• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Collaborating Centre for Nursing and Supportive Care (UK). The Management of Inadvertent Perioperative Hypothermia in Adults [Internet]. London: Royal College of Nursing (UK); 2008 Apr. (NICE Clinical Guidelines, No. 65.)

Cover of The Management of Inadvertent Perioperative Hypothermia in Adults

The Management of Inadvertent Perioperative Hypothermia in Adults [Internet].

Show details

11TREATMENT

Clinical Question: Are warming devices/mechanisms effective in treating IPH in adults in the different phases of perioperative care?

SELECTION CRITERIA

Selection criteria were as outlined in the general methods section.

Types of intervention

The following interventions were to be considered:

  1. Active warming mechanisms
    The following types of warming mechanisms were considered under active warming:
    • Electric blanket
    • Water mattress
    • Radiant heating
    • Warmed blankets
    • Heating gel pad.
  2. The following mechanisms were considered under thermal insulation:
    • Reflective blanket
    • Reflective clothing.

Other types of heat loss prevention, such as cotton sheets, cotton blankets, or wool blankets were considered to be ‘usual care’, unless the blankets were actively warmed.

Perioperative phase

Treatment of hypothermia could take place in any of the perioperative phases, but the phases were treated separately.

Types of comparison

The following comparisons were to be included:

Subgroup analyses were planned by type of warming device, power, duration of warming and degree of hypothermia.

Characteristics of clinical studies included in the review (Appendix C)

Eleven studies were included in this review (Alfonsi 2003; Bredahl 1995; Giuffre 1991; Hershey 1997; Jackson 1997; Karayan 1996; Lennon 1990; Stevens 2000; Summers 1990; Vanni 2003; Weyland 1994).

An additional study (Bräuer 2004) was included as indirect evidence, and is presented separately. The indirect population comprised cardiac patients, in the post-bypass stage after rewarming, who then underwent inadvertent hypothermia (‘after drop’).

The three studies excluded from the review are listed in Appendix E.

Study details

A total of 676 patients were included in the eleven studies. There were 50 additional patients in the indirect study, Bräuer (2004). The total number of patients in each study ranged from 18 (Alfonsi 2003) to 144 (Hershey 1997). Eight studies had fewer than 20 patients in the intervention arm (Alfonsi 2003; Bredahl 1995; Jackson 1997; Karayan 1996; Lennon 1990; Vanni 2003; Weyland 1994; Bräuer 2004, indirect) and two of these had less than 20 patients overall (Alfonsi 2003; Karayan 1996).

No studies were conducted in the UK, four studies were conducted in the US (Giuffre 1991; Hershey 1997; Lennon 1990; Summers 1990); two studies in France (Alfonsi 2003; Karayan 1996); two in Germany (Weyland 1994; Bräuer 2004, indirect), and one each in Denmark (Bredahl 1995); Brazil (Vanni 2003); South Africa (Jackson 1997) and Australia (Stevens 2000).

Mainly the studies did not state the source of funding (if any), but one (Summers 1990) was part funded by grant from Augustine Medical (forced air warming device manufacturer).

Five studies had more than two randomised groups: Guiffre (1991) had three arms; Hershey (1997) had three arms; Vanni (2003) had three arms; Weyland (1994) had three arms; Bräuer (2004, indirect) had five arms. Overall there were 18 direct study comparisons and ten indirect comparisons.

One study (Vanni 2003) treated the patients in the intraoperative phase or in both pre and intraoperative phases. One study treated the patients in the intraoperative phase (Karayan 1996). The other nine studies investigated treatment of IPH in PACU or ICU.

Participants

The age of the patients ranged from 16 to 86 years with a mean age (where given) ranging from 31 to 66 years; one study only included patients over 50 years (Bredahl 1995), and one study excluded patients over 60 years (Hershey 1997). Two studies were carried out exclusively in men (Alfonsi 2003; Bräuer 2004, indirect); one study was exclusively in women (Vanni 2003). Two studies did not state the gender (Karayan 1996; Lennon 1990). BMI was not stated in any study, although two (Alfonsi 2003; Karayan 1996) reported that none of the patients were obese, and two studies (Weyland 1994; Bräuer 2004, indirect) stated that the body weight was within -10% and +30% of normal.

Three studies included patients with ASA I to II status (Alfonsi 2003; Bredahl 1995; Vanni 2003); one studies had patients with ASA II to III (Karayan 1996) and one had patients with ASA I to III (Weyland 1994). In the indirect study, Bräuer (2004), the patients were ASA III and the other studies did not state the ASA status.

Generally, the studies gave insufficient information about the surgery and anaesthesia. Eight studies reported the type of surgery:

The grade of surgery was classified only for two studies (Alfonsi 2003; Karayan 1996) and was grade 2.

Five studies stated the surgery was elective (Bredahl 1995; Karayan 1996; Vanni 2003; Weyland 1994; Bräuer 2004, indirect). Six studies stated the duration of surgery:

Patients had general anaesthesia in seven studies (Hershey 1997; Lennon 1990; Jackson 1997; Karayan 1996; Vanni 2003; Weyland 1994; Bräuer 2004, indirect); combined general and regional in two (Alfonsi 2003; Bredahl 1995), and a mixture of general and/or regional in one (Stevens 2000). Two studies did not mention the type of anaesthesia (Giuffre 1991; Summers 1990). Duration of anaesthesia was more than 60 minutes in six studies (Alfonsi 2003; Bredahl 1995; Giuffre 1991; Summers 1990; Vanni 2003; Bräuer 2004, indirect) and not stated in the rest.

Patients were included if they had hypothermia, as defined by the authors, however, the degree and definition of hypothermia varied, as did the phase in which it occurred and the means of measuring temperature.

