Investigation of the HotDog,polymer resistive patient warming device

Active patient warming devices are often used during surgery, with two common classes: forced-air and resistive heating. ¹ The HotDog patient warmer (models B104 and U102; Augustine Biomedical & Design LLC, Eden Prairie, Minnesota, USA) falls into the latter category, using an electric current to heat a resistive polymer blanket. We were concerned by two reports submitted to the Database of Adverse Event Notifications (medical devices), ² describing patient burns associated with use of the HotDog warmer. This study investigated the accuracy of the HotDog temperature control when used according to manufacturer’s instructions in a non-faulty device. Our secondary aim was to investigate the temperatures and gradients across the HotDog blanket if the device were inadvertently used outside of the manufacturer’s instructions.

We utilised two mannequins for testing, a 23-kg paediatric Lifecast and a 58-kg adult SimManALS. We had access to two HotDog patient warmers (A and B), both in routine clinical use, and initially ran duplicate trials with each model to avoid the potential impact of a single device error. Vernier surface temperature sensors (Go Direct Surface Temperature [order code GDX-ST], Vernier, Beaverton, Oregon, USA), accurate to ±0.5°C, were placed on the ‘bony’ prominences of each mannequin to allow for continuous temperature monitoring throughout each trial. The mannequins were placed between the two components of the HotDog warmer, the top blanket and bottom mattress, and positioned supine in accordance with the device instruction manual. ³ The mattress and top blanket have independent temperature controls, and both were set to 39°C to allow for one hour of incubation prior to each test. The top blanket was then removed, and the mannequin was covered with surgical drapes to simulate typical operating conditions. Data collection, by surface temperature sensors and infrared images, was limited to the bottom mattress alone. The first tests were run with adult and paediatric mannequins in the supine position with the warmer used exactly according to the device instructions. The second tests, again supine, were done with the mannequins 5 cm off the sensor. It was evident that HotDog A attained temperatures outside the set temperature using the paediatric mannequin during these supine tests. We were concerned HotDog A was potentially faulty and ceased testing with it at this point. This device was returned to the hospital’s operating facility. The hospital elected to withdraw all HotDog devices, including this unit, pending a review. Both the Australian supplier of the HotDog patient warmer and the therapeutic goods administration were notified.

For testing in the lateral position, we placed the paediatric mannequin deliberately off the mattress temperature sensor, again only by 5 cm. Unfortunately, we did not have access to an adult mannequin that could be placed in a clinically realistic lateral position, and elected to test the paediatric model only. A thermal image was captured of the HotDog blanket after each trial using the FLIR ONE Pro Thermal Camera (model FLIR ONE Pro—iOS; Teledyne FLIR LLC, Wilsonville, Oregon, USA), to provide a visual representation of surface temperature gradients (see online supplementary material).

In three of the seven tests run, temperatures at least 2°C greater than the device setting of 39°C were recorded by the surface temperature sensors. All three of these cases occurred during the paediatric tests, with a maximum temperature of 41.4°C in 225 minutes recorded on HotDog A with the mannequin in the supine position. At the time this test was ceased the temperature was still rising, albeit with our potentially faulty device. While operating 2.4°C hotter than the set temperature, this is below a potential burn temperature (43°C) and remains clinically acceptable. ⁴ In the paediatric supine trials run with HotDog B the temperature plateaued at a very safe 36.8°C. However, when the supine paediatric mannequin was deliberately placed off the sensor, simulating a simple human factors error, the highest temperature recorded was 44°C, achieved in 135 minutes (Figure 1). This exceeds safe operating temperatures. When the same mannequin was placed laterally a temperature of 44.4°C was achieved in 120 minutes. Pleasingly, the adult models behaved consistently with device set temperatures, reaching a steady state between 75 and 120 minutes, with a maximum temperature of 38.9°C regardless of which HotDog device, A or B, was used.

Figure 1.

Temperature as recorded by the Vernier surface temperature sensors (accurate to ±0.5°C) placed on the bony prominences of a paediatric Lifecast mannequin placed deliberately off sensor on a HotDog patient warming device. Temperature was continuously monitored and recorded in 15-minute intervals; the trial was ceased after 135 minutes.

The highest temperature captured by the infrared camera was 36.3°C and the lowest 24.5°C. While the accuracy of the FLIR infrared camera for absolute temperature is poor, the infrared images do demonstrate well the temperature gradients across the mattress in patient contact areas. Although the temperatures were not outside a clinically safe range, the company’s advertising literature claims temperature gradients do not occur. ⁵ This can be seen in Figure 2 which shows the infrared images captured of a HotDog mattress that had been set at 39°C and for 60-minutes versus a HotDog that had just completed a 120-minute trial in a paediatric model.

Figure 2.

(a) HotDog patient warmer with temperature set at 39°C and incubated for 60 minutes and (b) HotDog patient warmer with temperature set at 39°C following trial with paediatric Lifecast mannequin placed off sensor for 135 minutes.

Our simple tests suggest that the HotDog device can achieve potential patient burn temperatures in a paediatric model, particularly if simple errors in setup occur. Even a 5 cm misalignment of the patient and sensor can result in a device temperature outside safety limits. The mattress has no associated warning on the device regarding the consequences of incorrect positioning and the manufacturer directs the sensor is to be covered with a light sheet at all times. This makes the sensor difficult to visualise, particularly under surgical drapes, and vulnerable to human error, leading to misalignment of the patient and sensor. Clearly, contact between the patient and the mattress temperature sensor is a critical safety mechanism by which the HotDog regulates temperature.

We excluded HotDog A as it operated outside parameters on initial tests; however, it is possible and even plausible that it was not a faulty device. The HotDog instruction manual indicates that if the temperature at the sensor exceeds one degree above the set point, the control will alarm and automatically power off the device. The HotDog never alarmed and powered off during any of our trials. Following our testing, we cannot recommend any method by which to ascertain if a device, outside of the sensor point itself, is faulty or exceeding safe temperature ranges during routine clinical use. We can say that the HotDog B performed safely with supine mannequins, with no recordable faults during our testing. However, it reached unsafe temperatures with a paediatric mannequin placed laterally and with a deliberate 5 cm misplacement error. Pleasingly, unsafe temperatures did not occur in the adult model, even when placed deliberately off the sensor. Many more HotDogs would need to be tested to explore the temperature range delivered by these devices. We acknowledge that a mannequin is not an ideal test bed for thermodynamic studies and patients add a layer of complexity that we are yet to test. We have nevertheless shown this device can operate outside its set temperature and, in some conditions, drift into a potential ‘burn to patient’ range.

Given the above we would recommend caution using the HotDog device, especially in paediatrics. Good contact between the mattress temperature sensor and the patient must always be maintained. Even small patient movements (e.g. 5 cm) would mandate reaffirmation of correct patient/device sensor contact. We are reporting this initial testing early, with an abundance of caution, hopefully to prevent a paediatric burn.

Supplemental Material

sj-pdf-1-aic-10.1177_0310057X231181405 - Supplemental material for Investigation of the HotDog, polymer resistive patient warming device

Supplemental material, sj-pdf-1-aic-10.1177_0310057X231181405 for Investigation of the HotDog, polymer resistive patient warming device by Teneal E Baxter in Anaesthesia and Intensive Care