Elective or semi-elective procedures requiring sedation, analgesia, or anaesthesia are commonly performed in neonates admitted to neonatal intensive care units (NICU) or neonates admitted for elective procedures. In the NICU, these procedures may include endotracheal intubation, eye examination for retinopathy of prematurity and for some surgeries like ligation of patent ductus arteriosus or laser therapy for retinopathy of prematurity. Elective procedures in neonates include other scheduled surgical or investigative procedures requiring anaesthesia. Under most circumstances in the current practice environment, sedative, analgesic and/or anxiolytic agents are used for such events. These agents are perceived to reduce pain, stress and haemodynamic instability associated with these procedures and facilitate completion of the procedure in a timely manner (Oei 2002). However, all of these agents are associated with side effects. Major complications of the use of any such agent include hypotension and respiratory depression necessitating provision of respiratory support or escalation of respiratory support and ongoing intensive care.

Commonly used agents include combinations of neuromuscular blocking agents such as suxamethonium or pancuronium, anxiolytics such as diazepam or midazolam and analgesics such as morphine or fentanyl. At times, these are combined with atropine to counteract chronotropic side effects. For short-term sedation and anaesthesia, propofol has been widely used in adult and most paediatric intensive care units. Its rapid onset of action and rapid termination of effects when discontinued has attracted a wide interest and use. Propofol is a lipophilic anaesthetic agent that has rapid distribution from blood to central nervous system and fat and rapid redistribution (Allegaert 2007). Propofol is metabolised mainly via glucuronidation. Changes in the body fluid composition can affect the metabolism and clearance of propofol. Serious side effects have been reported. Parke 1992 reported deaths of five paediatric patients following prolonged infusion of propofol. A constellation of symptoms including lactic acidosis, myoglobinuria, bradyarrhythmia and myocardial failure was noted. Bray 1998 reviewed these and subsequent reported cases of side effects and coined the term “propofol infusion syndrome”; however, none of the patients had common characteristics except that most of them had respiratory infection as the cause of admission to a paediatric intensive care unit. Vasile 2003 reported that impairment of fatty acid metabolism accompanied by use of catecholamines and/or steroids in patients with acute neurological disease or inflammatory process was responsible for myocytolysis and subsequent deaths in patients who had propofol infused at rates greater than 5 mg/kg/hr for greater than 48 hours. A survey of paediatric intensive care units revealed that 47% of units routinely used propofol for induction and maintenance of anaesthesia (Playfor 2004). Reported side effects include persistent hypoxaemia, bradycardia, hypotension, and clonic convulsion (Gelber 1997). In contrast to Parke 1992, there are also reports and anecdotal experiences of the successful use of propofol without complications in paediatric patients (Crawford 2003; Sloan 2003).

Propofol is a lipophilic anaesthetic agent. Use of propofol for maternal anaesthesia during cesarean section led to the observation that propofol is transferred across the placenta (Sanchez-Alcaraz 1998) and secreted in breast milk. Case reports have identified neonatal respiratory depression associated with the maternal propofol administration during cesarean section. However, these effects were transient and required brief interventions only (Sanchez-Alcaraz 1998; Van de Velde 2004). Maturational changes in body composition are highest in the neonatal period. Allegaert 2007, Allegaert 2007a and Rigby-Jones 2002 observed that the distribution of propofol in neonates is markedly different from the distribution in children and adults. Among neonates, preterm neonates and neonates who had cardiac surgery were at the most risk for accumulation due to their slow clearance of propofol during both bolus and continuous infusion (Rigby-Jones 2002). This may lead to longer recovery periods compared to adults and paediatric patients.

Based on the experience in the paediatric population, propofol has been used in neonates (Golden 2001; Lim 2005). In a neonatal swine model, propofol infusion was not associated with alteration of systolic and diastolic function of left ventricle (Graham 1998). Case reports of the successful use of propofol in neonates (Bacon 1994; Golden 2001), especially for neonates with oropharyngeal complications, have raised the issue of the potential for the use of propofol in neonates. However, serious side effects associated with propofol use have been reported (Gelber 1997; Veyckemans 2001). The office of Food and Drug Administration reviewed concerns regarding neurotoxicity associated with use of certain anaesthetic agents in neonates and issued a cautionary note regarding the safety of these agents (Mellon 2007). Hatch 1999 agreed with a cautious approach in adopting propofol for neonatal anaesthesia as suggested by Bray 1998; however, Reed 1996 questioned the mechanism of these events and called for rigorous studies. This controversy still exists. If identified as effective without significant side effects, propofol could play a significant role in anaesthesia for neonates.

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Electronic searches

Searching other resources

We collected and analysed data in accordance with the standard methods of the Cochrane Neonatal Review Group.

