ACT24 (1.60)£
ACT48 (-0.69) / none
Tessitore et al.:ACT / 10 / PB: (Dry) ACT(1.75)
(Wet) ACT (3.00)£
Between-group ES / Catecholamines: insufficient data / CMJ: PASimm (-0.21)
(Dry) ACTimm (-0.30)
(Wet) ACTimm (-0.57)*
24 & 48 h: all trivial
Bouncing Jump: all trivial
10m: all trivial

Abbreviations: PSY; psychological variables, BIO; biological variables, PERF; performance variables, PAS; passive recovery, CWI; cold water immersion, CWT; contrast water therapy, CWS; contrast water shower, CG; compression garments, ACT; active recovery, DOMS; delayed onset of muscle soreness, PF; perceived fatigue, PB; perceived benefit, POMS; profile of mood states, CK; creatine kinase, LDH; lactate dehydrogenase, Thigh: thigh circumference, BLa: peak blood lactate, VO2; oxygen consumption, HR; heart rate, CMJ; countermovement jump, RSA; repeated sprint ability, RA; repeated agility, TT; time trial, YYRT; Yo-Yo Intermittent Running Test, TorqueLE; Leg Extension Peak Torque, subscript indicates time of measurement

*=small difference from PAS (d=±0.20 - ±0.60)

**=moderate difference from PAS (d=±0.61 - ±1.20)

£=large difference from PAS (d=±1.21 - ±2.00)

££= very large difference from PAS (d=±2.01 - ±4.00)

°separate table

Compression Garments

A summary of results involving application of CG only are displayed in Table 2. A moderate difference in effect on DOMS was noted (1.11), while trivial to small differences in effect were calculated for BIO variables. Only one study provided pre-test values for PERF variables (6). It should be noted that another study examined the effects of CG alongside CWI and will be discussed later (Table 4) (18).

Table 2. Summary of Results: CG

Study/Modality / N / PSY (d) / BIO (d) / PERF (d)
Davies et al. :CG / 11 / DOMS: PAS48 (0.75)
CG48 (1.86)**
24h: trivial / LDH: PAS48: -0.06
CG48: 0.35*
24h: trivial
CK: all trivial
Thigh: all trivial / 5-0-5: PAS48: 0.41
CG48: 0.17*
CMJ: all trivial
5m: PAS48 (0.22)
CG48 (0.55)*
10m: PAS48 (0.52)
CG48 (0.76)*
20m: trivial
Hamlin et al.:CG / 22 / none / CK: PAS24 (1.41)
CG24 (0.70)*
BLa & HR: no pre values / Fatigue%: no pre values
3km run: no pre values

Abbreviations: PSY; psychological variables, BIO; biological variables, PERF; performance variables, PAS; passive recovery, CWI; cold water immersion, CWT; contrast water therapy, CWS; contrast water shower, CG; compression garments, ACT; active recovery, DOMS; delayed onset of muscle soreness, PF; perceived fatigue, PB; perceived benefit, POMS; profile of mood states, CK; creatine kinase, LDH; lactate dehydrogenase, Thigh: thigh circumference, BLa: peak blood lactate, VO2; oxygen consumption, HR; heart rate, CMJ; countermovement jump, RSA; repeated sprint ability, RA; repeated agility, TT; time trial, YYRT; Yo-Yo Intermittent Running Test, TorqueLE; Leg Extension Peak Torque, subscript indicates time of measurement

*=small difference from PAS (d=±0.20 - ±0.60)

**=moderate difference from PAS (d=±0.61 - ±1.20)

£=large difference from PAS (d=±1.21 - ±2.00)

££= very large difference from PAS (d=±2.01 - ±4.00)

°separate table

Hydrotherapy

Since most of the included studies involving hydrotherapy examined both CWI and CWT, results are combined and displayed in Table 3. Large and very large differences in effect were noted for both CWI and CWT for PSY variables. However, only one study investigated BIO variables, finding few small effects (11).In PERF variables, many small effects were reported, with large positive effects on repeated sprint ability (RSA) in one study (9). Interestingly, a detrimental effect was also noted for CWI on RSA (12).