Four studies recorded the core temperature at the tympanic membrane (Alfonsi 2003; Stevens 2000; Summers 1990; Vanni 2003); one recorded temperature at the pulmonary artery (Karayan 1996); two gave oesophageal temperatures (Weyland 1994; Bräuer 2004, indirect); two rectal (Bredahl 1995; Jackson 1997) and three oral (Giuffre 1991; Hershey 1997; Lennon 1990). The GDG regarded rectal and oral temperatures as only partially adequate measures of core temperature, except when sufficient detail was given for the measurement of oral temperatures and so these studies were included but regarded with caution. Hershey (1997) stated that the oral thermometer measurements correlated moderately well with tympanic temperatures in a previous study.

Two studies (Alfonsi 2003; Karayan 1996) described mild hypothermia (35.0 to 35.9°C); one (Guiffre 1991) was moderate (34.0 to 34.9°C); two were mild and moderate (Stevens 2000 excluded patients less than 34.5°C and Vanni 2003 reported mean temperatures between 34.9 and 35.2°C) and the rest did not state explicitly.

In the Karayan (1996) study, patients in the intervention group received forced air warming when their intraoperative core temperature dropped below 36.0°C; in practice, this was two hours after induction. The Vanni (2003) study was not designed as a trial to treat IPH, rather the intention was to prevent IPH. However all groups were hypothermic before forced air warming started. The authors attributed this drop in temperature to the premedication (7.5mg midazolam IM, 30 minutes before admission to the theatre, at which time patients were randomised to treatments). The GDG was not wholly convinced by this explanation and noted that this dose and route of administration is not used in the UK. The other studies had inclusion criteria for patients having temperatures below 36.0°C.

The three studies measuring tympanic membrane temperatures included patients with temperatures less than 36.0°C (Summers 1990; Stevens 2000 – implied inclusion) or had a final intraoperative temperature of 35.1°C. The studies recording oesophageal temperatures included patients with temperatures less than 35.5°C; those with rectal temperatures had to be less than 35.5°C (Bredahl 1995) or 35.9°C (Jackson 1997). One of the studies measuring oral temperatures required an inclusion temperature of 35°C or less (Guiffre 1991; no places of decimal), another was less than 35.0°C (Lennon 1990) and Hershey (1997) had an entry requirement of less than 36°C (no places of decimal).

Two studies had patients who were ventilated in ICU (Weyland 1994; Bräuer 2004, indirect); one study had about 30% of the patients ventilated (Hershey 1997); two studies had no patients ventilated (Alfonsi 2003; Lennon 1990 [exclusion criterion]). One other study stated the patients were in ICU (Jackson 1997).

Interventions

There was a range of interventions used:

  • Forced air warming device in nine studies, all had full body covering unless otherwise stated:
  • Electric blanket (50W), used until the temperature reached 37°C (Weyland 1994)
  • Radiant heaters in five comparisons:
    • Aragona Thermal Ceilings CTC X overhead heater, power setting 1kW, with the heater placed 75cm above the patient’s chest (Weyland 1994; Bräuer 2004, indirect);
    • Aragona Thermal Ceilings CTC X overhead heater, power setting 500W, about 60cm above the chest (Bredahl 1995);
    • Radiant heater with two radiant lights placed 71cm above the patient’s skin (Guiffre 1991);
    • Self-assembled combination of four halogen lamps (each 160W) placed 65cm above the patient’s body surface (indirect Bräuer 2004).
  • Head covering in one study (Hershey 1997)

The temperature settings and durations of warming were as follows:

The Vanni (2003) study gave the patients forced air warming either 60 minutes before induction and during the intraoperative period, or intraoperatively only. Karayan (1996) gave forced air warming intraoperatively, two hours after induction of anaesthesia. Most of the other studies using forced air warming devices warmed the patients until they reached a specified temperature, but Summers (1990) seemed to restrict the warming period to one hour, and the durations for Lennon (1990) and Jackson (1997) were 90 minutes and three hours respectively.

Mostly the duration of radiant heating was until a specified temperature was reached, but the Bredahl (1995) study heated the patients for 2 hours and the power was decreased if the skin temperature exceeded 37°C.

Comparators

Several studies used heated blankets as a comparator. Two (Guiffre 1991; Stevens 2000) specified that the blankets were changed on a regular basis (e.g. every 15 minutes); one changed the blankets as needed (Summers 1990) and the others did not state if the blankets were changed (Hershey 1997; Lennon 1996). One study (Lennon 1996) stated that the blankets were warmed to 37°C, one reported that the blankets were stored at 66 to 77°C (Giuffre 1991) and the others did not report the temperature. The GDG noted that the procedure of changing blankets was not carried out in the UK.

Comparisons

The following comparisons were reported:

I. Intervention in the preoperative phase

  • Active warming (preoperatively) plus active warming (intraoperatively) versus active warming (intraoperatively) (Vanni 2003) [cross-phase].

II. Intervention in the pre and intraoperative phases

III. Intervention in the intraoperative phase

IV. Intervention in the postoperative phase

There were no studies that simply investigated a thermal insulation mechanism versus usual care.

More specifically the comparisons were:

I. Preoperative phase

A. Active warming versus usual care
A1. Active warming (preoperatively) plus active warming (intraoperatively) versus active warming (intraoperatively) [cross-phase]
  • Forced air warming for 60 minutes pre-induction (pre) plus forced air warming (intra) versus usual care (pre) + forced air warming (intra) (Vanni 2003)

II. Pre and intraoperative phase

A. Active warming versus usual care
  • Forced air warming (full body) versus usual care (cotton sheet), from 60 minutes pre-induction; all patients received room temperature fluids at 8 to 10ml/kg/h (Vanni 2003).