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

We examined 427 citations to identify eligible studies. We retrieved seven articles for full assessment and identified only one randomised controlled trial (Ghanta 2007) as eligible for inclusion in the review. Two studies (Fabrice 2010; Silva 2008) were identified as ongoing from Clinicaltrials.gov websites. A study by Silva 2008 was reported to have recruitment complete in the year 2009. The authors reported that their manuscript is soon be published.

Ghanta 2007 is described below in the table Characteristics of included studies. This study randomised neonates to a propofol group and a morphine-atropine-suxamethonium group. Median gestational age of neonates was 27 weeks in both groups. Time required for successful intubation, heart rate and blood pressure at five minutes were reported outcomes. Data on mean and SD for time taken for intubation, preparation and recovery and rates of metabolic acidosis (defined as base deficit > 2 units) were obtained by contacting Dr Ju Lee Oei (Senior Investigator on this trial).

Ghanta 2007 was an open-label randomised controlled trial. The details of risk of bias assessment are described in the table Characteristics of included studies below. As this was an open randomised trial, investigators ensured independent outcome assessment. It is unclear whether the person assessing outcome was masked to group allocation or not as this person may be prone to personal bias. It is also not clear whether it was only one person who obtained data on all neonates or whether multiple personnel were involved as this may result in differential coding of outcome. Notwithstanding these points, the study was of good methodological quality.

Due to the inclusion of only one study (Ghanta 2007) in this review, we have not provided detailed subgroup analyses at this stage. The study included in this review falls into Comparison 2, Subgroup 1 for all three categories. The study compared propofol with a combination of morphine-atropine-suxamethonium for endotracheal intubation in preterm infants. Thirty-three infants were enrolled in propofol group and 30 infants were in morphine-atropine-suxamethonium group. Most intubations were via nasal route.

Primary outcomes

Success of achievement of proper sedation/anaesthesia as measured by:

1. Completion of the targeted objective as measured by:

1. Need for repeated attempts to achieve primary goal: There was no statistically significant difference in the number of infants who required multiple intubation attempts (39% in propofol group versus 57% morphine-atropine-suxamethonium group; RR 1.40, 95% CI 0.85 to 2.29, RD 0.17, 95% CI -0.07 to 0.42, Figure 1). The number of intubation attempts did not differ statistically between groups (median two attempts, interquartile range one to three attempts for morphine-atropine-suxamethonium group versus median one attempt, interquartile range one to two attempts for propofol group; P = 0.082).
2. Change in heart rate or number of episodes of bradycardia (as defined by investigators): It is reported that there was no difference in the heart rate at and during procedure between two groups. The data were presented in box-whisker plot only. Heart rate decreased in both groups during the procedure; however, according to the study authors there was no statistically significant difference between groups.
3. Change in oxygen saturation or episode of hypoxaemia requiring intervention (as defined by investigators): Median (interquartile range) intra-procedural oxygen saturations were 80% (67% to 88%) in the propofol group and 60% (43% to 82%) in the morphine-atropine-suxamethonium group (P = 0.019). Median oxygen saturations at recovery were also significantly higher in the propofol group (median (interquartile range) 95% (92% to 98%) compared to 92% (90% to 96%) in morphine-atropine-suxamethonium group (P = 0.008);
4. Change in blood pressure or hypotension episodes requiring intervention (as defined by investigators): Blood pressure increased during the procedure in both groups. It is reported that there was no difference in the difference of blood pressure at baseline and during procedure between two groups. The data were presented in box-whisker plot only.

2. Time to complete the procedure as defined by time of administration of the agent to time to complete intubation (time for successful completion of procedures among neonates who had completion of procedure in the first attempt and among those who required more than one attempt):

a. For all intubation attempts combined: Time was shorter in propofol group (median 120 seconds, interquartile range 60 to 180 seconds) versus morphine-atropine-suxamethonium group median 260 seconds, interquartile range 60 to 435 seconds, P = 0.007).

b. For intubation successful at first attempt: There was no difference between groups (median 60 seconds, interquartile range 60 to 120 seconds in propofol group versus median 60, interquartile range 52 to 120 seconds in morphine-atropine-suxamethonium group P = 0.641).

c. For intubation that required mor than one attempt: Time was in propofol group (median 180 seconds, interquartile range 120 to 300 seconds in propofol group versus median 360, interquartile range 300 to 750 seconds in morphine-atropine-suxamethonium group P < 0.001).

3. Time for preparation of the anaesthetic agents (time from onset of preparation of medication to administration of medication): The time required to prepare drugs for study was shorter in the propofol group (median 180 seconds, interquartile range 180 to 210 seconds in propofol group versus median 960 seconds interquartile range 900 to 1200 seconds in the morphine-atropine-suxamethonium group; P .001).

