Convection-based vs conduction-based cooling in rehabilitation: A cadaveric porcine study of surface and intra-articular temperature reduction

Abstract

BACKGROUND

Postoperative inflammation after joint injury or surgery contributes to pain, edema, and delayed functional recovery. Cryotherapy is commonly used to alleviate pain and modulate inflammatory responses; however, the effectiveness of various cooling modalities in reducing temperatures in deeper periarticular and intra-articular tissues remains uncertain.

AIM

To compare the cooling performance of a convection-based cryotherapy system with two circulating-pad cooling systems in reducing temperature at various tissue depths.

METHODS

A cadaveric porcine knee model was used to assess comparative thermal transfer in periarticular tissues and the joint space. Thermocouples were placed at three sites: Subdermal, intramuscular, and intra-articular. Four experimental cooling conditions were evaluated by using: SootheAway 1 (SA1), SootheAway 2 (SA2), ThermoCuff (convection-based), and ambient control. Each cooling session lasted 60 minutes using a target coolant/air temperature of 35 °F (1.6 °C). The prespecified primary endpoint was the change in temperature at 20 minutes (ΔT20 = T20 - T0).

RESULTS

ThermoCuff produced a larger observed early temperature reduction through the tested compartments. At the subdermal level, ΔT20 was -14.8 °C with ThermoCuff compared with -8.4 °C for SA2, -3.9 °C for SA1, and -1.8 °C for ambient exposure. Intramuscular cooling was modest with ThermoCuff (-1.8 °C) and SA2 (-1.3 °C), whereas SA1 showed warming (+6.5 °C). At the intra-articular level, ThermoCuff achieved the highest temperature reduction (-10.0 °C), compared with -3.9 °C for SA2 and -3.3 °C for ambient exposure; SA1 showed warming (+4.8 °C).

CONCLUSION

In the cadaveric porcine knee model, the convection-based system achieved better early cooling than the circulating-pad systems, especially at the subdermal and intra-articular levels. These preliminary preclinical findings support the hypothesis that cooling modality might affect the depth and magnitude of thermal transfer under controlled laboratory conditions and require in vivo confirmation before clinical translation.

Key Words: Cryotherapy; Convection-based cooling; Conductive cooling; Intra-articular temperature; Cadaveric porcine model; Rehabilitation

Core Tip: Thermal transfer from various cryotherapy systems was evaluated using a cadaveric porcine knee model. Subdermal, intramuscular, and intra-articular temperatures were measured with thermocouples during cooling using ThermoCuff, two SootheAway circulating-pad systems, and an ambient control. The primary endpoint was the change in temperature at 20 minutes. ThermoCuff was relatively better in achieving early cooling, particularly in subdermal and intra-articular sites, with intra-articular temperature decreasing by about 10.0 °C. Intramuscular cooling was modest across all systems. These preliminary results indicate that convection-based cryotherapy may achieve apparently higher deep-tissue cooling than circulating-pad systems; however, in vivo validation is required before clinical application.

INTRODUCTION

Postoperative inflammation following joint injury or surgical intervention plays a critical role in tissue healing, but can also lead to secondary complications if prolonged[1]. Furthermore, it contributes to increased vascular permeability and fluid extravasation, resulting in edema and joint effusion[2]. The accumulation of intra-articular fluid elevates intra-compartmental pressure, stimulates nociceptors, and exacerbates pain, potentially delaying early mobility[3,4].

Moreover, pain and swelling trigger reflexive arthrogenic muscle inhibition, particularly in periarticular muscles; a clear example is quadriceps inhibition after anterior cruciate ligament reconstruction or total knee arthroplasty, which delays functional recovery and impairs joint stability[5,6]. That is why effective modulation of postoperative inflammation is a cornerstone of rehabilitation, aiming to reduce edema and pain while facilitating early muscle activation and mobility restoration[7].

Cryotherapy refers to the controlled application of cold to reduce and modulate inflammatory responses that arise from both physiological stressors, such as exercise, and pathological conditions, including postoperative or postinjury inflammation[8]. It elicits physiological responses that actively oppose inflammation, which justifies the alternative term cryostimulation and explains its widespread use across a broad spectrum of clinical and athletic settings - from enhancing recovery after physical exertion and providing analgesia following injury to supporting muscle regeneration and alleviating symptoms of inflammatory or autoimmune disorders[9-11].