III. Intraoperative phase

A. Active warming versus usual care
  • Forced air warming (full body) versus usual care (cotton sheet) from induction
    • All patients received room temperature fluids at 8 to 10ml/kg/h (Vanni 2003).
  • Forced air warming (upper body) versus usual care (warm cotton sheet) from when the patients became hypothermic (2 hours after induction)
    • All patients received warmed fluids, at a volume of 3.1 and 3.8 litre (Karayan 1996).

IV. Postoperative phase

A. Active warming versus usual care
  • Forced air warming versus usual care (two direct and two indirect studies):
  • Radiant heater versus usual care
    • Radiant heater (Aragona Thermal Ceilings; 1kW, 75cm from chest) versus standard hospital blanket (Weyland 1994);
    • Radiant heater (Aragona Thermal Ceilings; 1kW, 75cm from chest) versus standard polyester filled hospital blanket; insulation value 1.7 clo (Bräuer 2004, indirect);
    • Radiant heater (self assembled): 4 Hydrosun 500 halogen lamps (4x160W; 60cm from chest) versus standard polyester filled hospital blanket; insulation value 1.7 clo (Bräuer 2004, indirect).
  • Electric heating blanket versus usual care (standard hospital blanket; Weyland 1994).
B. Active warming 1 versus active warming 2
B1. Active warming 1 versus active warming 2 (with no additional warming)
  • Forced air warming blanket versus warmed blankets (three studies):
  • Radiant heater versus electric blanket (one study):
    • Radiant heater (Thermal Ceilings; 1kW; 75cm from chest) versus electric blanket (50W, placed between two standard hospital blankets) (Weyland 1994)
B2. Active warming 1 versus active warming 2 (with additional warming mechanisms in both groups)
  • Forced air warming versus warmed blanket (two direct and two indirect studies)
    • Forced air warming blanket (Bair Hugger®) versus warmed blanket (changed every 15 minutes, temperature not stated) (Stevens 2000)
      • Both groups had a head covering, which was not said to be warmed;
    • Forced air warming blanket (Bair Hugger®; medium, presumed 37°C) versus warmed cotton blanket (changed every 20 minutes, stored 66 to 77°C)
      • Both groups had a warmed head covering which was replaced every 20 minutes (Giuffre 1991);
  • Radiant heater versus warmed blankets (one study)
    • Radiant heater ( 2 radiant lights 71cm from skin) versus warmed cotton blanket (changed every 20 minutes, stored 66 to 77°C)
      • Both groups had a warmed head covering which was replaced every 20 minutes (Giuffre 1991).
C. Active warming 1 (subtype 1) versus active warming 1 (subtype 2)
  • Forced air warming blanket 1 versus forced air warming blanket 2
    • Full body forced air warming blanket (Bair Hugger®, max setting) versus full body forced air warming blanket (Warm Touch®, max setting) (Bräuer 2004, indirect);
  • Radiant heater 1 versus radiant heater 2
    • Radiant heater (Aragona Thermal Ceilings; 1kW, 75cm from chest) versus radiant heater (self assembled): 4 Hydrosun 500 halogen lamps (4x160W; 60cm from patient’s chest) (Bräuer 2004, indirect).
D. Active warming versus thermal insulation
  • Radiant heater (Thermal Ceiling; 500W; about 65cm above body surface) versus reflective blanket (type not stated) plus 3 cotton blankets (Bredahl 1995)
E. Thermal insulation versus usual care
E1. Thermal insulation versus usual care with active warming in both groups
  • Reflective blanket (type not stated) plus reflective head covering (thermal insulation) versus usual care (Hershey 1997):
    • Both groups had two warmed thermal blankets (not stated to be changed; temperature not stated; active warming).
F. Thermal insulation 1 versus thermal insulation 2, with active warming in both groups
F1. Thermal insulation 1 versus thermal insulation 2, with active warming in both groups
  • Reflective blanket plus reflective head covering (thermal insulation) versus reflective blanket (thermal insulation) (Hershey 1997):
    • Both groups had two warmed thermal blankets (not stated to be changed; temperature not stated; active warming).

Outcomes

The studies measured the following outcomes:

Primary outcomes

Two studies recorded the number of patients with IPH (Karayan 1996; Vanni 2003), but several measured the core temperature at different times. The GDG decided that the most useful outcome measures, where given, were the rate of increase in temperature and the time taken to reach normothermia.

Four studies recorded the core temperature at the tympanic membrane (Alfonsi 2003; Stevens 2000; Summers 1990; Vanni 2003); one measured pulmonary artery temperatures (Karayan 1996); two measured oesophageal temperatures (Weyland 1994; Bräuer 2004, indirect); two rectal (Bredahl 1995; Jackson 1997) and three oral (Giuffre 1991; Hershey 1997; Lennon 1990).

METHODOLOGICAL QUALITY OF INCLUDED STUDIES

An adequate method of sequence generation was recorded in four studies (Alfonsi 2003, computer generated; Giuffre 1991, shuffled envelopes; Hershey 1997, random numbers table; Summers 1990, coin toss); there was an inadequate method in one study (Stevens 2000; alternation) and the method was unclear in the remaining studies.

A partially adequate method of allocation concealment was reported in three studies (Alfonsi 2003: sequentially numbered opaque envelopes; Lennon 1990, sealed envelopes; Vanni 2003: sequentially numbered opaque envelopes); allocation concealment was inadequate in one study (Stevens 2000; alternation) and unclear in the remaining studies.

Two studies reported that the outcome assessors were blinded for shivering (Alfonsi 2003; Vanni 2003) and two that they were not (Hershey 1997; Jackson 1997), the others did not say. It was unlikely that the patients were blinded, except for studies in ICU (Weyland 1994; Bräuer 2004, indirect).

