Ketamine-propofol (Ketofol) for procedural sedation and analgesia in children: a systematic review and meta-analysis

Objectives The aim of this review is to elucidate the efficacy and side effects of ketofol in comparison to other anaesthetic agents during procedural sedation and analgesia. Method The Cochrane Central Register of Controlled Trials (1996 to Feb 2019) and MEDLINE (1966 to Feb 2019) were searched, including the related randomised control trials and reviewed articles to find unpublished trials or trials not obtained via electronic searches. Inclusion criteria for the studies included comparing recovery time, recording clinician satisfaction, and assessing the adverse effects of ketofol. Results Eleven trials consisting of a total of 1274 patients met our criteria and were included in this meta-analysis. Five trials compared ketofol with a single agent, while six trials compared ketofol with combined agents. While comparing between ketofol and a single agent (either ketamine or propofol), ketofol showed significant effect on recovery time (MD: -9.88, 95% CI: − 14.30 to − 5.46; P = 0.0003; I2 = 92%). However, no significant difference was observed while comparing ketofol with combined agents (RR: 0.75, 95% CI: − 6.24 to 7.74; P < 0.001; I2 = 98%). During single-agent comparison, ketofol showed no significant differences in terms of clinician satisfaction (RR: 2.86, 95% CI: 0.64 to 12.69; P = 0.001; I2 = 90%), airway obstruction (RR: 0.72, 95% CI: 0.35 to 11.48; P = 0.81; I2 = 0%), apnoea (RR: 0.9, 95% CI: 0.33 to 2.44; P = 0.88; I2 = 0%), desaturation (RR: 1.11, 95% CI: 0.64 to 1.94; P = 0.28; I2 = 21%), nausea (RR: 0.52, 95% CI: 0.91 to 1.41; P = 0.2; I2 = 38%), and vomiting (RR: 0.63, 95% CI: 0.25 to 1.61; P = 0.18; I2 = 42%). During comparison with combined agents, ketofol was more effective in reducing hypotension (RR: 4.2, 95% CI: 0.2 to 0.85; P = 0.76; I2 = 0%), but no differences were observed in terms of bradycardia (RR: 0.70, 95% CI: 0.14 to 03.63; P = 0.09; I2 = 53%), desaturation (RR: 1.9, 95% CI: 0.15 to 23.6; P = 0.11; I2 = 61%), and respiratory depression (RR: 1.98, 95% CI: 0.18 to 21.94; P = 0.12; I2 = 59%). Conclusion There is low certainty of evidence that ketofol improves recovery time and moderate certainty of evidence that it reduces the frequency of hypotension. There was no significant difference in terms of other adverse effects when compared to other either single or combined agents. Trial registration PROSPERO CRD42019127278.


