Incidence and Risk Factors for Postoperative Hypothermia After Orthopaedic Surgery

John P. Kleimeyer, MD; Alex H. S. Harris, PhD; Joseph Sanford, MD; William J. Maloney, MD; Bassam Kadry, MD; Julius A. Bishop, MD


J Am Acad Orthop Surg. 2018;26(24):e497-e503. 

In This Article


Postoperative hypothermia is a morbid and often preventable sequela of surgery, yet its incidence and relevant risk factors in orthopaedic surgery are poorly understood. Because unintended perioperative hypothermia plays a greater role in patient safety and quality measures,[1] improved identification of at-risk populations is warranted. In this study, the incidence of postoperative hypothermia defined as <36.0°C was 8.4% across a large and diverse sample of patients and procedures. Intraoperative hypothermia was found in 72.5% of procedures. Intraoperative hypothermia, lower preoperative temperature, female sex, adult reconstruction, and hip surgery were found to be markedly associated with postoperative hypothermia. Older age and lower BMI were also markedly associated with postoperative hypothermia, whereas the case duration and arthroscopy were correlated with lower postoperative temperature.

This study is the first to identify intraoperative hypothermia as both a common event and a strong risk factor for postoperative hypothermia in a general orthopaedic population. This finding is intuitive but important because intraoperative hypothermia is easily identifiable and treatable. The association between lower preoperative temperature and postoperative hypothermia also supports prewarming to prevent this complication.[37] The effect size of age also increases per year, with a 67% increase in the odds of postoperative hypothermia above age 85. Providers should be aware of this relationship and provide further attention to this high-risk elderly population to avoid complications of hypothermia. Adult reconstruction and hip procedures are associated with postoperative hypothermia with a large effect size. However, these procedures may be associated with larger surgical exposures, higher blood losses, decreased passive rewarming surface area, or other unrecorded factors. Furthermore, trauma was found to have the lowest rate of postoperative hypothermia. Patients with hip fracture were previously found to have high rates of hypothermia;[17,18] however, patients in these studies were not compared with patients undergoing hip or trauma procedures. The hip region may be a confounding factor given its strong effect in driving postoperative hypothermia. Patients sustaining trauma may also be exposed to more rigorous rewarming not recorded in this study.

The rate of postoperative hypothermia found in this study is similar to previously published rates of postoperative hypothermia in specific populations. Rates after arthroscopic hip surgery have been reported at 2.7%, versus 6.2% for all arthroscopy in the current study.[25] Perioperative hypothermia in hip reconstruction has been found at similar rates compared to intraoperative hypothermia in our study.[26] High rates have been reported in adult reconstruction at 26.3% and 28.0% in primary total hip and total knee arthroplasty, respectively.[21] Intraoperative hypothermia has additionally been demonstrated in specific populations, from 17.0% in 1,525 cases of hip fractures[27] to 37.0% in 2,580 cases of hip and knee arthroplasty,[28] associated with deep surgical-site infection in the hip fracture population. Previously documented risk factors specific to orthopaedic patients have included case duration, low BMI, age, sex, and intraoperative hypotension.[24,25,27,28] These previous studies have been limited to arthroscopic procedures and include factors such as irrigation fluid temperature that are not broadly generalizable.[23,29] In this study, we found an association of the case duration with decreased postoperative temperature but no association with postoperative hypothermia, which indicates that this was not a strong factor. Results of nonorthopaedic studies have also identified risk factors such as major surgery, blood transfusion, higher ASA grade, and general versus regional anesthesia.[30,31] The rate of local anesthetic and monitored anesthesia care compared with general anesthesia was low in this study, possibly indicative of fewer routine minor or outpatient procedures.

This study has several limitations. Although procedures were reviewed retrospectively, all data were recorded prospectively and all relevant cases were included. The primary outcomes of this study therefore should be accurate. Many procedures also included incomplete covariate data, and therefore several subspecialty or anatomic groups, such as hand surgery, were underrepresented. Because no cases of postoperative hypothermia were found among these patients (n = 30), they were not included in the regression analysis, and our results are therefore not generalizable to these populations. Similar rates of postoperative hypothermia were found in the excluded procedures, and the differences are not expected to cause a meaningful change in the study outcomes. BSA was not included in the study because it was highly correlated with BMI. BMI was chosen as a candidate predictor because it is more commonly evaluated in orthopaedic practice and is referenced in preference to BSA in studies evaluating hypothermia in the orthopaedic populations.[24,25,27,28] Various measurements of temperature were used and were not distinguished in analysis. Although peripheral thermometers have been found to have questionable accuracy,[32] 89.7% of measurements in this study were made from a central source. Data on a number of potentially relevant variables were not available, including ambient room temperature, duration of hypothermia, rewarming strategies, and postoperative outcomes. Our institution's standard practice is to use passive warming measures including forced air; however, use in specific cases was not recorded. Furthermore, rewarming strategies have demonstrated variable efficacy in previous reports.[33] Although the influence of postoperative hypothermia on the outcome warrants additional research, it is beyond the scope of this study. Finally, we recognize that procedure coding data may not accurately identify all procedures.[34–36] However, our CPT codes were assigned based on the electronic medical record at a single institution and are therefore likely to be more accurate than those obtained from a large administrative database.

Postoperative hypothermia is common after orthopaedic surgery. This study identified several variables that are associated with an increased risk of postoperative hypothermia in a large and diverse orthopaedic patient population. In addition, we demonstrated that intraoperative hypothermia is frequent and strongly associated with postoperative hypothermia, representing a clear opportunity for prevention. Aggressive preventive measures, such as preoperative and active warming, have been established as safe and effective and should be used in high-risk patient groups.[1,37,38] Additional research is indicated to explore the efficacy of hypothermia prevention protocols and to better understand the effect of postoperative hypothermia on patient outcomes.