The Effect of Patient Warming During Caesarean Delivery
The Effect of Patient Warming During Caesarean Delivery
For this meta-analysis, we analysed randomized controlled trials comparing active warming techniques (specifically forced air warming or warmed fluid) to no warming before and during elective Caesarean delivery, and followed PRISMA guidelines. We conducted a literature search with no language restriction on January 16, 2014 and repeated the search on August 27 and December 3, 2014. Searches were performed in PubMed (1950 to August 2014), Ovid EMBASE (1970 to December 2014), Ovid MEDLINE (1950 to December 2014), Scopus (1960 to December 2014), EBM Reviews Cochrane Central Register of Controlled Trials 2nd Quarter 2014, clinicaltrials.gov, and CINAHL (December 2014). We consulted the clinical trials registry (www.clinicaltrials.gov) on August 27, 2014 to identify any unpublished studies. The search strategy consisted of a combination of subject headings (obstetric, Caesarean) and keywords/key phrases (temperature, warming, Caesarean) for each of MEDLINE, EMBASE, and CINAHL searched in specified fields (such as ti=title/ab=abstract). In the event that a database did not index articles, we conducted keyword searching in the entire record (see Appendix 1 for detailed PubMed search criteria). Reference lists of all identified studies were also checked.
All randomized controlled trials utilizing forced air warming or warmed fluid were considered. We included studies comparing groups that commenced warming from within 30 min of neuraxial anaesthesia placement up to and including warming in the post anaesthetic care unit. We excluded studies using general anaesthesia and other methods that may minimize perioperative hypothermia including various intrathecal opioid doses, leg wrapping, warmed intrathecal drugs, different anaesthetic techniques, and increased ambient temperature. Studies were also excluded if they did not report maternal or neonatal outcomes. The quality of studies included in the meta-analysis was reviewed using the Cochrane Collaboration's tool for assessing risk of bias. Areas of methodological quality assessed included concealment of allocation, random sequence generation, blinding of the assessors and participants, and accounting for all subjects. Overall quality was graded as low (high risk of bias), high (low risk of bias), or unclear risk of bias for each domain entry using a standardized tool. At least two individuals extracted the study data independently utilizing a standardized review protocol and recorded the information on a data collection sheet. Differences were resolved by re-examination of the original manuscripts and by discussion with a third investigator. The data were then entered into a computer by one of the authors (Y.C.) and checked by a second investigator (P.S.).
The primary outcome was the maximum temperature change in the perioperative period. For the purposes of this study, the perioperative period was defined as the time from 30 min before anaesthesia to 15 min after arrival on the post anaesthetic care unit. Secondary outcomes included (1) temperature at the end of surgery or on admission to the post anaesthetic care unit (2) shivering (3) nausea and vomiting (4) thermal comfort (5) hypothermia (6) hypotension (7) vasopressor use (8) neonatal temperature at delivery (9) umbilical cord blood pH and (10) Apgar scores at 1 and 5 min.
Data were analysed using the Review Manager software (RevmanVersion 5.3.5 Copenhagen: the Nordic Cochrane Centre, The Cochrane Collaboration, 2014), CMA (comprehensive meta-analysis, Version 2, 2005), and R routine metacont (R package Meta). We calculated pooled estimates for all studies combined and also performed a subgroup analysis according to warming modality used (forced air warming or fluid warming). We compared subgroups using the Q test. For dichotomous outcomes, the risk ratio (RR) and 95% confidence interval (CI) were calculated (a RR<1 favoured warming). In addition, the number needed to treat (NNT) was calculated for statistically significant dichotomous outcomes. For continuous data, the standardized mean difference (SMD) or mean difference (MD) and 95% CI were determined. The MD was used for all continuous outcomes except when the data available from the included studies were in different formats. This applied to the outcome of temperature change, where data was available either as a mean (sd) temperature change or as baseline temperature and post intervention temperature, and the outcome of thermal comfort where two different scales were used by the included studies. The percentage of heterogeneity was assessed with the I statistic. Significant heterogeneity was assumed to be present if I>50%. For the primary outcome we explored significant heterogeneity, by performing sensitivity analyses, excluding studies with methodological differences according to type of neuraxial technique or site of forced air warming. Publication bias for the primary outcome was assessed using funnel plots and Egger's test. In case of funnel plot asymmetry, a contour-enhanced funnel plot was examined to further assess for publication bias. A P value <0.05 was considered statistically significant. All data were combined and analysed using the DerSimonian-Laird random effects model.