Most of the studies demonstrated baseline comparability. Two studies were not comparable for the length of time in the theatre. In Summers (1990), the warming group was longer by 35 minutes, and in Hershey (1997) the reflective blanket group mean was 32 and 49 minutes respectively longer than the reflective blanket + hat and control groups. One study (Stevens 2000) was not comparable for the proportion of orthopaedic patients (more in control group: 3.6 versus 13.2%).

For the preoperative and intraoperative treatment studies (Karayan 1996; Vanni 2003) the temperature at baseline was comparable between the groups. For the postoperative treatment studies, the temperature on arrival in PACU/ICU was reported for all studies except Stevens (2000); Weyland (1994); Bräuer (2004), indirect. All except Summers (1990) showed comparable temperatures at baseline (Figure 1). This study had a significantly lower mean baseline temperature on arrival in PACU for the intervention group (0.38°C). We note that two of the comparisons in Hershey (1997) showed a difference in baseline of 0.2°C, but this was not statistically significant, and there was a difference in the median of 0.3°C for Bredahl (1995), which the authors said ‘did not yield intergroup differences’.

Figure 1. Baseline core temperatures.

Figure 1

Baseline core temperatures.

Three studies described an a-priori power calculation (Alfonsi 2003; Bredahl 1995; Hershey 1997). In Alfonsi (2003), the power calculation required 9 patients per group to achieve a difference of 0.4°C. The Bredahl (1995) study required 13 patients in each group to achieve a core temperature change of 0.5°C. The Hershey (1997) study calculated a sample size of 48 per group was required to detect a medium effect size of F=0.25.

One study reported more than 20% of dropouts for one outcome (6/15 (40%) of the forced air group did not have the temperature recorded at 90 minutes) (Lennon 1990). In the Stevens (2000) study 3/60 (5%) of the forced air group and 4/60 (7%) of the control group had incomplete data. The Hershey (1997) study had missing data for 2/48 (4%) in the reflective blankets and reflective blanket + hat groups.

Overall, three studies were regarded as having potential for bias (Stevens 2000, allocation concealment; Hershey 1997 and Summers 1990, baseline comparability). These were treated with caution and examined in sensitivity analyses. The Lennon (1990) outcome at 90 minutes was also potentially biased.

RESULTS

I. Treatment in the preoperative phase

A. Active warming versus usual care (cross phase comparison)

One study (Vanni 2002) in 20 patients investigated the additive effect of preoperative warming to intraoperative warming for the treatment of IPH.

1. Core temperature at different intraoperative times

Data were extracted from a graph for a series of intraoperative times (Figure 2). The confidence intervals at 30, 60 and 120 minutes were too wide to determine if there was a difference between interventions.

Figure 2. Core temperature: intraoperative temperature; active warming (pre and intraoperative) versus active warming (intraoperative).

Figure 2

Core temperature: intraoperative temperature; active warming (pre and intraoperative) versus active warming (intraoperative). NB: Scale −4 to +4

2. Core temperature: lowest intraoperative temperature

The lowest intraoperative measurements were extracted from graphs (Figure 3) and were found at 30 minutes for the pre and intraoperative warming group and 120 minutes for the intraoperative warming group. The confidence interval is too wide to determine if there was a difference between interventions..

Figure 3. Core temperature: lowest intraoperative temperature; active warming (pre and intraoperative) versus active warming (intraoperative).

Figure 3

Core temperature: lowest intraoperative temperature; active warming (pre and intraoperative) versus active warming (intraoperative).

3. Core temperature: end of surgery

At the end of surgery, there was no significant difference in the core temperature at the end of surgery, although the confidence interval is fairly wide (Figure 4).

Figure 4. Core temperature: end of surgery; active warming (pre and intraoperative) versus active warming (intraoperative).

Figure 4

Core temperature: end of surgery; active warming (pre and intraoperative) versus active warming (intraoperative).

4. Time to reach 36.0°C

The time to reach 36.0°C was estimated from a graph. For the group that was warmed preoperatively, it took between 60 and 75 minutes for the core temperature to exceed 36.0°C (36.5°C was reached) from 34.9°C. Once the temperature was at 36.5°C it did not fall below 36.0°C intraoperatively during further warming.

5. Shivering

Vanni (2003) evaluated shivering as absent, mild (when only detected by ECG artefacts) or severe (when clinically obvious). Only mild shivering was observed in Vanni (2003), and the GDG decided that shivering evaluated with ECG artefacts was not an appropriate method of assessment, because other involuntary movements (e.g. in those with Parkinson’s disease) may be recorded. Therefore the incidence of mild shivering was not considered for this study, and there was no incidence of severe shivering in either group.

II. Treatment in the pre and intraoperative phases

A. Active warming versus usual care

One study (Vanni 2003) in 20 patients compared full body forced air warming in both the pre and intraoperative phases versus usual care for the treatment of IPH; all patients received room temperature fluids at 8 to 10ml/kg/h.

1. Core temperature at different intraoperative times

Vanni (2003) reported a series of intraoperative temperature measurements in 20 patients and data were extracted from graphs.

The analysis showed a large significant difference between interventions at all durations, with the patients warmed in the pre and intraoperative phases having higher mean core temperatures than those given usual care. However, at all times the confidence intervals are wide.

Figure 5. Core temperature intraoperatively – forced air warming in the pre and intraoperative phases versus usual care.