Table 3. Summary of Results: Hydrotherapy

Study/Modality / N / PSY (d) / BIO (d) / PERF (d)
Elias et al. :CWI, CWT / 24 / DOMS: PAS1 (5.62)
PAS24 (6.75)
PAS48 (4.25)
CWI1 (3.14)££
CWI24 (3.43)££
CWI48 (0.71)££
CWT1 (5.71)
CWT24 (6.00)**
CWT48 (2.57)£ / none / RSA: PAS24 (1.97)
PAS48 (0.97)
CWI24 (0.09)£
CWI48 (0.00)**
CWT24 (0.79)£
CWT48 (0.44)*
Various Jump Variables(9)
Higgins et al.:CWI, CWT2011 / 26 / none / none / RSA: PAS4w (0.23)
CWI4w (-0.55)**
CWT4w (0.30)
300m: PAS4w (-0.23)
CWI4w (-0.29)
CWT4w (-0.49)*
Higgins et al.:CWI, CWT2013 / 24 / DOMS: PAS1 (-1.23)
PAS48 (-1.05)
PAS72 (-0.30)
PAS96 (-1.02)
PAS144 (-0.94)
CWI1 (-1.16)
CWI48 (-0.62)*
CWI72 (-0.57)*
CWI96 (-0.66)
CWI144 (-1.16)*
CWT1 (-1.86)**
CWT48 (-1.98)**
CWT72 (-1.60)£
CWT96 (-1.66)**
CWT144 (-2.39)£ / Thigh:1h, 48h: all trivial
PAS72 (0.04)
PAS96 (0.48)
PAS144 (0.30)
CWI72 (0.38)*
CWI96 (0.72)*
CWI144 (0.56)*
CWT72 (0.35)*
CWT96 (0.48)
CWT144 (0.53)* / CMJ: 1h: all trivial
PAS48 (0.08)
PAS72 (0.37)
PAS96 (0.26)
PAS144 (-0.45)
CWI48 (-0.13)*
CWI72 (-0.19)*
CWI96 (-0.19)*
CWI144 (-0.38)
CWT48 (-0.18)*
CWT72 (0.08)*
CWT96 (-0.23)*
CWT144 (-0.28)
Flex: 1-96h: all trivial
PAS144 (0.27)
CWI144 (0.07)*
CWT144 (0.11)
Juliff et al.:CWT, CWS / 10 / PF: PASimm (1.00)
PAS5 (1.57)
PAS24 (0.57)
CWTimm (1.00)
CWT5 (0.86)**
CWT24 (0.75)
CWSimm (0.25)**
CWS5 (0.83)**
CWS24 (0.33)*
PB: PASpost (0.36)
CWTpost (-0.13)*
CWSpost (-1.87)££ / none / RA: PASimm (0.27)
PAS5 (-0.09)
PAS24 (0.82)
CWTimm (0.27)
CWT5 (-0.06)
CWT24 (0.27)*
CWSimm (0.15)
CWS5 (-0.46)*
CWS24 (0.23)*
Robey et al.:CWT / 20 / DOMSe: PASimm (2.10)
PAS24 (1.3)
PAS48 (0.90)
PAS72 (0.50)
CWTimm (2.44)*
CWT24 (1.78)*
CWT48 (1.33)*
CWT72 (1.22)**
DOMSc: PASimm (1.90)
PAS24 (2.50)
PAS48 (2.20)
PAS72 (1.30)
CWTimm (2.75)**
CWT24 (3.00)*
CWT48 (2.75)*
CWT72 (1.88)* / none / Torquec: all trivial
Torquee: 24,48h: all trivial
PAS72 (-0.27)
CWT72 (0.14)*
2km row: all trivial
Rupp et al.:CWI / 22 / PF: PAS24 (0.00)
PAS48 (2.00)
CWI24 (-0.21)*
CWI48 (1.91) / none / CMJ: PASimm (0.28)
PAS24 (-0.13)
PAS48 (-0.19)
CWIimm (0.07)*
CWI24 (-0.09)
CWI48 (-0.09)
YYRT: all trivial

