Perioperative hypothermia is a common and important problem.¹ It is defined as a core body temperature of < 36 °C or a decrease of > 1 °C from baseline and occurs in both general anaesthesia (GA) and neuraxial anaesthesia (NA).¹ During GA, hypothermia is commonly encountered because of a widening of the thermoregulatory inter-threshold range, a decrease in the shivering threshold and redistribution of blood from core to peripheral vasculature.² Patients under NA develop hypothermia because of a resetting of the shivering and vasoconstriction threshold, combined with vasomotor paralysis, resulting in central to peripheral redistribution of heat.^3,4

Current anaesthetic guidelines recommend the use of core temperature (T_c) monitoring in patients undergoing any anaesthesia for longer than 30 min duration.⁵ However, T_c is not routinely monitored in patients undergoing NA. This is because of accurate T_c monitors being inappropriately invasive for NA and readily available non-invasive devices being unreliable at lower T_c.^6,7 Inadequate monitoring, coupled with the complications of perioperative hypothermia, demonstrates that hypothermia needs to be recognised through accurate temperature monitoring and addressed.^4,6

Lower limb orthopaedic surgery patients commonly undergo NA, where patients may be vulnerable to hypothermia. Orthopaedic operating theatres are particularly strict with maintaining cooler ambient temperatures. Most theatres aim for an ambient temperature of 18 °C – 21 °C to maintain personnel comfort and limit surgical site infections (SSI) despite local infrastructure policies recommending temperatures to be maintained in the range of 18 °C – 24 °C.⁸ In obstetric patients receiving NA, low ambient temperatures have been shown to worsen patient hypothermia, which may persist for up to 8 h into the postoperative period.¹ The rationale for lower ambient temperatures in orthopaedic theatres relates to the prevention of sepsis, yet hypothermia may predispose patients to infection secondary to immune suppression. The incidence and severity of hypothermia in the orthopaedic patient population have not been adequately quantified. We aimed to prospectively determine the incidence of clinically significant perioperative hypothermia (T_c < 35 °C) in orthopaedic patients undergoing NA to address this knowledge gap and inform future anaesthetic practice.

Data collection occurred between 22 May 2023 and 01 December 2023. During this period, 160 participants were recruited. No participants refused consent. Six recruited participants were subsequently excluded from analysis as they did not meet study criteria (4 converted to GA; 1 < 18 years; in 1 no T_c measurements recorded). A total of 1822 measurements were recorded from the remaining 154 participants of a maximum of 1831 expected measurements (99.5% completion rate).

Hypothermia (T_c < 35 °C) was found in 69 out of 154 participants (45.5%; 95% CI: 37.7–53.4). Participant and management characteristics are summarised in Table 1, grouped by the primary outcome.

Severe hypothermia (T_c < 34 °C) was found in 22 out of 154 participants (14.3%; 95% CI: 9.6–20.8). The mean lowest T_c recorded in participants was 35.0 °C (s.d.: 1.02). Univariable comparisons showed a significant difference in mean blood loss between normothermic and hypothermic participants (124.1 mL [s.d. 114.7; 95% CI: 98.4–149.8] vs. 199.9 mL [s.d.: 180.1; 95% CI: 156.3–243.5]; p < 0.01). This difference was greater in normothermic versus severely hypothermic participants (138.3 mL [s.d.: 125.4; 95% CI: 116.2–160.4] vs. 284.3 mL [s.d.: 230.3; 95% CI: 179.5–389.1]; p < 0.01). Spinal-to-skin incision was significantly longer in the hypothermic versus normothermic group, as shown in Table 1. Mean ambient theatre temperature was 18.2 °C (s.d.: 2.75). We did not perform multivariable analysis.

The mean changes in T_c for the cohort are illustrated in Figure 1.

This prospective observational study found that hypothermia (T_c < 35 °C) was present in almost half of participants undergoing orthopaedic surgery under NA, with one in seven participants experiencing severe hypothermia (T_c < 34 °C).