Figure 5

Core temperature intraoperatively – forced air warming in the pre and intraoperative phases versus usual care. NB: Scale −4 to +4

2. Core Temperature – lowest intraoperative temperature measured

The lowest temperatures for the treatment group and control group were at 30 and 120 minutes in Vanni (2003). The confidence interval is wide, but there is a large statistically significant difference between groups; mean difference: 1.14°C (95% CI 0.25, 2.03) for a core temperature of 35.1°C for the control group.

Figure 6. Lowest intraoperative temperature.

Figure 6

Lowest intraoperative temperature. NB: Scale −4 to +4

3. Core temperature – final intraoperative temperature (Figure 7)

The Vanni (2003) study in 20 patients reported the core temperature at the end of the intraoperative period (duration of surgery was 167 minutes (SD 57) for the control group and 175 min (SD 66) for the intervention group). There is a large statistically significant difference between groups; but the confidence interval is wide.

Figure 7. Final intraoperative temperature.

Figure 7

Final intraoperative temperature. NB: scale −4 to +4

4. Incidence of hypothermia at the end of anaesthesia

One study (Vanni 2003) in 20 patients reported the incidence of hypothermia (core temperature less than 36.0°C) at the end of anaesthesia (duration of surgery was 167 minutes (SD 57) for the control group and 175 minutes (SD 66) for the intervention group. There was a very large effect, with all patients being hypothermic in the control group and two in the intervention group. The confidence interval is very wide.

Figure 8. Incidence of hypothermia at end of anaesthesia.

Figure 8

Incidence of hypothermia at end of anaesthesia. NB: Scale 0.001 to 1000.

5. Incidence of shivering

Vanni (2003) categorised shivering as absent, mild (when only detected by ECG artefacts) or severe (when clinically obvious). Shivering was assessed by an independent observer blinded to the study treatment. The GDG suggested that as any involuntary movement (e.g. in those with Parkinson’s disease) would be recorded by ECG artefacts, this would not be an accurate method of assessing shivering. Therefore only shivering classified severe was considered for this study. However, no patients experienced severe shivering in this study.

6. Time to reach 36.0°C

The time to regain a core temperature of 36°C was reported as 57 (SD 15) minutes in the control group and from 15 to 30 minutes in the three patients in the warming groups (pre plus intraoperative phase = 2 patients; intraoperative only = 1 patient).

III. Treatment in the intraoperative phase

Two studies (Karayan 1996; Vanni 2001) with 38 patients compared the effectiveness of forced air warming compared with usual care in patients undergoing surgery under general anaesthesia.

In the Karayan (1996) study, 18 patients undergoing abdominal aortic surgery received upper body (equivalent to 24% of the body surface area) forced air warming (set at ‘high’); the forced air blanket was covered with additional 2 cotton sheets and the usual care group received a warm cotton sheet. All patients received warmed IV fluids.

In the Vanni (2001) study, 20 patients undergoing abdominal surgery lasting at least 2 hours received forced air warming (42 to 46°C); the blanket covered the thorax, shoulders, arms and hands and was covered by an additional cotton sheet. Patients in the usual care group received two cotton sheets covering the thorax, shoulders, arms and hands. All patients received fluids kept at the theatre temperature (21.5 to 22°C) before infusion.

In Vanni (2001), patients were hypothermic before induction of anaesthesia (35.2°C [SD 1.2]; 35.1°C [SD 1.1]). In Karayan (1996), both groups were above 36.0°C before induction of anaesthesia.

In the Karayan (1996) study, warming (or no warming) commenced when patient’s core temperature fell below 36.0°C. The core temperature in the patients randomised to the forced air warming group fell below 36.0°C at 2 hours after induction.

The delays in activating the warming system were for the following reasons:

  • Use of a warming system before induction would require a cover on the lower limbs: the authors considered this a risk in patients with aorto-iliac occlusive disease because of the risk of burning.
  • Insertion of invasive monitoring in the upper part of the body would have been precluded.

1. Core temperature at different intraoperative times

For the Karayan (1996) study the following results are presented: 1 hour after induction (although no warming at this stage - shown for completeness); 2 hours after induction (when warming commenced); 3 hours after induction (1 hour of warming), 4 hours after induction (2 hours of warming), 5 hours after induction (3 hours of warming) and 6 hours after induction (4 hours of warming). Results for Vanni (2001) are presented at 30 minutes and combined with Karayan (1996) at 1 hour of warming and 2 hours of warming.

The mean difference was not significant at 30 minutes (Vanni 2003). In the meta-analysis of the two studies with 38 patients at 60 and 120 minutes, there was a significantly higher mean temperature for the warmed group; 60 minutes: WMD 0.81°C (95% CI 0.36, 1.26; and 120 minutes: WMD 1.22°C (95% CI 0.74, 1.69). In each case the confidence interval is fairly wide.

At 3 hours and 4 hours, data are available from one study (Karayan 1996) with 18 patients: there was a significantly higher mean temperature for the warmed group, although the confidence interval is wide.

Figure 9. Core temperature at different intraoperative times.

Figure 9

Core temperature at different intraoperative times.

2. Incidence of hypothermia

The Vanni (2003) study reported the incidence of hypothermia on arrival in the recovery room. This showed that all 10 patients in the usual care group had hypothermia, compared with only 1 of 10 in the warmed group.

Karayan (1996) also reported that all unwarmed patients were hypothermic at end of surgery. Mean core temperature for the warmed group at the end of surgery was 36.5ºC (SD 0.3ºC). Then assuming that none of these patients were hypothermic, the comparison in this study becomes: 0/9 versus 9/9 patients hypothermic, and the two studies can be combined. Meta-analysis showed a highly significantly lower incidence of hypothermia for the warmed group; Peto OR 0.03 (95%CI 0.01, 0.09). The confidence interval is very wide.

Figure 10. Incidence of hypothermia.

Figure 10

Incidence of hypothermia.