4. Time to achieve sleep/muscle relaxation (time from receipt of medication to achievement of sleep/muscle relaxation): There was no statistically significant difference in the time to achieve sleep or muscle relaxation between groups (median 60 seconds, interquartile range 60 to 120 seconds for morphine-atropine-suxamethonium group versus median 60 seconds, interquartile range 30 to 60 seconds in propofol group; P = 0.087)

5. Time to recovery to prior clinical status (time from completion of procedure to return of clinical status): There was a reduction in time to recover to previous clinical status with propofol compared to control group (median 780 seconds, interquartile range 360 to 1110 seconds versus median 1425 seconds, interquartile range 645 to 2250 seconds, P = 0.002)

6. Safety of propofol (occurrence of these side effects during or within 24 hours of stoppage of medication) (as defined by investigators):

1. incidence of sustained hypoxaemia > 1 minute (saturation below target limit): data not collected;
2. incidence of hypotension (need for fluid bolus or inotropes to maintain blood pressure): data not collected;
3. incidence of metabolic acidosis (need for correction of acidosis by fluid bolus or sodium bicarbonate): not reported in published paper, but study authors provided data on rates of base deficit > 2 units between groups. There was no difference in the incidence between two groups (RR 1.19, 95% CI 0.95 to 1.48); Figure 2
4. incidence of lactic acidosis (serum lactate > 2 mmol/l): there was no statistically significant difference between groups in the incidence of lactic acidosis (RR 0.40, 95% CI 0.02 to 9.06 and RD -0.06, 95% CI -0.20 to 0.09); Figure 3
5. incidence of rhabdomyolysis (myoglobinuria, rising creatinine): data not collected;
6. incidence of hyperkalaemia (serum potassium > 5.5 mmol/l): data not collected;
7. incidence of renal failure (serum creatinine > 125 umol/l): data not collected;
8. incidence of seizures: data not collected;
9. mortality (all cause mortality assessed at < 28 days of completion of therapy): data not collected.

Only one randomised controlled trial of 63 neonates has evaluated the efficacy and safety of propofol for intubation in neonates. This trial demonstrated the feasibility of giving propofol to neonates with a significant reduction in duration for preparation, procedure time and post-procedural recovery. The study authors did not observe any increased incidence of side effects attributable to propofol. However, it must be recognised that the number of infants enrolled in this trial is too small to detect all but extreme problems with safety and authors did not systematically looked for all adverse effects.Two studies are ongoing, one comparing propofol with sevoflurane and the other comparing propofol and ramifentanil versus midazolam and ramifentanil.

Overall, the data are insufficient to make any practice recommendations. However, the preliminary data on efficacy would suggest that further studies are needed. With a large number of infants receiving one or other form of stressful procedure throughout the world, it is feasible to support a multicentre undertaking.

The only study included in this review was an unblinded study; however, the study authors ascertained outcome in a non-biased way and the overall risk of bias was low. The sample size of this study was small.

We have obtained certain data from the authors; however, validity of these data is good as we obtained the raw data file and performed analyses.

Recently, the impact of propofol infusion on heart rate, oxygen saturation and blood pressure in neonates has been studied. A transient drop in heart rate and oxygen saturations were observed, but at the same time blood pressure was reduced for approximately one hour (Vanderhaegen 2009). The clearance of propofol was affected by postnatal age, concurrent medication and presence of cardiomyopathy. Propofol metabolism is slower in neonates due to age-dependent maturation of glucuronidation (Allegaert 2008). A recent review of three pharmacokinetic studies on the use of propofol in neonates concluded that a careful approach is necessary as marked inter-individual variability and reduced clearance of propofol is observed in neonates (Allegaert 2009). Repeated administration of boluses or continuous infusion of propofol in the first week after birth was discouraged. No other systematic review has evaluated this comparison. Another important point is that propofol is a hypnotic agent without analgesic effects (Papoff 2008). Their experience with 21 patients who received fentanyl and propofol for neonatal intubation indicated feasibility of combining a hypnotic agent with an analgesic agent. The American Academy of Pediatrics suggests administration of an analgesic or anaesthetic dose of hypnotic for procedural discomfort/pain. Future studies should take this into consideration.

No practice recommendation can be made based on the available evidence regarding use of propofol in neonates.

Further research is needed in two domains. First, further data on the pharmacokinetics of propofol in neonates are needed. A safe dose of propofol could be identified in neonates undergoing elective intubation who would be given escalating doses of propofol. Blood could be sampled at predefined times after administration in order to correlate propofol levels with clinical findings including depth of relaxation, anaesthesia and ease of procedural completion. Second, once a relatively safe dose is identified, randomised controlled trials assessing the safety and efficacy (successful completion of intended procedure, ease of performance of procedure, etc.) are needed. It would be appropriate to compare the safety and efficacy of propofol with other agents (mostly combinations of sedative, anxiolytic and a muscle relaxant).

Risk of bias table