Cooling transmission during cryotherapy occurs through different mechanisms, primarily via conduction or convection[12]. Conduction cooling transfers heat through direct contact between a cold surface and the skin. In contrast, convection cooling removes heat by circulating a cooled medium such as air or a liquid around the target area, enabling continuous heat exchange without direct tissue contact[12].

Although both conduction and convection methods aim to reduce inflammation and pain, conduction-based approaches (such as ice packs, cold gels, and compression ice) are considered limited regarding achieving sufficient intra-articular temperature reduction due to limited tissue penetration[13-15]. They are also associated with shallow cooling depth, a risk of frostbite or skin irritation, and a short duration of therapeutic effect[16,17]. These limitations underscore the need for more efficient alternatives, such as convection-based cooling systems, which are suggested to provide deeper, faster, and more sustained cooling within joint structures; however, convection-based cold devices for local use in humans are scarcely described (in contrast to whole-body and Whirlpool cryotherapy)[12,18,19].

Therefore, we aimed to evaluate the effectiveness of a convection-based cryotherapy modality for intra-articular cooling with a traditional conduction-based method. The evaluation employed a porcine cadaveric knee model with temperature monitoring at subdermal, intramuscular, and intra-articular sites. The explicit hypothesis was that convection-based cooling using ThermoCuff would reduce intra-articular (synovial) temperature more rapidly and to a greater extent than traditional conduction-based cooling methods.

MATERIALS AND METHODS

Experimental model and setup

A fresh cadaveric porcine hind limb was used as a surrogate model to evaluate comparative thermal transfer behavior in periarticular soft tissues (skin and periarticular muscles) and the knee joint space. A pig leg specimen was obtained from a freshly slaughtered, healthy animal following veterinary inspection at Assiut slaughterhouse, Egypt. The specimen was purchased from the owner, preserved in ice, and transported directly to the laboratory for experimentation.

Cooling was applied using either the SootheAway circulating-pad cooling system (Figure 1A) or the ThermoCuff forced-air convection cooling system (Figure 1B). For the ThermoCuff application, the cuff was wrapped around the specimen with the inlet and outlet ports kept unobstructed. The cuff position was maintained using silicone bands, taking care not to compress or occlude airflow channels. Furthermore, before each run, airflow through the inlet and outlet ports was inspected to confirm continuous forced-air movement through the cuff.

Figure 1 Experimental setup and cooling systems used in the study. A: Cooling was applied using the SootheAway circulating-pad cooling system; B: Using the ThermoCuff forced-air convection cooling system, which consisted of a full-wrap cuff (12 inches × 16.5 inches) with dual 2-inch inlet and outlet ports, secured with silicone bands during testing; C: After each test, the specimen was re-warmed in a hot water bath to re-establish the target physiological starting temperature before the next trial. ChatGPT was used to enhance the quality of the figures, without distortion or manipulation.

The thermocouple function was verified prior to data collection using two reference conditions: Ice-water and warm-water within the expected experimental temperature range. The probes were allowed to stabilize before recording, and readings were checked for consistency with the reference conditions. Although formal laboratory calibration against a certified reference thermometer was not performed, probes exhibiting unstable or implausible values were repositioned or excluded prior to recordings.

For standardization of probe placement, we used predefined anatomical locations where we checked orientation and measured insertion depths for each compartment (subdermal, intramuscular, and intra-articular) using external tape measurements, and the same probe trajectories were maintained as possible across conditions (we admit that minor variation in probe position between repeated runs cannot be excluded).

Before each experiment, the specimen was re-warmed in a thermostatically controlled warm-water bath until the target starting temperature range was reached. After removal from the bath, the specimen was dried, remounted, and allowed to equilibrate until thermocouple readings stabilized (defined as the absence of meaningful drift in the recorded temperature over consecutive readings before initiation of the cooling session), then mounted vertically on a stand for testing (Figure 1C). It is worth noting that, owing to the non-perfusion nature of the cadaveric porcine model, complete thermal homogenization between the superficial, intramuscular, and intra-articular compartments could not be guaranteed.