3. Time to reach 36.0°C

The time to regain a core temperature of 36.0ºC was reported in Vanni (2003) as 57 (SD 15) minutes in the control group and from 15 to 30 minutes in the three patients in the warming groups (pre and intraoperative phase = 2 patients; intraoperative only = 1 patient).

In Karayan (1996), the time to regain a core temperature of 36.0°C was around 3 hours in the warmed group while the control group had not regained a core temperature of 36.0ºC at the last measurement at 4 hours.

IV. Treatment in the postoperative phase

A. Active warming versus usual care

1. Core temperature in the postoperative period

Two studies with 18 and 20 patients recorded the core temperature (Jackson 1997, rectal; Alfonsi 2003, tympanic temperature); temperatures in the Jackson (1997) study were recorded at various times postoperatively. Generally, the confidence interval is fairly wide, but at longer times (60 minutes and above) the mean temperature is significantly higher for the active warming group. At 60 minutes, the mean control group temperature was still below 36.0°C, but that for the warmed group was above. It is noted that rectal temperatures were measured for the Jackson (1997) study. There was no significant difference in tympanic temperature at 37 minutes in the Alfonsi (2003) study.

Figure 11. Core temperature.

Figure 11

Core temperature. NB: Scale −4 to +4

2. Time taken to increase the temperature

One study (Weyland 1994) in ICU patients reported individual patient data for the time taken to increase the temperature from 35.0ºC to 35.5ºC; from 35.5ºC to 36.0ºC and from 36.0°C to 36.5°C. These data were extracted from a graph. In this study, 12 patients were allocated to radiant heater, of whom 10 had temperatures that fell to 35.0ºC; these patients took a mean of 25.1 minutes (SD 8.6) to regain a temperature of 35.5ºC; a further 23.3 minutes (SD 5.1) to regain 36ºC; and a further 25.2 minutes (SD 6. 8) to raise the temperature to 36.5°C. 12 patients were allocated to an electric blanket, of whom 11 had temperatures that fell to 35ºC; these patients took a mean of 54.4 minutes (SD 34.0) to regain a temperature of 35.5ºC and a further 44.8 minutes (SD 13.0) to regain 36ºC; and a further 41.9 minutes (SD 22.5) to raise the temperature to 36.5°C. These results compared with the 11 patients in the control group, of whom 10 had temperatures that fell to 35.0ºC; these patients took a mean of 59.7 minutes (SD 40.3) to regain a temperature of 35.5ºC; a further 51.2 minutes (SD 20.2) to regain 36ºC and 11 took a further 48.5 minutes (SD 25.6) to raise the temperature to 36.5°C.

Figure 12. Time to raise the temperature.

Figure 12

Time to raise the temperature.

The meta-analysis of the two comparisons showed some heterogeneity in the time to raise the temperature from 35.5 to 36.0°C (I2=61%; p=0.11), and a subgroup analysis by type of active warming was carried out.

Figure 13a. Time to raise temperature (subgroup analysis for radiant heaters).

Figure 13a

Time to raise temperature (subgroup analysis for radiant heaters).

For the radiant heater (1000W) versus usual care, in 23 patients, there was a statistically significant difference between interventions and the difference in time taken to raise the temperature by 0.5°C was similar for the different initial temperatures.

Figure 13b. Time to raise temperature (subgroup analysis for electric blankets).

Figure 13b

Time to raise temperature (subgroup analysis for electric blankets).

For the electric blanket (50W) intervention versus usual care, in 23 patients, there was no significant difference between electric blanket and usual care in the time taken to raise the temperature from 35.0°C to 35.5°C; 35.5°C to 36.0°C or 36.0°C to 36.5°C.

3. Rate of temperature change (°C/h)

The indirect study Bräuer (2004) in post-bypass patients recorded the median rate of increase of temperature, giving p values for the difference. These were converted to standard errors and used in the generic inverse variance option of the Review Manager software. There were 20 patients in each comparison.

Bräuer (2004) compared two forced air warming blankets, and two radiant heaters with a polyester filled blanket, and found statistically significant differences, compared with usual care, for both forced air warming blankets, but only for one radiant heater – a self assembled set of four 160W lamps. These differences in the median temperature were clinically significant.

Figure 14. Rate of warming.

Figure 14

Rate of warming.

4. Incidence of shivering

Two studies with 55 patients (Jackson 1997; Weyland 1994) assessed shivering in the recovery room, Jackson for different time periods postoperatively and Weyland over the whole monitoring period (Figure 15). Generally, the confidence interval is too wide to draw conclusions.

Figure 15. Incidence of shivering.

Figure 15

Incidence of shivering.

Postoperative complications

5. Patient’s perception of cold

One study with 18 patients (Alfonsi 2003) reported the patient’s perception of cold at the end of the forced air warming period (Figure 16).

Figure 16. Patient’s perception of cold.

Figure 16

Patient’s perception of cold.

There was a statistically significant difference in the number of patients perceiving that they were cold, although the confidence interval is wide. The relative risk was 0.38, i.e. about 3 times the risk for the control patients. This corresponds to a number needed to treat of 2 (95% CI 2, 6), for a control group risk of 89%.

6. Patient’s perception of pain

One study with 18 patients (Alfonsi 2003) reported the patient’s perception of pain at the end of the forced air warming period (Figure 17). There was no significant difference in the number of patients perceiving pain, although the confidence interval is fairly wide.

Figure 17. Patients’ perception of pain.

Figure 17

Patients’ perception of pain.

We considered it useful to have an estimate of how long it took patients to warm to 36.0°C in PACU (or ICU) under usual care. Therefore, the mean durations were estimated from the above studies. We note these are observational data based on very small numbers of patients.