Most participants in this study were actively warmed intraoperatively using forced air warming devices (97%) and warmed fluids (73%). Despite these standard measures, there was still a significant incidence of hypothermia. This suggests that current, routine intraoperative warming measures are ineffective at preventing perioperative hypothermia. Although most studies assessing hypothermia and NA are not specific to orthopaedic surgery, an incidence as high as 90% has been found.¹¹ Use of active warming has been shown to decrease the incidence of hypothermia, and in those patients where hypothermia does develop, a faster recovery to normothermia, when under NA, compared to GA.^11,12

The incidence of shivering among participants did not correlate with hypothermia, as only 13% of participants had this positive finding, and there was no difference between the normothermic and hypothermic groups, either during surgery or postoperatively. This suggests that shivering is not a good surrogate marker of hypothermia during NA.

There was a clear association between increasing participants’ age and hypothermia, with a mean difference of 16 years between the normothermic and hypothermic groups. Older patients are more vulnerable to develop hypothermia as a result of the physiological changes associated with ageing, including degenerative body changes, decreased reserve, low metabolic rates, cold intolerance and altered compensatory abilities.¹³ This group may represent a high-risk group that could be prioritised for preventative measures in clinical practice.

Blood loss between the normothermia (124 mL) and hypothermia (199 mL) groups was statistically significant, with an increased blood loss of approximately 60% seen in the hypothermic group. Practically, this represents only 75 mL because of low total blood loss amounts. The ‘severe hypothermia’ group, however, showed blood loss of 150 mL more than those with T_c > 34 °C. Blood loss of 150 mL is more likely to have a significant clinical impact, supporting the avoidance of severe hypothermia as an important clinical target. A meta-analysis looking at the effect of perioperative hypothermia on surgical blood loss found that even mild hypothermia resulted in increased blood loss by approximately 16% (95% CI: 4–26).¹⁴ Our study showed more significant blood loss than previous studies although this was not a primary outcome, and determination of blood loss relied on a combination of objective and subjective measures. Prospective research designed specifically to assess this outcome would be required to address this knowledge gap.

We also found an association between hypothermia and increasing time from NA-to-skin incision on univariable comparison. The hypothermia group took approximately 4 min longer until skin incision occurred. Hypothermia is often assumed to occur as a result of prolonged surgical duration, but findings from this study support the physiological principle of redistribution of heat from central to peripheral areas soon after insertion of NA, caused by vasomotor paralysis.^3,6 During this initial period post-NA insertion, patients are left exposed prior to cleaning and application of surgical drapes. Patient exposure should be limited in duration and prevented where possible to avoid worsening the risk of hypothermia during this phase of surgery.

Average ambient temperature measured in this study was 18.2 °C, which is on the lower end of accepted ranges, falling marginally within acceptable local infrastructure policy (recommended 18 °C – 24 °C).⁸ Orthopaedic operating theatre temperatures are kept low in an attempt to reduce the risk of SSIs. Maintenance of normothermia is recommended as an essential step in the prevention of SSIs by both the National Institute for Health and Care Excellence (NICE) and Centers for Disease Control and Prevention (CDC).^15,16 Normothermia promotes faster wound healing through increased collagen production at the incision site.¹⁷ Hypothermia impairs immune function by reducing macrophage function and platelet adhesion, resulting in delayed wound healing and increases risk of re-contamination of surgical wounds.¹⁸ Given the significant incidence of hypothermia seen in our study, it is plausible that increasing operating room temperature to at least 20 °C – 22 °C could result in a reduced incidence of hypothermia while still maintaining ambient temperatures that fall within operating standards.

