Critique of Yang et al.'s study on radiant heat for postoperative hypothermia treatment

Hypothermia Treatment Using Radiant Heat

Hypothermia

Critique of Postoperative Hypothermia Treatment Using Radiant Heat by Yang et al. (2012)

Critique of Postoperative Hypothermia Treatment Using Radiant Heat by Yang et al. (2012)

Perioperative hypothermia has been shown to increase the risk of post-surgery complications, thereby threatening the health of patients and increasing the cost of care (reviewed by Hooper et al., 2009). Guidelines for active warming measures have been suggested by a number of experts, which attempt to address the health risk perioperative hypothermia represents. Should patients arrive in the post-anesthetic care unit (PACU) in a hypothermic state, the conventional approach is to cover patients with pre-warmed hospital blankets; however, forced air-warming has been shown to be more effective across patient populations.

Yang and colleagues (2012) recently published the results of a study examining the relative efficacy of warmed blankets and a radiant heat source for achieving a normothermal state in the PACU. This report will examine all aspects of this study in detail and provide a critical analysis, thereby determining whether their findings are sufficiently rigorous to complement the existing evidence base supporting the implementing of active warming measures in the PACU.

Substantive and Theoretical Dimensions

Hypothermia has long been considered a significant health risk to postoperative patients (Pikus and Hooper, 2010); so much so that the American Society of PeriAnesthesia Nurses (ASPAN) issued formal guidelines in 1998 that have been updated as research findings have become available (Hooper et al., 2009). This continues to be an active area of research and therefore investigating the relative effectiveness of warm blankets and radiant heaters is clinically relevant. Should the researchers discover a better method of warming patients postoperatively, then the benefits to surgical patient outcomes could be significant. This explains why Yang and colleagues (2012) tracked the cost of care and hospital length of stay for patients enrolled in their study.

The use of radiant heaters is one method of many suggested by the ASPN, but forced air warming is considered the gold standard for perioperative treatment of hypothermia (Hooper et al., 2009). Yang and colleagues (2012) mentioned forced air warming as a common method for achieving normothermia in the PACU, but viewed this option as problematic because of the contamination issues such devices create. In countries with a relatively underdeveloped health care system, forced air warming may not be a viable choice because the devices require so much attention. Although not discussed directly by the authors, radiant heat would be one method of warming that would help minimize contamination risks and require little maintenance. The authors of this study may have also felt unconstrained to try methods they have experience with, since this research took place in Taiwan. Radiant heat is therefore a clinically relevant method for patient warming because it reduces the contamination risk of forced air warming and is more conducive to a care environment lacking the resources common to Western socities.

The research methods employed in this study are straightforward and directly address the question of whether radiant heat is more effective than warmed blankets for hypothermic postoperative patients. The primary outcome measure was changes in patient tympanic temperature. Therefore, the independent variables were warmed hospital blankets and radiant heat, while the dependent variable was tympanic temperature. The independent and dependent variables are directly relevant to the question being asked and have important clinical significance.

In framing their question, Yang and colleagues (2012) reviewed the relevant literature and presented the findings from selected studies in the introduction. Their introduction helped frame the research question being investigated sufficiently for the reader to easily comprehend both the rationale for the study and why the study needed to be performed. However, it was far from comprehensive. A recent multi-study review of postoperative warming techniques (Pikus and Hooper, 2010) was not cited, nor was several relevant studies mentioned in this review. In fact, the studies cited by Yang and colleagues (Giuffre et al., 1991 and Weyland et al., 1994) were not cited by Pikus and Hooper, and vice versa (Brauer et al., 2004 and Villamaria et al., 1997). Pikus and Hooper limited their analysis to studies conducted since the release of the ASPAN guidelines in 1998 and therefore would not have included the articles cited by Yang and colleagues; however a good explanation for why Yang and colleagues did not cite the same articles that Pikus and Hooper did is lacking. The literature review by Yang and colleagues is therefore missing at least two relevant studies that should have been discussed.

Ethical Dimensions

The research study was approved by the relevant institutional review board for the hospital where the study took place (Yang et al., 2012). All patients were required to sign an informed consent form before inclusion in the study. Physicians and other medical care personnel responsible for the patients' care were not blinded to either their enrollment status or postoperative hypothermia treatment modality. How confidentiality was maintained beyond this point was not mentioned. In terms of patient vulnerability, this patient group would not be more vulnerable than other patients undergoing a surgical procedure under general anesthesia. Although the average age was between 61 and 65 years, this patient group was otherwise quite healthy.

Interpretive Dimensions

Yang and colleagues (2012) found that radiant heat warmed postoperative patients significantly (p < .001) faster (1.83° C/hr) than warmed cotton blankets (1.03° C/hr). This difference is quite large and represents a substantial improvement over warming rates found by other research groups (Yang et al., 2012; Pikus and Hooper, 2010). Yang and colleagues also divided the study groups into three temperature ranges, to better understand whether warming rates were a function of tympanic temperature upon admission into the PACU. What they found was that patients with lower body temperatures warmed at a slower rate (1.5° C/hr) than patients who were within 0.5° of the discharge temperature (36° C) upon admission to the PACU.