Cooling conditions

Four experimental conditions were evaluated under the same setup.

SootheAway 1: Circulating pad cooling system: The pad was fully wrapped around the knee joint and secured with elastic straps to ensure uniform contact pressure. Coolant circulation was maintained at a steady flow rate of 9.6 cubic feet per minute.

SootheAway 2: Circulating-pad cooling system (updated model): An enhanced circulation mechanism and modified pad design intended to improve surface contact and coolant flow [the same flow was used as in SootheAway 1 (SA1)].

ThermoCuff: Forced-air/ventilation-based cooling system (convection-based): A full wrap cuff (12” × 16.5”) with dual 2-inch inlet and outlet ports was secured using silicone bands.

Ambient: No active cooling (control).

For active cooling conditions, the target coolant/air temperature was 35 °F (1.6 °C). Each cooling session lasted 60 minutes, with continuous temperature recording. To reflect common cryotherapy application durations in rehabilitation practice, the prespecified primary endpoint was the temperature change at 20 minutes.

Each condition was repeated in three runs to assess within-model repeatability. Because repeated runs were performed in a cadaveric model (single specimen), these runs were considered technical replicates rather than independent biological replicates.

Temperature was monitored continuously using thermocouples (stainless-steel needle thermocouple probes) placed at three predefined sites: Subdermal (immediately beneath the skin), intramuscular (within periarticular muscle tissue), and intra-articular (retropatellar/behind the patella) (Figure 2). Each probe consisted of a rigid metallic needle shaft connected to an external temperature-reading/data-logging system, featuring a measurement range exceeding the experimental temperatures and a display resolution of 0.1 °C. Temperature was continuously recorded during each cooling session and documented in both Celsius and Fahrenheit. Probe locations were kept consistent across experimental conditions to ensure comparability between runs. Prior to data acquisition, thermocouples were functionally verified using reference temperature points. Probe readings were checked in both ice-water and warmed water baths before experimental recording. Readings were allowed to stabilize before data collection, and probes displaying unstable or implausible values were repositioned or excluded. This verification procedure ensured that the thermocouples operated within the expected temperature range for the study.

Figure 2 Placement of temperature probes for skin, intramuscular, and intra-articular measurements. A: Skin temperature was measured using a probe inserted sub-dermally at a depth of approximately 1.75 inches, positioned parallel to the skin surface; B: Muscle temperature was recorded using an intramuscular probe inserted approximately 3.5 inches into the quadriceps muscle; C: Intra-articular knee temperature was measured by inserting a probe 125 inches into the joint space behind the patella. Tape measures were used to standardize probe insertion depth and ensure consistent placement across measurements. These probe positions enabled simultaneous evaluation of thermal responses in the superficial, intramuscular, and intra-articular regions during the experiment. ChatGPT was used to enhance the quality of the figures, without distortion or manipulation.

Data handling and endpoint definition

The primary endpoint was the change in temperature at 20 minutes (ΔT20) at each tissue depth, calculated as ΔT20 = T (20 minutes) - T0 (baseline). With this convention, negative values indicate cooling and positive values indicate warming.

Statistical analysis

Given the exploratory design, cadaveric single-specimen model, and technical-replicate structure, analyses were primarily descriptive. For each cooling condition and tissue depth, baseline temperature (T0), 20-minute temperature [T (20 minutes)], and ΔT20 were summarized. Formal inferential testing was not emphasized because repeated runs in a single specimen do not provide independent biological replication.

RESULTS

Temperature was continuously recorded for 60 minutes under all conditions; however, the 20-minute time point served as the primary comparative endpoint, representing a typical cryotherapy session duration in rehabilitation settings.

The magnitude of the early temperature change varies across measured compartments, depending on the cooling modality and tissue depth. Under the tested conditions, the convection-based ThermoCuff induced larger observed early reductions in temperature than the circulating-pad systems, especially at the subdermal and intra-articular levels (Figures 3 and 4).