(Continued from previous page)
Trial registration: PROSPERO CRD42019127278.
Keywords: Children, Paediatric, Ketofol, Ketamine-propofol, Procedural sedation and analgesia, Systematic review, Meta-analysis Background Procedural sedation and analgesia (PSA) is a treatment strategy involving the administration of agents with sedative, analgesic, or dissociative properties to suppress a patient's consciousness to varying degrees. It facilitates the completion of painful procedures, while ensuring the safety and comfort of the patient without compromising airway patentability [1]. The demand of PSA outside of the operation theatre (OT) has grown exponentially for both diagnostic and therapeutic purposes, especially in the emergency department (ED) and intensive care unit (ICU) [2], as the clinician develops confidence in managing the sedated patients, especially the paediatric population [3]. Various types of unpleasant procedures require optimal sedation and analgesia in order to increase the procedure's success rate, clinician satisfaction, and to reduce the discomfort and anxiety of the patients, especially the paediatric population. As a standard of healthcare, the American College of Emergency Physicians recommends effective management of anxiety and discomfort associated with such procedures [1]. In an era of increased emphasis on maintaining quality and minimising costs, PSA outside the OT always offers significant advantages in terms of delivering timely and cost-effective care to the patient [4].
Ideal PSA agents should have predictable effects on the regulation of the sedation, analgesia, amnesia, and motor function. In addition, it is beneficial to administer an agent with fast induction, steady control, quick recovery time, and minimal side-effects. Unfortunately, despite the availability of numerous anaesthetic agents, no single agent demonstrates superiority over the others. Thus, various combinations of different sedative, analgesic, or dissociative agents have been considered in order to optimise the desired effect and minimise the adverse effect [5]. However, PSA in the paediatric population remains a challenge, as the sedation risk is higher in children [3]. The anatomical and physiological differences between the paediatric and adult airways make it potentially more difficult and complicated to manage the former. Medical problems affecting the airway, breathing, circulation, and neurological function play an important role in increasing the risk of PSA. The Pediatric Sedation Research Consortium (http://www.pedsedation.org) in the United States of America (USA) is a multi-centre registry involving over 30 centres that practice PSA [6]. It reported that there were a total of 1020 adverse events over 30,000 PSA procedures undertaken outside OT with no death, one cardiac arrest, and required cardiopulmonary resuscitation. Less serious events were more commonly associated with desaturation, which occurring 157 times over 10,000 sedation procedures. Unexpected apnoea, excessive secretions, and vomiting had frequencies of 24.0, 41.6, and 47.2 per 10,000 procedures, respectively [7].   A large number of ED in the USA have adopted ketofol as their primary sedation regime [8]. Both drugs exhibit a synergistic reaction to each other and theoretically offset each other's adverse effects. A combination of the drugs allows a smaller dose of each one, thus potentially improving the quality, safety, and duration of recovery time [9]. A recent study had shown that ketofol reduces complications in PSA in adults compared to propofol alone [5]. But the superiority and safety of ketofol in the paediatric population are still debatable. If it is proven to be a safe and effective anaesthetic agent in PSA in children, it should be considered as a primary choice. This study was conducted to elucidate the efficacy and role of ketofol in comparison to other anaesthetic agents in PSA in paediatric patients. We used the search strategy in Appendix for CENTRAL and MEDLINE searches. We restricted to the publications in English language. We searched the reference list of known randomised control trials (RCTs) and reviewed papers to identify unpublished trials or trials that are not found by electronic searches. We searched the World Health Organization International Clinical Trials Registry Platform (ICTRP) (http:/www.who.int/ictrp/en/) and www.clinicaltrials.gov for on-going trials. All the RCTs or clinical control trials comparing ketofol with other anaesthetic agents were included in the study. We included blinded and open-labelled studies including patients below 18 years old who underwent procedural sedation. The primary outcomes were recovery time and satisfaction of the clinician. Secondary outcomes included adverse events (such as nausea, vomiting, airway obstruction, apnoea, desaturation, respiratory depression, hypotension, and bradycardia) and hemodynamic parameters (heart rate, respiratory rate, and blood pressure). The clinical outcomes were defined based on the assessment of the managing clinicians.
We evaluated the search titles and abstracts and accessed full-text articles when they appeared to meet the eligibility requirements, or when the title or the abstract provided insufficient information to determine the eligibility. We independently assessed the validity of the trials and reported the reasons for the exclusion. We resolved any disputes through discussions between the review authors. If clarification was required, we contacted the authors of the publications. Using data extraction form, from each of the selected trials, we extracted data on study setting, participant characteristics (age, sex, and ethnicity), methodology (number of participants randomised and analysed), type and dosage of anaesthetic agent, recovery time, clinicians' and parents' satisfaction, the occurrence of related adverse events, and hemodynamic parameters, independently.
We evaluated the parameters of risk of bias, such as random sequence generation, allocation concealment,  blinding of participants and personnel, blinding of outcome assessors, completeness of data, selectivity of outcome reporting, and other biases [10]. We performed the risk of bias assessment independently and resolved any disagreements by discussion. We assessed the quality of evidence for primary and secondary outcomes according to the Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology [11] for risk of bias, inconsistency, indirectness, imprecision, and publication bias; classified as very low, low, moderate, or high. We measured the treatment effect for dichotomous outcomes using risk ratios (RRs) and absolute risk reduction, and for continuous outcomes, we used mean *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio.