We did not measure SSIs in this study, but it is an important area of concern in hypothermic patients. Studies testing the association between perioperative hypothermia and the incidence of SSIs have shown varying results. A recent systematic review found that the nature of the surgical intervention influenced the likelihood of development of SSIs. Hypothermic patients undergoing ‘clean’ operations (elective plastic surgery and shoulder arthroplasty) had a reduced risk of developing SSIs because of the elective nature of the procedure and likelihood of the wound remaining uncontaminated.^19,20 In contrast, hypothermic patients with contaminated wounds or undergoing emergent post-traumatic surgery showed a higher incidence of SSIs.¹⁹ In our study, most participants were undergoing procedures following traumatic injuries, suggesting that hypothermia should be avoided to reduce the risk of preventable SSIs. Further studies focused on orthopaedic emergency procedures performed under NA, hypothermia and the incidence of SSIs are still needed.

Given the high incidence of hypothermia seen in our study, temperature monitoring in orthopaedic patients undergoing NA is necessary. Despite this, routine monitoring is not performed by anaesthetists in our setting because of a lack of cost-effective and appropriate, non-invasive temperature monitoring devices.^6,21 Oral and infrared thermometers are affordable and reusable and therefore ideal for resource-constrained environments. However, their use is limited in the perioperative period because of concerns regarding their reliability and validity. A recent study by Vawda et al. showed poor agreement in measurements between oral, infrared and the Dräger Tcore™ temperature monitoring systems in obstetric patients undergoing spinal anaesthesia, suggesting that they are not interchangeable.²² The Dräger Tcore™ non-invasive temperature monitor that was used in our study uses dual heat flux technology and provides a continuous, accurate reading when compared to the gold-standard Swan-Ganz pulmonary artery catheter, and has been validated for use in the perioperative setting.^23,24 However, these studies suggest that concerns remain regarding the reliable detection of hypothermia and the precision of readings. The sensor is also single use, and cost precludes its routine use in the state sector. However, it is possible that the use of accurate temperature monitoring may assist in reducing the incidence of perioperative hypothermia and potential complications such as SSIs and increased blood loss, which may result in lower healthcare costs. Further research is warranted to find an appropriate T_c monitor for NA in low- and middle-income settings and to assess the cost-effectiveness of routine T_c monitoring.

There were some limitations of this study. This was a single-centre study, which may reduce generalisability. However, as a busy regional hospital, the case load and type are likely to apply to many other South African hospitals. We also believe that the lower ambient operating room temperatures, which may have contributed to the high incidence of hypothermia seen in this study, are commonly seen in other settings. Anaesthesia practice was largely left at the discretion of the attending anaesthetist on the day of surgery, which could have been a confounding variable; however, institutional practices are determined by local and international guidelines under the supervision of a specialist anaesthesiologist. The associations discussed were based on univariable analysis and should therefore be treated as exploratory in nature. Finally, we were unable to quantify outcomes such as the incidence of sepsis because of resource constraints, limiting the ability to assess the clinical impact of hypothermia beyond the operating theatre.

This study focused on quantifying the problem of perioperative hypothermia in orthopaedic NA and was not powered to determine the clinical impact of hypothermia. Given the incidence and severity of the hypothermia that has been demonstrated, further research should examine both preventative measures aimed at reducing hypothermia and at quantifying clinically relevant complications of perioperative hypothermia including SSIs, bleeding and perioperative morbidity and mortality. Research related to the impact of hypothermia on patients’ experience is also lacking and should be considered in the future. Novel techniques for pre- and intra-operative warming, with particular focus on resource-limited settings, could be explored further.

This study found that hypothermia was present in almost half of participants undergoing orthopaedic surgery under NA, and one in seven experienced severe hypothermia. Temperature monitoring should be considered essential in all patients undergoing NA for orthopaedic procedures although T_c monitoring may not always be available or practical. Preventing unnecessarily low ambient theatre temperatures should be considered as a strategy to prevent hypothermia. Avoiding prolonged patient exposure following insertion of NA should also be prioritised. Further research into effective prevention of hypothermia and the potential clinical impact of perioperative hypothermia in orthopaedic procedures under NA is warranted.