Figure 3 Temperature response from baseline to 20 minutes by tissue depth during the experiment. Temperature responses from baseline to the prespecified 20-minute endpoint are presented for three monitored compartments. A: Subdermal tissue; B: Intramuscular tissue; C: The intra-articular space. Four experimental conditions were evaluated: SootheAway 1, SootheAway 2, ThermoCuff, and ambient control. Temperatures are reported in degrees Celsius (°C). SA1: SootheAway 1; SA2: SootheAway 2; TC: ThermoCuff; AMB: Ambient control.

Figure 4 Temperature change at the 20-minute primary endpoint stratified by tissue depth and cooling condition. The ΔT20, defined as T (20 minutes) - T0, for subdermal tissue, intramuscular tissue, and the intra-articular or retropatellar knee joint space. A: Subdermal tissue; B: Intramuscular tissue; C: The intra-articular or retropatellar knee joint space. Negative values represent cooling, and positive values represent warming. ThermoCuff produced the largest early reductions in subdermal and intra-articular temperatures under the tested cadaveric conditions. In contrast, intramuscular temperature changes were less pronounced across all conditions. SA1: SootheAway 1; SA2: SootheAway 2; TC: ThermoCuff; AMB: Ambient control.

Subdermal (skin-level) temperature response

At the subdermal site, ThermoCuff produced a relatively higher temperature reduction at 20 minutes, followed by SA2 and SA1, whereas the ambient condition exhibited only a minor decrease (Table 1). Specifically, ΔT20 values were -14.8 °C for ThermoCuff, -8.4 °C for SA2, -3.9 °C for SA1, and -1.8 °C for ambient exposure. These findings indicate that the convection-based system has the potential to achieve greater early superficial cooling under the tested preclinical conditions.

Table 1 Subdermal temperature response at baseline and 20 minutes.

Condition	T0 [°C (°F)]	T20 [°C (°F)]	ΔT20 [°C (°F)]
SootheAway 1	42.5 (108.5 °F)	38.6 (101.5 °F)	-3.9 (-7.0 °F)
SootheAway 2	35.6 (96.1 °F)	27.2 (81.0 °F)	-8.4 (-15.1 °F)
ThermoCuff	35.4 (95.7 °F)	20.6 (69.1 °F)	-14.8 (-26.6 °F)
Ambient	40.8 (105.4 °F)	39.0 (102.2 °F)	-1.8 (-3.2 °F)

ΔT20 = T (20 minutes) - T0; negative values indicate cooling.

Intramuscular temperature responses

At the intramuscular site, temperature changes were smaller than those observed subdermally, consistent with slower heat transfer to deeper tissues. ThermoCuff and SA2 produced modest intramuscular temperature reductions at 20 minutes (ΔT20 = -1.8 °C and -1.3 °C, respectively), whereas SA1 exhibited an increase in intramuscular temperature (ΔT20 = +6.5 °C) (Table 2). The ambient condition showed minimal change (ΔT20 = -0.4 °C). The intramuscular warming pattern observed with SA1 suggests variability in deeper tissue thermal responses under the tested conditions and warrants cautious interpretation (Table 2).

Table 2 Intramuscular temperature response at baseline and 20 minutes.

Condition	T0 [°C (°F)]	T20 [°C (°F)]	ΔT20 [°C (°F)]
SootheAway 1	24.6 (76.3 °F)	31.1 (88.0 °F)	+6.5 (+11.7 °F)
SootheAway 2	33.5 (92.3 °F)	32.2 (90.0 °F)	-1.3 (-2.3 °F)
ThermoCuff	29.6 (85.3 °F)	27.8 (82.0 °F)	-1.8 (-3.2 °F)
Ambient	31.4 (88.5 °F)	31.0 (87.8 °F)	-0.4 (-0.7 °F)

ΔT20 = T (20 minutes) - T0; negative values indicate cooling.

Intra-articular temperature response

At the intra-articular site, ThermoCuff produced a relatively greater temperature reduction at 20 minutes, with temperature decreasing from 37.2 °C to 27.2 °C (ΔT20 = -10.0 °C). SA2 showed a smaller but measurable cooling effect, decreasing from 37.2 °C to 33.3 °C (ΔT20 = -3.9 °C). The ambient condition also demonstrated a modest reduction in intra-articular temperature, from 33.9 °C to 30.6 °C (ΔT20 = -3.3 °C). By contrast, SA1 was associated with an increase in intra-articular temperature, rising from 28.5 °C at baseline to 33.3 °C at 20 minutes (ΔT20 = +4.8 °C), indicating warming rather than cooling (Table 3).

