Search results and characteristics of the studies
The flow chart (Figure 1) shows the process by which the selected studies were searched. A total of 361 potential articles were identified. We reviewed 30 full-text articles, after screening the titles and the abstracts. A total of 16[10–25] studies including 1485 patients met our selection criteria and were included in the analysis (Table 1).
In 15 trials, participants were adults. One trial [23] included children aged 5–12 years. Participants in seven trials [10,12,13,18,19,23,24] underwent operations under general anesthesia; participants in nine trials [11,14,15,16,17,20,21,22,25] were under spinal anesthesia. In 13 trials[10,12,14,15,16,17,18,19,20,21,22,23] ketamine was compared to placebo; in 4 trials[12,13,23,24] ketamine was compared to pethidine; in the other 4 trials[11,15,16,22] ketamine was compared to tramadol. Ketamine was also compared to ondansetron in 4 trials[10,15,17,21].
The administration time and routes were different among included trials. In 10 trials [11,14,15,16,17,20,21,22,23,25] the intervention drugs were given immediately after induction of anesthesia or intrathecal injection; in five trials [10,12,18,19,24] drugs were administrated before completion of the surgical procedure; in one trial [13] patients received study drugs before wound closure. Study drugs were given as an IV bolus in 14 trials [10–22,24]; in two trials, patients received the study drugs epidurally [25] or intramuscularly [24]. In four trials [14,15,18,19], patients underwent orthopedic surgery; patients in two studies [21,22] underwent abdominal surgery; patients in two studies[16,17] underwent cesarean section; patients in two studies[20,25] underwent urological surgery; in three trials, participants received ENT surgery[10], endoscopic sinus surgery[12], or tonsillectomy surgery[23].
Regarding measurement of intensity of shivering in a patient, 13 trials [10,13–17,19–25] utilized a scale with variation points ranging from 0–4: 0 = no shivering; 1 = piloerection or peripheral vasoconstriction, but no visible shivering; 2 = muscular activity in only one muscle group; 3 = muscular activity in more than one muscle group, but not generalized; 4 = shivering involving the whole body. One trial [12] applied a 0–3 scale for evaluating the intensity: 0 = no shivering; 1 = mild fasciculation of face or neck muscles; 2 = visible tremor involving more than one muscle; 3 = gross muscular activity involving the entire body. Two studies [11,18] did not assess the intensity of postanesthetic shivering.
Assessment of the risk of bias in the included studies
Two authors (Y. Z., A.M.) independently assessed the following domains using the Cochrane ‘Risk of bias’ tool:
• Sequence generation
• Allocation concealment
• Blinding of participants, personnel
• Blinding of outcome assessment
• Incomplete outcome data
• Selective outcome reporting
• Other bias
We completed ‘Risk of bias’ figures for each included study (Figure 2). See more details in Appendix.
Publication bias
Begg’s test showed that there was no publication bias for the primary outcome (p = 0.055).
Effects of interventions
Primary outcome
Ketamine vs placebo The incidence of postanesthetic shivering was compared between ketamine and a placebo in 13 trials including 1166 patients (Figure 3). Ketamine has been shown to significantly decrease the incidence of shivering (pooled OR = 0.13; 95% CI: 0.06 to 0.26, P<0.00001). There was significant and prominent heterogeneity for this outcome (I2 = 74%). Begg’s test showed that there was no risk of publication bias (P = 0.06). A subgroup analysis was performed to explore the evidence-based reason. In subgroup analysis of anesthetic methods, the heterogeneity was 67% in the GA (general anesthesia) group (Figure 4.). Sensitivity analysis was performed; a trial[22] was removed which utilized an air forced warmer intraoperatively, which showed a similar result favoring ketamine (pooled OR = 0.18; 95% CI: 0.09 to 0.37) and decreased heterogeneity (I2 from 67% to 21%). Ketamine reduced the incidence of postanesthetic shivering in general anesthesia (pooled OR = 0.13; 95% CI: 0.06 to 0.26) and in spinal anesthesia (pooled OR = 0.08; 95% CI: 0.03 to 0.18). (Figure 5) shows the subgroup analysis based on the dose of ketamine used in the included trials. Ketamine reduced the incidence of postanesthetic shivering at the dose of 0.25mg/kg (pooled OR = 0.12; 95% CI: 0.03 to 0.52) and at the dose of 0.5mg/kg (pooled OR = 0. 14; 95% CI: 0.07 to 0.28). We performed a subgroup analysis based on the type of surgery, as this may influence the incidence of postanesthetic shivering (Figure 6). Ketamine significantly lowered the incidence of postanesthetic shivering in patients after orthopedic surgery (pooled OR = 0.32; 95% CI: 0.13 to 0.77). Among patients undergoing abdominal, cesarean section, urological, ENT or endoscopic surgery, ketamine also reduced the incidence of postanesthetic shivering.
Ketamine vs other pharmacological interventions A total of four studies [12,13,23,24] investigated the effect of ketamine on the prevention of shivering compared with pethidine. The pooled analysis showed a definite difference in favor of pethidine (pooled OR = 4.38; 95% CI: 1.76 to 10.92). No significant difference in postanesthetic shivering was found between ketamine and other pharmacological interventions (Figure 7).
Secondary outcomes
Ketamine vs placebo Except for the other side effects (hypotension, bradycardia, hallucination), there was no significant difference in the incidence of postoperative nausea and vomiting (PONV) between ketamine and the placebo (pooled RR = 0.72; 95% CI: 0.48 to 1.08) (Table 3). Ketamine reduced the incidence of hypotension and bradycardia compared with the placebo (pooled RR = 0.28; 95% CI: 0.17 to 0.47; pooled RR = 0.18; 95% CI: 0.05 to 0.65). The incidence of hallucination was more significant and prevalent in the patients who received 0.5mg/kg ketamine (Table 4); there was only one episode of hallucination in patients receiving 0.25mg/kg. Seven studies [14,15,16,20,21,22,25] reported the sedation score of patients on a five-point scale: 1 = fully awake and oriented, 2 = drowsy, 3 = eyes closed but arousable to command, 4 = eyes closed but arousable to mild physical stimulation, and 5 = eyes closed but unarousable to mild physical stimulation. A pooled analysis was not performed because of the lack of uniform sedation score scales in the trials. All of these studies showed that the patients in the ketamine group were more sedated compared to the placebo group. Five trials [14,16,17,19,23] reported a significant decrease in core temperature in both the ketamine and placebo groups compared to the baseline temperature of participants. However, it was not significant between groups, at any time point. Three trials [20,21,22] reported a significant difference in core temperature between ketamine and the placebo; a greater decrease in temperature was found in the placebo group.
Ketamine vs other pharmacological interventions No significant difference was found in the incidence of PONV between ketamine and pethidine (pooled OR = 0.88; 95% CI: 0.38 to 2.07). Compared to tramadol, the difference in the incidence of PONV and hypotension is not significant (OR = 0.57; 95% CI: 0.18 to 178; OR = 0.90; 95% CI: 0.36 to 2.24). The incidence of PONV was higher in the ketamine group than the ondansetron group (OR = 4.49; 95% CI: 1.24 to 16.21). However, ketamine showed a lower incidence of hypotension compared to ondansetron (OR = 0.09; 95% CI: 0.00 to 3.23). Core temperature changes were reported graphically; there was no significant difference between ketamine and other pharmacological interventions.
Summary of findings and quality of evidence The Summary of findings with GRADE recommendations are shown in Table 2.