Table 1. Time to raise temperature for the usual care group for different studies.

Table 1

Time to raise temperature for the usual care group for different studies.

B. Active warming 1 versus active warming 2

B1. Forced air warming versus warmed blanket

One often-used treatment for postoperative hypothermia is warmed blankets: these may be regularly changed, changed as needed, or not changed, but all are methods of active warming of the patients. Four studies compared forced air warming with warmed blankets (Giuffre 1991; Lennon 1990; Stevens 2000; Summers 1990).

1. Core temperature postoperatively

Two studies assessed the core temperature at different times postoperatively (Summers 1990; Lennon 1990). Lennon (1990) recorded oral temperatures; Summers (1990) had a baseline discrepancy of 0.38°C (higher for the warmed blanket). Both studies used a forced air warming device: the setting was 43°C for Lennon (1990), but was not reported for Summers (1990). Lennon (1990) warmed blankets to 37°C, but did not state if the blankets were changed; Summers (1990) did not state the temperature, but changed the blankets as needed. The results at different times are shown in Figure 18.

Figure 18. Temperature for different times.

Figure 18

Temperature for different times. NB: Scale −4 to +4°C

The two studies show significant heterogeneity at all times except 15 minutes. The Lennon (1990) study in 30 patients measured oral temperatures, which may not be closely related to core temperature; Summers (1990) had a baseline difference that was comparable or bigger than the difference in effect. This study also had significant difference in the time spent in theatre (the forced air warming group was longer by 35 minutes). It was decided to treat Summers (1990) as confounded, and draw tentative conclusions only from the Lennon (1990) study, even though this was not the best method of measuring temperature. This study showed that the forced air warming device was significantly more effective at rewarming than a warmed blanket that probably was not changed. We note that the oral temperature of the warmed blanket patients was low (below 35.0°C), indicating moderate hypothermia, and that the control group did not reach 36.0°C even after 75 minutes.

2. Time taken to increase the temperature

Two studies (Giuffre 1991; Stevens 2000) reported the time taken to increase the temperature to 36.0°C. The initial mean temperatures for the Guiffre (1991) study were 34.40°C (SD 0.42) and 34.43°C (SD 0.43) for the intervention and control groups respectively. Initial temperatures were not reported for Stevens (2000). The Guiffre (1991) study also employed a warmed head covering in all patients and recorded oral temperatures; the Stevens (2000) study recorded tympanic temperatures. For the forced air warming intervention, the setting was said to be medium, presumed to be 37°C for the Guiffre (1991) study, and ‘high’ for Stevens (2000). Both studies replaced the control group warmed blanket regularly (every 15 to 20 minutes); blankets in the Guiffre (1991) study were stored at 66 to 77°C and the temperature was not stated for Stevens (2000).

The results are shown in Figure 19. There was a significantly shorter time to 36.0°C for the forced air warming device in the Guiffre (1991) study, but this difference was not found in the Stevens (2000) study. Meta-analysis showed significant heterogeneity (I2=77%, p=0.04), but there is insufficient evidence to decide the cause of the heterogeneity. However, we note that the Stevens (2000) study had potential for bias because alternation was used to assign the treatments, although the Guiffre (1991) study only recorded oral temperatures and the confidence interval was wide.

Figure 19. Time to 36.0°C.

Figure 19

Time to 36.0°C.

3. Rate of temperature change (°C/h)

One study (Lennon 1990) in 30 patients recorded the rate of change in temperature, giving p values for the difference. These were converted to standard errors and used in the generic inverse variance option of Review Manager. Lennon 1990 found a statistically significant difference of 0.70°C/h (95%CI 0.25, 1.14) for forced air warming compared with blankets warmed to 37°C (figure 20).

Figure 20. Rate of increase in temperature.

Figure 20

Rate of increase in temperature.

4. Time to discharge from PACU

One study with two comparisons (Guiffre 1991) recorded the time to discharge from PACU. This was the time when the nurse judged the patient ready to leave rather than when the patient actually left. This study also employed a warmed head covering in all patients and recorded oral temperatures. There was no significant difference between interventions, although the confidence interval is fairly wide (Figure 21).

Figure 21. Time to discharge from PACU (minutes).

Figure 21

Time to discharge from PACU (minutes).

5. Incidence of shivering

Two studies with 121 patients (Lennon 1990; Summers 1990) assessed shivering in the recovery room, Lennon (1990) at different time periods and Summers (1990) over the whole monitoring period (Figure 22). Generally, the confidence interval was too wide to draw conclusions, although the Lennon (1990) study showed a significantly less shivering for the forced air warming group at 45 minutes and borderline significance at 15 minutes.

Figure 22. Incidence of shivering.

Figure 22

Incidence of shivering.

The Summers (1990) study also reported the duration of shivering (Figure 23). There was no significant difference between groups, although the confidence interval is fairly wide. We also note that the Summers (1990) study had significant differences in baseline characteristics (time spent in the operating room).

Figure 23. Duration of shivering.

Figure 23

Duration of shivering.

Postoperative Complications

6. Patient’s thermal comfort

One study with 91 patients (Summers 1990) reported the patient’s thermal comfort 30 minutes after forced air warming commenced and at the time of discharge (probably 60 minutes) (Figure 24). The scale used was the Christoph comfort scale, but it was not clear what this was. However, the study authors commented that the patients’ perception of comfort was greater for the forced air warming group, which indicates that a higher value on the scale is an improvement. A statistically significant difference was found in favour of the forced air warming group, but its magnitude is unclear. We also note that the Summers (1990) study had significant differences in baseline characteristics (time spent in theatre).

Figure 24. Thermal comfort.

Figure 24

Thermal comfort.