Table 3 Intra-articular temperature response at baseline and 20 minutes.

Condition	T0 [°C (°F)]	T20 [°C (°F)]	ΔT20 [°C (°F)]
SootheAway 1	28.5 (83.3 °F)	33.3 (91.9 °F)	+4.8 (+8.6 °F)
SootheAway 2	37.2 (99.0 °F)	33.3 (91.9 °F)	-3.9 (-7.0 °F)
ThermoCuff	37.2 (99.0 °F)	27.2 (81.0 °F)	-10.0 (-18.0 °F)
Ambient	33.9 (93.0 °F)	30.6 (87.1 °F)	-3.3 (-5.9 °F)

ΔT20 = T (20 minutes) - T0; negative values indicate cooling.

DISCUSSION

Using the prespecified 20-minute endpoint, the convection-based ThermoCuff demonstrated the largest observed descriptive early cooling effect at subdermal and intra-articular measurement sites in this cadaveric porcine model. Differences at the intramuscular level were smaller in magnitude, and one circulating-pad condition (SA1) showed warming rather than cooling, indicating variability in deeper-tissue thermal behavior under the tested conditions. Because this study used a cadaveric single-specimen model without physiological perfusion or thermoregulation, these findings should be interpreted cautiously as preclinical comparative thermal-transfer data rather than direct evidence of clinical effectiveness. Although the results we obtained are encouraging, we acknowledge that they are preliminary and should be interpreted with inherent limitations in mind.

First, we used a single fresh porcine hind limb as an alternate model for periarticular and intra-articular thermal conduction. Although this model facilitated controlled comparisons between cooling modalities, it does not fully replicate the biological conditions of living human tissue. The absence of active blood flow, tissue perfusion, metabolism, inflammatory reaction, and thermoregulatory mechanisms may affect both the rate and extent of cooling. In vivo, perfusion may dissipate or redistribute heat, potentially attenuating or modifying the temperature reductions observed in a cadaveric model. Second, considering that all repeated measurements were performed on a single cadaveric porcine specimen, the repeated runs should be interpreted as technical replicates rather than independent biological replicates. Therefore, the observed differences are presented descriptively and should not be interpreted as statistically powered evidence of superiority between cooling systems; however, the results should be considered with the understanding that what we found is preliminary and still needs confirmation in the clinical setting. Third, we designed this study as an exploratory comparative experiment, and formal inferential statistical analysis was not emphasized, which further reduced the power of the results. Fourth, despite efforts to re-warm the specimen before each run, baseline temperatures varied across conditions and tissue depths, potentially influencing the absolute magnitude of the temperature change. As a result, the primary endpoint was reported as the change from baseline at 20 minutes rather than as absolute temperature alone. Fifth, despite all precautions, baseline temperatures varied across compartments and conditions, which might be attributed to the different thermal behavior of each compartment, absence of active perfusion and thermoregulation mechanisms leading to differences in the rates at which each tissue reaches thermal equilibrium, and, owing to using the same specimens, residual thermal effects from preceding cooling sessions may have persisted in deeper tissues despite rewarming. Finally, anatomical and biological differences between the porcine and human knee limit the generalizability of these findings. Validation in larger studies using multiple specimens and, ultimately, in clinical in vivo settings is necessary.

The depth of cooling achieved during traditional cold packs is generally limited, as tissues located more than two cm below the surface experience minimal temperature reduction from superficial cold packs. This presents a challenge in effectively cooling intra-articular and peri-bone regions[20]. In the current preclinical study, we evaluated the cooling effect at three anatomical locations: Skin, intramuscular, and intra-articular. The main reason for this selection is that reductions in skin and deep tissue temperatures are not strongly correlated, making skin temperature an unreliable indicator of deep tissue cooling[21]. Temperature changes vary considerably across tissue layers, with the skin responding first and cooling most rapidly[22]. Because muscle lacks thermal receptors, the skin plays a crucial role in initiating thermoregulatory responses in clinical settings. During cooling, deep layers such as the muscle tissue lose heat primarily to the superficial muscle layers, while warm blood is redirected to deeper tissues, slowing their rate of cooling[21].

