Efficacy of Various Cooling Techniques During Exercise in Persons With Spinal Cord Injury: A Pilot Crossover Intervention Study

Two persons with complete (AIS A) SCI were recruited and enrolled. Levels of injury were classified into tetraplegia (TP) and paraplegia (PP). One age- and sex-matched able-bodied (AB) male was also recruited. None of these persons regularly performed scheduled exercise activities in thermoneutral or warm environments so heat acclimation (which requires 7–14 consecutive days of exercise) was not likely to be a confounder. Each participant was asked to fast for 3 hours prior to exercise and not exercise the day before or of the study. None of the participants were taking anticholinergics, which may have the potential to alter sweating.

At least 3 hours prior to the study, each subject ingested a US Food and Drug Administration (FDA)–approved thermometer pill (CorTemp) to measure Tc at 5-minute intervals throughout the study. On arrival to the lab, participants were asked to empty their bladder to prevent any interruption to the protocol and limit the occurrence of autonomic dysreflexia in the TP subject who would be at risk. The only clothing worn was shorts to allow for the maximum amount of ambient heat input through the skin while at the same time enabling optimal heat dissipation through evaporative or conductive heat loss. After entering an air-tight sealed room heated to 90°F via space heaters with continuous temperature monitoring via sling psychrometer, participants rested for 10 minutes to acclimate to the ambient temperature. Participants then began to exercise on a wheelchair treadmill for 30 minutes at a speed of 30 revolutions per minute counted by a metronome. This protocol was conducted under five different conditions, which were administered in a random order on five different days: (1) cool (C) = 75°F without cooling intervention; (2) hot (H) = 90°F without cooling intervention; (3) 90°F wearing a cooling vest (CV); (4) 90°F applying a water spray (WS) every 5 minutes over face, neck, arms and legs; and (5) 90°F drinking an ice slurry (IS) every 5 minutes. Each exercise session was conducted at the same time of day to control for circadian variation in Tc with 7 days between sessions to avoid heat acclimation. By controlling the ambient environment, randomizing order of sessions, and using a repeated measures design in persons with motor complete (AIS A) injury, we aimed to objectively compare the efficacy of each intervention in a rigorously designed protocol in persons with distinctly varied degrees of intact central sympathetic control, that is, connection from the hypothalamic thermoregulatory center to the efferent sympathetic chain.

This study was approved by the local institutional review board. All subjects provided informed consent to participate.

Change in Tc (ΔT_C) from 0 to 30 minutes was 0.56, 0.28, and 0.27°F in the TP, PP, and AB participant, respectively. As expected, the TP participant demonstrated twice the ΔT_C as compared to the PP and AB participants. Comparison of the results demonstrating ΔT_C for H, CV, WS, and IS interventions are presented in Figure 1.

ΔT_C of the H condition of all participants is depicted in Figure 1a. A progressive rise in T_C was noted in the TP participant (ΔT_C of 1.74°F), whereas there was no appreciable change in Tc in the PP participant (ΔT_C of 0.32°F) and a small lowering of Tc in the AB participant (ΔT_C of −0.6°F). No sweat responses were observed over the entire body surface area of TP, whereas sweating over the abdomen and legs was seen in PP.

ΔT_C in the CV condition is depicted in Figure 1b. Interestingly, T_C increased over the course of study in both the TP (ΔT_C of 0.95°F) and PP (ΔT_C of 0.68°F) participants, whereas Tc was initially increased and then returned to baseline levels in the AB participant.

ΔT_C in the IS intervention is depicted in Figure 1c. The rise in T_C appeared to be attenuated under the condition of IS in the TP participant compared to the H condition (ΔT_C of 0.5°F vs 1.74°F, respectively), whereas Tc continuously fell during the IS condition in the AB participant and was fairly well maintained throughout the IS intervention in the PP participant.

ΔT_C in the WS cooling intervention is depicted in Figure 1d. Unlike all the other conditions where ΔT_C was highest in persons with SCI, in this condition, ΔT_C was highest in the AB at 0.45°F. Meanwhile, ΔT_C of TP and PP remained essentially unchanged at 0.03°F and −0.09°F, respectively.