We also recorded the time to raise the temperature to 36.0ºC for the warmed blankets group. We note these are observational data.

Table 2. Time to raise the temperature for warmed blankets groups for different studies.

Table 2

Time to raise the temperature for warmed blankets groups for different studies.

B2. Radiant heating versus warmed blankets

One study compared radiant heating with warmed blankets (Giuffre 1991).

1. Time taken to increase the temperature

One study (Giuffre 1991) reported the time taken to increase the temperature to 36.0°C. The initial mean temperatures for the Guiffre (1991) study were 34.46 (SD 0.42) and 34.43°C (SD 0.43) for intervention and control groups respectively. The Guiffre (1991) study also employed a warmed head covering in all patients and recorded oral temperatures. The control group had warmed blankets replaced regularly (every 15 to 20 minutes); blankets were stored at 66°C to 77°C. There was no significant difference between groups, but the confidence interval is fairly wide.

Figure 25. Time to 36.0°C.

Figure 25

Time to 36.0°C.

2. Time to discharge from PACU

One study (Guiffre 1991) recorded the time to discharge from PACU. This was the time when the nurse judged the patient ready to leave rather than when the patient actually left. This study also employed a warmed head covering in all patients and recorded oral temperatures. There was no significant difference between interventions, although the confidence interval is fairly wide (Figure 26).

Figure 26. Time to discharge from PACU (minutes).

Figure 26

Time to discharge from PACU (minutes).

B3. Radiant heat versus electric blanket

One study in 24 patients (Weyland 1994) compared a radiant heater (1000W) with an electric blanket (50W). Individual patient data were extracted from a graph. The oesophageal temperature was recorded.

1. Time to raise temperature

Weyland (1994) recorded the time to increase the temperature from 35.0°C to 35.5°C; 35.5°C. to 36.0°C and 36.0°C to 36.5°C. In this direct comparison, warming was significantly faster (17 to 29 minutes) for the radiant heater compared with the electric blanket, for all ranges of temperature.

Figure 27. Time to raise temperature.

Figure 27

Time to raise temperature.

2. Incidence of shivering

Weyland (1994) reported shivering over the warming period and showed significantly less shivering for the radiant heater, although the confidence interval is wide. The risk of shivering is 1/5th that of the electric blanket, and the NNT is 2 (95% CI 2, 4) for a control group risk of 75%.

Figure 28. Incidence of shivering.

Figure 28

Incidence of shivering.

B4. Forced air warming versus radiant heating

One indirect study in post-bypass patients compared two forced air warming devices (Bair Hugger® and Warm Touch®, setting and flow maximal for both) and two radiant heaters (Thermal Ceilings 1000W and self assembled 4 x 160W) (Bräuer 2004). Oesophageal temperatures were raised from below 35.5ºC to above 37.5ºC.

1. Rate of change of temperature

The study reported medians with 10th and 90th percentiles. These are reported in the table below. The authors reported no significant difference between any of these active interventions.

Table 3. Rate of change of temperature.

Table 3

Rate of change of temperature.

C. Active warming versus thermal insulation

One study (Bredahl 1995) compared a radiant heater (500W) with a reflective blanket in 30 patients who had undergone combined general and regional anaesthesia. Both arms of the study had warmed IV fluids and the rectal temperature was measured.

1. Core temperature after two hours in PACU

The Bredahl (1995) study reported the median and interquartile range at 15 minute intervals in a graphical form, and it is clear that the rate of change of median rectal temperature is greater for the radiant heating group compared to the reflective blanket group. The authors reported that the increase in median temperature over two hours was significantly greater for the radiant heating group (1.6ºC in the radiant heater group; 0.9ºC in the thermal insulation group; p<0.05).

2. Shivering

There was little difference in the number of patients shivering, either over the whole warming period (Figure 29) or at any time during it. The confidence interval is fairly wide.

Figure 29. Incidence of shivering.

Figure 29

Incidence of shivering.

D. Thermal insulation versus usual care

One study with 144 patients (Hershey 1997) compared different types of thermal insulation as an adjunct to the use of two warmed thermal blankets. The study did not mention if the blankets were changed, and their warming temperature was not stated. Patients were randomised to one of the following three interventions:

  • Two warmed thermal blankets (group 1; control group);
  • Reflective blanket plus two warmed thermal blankets (group 2);
  • Reflective blanket + reflective head covering + two warmed thermal blankets (group 3).

There were some reservations about the study: it measured oral temperatures, and the reflective blanket group (group 2) spent longer in theatre than either of the other groups (mean duration in theatre 184 minutes for control group; 233 minutes for group 2 and 201 minutes for group 3; SDs not given).

1. Time to raise the temperature

Hershey (1997) reported the time taken to reach 36.0ºC from an initial mean temperature of 34.8ºC or 35.0ºC (Figure 30). There was little difference between interventions, and the addition of a reflective blanket and hat did not appear to help in reducing the time taken to reach 36.0ºC.

Figure 30. Time to raise temperature.

Figure 30

Time to raise temperature.

Summary of times to raise the temperature for various warming mechanisms

Table 4 summarises the times taken to raise the temperature for different warming mechanisms for all the trials. We note that these are observational data, usually in small numbers of patients, but are included as an indication of how long it takes to rewarm hypothermic patients in PACU and ICU.

Table 4. Time to raise temperature for patients given forced air warming and radiant heat.

Table 4

Time to raise temperature for patients given forced air warming and radiant heat.

Presumably this was a misprint and should have read 37°C

Footnotes

#

Presumably this was a misprint and should have read 37°C

Copyright © 2008, National Collaborating Centre for Nursing and Supportive Care.
Bookshelf ID: NBK53777
PubReader format: click here to try

Views

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...