Most other research groups discovered warming rates well below those found by Yang et al. (2012), therefore it seems reasonable to question the use of tympanic temperature when the radiant heat source is placed next to the head. If there is a causal relationship between heat source placement and tympanic temperature, then this has important safety implications in terms of PACU discharge decisions. The use of tympanic temperature determined with an infrared probe is not recommended by ASPAN guidelines for perioperative warming (Hooper et al., 2009). Based on these guidelines, in situations when the core temperature cannot be determined without putting patient safety at risk, then oral temperature should be used to approximate core temperature. Since the patients in this study had back surgery, obtaining an oral temperature would have been feasible.

What Yang and colleagues (2012) claimed to have shown is that patients were warmed faster when subjected to radiant heat, compared to patients warmed with hospital blankets. This finding is clinically relevant because the need for active warming systems for perioperative patients has already been recognized by the ASPAN (Hooper et al., 2009), therefore much of the research being done in this area concerns the relative efficacy of the different active warming approaches. While the use of warmed blankets remains the conventional approach for treating hypothermic surgery patients, forced air warming remains the recommended approach. Yang and colleagues (2012) dismiss forced air warming as too much of a contamination risk and therefore did not include it in their study. In essence, they used the conventional warmed blankets as the control for assessing whether radiant heat was more effective. To ask this question in 2012, when active warming measures have already been shown to be more effective than warmed blankets, seems a bit redundant. When combined with the questionable outcome measure of tympanic temperature, the clinical relevance of these findings is limited.

The other outcome measures included in the study were wound infection rate, length of hospital stay, and hospitalization costs (Yang et al., 2012). Wound infection rates are important because hypothermia can induce vasoconstriction, thereby causing localized tissue hypoxia and increasing the risk of wound infection (reviewed by Hooper et al., 2008). Neutrophil function can also be impaired, which would interfere with immune surveillance of the wound. The length of hospital stays, and therefore the costs of care, can be increased by as much as 20% if the wound does not heal appropriately. Hypothermia has also been shown to increase the risk of pressure ulcers, reduce the efficacy of drugs, and increase the need for blood transfusions in peroperative patients.

The study by Yang and colleagues (2012) revealed that no infections were detected in any of the patients enrolled in their study. The length of hospital stay did differ between the two groups, but did not reach statistical significance. There was a significant amount of variability between the two experimental groups, which may have prevented a significant finding. Increasing the sample size might reveal that radiant heat is associated with a shorter hospital stay. Medical costs were also lower for radiant heat patients, but again the difference was not statistically significant. These results suggest that postoperative patients recovering in the PACU may benefit more from radiant heat than pre-warmed blankets, in terms of length of hospital stay and hospitalization costs.

Yang and colleagues (2012) conclude that their findings provide an evidenced-based rationale for choosing radiant heat to warm postoperative patients having undergone spinal surgery. While admitting that the patient population that could benefit from their research is limited to patients undergoing spinal surgery, they suggest radiant warming should be an option considered by medical facilities where forced-air warming is too expensive or labor intensive to implement because it represents too much of a contamination risk.

Based on the above analysis there are several threats to both the internal and external validity of the study by Yang and colleagues (2012). Selection bias was potentially introduced by the lack of patient randomization, which may explain why the two groups differed in terms of comorbidity and perioperative solution administration. The latter two confounding variables were not controlled for in the data analysis and could have increased the difference between the two groups in favor of radiant heat subjects. The authors were remiss in not mentioning this in the discussion section. The primary threat to external validity was mentioned by the authors in the discussion section, which was the uniform nature of the patient population used in the study and therefore its lack of generalizability to other surgical patient populations.

Yang and colleagues (2012) failed to discuss adequately the considerable limitations of their findings in the discussion and conclusions sections. They never addressed the potential confounding factor of placing a radiant heat source adjacent to the tympanic membrane, which was the primary outcome measure for the study. When this deficiency is combined with the reduced comorbidity and perioperative solution administration of the radiant heat patient group, this may explain why they found a very high warming rate for radiant heat. This deviation from previously published heating rates was never addressed in the results, discussion, or conclusion sections.

Yang and colleagues (2012) conclude that their findings show that radiant heat is superior to pre-warmed hospital blankets for treating postoperative hypothermia. However, in light of the above discussion, the only conclusion that seems reasonable to make is that radiant heat is at least as effective as pre-warmed hospital blankets. Given the trend towards shorter hospital stays and reduced hospitalization costs, it seems worthwhile to conduct a better controlled and larger study.

Presentation and Stylistic Dimensions

The research article by Yang and colleagues (2012) was both easy and pleasant to read. The format was the traditional Introduction, Methods, Results, and Discussion, which made it easy to navigate and understand what the study was about. All sections were concise and lacked jargon, but the introduction and discussion sections would have benefitted from being more comprehensive as mentioned above. The data tables were easy to read and understand, while providing a lot of information. Overall, the data and findings were well presented in a traditional style.