ΔT_C over all conditions by participants is presented in Figure 2. As anticipated, the participant with TP had the greatest degree of thermoregulatory dysfunction in the H condition. All three cooling interventions mitigated the rise in T_C in the TP subject, whereas only WS and IS mitigated T_C rise in the PP subject. Meanwhile, none of the cooling interventions provided an additive core temperature cooling effect in AB.

Figure 3 depicts that the slope of the WS condition was the least of all conditions, indicating it be most efficacious at attenuating ΔT_C in TP.

This is the first study to compare efficacy of more than two cooling interventions during exercise under heat stress in a controlled setting in persons with complete (AIS A) cervical and thoracic injury compared to an intact individual.

Not surprisingly, the person with TP had the largest rise in T_C during exercise under thermoneutral (75°F) and heat stress (90°F) conditions without any cooling intervention. The degree of T_C rise from baseline was highest in TP followed by PP and AB, respectively, which confirms previous findings that the balance of heat gain and heat loss via sweating and vasodilation is proportional to the neurological level of SCI.5,23,27 

Heat energy is constantly transferred from hotter sources to cooler ones (ie, down a thermal gradient). Conductive cooling involves the transfer of heat by direct contact with the skin. Physiologically, when wearing a topical garment that is cooler than the skin temperature, conductive cooling will occur. This response may be enhanced by the increased skin blood flow as occurs with dynamic exercise, thus cooling the peripheral blood and attenuating T_C increases. In persons without SCI, ice vests (<32°F) are comparably more efficacious than cooling vests (50–68°F).14 

In our study, CV was overall the least effective cooling intervention among the three subjects tested (Figure 3). Most notably, the ΔT_C of PP and AB was higher with CV than in the H condition, suggesting the CV worsened heat dissipation. We hypothesize that (1) CV may attenuate or fully prevent evaporative cooling over skin surface area it covers; and/or (2) local cutaneous vasoconstriction of skin surface areas in direct contact with the CV may reduce delivery of heat to the body surface and hence reduce whole body cooling via conduction. CV only mitigated the rise in T_C in the TP participant. For this reason, future studies documenting the effectiveness of cooling devices in the SCI population should stratify by lesion level (ie, TP vs PP).

Two studies have investigated the efficacy of the cooling vests (50–68°F) in persons with SCI. A randomized crossover study of 10 men with motor complete (AIS A and B) paraplegia (levels T4-T12) exercising in ambient temperatures with CV demonstrated that the vest reduced skin temperature and perceived thermal sensation but did not affect T_C.29 A subsequent experiment of exercise in a normothermic field-based setting (ie, wheelchair rugby and basketball) further showed no significant decrease in T_C with use of a cooling vest in a mixed group of persons with tetraplegia and paraplegia of varying levels of completeness.7 

In contrast, ice vests (<32°F) in a group of individuals with tetraplegia with varying degrees of completeness of injury have been shown to reduce T_C when applied during or before intermittent sprint exercise in the heat.30,31 

Internal heat loss occurs when a cold (<50°F) fluid is ingested and equilibrates toward T_C. Heat loss is not magnified with ingestion of IS during heat stress in persons without SCI, as it transiently decreases evaporative and convective cooling through sweating and vasodilation, respectively.32 However, in persons who have physiological disruptions in sweating, such as those with SCI, it may be hypothesized that the cooling effects of the IS are “disproportionally greater.”32 A recent follow-up study reported that cold water ingestion improves heat tolerance in a persons with multiple sclerosis (MS) whose vasomotor and sudomotor activity is impaired from neurological disruption of the thermoregulatory pathways.33 

In the present study, IS provided an additive cooling benefit compared to H condition in the persons with SCI (TP and PP), but not the AB, which supports the aforementioned phenomena and parallels the MS study.

With evaporative cooling, the heat required to effect a phase change from liquid to vapor is lost to the environment, thus sweat evaporation is responsible for removing heat from the skin surface. Evaporative cooling is the main method of heat dissipation during exercise in high ambient heat in persons without SCI.25 In persons with SCI, previous studies have demonstrated that persons with lesions above T6-T8, with intact sweat responses on more than 50% of total body skin surface area, are able to effectively maintain T_C during heat stress nearly as effectively as persons without SCI.26,27 This is consistent with literature from persons without SCI that defines a minimum cutoff of 40% skin surface area of body cooling that is required to prevent rise on T_C during heat stress.34 Furthermore, whole body sweat rates in some individuals with paraplegia are similar to matched persons without SCI, which may also explain this ability to maintain T_C during high heat strain conditions.35,36 We hypothesize that, if WS is applied to skin surface areas that do not sweat, WS would cause evaporative cooling and attenuate the rise in T_C during exercise or high ambient heat exposure. As such, providing artificial sweat in the form of WS to persons with high paraplegia (above T6) may be a promising intervention.