Abbreviations: PSY; psychological variables, BIO; biological variables, PERF; performance variables, PAS; passive recovery, CWI; cold water immersion, CWT; contrast water therapy, CWS; contrast water shower, CG; compression garments, ACT; active recovery, DOMS; delayed onset of muscle soreness, PF; perceived fatigue, PB; perceived benefit, POMS; profile of mood states, CK; creatine kinase, LDH; lactate dehydrogenase, Thigh: thigh circumference, BLa: peak blood lactate, VO2; oxygen consumption, HR; heart rate, CMJ; countermovement jump, RSA; repeated sprint ability, RA; repeated agility, TT; time trial, YYRT; Yo-Yo Intermittent Running Test, TorqueLE; Leg Extension Peak Torque, subscript indicates time of measurement

*=small difference from PAS (d=±0.20 - ±0.60)

**=moderate difference from PAS (d=±0.61 - ±1.20)

£=large difference from PAS (d=±1.21 - ±2.00)

££= very large difference from PAS (d=±2.01 - ±4.00)

°separate table

It should be noted that a study by Montgomery et al. was analyzed, in which investigators examined effects of both CG and CWI, whose results are displayed in Table 4. Here, moderate and large effects were reported for PSY variables, however, mainly small effects were noted for PERF variables. Interestingly, a moderate detrimental effect was noted for CG on 20 m sprint time.

Table 5. Montgomery et al. Results

Study/Modality / N / PSY (d) / BIO (d) / PERF (d)
Montgomery et al.: CWI, CG / 29 / DOMS: PASpost (5.40)
CWIpost (3.80)£
CGpost (3.60)**
PF: PASpost (1.78)
CWIpost (0.80)**
CGpost (1.33)* / none / CMJ: PASpost (-0.29)
CWIpost (-0.67)*
CGpost (-0.58)*
Flex: PASpost (-0.81)
CWIpost (-0.32)*
CGpost (-1.07)*
20m: PASpost (0.36)
CWIpost (0.15)*
CGpost (1.57)**
Line Drill: PASpost (0.28)
CWIpost (-0.11)*
CGpost (-0.66)*
Agility: PASpost (0.56)
CWIpost (0.54)
CGpost (0.17)*

Abbreviations: PSY; psychological variables, BIO; biological variables, PERF; performance variables, PAS; passive recovery, CWI; cold water immersion, CWT; contrast water therapy, CWS; contrast water shower, CG; compression garments, ACT; active recovery, DOMS; delayed onset of muscle soreness, PF; perceived fatigue, PB; perceived benefit, POMS; profile of mood states, CK; creatine kinase, LDH; lactate dehydrogenase, Thigh: thigh circumference, BLa: peak blood lactate, VO2; oxygen consumption, HR; heart rate, CMJ; countermovement jump, RSA; repeated sprint ability, RA; repeated agility, TT; time trial, YYRT; Yo-Yo Intermittent Running Test, TorqueLE; Leg Extension Peak Torque, subscript indicates time of measurement

*=small difference from PAS (d=±0.20 - ±0.60)

**=moderate difference from PAS (d=±0.61 - ±1.20)

£=large difference from PAS (d=±1.21 - ±2.00)

££= very large difference from PAS (d=±2.01 - ±4.00)