GRADE (Working Group grades of evidence)
High certainty: We are very confident that the true effect lies close to that of the estimate of the effect.
Moderate certainty: We are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.
Footnote: a Duration of treatment varies following different procedures, and thus, required different doses of intervention b Small sample size c Subjective outcome with different types of population Fig. 5 Comparison of ketofol vs. single-agent control with respect to clinician satisfaction differences (MDs), both with 95% confidence intervals (CIs). We reported the results of the random-effects model. We examined for a unit of analysis errors on the included trials. Unit of analysis errors can occur when trials randomise participants to intervention or control groups in clusters but analysed the results using the total number of individual participants. We adapted the test results showing a unit of analysis errors based on the mean cluster size and coefficient of the intracluster correlation, if any [10]. We intended to contact the original authors of the trials to request missing or improperly reported data. If missing data were not available, we analysed the available data. We undertook metaanalyses using Review Manager (RevMan) version 5.3.5 (Nordic Cochrane Centre, Cochrane Collaboration).
In two measures, we tested the existence of the heterogeneity. First, through examining populations, conditions, interventions and results, we measured the apparent heterogeneity at face value. Second, we analysed statistical heterogeneity using the statistic I 2 . We used the heterogeneity analysis guide, as follows: May not be significant between 0 and 40%, moderate heterogeneity between 30 and 60%; substantial heterogeneity between 50 and 90%, and considerable heterogeneity between 75 and 100% [10]. If there were enough trials, we intended to use a funnel plot to explore the possibility of reporting biases or small trial biases, or both. We conducted a sensitivity analysis to analyse the influence of bias risk for sequence generation and allocation concealment.
The assessment of the risk of bias is shown in Figs. 2 and 3. Figure 2 shows the proportion of studies with low, high, or unclear risk of bias for each risk of bias domain. Figure 3 shows the risk of bias summary for individual studies. All the trials described the method of randomisation used. For randomisation of participants, five trials used the closed envelope method [14,16,20,22,23]; four trials used the computer-generated method [13,17,18,21]; one trial used the coin toss method [19]; one trial used the 1:1 block randomisation with different strata based on the age and type of magnetic resonance imaging (MRI) [15]. Blinding of participants and personnel was not described in four trials [13,19,22,23]. Six trials analysed all the samples without any withdrawal [16,[19][20][21][22][23]. Five trials carried the intention to treat an analysis in which the participants were analysed according to the group that they were initially assigned [13-15, 17, 18]. All trials reported the outcomes as specified in their respective Methods sections. We detected no other potential source of bias.
A total of 11 trials measured the primary outcome, i.e., recovery time [13][14][15][16][17][18][19][20][21][22][23]. One study defined recovery time as time, in minutes, required for the patient to be conscious and responding to verbal stimuli, airway recovery as the return of gag reflex or cough, and motor recovery as a purposeful movement of limbs [21]; another study defined recovery time as the period needed by the patient to regain consciousness spontaneously [13]; four studies used Steward Recovery Score to define the recovery time but with different cut off points, of which two studies used score of 6 [20,22] while another two studies used score of 7 [16,19]; one study defined time from discontinuation of infusion and achievement of Ramsey Sedation Score of 3 as recovery time [14]; another study used Modified Vancouver Sedation Recovery Scale to determine the recovery time [23].
Three trials measured the satisfaction of clinicians [17,19,23]. Nine trials measured the secondary outcome, i.e., adverse effects [13-19, 21, 23]. Our protocol intended to report hemodynamic parameters as a secondary outcome. These were not analysed because they either involved comparison in a different unit or a non-comparable group [13][14][15]23].. Four trials were not included in the hemodynamic status data because their data were demonstrated in the graph [16,19,20,22].

GRADE (Working Group grades of evidence)
High certainty: We are very confident that the true effect lies close to that of the estimate of the effect.
Moderate certainty: We are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