In our study, WS applied to all exposed skin surface areas (ie, face, chest, abdomen, arms, and legs) was most effective at maintaining T_C during exercise in the heat in both participants with SCI, but the most profound benefit was in the TP participant who demonstrated no sweating responses over any skin surface area. Meanwhile, WS caused a rise in Tc in the AB participant, who had innate sweating capacity and whose Tc dropped most in the intervention without cooling (ie, H). This suggests no additional Tc cooling benefit of WS applied to skin surface areas with intact endogenous sweating responses, which, of note, has been previously suggested by other human and animal studies.17,37,38 

Only one study to date has tested the efficacy of WS in a mixed group (complete and incomplete injuries) of individuals with paraplegia (T3-T12) under heat stress.39 WS was “ineffective in attenuating the onset of heat strain during high-intensity arm exercise in a comfortable environment.”39 It should be noted that WS was applied to the back of head, neck, forearm, and face, which accounts for ~15% body surface area; this may have been too small an area, as mentioned earlier. Furthermore, only persons with paraplegia were studied. This group typically has intact sweating responses over these specific skin surface areas, and it is unknown if applying additional WS in regions with intact sweat responses would have an additive cooling effect, as previously mentioned. Moreover, outcome measures were not stratified by level of paraplegia such as high paraplegic (at/above T6) versus low paraplegic (below T6) to evaluate for differences between these groups with varying thermoregulatory capacities; this might have revealed a main effect of level of injury. Furthermore, as our data and past data suggest, persons with paraplegia do not struggle with progressive increases in T_C during exercise, so it may be harder to detect a significant cooling effect, regardless of the intervention, in this population.

Griggs et al combined precooling with an ice vest followed by WS during a 60-minute intermittent sprint exercise protocol and found the combination of precooling with an ice vest and WS significantly attenuated the rise in T_C in persons with tetraplegia. When compared to precooling with the ice vest without the WS during exercise, the attenuation in rise of T_C was less.40 Thus, WS as an efficient cooling technique for persons with tetraplegia appears to hold promise.

Anecdotally, in SCI athletic events, WS is observed to be the most frequently used external cooling intervention; possibly because it reproduces the evaporative cooling of endogenous sweat that is impaired in persons with levels of injury above T6.

To my knowledge, there are no reported studies examining effects of WS alone on T_C and exercise capacity in persons with TP, who have the greatest thermoregulatory dysfunction. This promising approach should be tested.

Intensity was not strictly controlled, however all participants propelled the wheelchair at a rate of 30 revolutions per second. Although it is intuitive that higher energy expenditure/intensity would likely increase heat generation, two recent studies have shown the level of injury to play a larger role in rising T_C during exercise. Griggs et al in 2015 concluded, despite similar external work (in TP vs PP), a marked increase in T_C in TP during exercise and recovery signifies that “thermoregulatory differences between the groups were predominantly due to differences in heat loss.”24,^(p469) In 2016 Griggs did a study to directly test whether a player's physical impairment or activity profile determined the amount of thermal train experienced during wheelchair rugby match play.41 Players with tetraplegia were under greater thermal strain during wheelchair rugby match play (compared AB persons) “as a result of their reduced heat loss capacity, due to their physical impairment and not because of their activity profile.”41 ^(p177) Despite these data, uncontrolled exercise intensity could still be considered a limitation in this study and should be noted.

This study is the first to compare three cooling interventions during exercise in a controlled heated setting among three persons with varying degrees of neurological control of sudomotor and vasomotor activity. WS attenuated the rise in T_C during heat stress most effectively when compared to the CV and IS interventions in both persons with SCI. These findings provide a rationale for a larger study in persons with SCI under exercise-induced heat stress using WS, being sure to cover skin surface areas that do not sweat. Given that those with complete (AIS A) TP demonstrate most thermal dysregulation, this population should be studied first to be able to detect an effect.