We set 20 minutes as the primary endpoint for evaluating temperature change. This is because there is typically a delay in the onset of cooling, as tissue temperature does not decrease immediately after cold application; therefore, short treatment durations of less than 15 minutes may not provide adequate cooling. Furthermore, a period of post-application cooling often occurs, during which tissues continue to lose heat for several minutes after the removal of the cold source. These factors collectively highlight the importance of optimizing cryotherapy duration and application techniques to achieve desired therapeutic effects[13].

Although the current study was preclinical and carried out on cadaveric porcine specimen, we believe that, based on the preliminary results we obtained, convection-based cooling systems might provide some of the following advantages (yet, these advantages are partially hypothetical and confirmation through robust clinical studies is still required).

First, a possible and potential clinical advantage of the convection-based cooling system might be its ability to achieve deeper intra-articular cooling than conventional conduction-based cryotherapy. Unlike traditional ice packs, which primarily affect superficial tissues, convection-based cooling actively transfers heat away from deeper joint structures through continuous airflow, allowing more efficient thermal exchange[1,2]. This could be explained by a greater intra-articular temperature reduction, corresponding to up to 160% greater cooling efficiency. Such deeper cooling is important in a clinical setting when using cryotherapy as a line of management, for example, in some joint conditions, because synovial inflammation, joint effusion, and ligamentous injuries originate within intra-articular structures, and superficial cooling alone may be insufficient to adequately modulate these processes[3,5].

Second, the relatively possible faster achievement of the cooling effect, where the target intra-articular cooling is achieved in approximately 20% of the time required by conduction-based devices[6]. The clinical implication of this issue is that rapid cooling may provide faster analgesia and earlier suppression of inflammatory mediators, particularly in the acute postoperative and post-injury phases. Faster temperature reduction has been associated with earlier pain relief and improved patient tolerance, allowing more efficient treatment sessions and potentially facilitating earlier initiation of rehabilitation interventions[7,9]. Furthermore, the enhanced cooling efficiency of convection-based systems may therefore provide relatively better clinical outcomes, including improved range of motion, decreased joint swelling, and improved functional recovery[12,21].

Third, in addition to its thermal effects, a convection-based cooling system provides controlled compression, which might enhance some of its therapeutic benefits. Compression levels ranging from mild (8-15 mmHg) to firm (30-40 mmHg) help improve venous return, reduce edema, and enhance lymphatic drainage. Combined cryotherapy and compression systems have previously been shown to provide superior outcomes compared with cryotherapy alone, particularly following orthopedic surgery such as ligament reconstruction and total knee arthroplasty[22]. The integration of convective cooling with compression may therefore provide additive therapeutic benefits.

Lastly, convection-based cooling systems are considered safer. Traditional ice packs can cause uneven cooling, excessive tissue temperature reduction, and frostbite injury if improperly applied[1]. In contrast, ThermoCuff provides controlled temperature regulation through regulated airflow, reducing the risk of cold-induced tissue injury while maintaining therapeutic effectiveness[23]. This improved safety profile makes it particularly suitable for postoperative and clinical settings requiring repeated use.

CONCLUSION

Keeping the study limitations and its preclinical nature in mind, it is apparent that, from this exploratory single-specimen cadaveric porcine knee model, convection-based cooling systems such as ThermoCuff consistently reduced temperatures most effectively and uniformly across the skin, muscle, and intra-articular regions, with a more pronounced decrease observed intra-particularly after 20 minutes. The potentially better cooling performance, combined with added compression, indicates that convection-based cryotherapy may accelerate recovery and improve functional outcomes in clinical settings. However, as this study used only a single cadaver and did not maintain physiological body temperature, these findings represent preliminary, preclinical comparisons. Further research involving multiple specimens and in vivo clinical models is necessary to investigate if the observed intra-articular temperature reductions translate into clinically meaningful improvements in pain, swelling, range of motion, or functional recovery.