Discussion
The concept of combination of ketamine and propofol is based on the synergistic effects and the benefits they provide while countering each-other's side effects. This review was designed to include all RCTs addressing the effectiveness of ketofol in PSA in the paediatric population compared to other analgesic agents. Ketofol showed a significant effect on recovery time compared to singleagent but showed no differences when compared to a combined agent. The subgroup analyses (different ratios of dosages) showed no differences. However, three studies with small sample sizes in single-agent analyses were not included in our analysis due to a non-usable format that discouraged the generation of an appropriate and meaningful conclusion. Ketofol showed no effect on clinician satisfaction and respiratory adverse events (airway obstruction, apnoea, desaturation, and respiratory depression) in comparisons with both single and combined agents. In adverse cardiovascular events, ketofol reduced the frequency of hypotension but showed no effect on bradycardia. There was no significant difference between the gastrointestinal adverse events (nausea and vomiting) of ketofol and control groups.
The quality of trial evidence was variable. Generally, there was a low or unclear risk of bias for most trials in most domains. During single-agent comparison, the certainty of the evidence for recovery time was low. We downgraded for inconsistency and imprecision as we noted substantial statistical heterogeneity and small sample size. During combined agent comparison, the certainty of the evidence for recovery time was moderate. We downgraded because of apparent inconsistencies. The quality of the evidence for both subgroup analyses for dosage ratios 1:1 and 1:2 was downgraded two levels due to heterogeneity and small sample size. A lowquality rating was assigned to the pooled estimates of effect for the outcome of clinician satisfaction due to small sample size and large heterogeneity.
For the outcome of adverse events of nausea, vomiting, airway obstruction, apnoea, and desaturation in singleagent comparison, the overall quality of the evidence was judged to be high. However, during combined agent comparison, the quality of the evidence for respiratory depression, desaturation, and bradycardia was judged to be low. The evidence for these outcomes was from small sample size trials, and our certainty in the evidence was reduced owing to the imprecision. We also noted the unexplained difference between study data, and thus, we downgraded for inconsistency. Evidence for hypotension was obtained from a few studies and few participants. Therefore, the evidence was imprecise, and this reduced our certainty in the evidence to moderate (Table 3). Clinical heterogeneity was anticipated, and therefore, we explored the statistical heterogeneity through subgroup analysis. Specifically, we analysed on the basis of the different ratios of dosages in the mixture where it was best reflecting the interventions present in the analysed data.
Prior to our meta-analysis, there is a lack of comprehensive review regarding the use of ketofol in PSA in the paediatric population. We found a systematic review and meta-analysis that compared ketofol with propofol alone in adults and children, which included six RCTs with a total of 932 patients [24]. In the review, two RCTs were peer-reviewed abstracts selected for presentation at international conferences, one of these RCTs included paediatric patients aged 3 to 18 years old, which was excluded from our study [25] due to the absence of a fulltext article. The above-mentioned review only included ED patients undergoing PSA for any painful non-elective procedures; whereas, the current review included both ED patients and those undergoing elective procedures. They concluded a significantly higher frequency of adverse respiratory events in cases involving ketofol compared to those involving propofol. However, our study showed no such difference compared to the control group. Their results with respect to recovery time were inconclusive owing to the non-usable data of the included RCTs.
Another systematic review and meta-analysis that compared ketofol to propofol included all relevant RCTs, regardless of patient's age, sex, location, publication year, and language [5]. The review reported outcomes of various adverse events but did not describe recovery time. Contrary to our findings, the outcome of respiratory complications that require intervention and bradycardia, was significant compared to the control group. However, the findings related to other outcomes of hypotension, nausea, and vomiting were similar to those obtained in the current review.

Limitations
This meta-analysis has several limitations. First, our study demonstrates significant heterogeneity for the outcome of recovery time due to multiple factors, including the variety among the patient populations, type and duration of different procedures, a different dosage of bolus and maintenance with a varying method of mixture. Most of the studies did not use standardised and quantifiable endpoints to define recovery time. However, exploration of an effect in subgroup analyses found no difference between them. Second, hemodynamic parameters that play an important role in monitoring patient who undergoes PSA were not analysed due to lack of presentable data with a different unit of measurement.
Additionally, small samples size ranging from 46 to 351 participants discouraged us from generating appropriate and meaningful analyses. Lastly, we conducted a thorough search and used two review authors to assess the study eligibility, extract the data, and assess the risk of bias in the included studies, thereby reducing potential bias in the review process. Although the search strategy used to identify potentially relevant studies was extensive, only the articles in English language were included in this review. Some relevant studies may have been skipped if they were conducted in other languages.

Conclusion
There is a low certainty evidence that ketofol improves the recovery time compared to single-agent and reduces the frequency of hypotension along with a moderate certainty of evidence for PSA in the paediatric population. There were no differences in other adverse effects during comparison with either single-agent or combined agent. A larger sample size would increase the certainty of this evidence. Our study can be a guide for practitioners to decide the choice of agents in PSA.
Future studies addressing this research question might benefit by focusing on some of the limitations that we encountered with current evidence. Firstly, in order to increase the certainty in effect for most outcomes in this review, larger sample size is required. More research in Fig. 15 Comparison of ketofol vs. combined agents control with respect to bradycardia this field would also enable a more precise exploration of subgroups. In particular, the favourable dosages of bolus and maintenance, ratio, and method of mixture for the different combinations of ketamine and propofol have an essential influence on their effects, such as recovery time, sedation level, and hemodynamic adaptation. Secondly, there is currently a lack of hemodynamic data, which plays a significant role in monitoring the patient. The subtle change of hemodynamic parameters might help to prevent complications during PSA, and plays a vital role in the choices of PSA by the physician. Finally, future studies should clearly define recovery time as well as sedation level using a standardised scoring system.