Methods
For this meta-analysis, we analysed randomized controlled trials comparing active warming techniques (specifically forced air warming or warmed fluid) to no warming before and during elective Caesarean delivery, and followed PRISMA guidelines. We conducted a literature search with no language restriction on January 16, 2014 and repeated the search on August 27 and December 3, 2014. Searches were performed in PubMed (1950 to August 2014), Ovid EMBASE (1970 to December 2014), Ovid MEDLINE (1950 to December 2014), Scopus (1960 to December 2014), EBM Reviews Cochrane Central Register of Controlled Trials 2nd Quarter 2014, clinicaltrials.gov, and CINAHL (December 2014). We consulted the clinical trials registry (www.clinicaltrials.gov) on August 27, 2014 to identify any unpublished studies. The search strategy consisted of a combination of subject headings (obstetric, Caesarean) and keywords/key phrases (temperature, warming, Caesarean) for each of MEDLINE, EMBASE, and CINAHL searched in specified fields (such as ti=title/ab=abstract). In the event that a database did not index articles, we conducted keyword searching in the entire record (see Appendix 1 for detailed PubMed search criteria). Reference lists of all identified studies were also checked.
All randomized controlled trials utilizing forced air warming or warmed fluid were considered. We included studies comparing groups that commenced warming from within 30 min of neuraxial anaesthesia placement up to and including warming in the post anaesthetic care unit. We excluded studies using general anaesthesia and other methods that may minimize perioperative hypothermia including various intrathecal opioid doses, leg wrapping, warmed intrathecal drugs, different anaesthetic techniques, and increased ambient temperature. Studies were also excluded if they did not report maternal or neonatal outcomes. The quality of studies included in the meta-analysis was reviewed using the Cochrane Collaboration's tool for assessing risk of bias. Areas of methodological quality assessed included concealment of allocation, random sequence generation, blinding of the assessors and participants, and accounting for all subjects. Overall quality was graded as low (high risk of bias), high (low risk of bias), or unclear risk of bias for each domain entry using a standardized tool. At least two individuals extracted the study data independently utilizing a standardized review protocol and recorded the information on a data collection sheet. Differences were resolved by re-examination of the original manuscripts and by discussion with a third investigator. The data were then entered into a computer by one of the authors (Y.C.) and checked by a second investigator (P.S.).
The primary outcome was the maximum temperature change in the perioperative period. For the purposes of this study, the perioperative period was defined as the time from 30 min before anaesthesia to 15 min after arrival on the post anaesthetic care unit. Secondary outcomes included (1) temperature at the end of surgery or on admission to the post anaesthetic care unit (2) shivering (3) nausea and vomiting (4) thermal comfort (5) hypothermia (6) hypotension (7) vasopressor use (8) neonatal temperature at delivery (9) umbilical cord blood pH and (10) Apgar scores at 1 and 5 min.
Data were analysed using the Review Manager software (RevmanVersion 5.3.5 Copenhagen: the Nordic Cochrane Centre, The Cochrane Collaboration, 2014), CMA (comprehensive meta-analysis, Version 2, 2005), and R routine metacont (R package Meta). We calculated pooled estimates for all studies combined and also performed a subgroup analysis according to warming modality used (forced air warming or fluid warming). We compared subgroups using the Q test. For dichotomous outcomes, the risk ratio (RR) and 95% confidence interval (CI) were calculated (a RR<1 favoured warming). In addition, the number needed to treat (NNT) was calculated for statistically significant dichotomous outcomes. For continuous data, the standardized mean difference (SMD) or mean difference (MD) and 95% CI were determined. The MD was used for all continuous outcomes except when the data available from the included studies were in different formats. This applied to the outcome of temperature change, where data was available either as a mean (sd) temperature change or as baseline temperature and post intervention temperature, and the outcome of thermal comfort where two different scales were used by the included studies. The percentage of heterogeneity was assessed with the I statistic. Significant heterogeneity was assumed to be present if I>50%. For the primary outcome we explored significant heterogeneity, by performing sensitivity analyses, excluding studies with methodological differences according to type of neuraxial technique or site of forced air warming. Publication bias for the primary outcome was assessed using funnel plots and Egger's test. In case of funnel plot asymmetry, a contour-enhanced funnel plot was examined to further assess for publication bias. A P value <0.05 was considered statistically significant. All data were combined and analysed using the DerSimonian-Laird random effects model.
Source...