°separate table

DISCUSSION

Active Recovery

In a study by Losnegard et al. (16), various BIO markers were assessed, noting a decrease in peak BLa in ACT compared to PAS (d=0.33 and d=1.08, respectively). Increases in VO2peak and VO2 in the ACT were also recorded (d=0.46 and d=0.82, respectively). Further, there was also a small effect on peak HR during a subsequent cross-country skiing 800 m time trial (TT) (d=0.22) in ACT. These small changes appear to translate slightly to performance, as the ACT group experienced a small attenuation in the 800 m TT (PAS: d=0.43, ACT: d=0.10). This recovery of PERF may be due toimproved muscle relaxation time (T2) resultant of increased plasma volume and muscle perfusion, combating the deleterious effects of acidification(21). Additionally, the increased blood flow circulates the BLa, which can be sequestered for fuel by distal skeletal muscle and the heart, including that needed for re-phosphorylation of ATP (16,21,29,32,33). Intracellular lactate efflux may also draw out protons, attenuating the muscle performance losses associated with acidification (29). However, one study has shown a decrease in post-recovery CMJ height, an ATP-PCr dominant activity (33). But, for repeated aerobic activity, it could be concluded that ACT keeps the athlete’s metabolic system at a heightened level, as slightly enhanced cardiovascular aspects were observedduring subsequent performance (16).Furthermore, it should be noted that this study did not examine PSY variables.

Suzuki et al. (32) assessed collegiate rugby players’ mood (POMS) following ACT, indicating increased levels of anger (ACT: d=0.92, PAS: d=-0.29), confusion (ACT: d=1.26, PAS: d=0.26), and fatigue (ACT: d=1.31, PAS: d=0.85) immediately post-recovery.However, ACT did experience reduced tension (d=-1.04) compared to PAS (d=-0.54) 48 hrs post recovery. Numerous BIO markers were also examined, including CK and LDH. CK and LDH were elevated to moderate-very large levels in the ACT group immediately post recovery, 24 hrs, and 48 hrs post recovery compared to PAS. Increases in serum CK levels are largely due to eccentric muscle damage, collisions, and/or AMPK activity, all of which are present in rugby match-play (2). CK plays a major role in energy maintenance, as it is critical in PCr re-synthesis (2). Thus, the increase in its efflux could produce decreased PERF, as this re-synthesis is hindered. LDH is also a key in energy production, allowing for BLa metabolism with an increase in this study noted as evidence of heavy anaerobic reliance (5).Thus, its increase may spare glucose during recovery, allowing for enhanced subsequent performance compared to PAS. This is supported by a decrease in BLa reported by Losnegard et al. (16), yet contradicted by Tourginha et al. (34) who reported no difference between PAS and ACT in BLa in minutes following a specific judo task. Moreover, the efflux of these enzymes into the blood may inhibit these energy production activities within the muscle cell. However, Suzuki et al.(32) does not provide any evidence of translation to PERF recovery, as no variables were assessed. Of the various other variables examined, neutrophil phagocytic activity is of particular interest, as its rate is greatly reduced immediately post recovery, 24 hours, and 48 hours post recovery in PAS (d=-1.33, d=-2.83, and d=-3.67) compared to ACT (d=0.38, d=-0.29, d=0).This could indicate enhanced immune function via ACT. It should be noted that in the Wyatt et al. (38) meta-analytic review of the overtraining syndrome, immune system function was depressed in athletes that were over-trained (i.e., under-recovered).

Only one study included in this meta-analysis measured a variable from all three categories (33). However, insufficient raw data was provided for urinary catecholamines, although authors reported no significant difference. Here, two modes of ACT were implemented (aquatic and dry land) for collegiate futsal players following match play. In PERF measures (CMJ height, bouncing jump height, and 10 m sprint time), only trivial to small effects were indicated. The most notable effect was seen in CMJ height post recovery for the ACTw (d=-0.57) versus PAS (d=-0.21) and ACTd (d=-0.30), indicating a small detrimental effect of aquatic ACT on PERF. Despite this, Tessitore et al. (33) did find large effect (d=1.75) of ACT on dry land and a very large effect (d=3.00) of aquatic ACT on participants’ perceived benefit, compared to PAS.

Ballmann et al. (3) examined redox status, specifically, genetic transcription factor activity during post-exercise recovery. This group found that nuclear factor (euthyroid-derived2)-like factor (NFE2L2) activity was blunted during ACT in a hypoxic environment. This is a genetic transcription factor that is heavily involved in cellular protective mechanisms (3). While the effect of altitude is beyond the scope of this study, this may provide direction for future investigations not only for ACT, but other recovery methods.