Prophylactic caffeine to prevent postoperative apnoea following general anaesthesia in preterm infants

Growing ex-preterm infants who undergo general anaesthesia for surgery at about term equivalent age may have episodes of apnoea, cyanosis and bradycardia during the early postoperative period.

The American Academy of Pediatrics defined infant apnoea as a pause in breathing of greater than 20 seconds, or one of less than 20 seconds and associated with bradycardia and/or cyanosis (AAP 2003). Recurrent episodes of apnoea are common in preterm infants and the incidence and severity increases at lower gestational ages. By term equivalent postnatal age infants have usually 'outgrown' their tendency to spontaneous apnoea. However, if some additional insult such as infection, or the administration of drugs that depress the central nervous system occurs (such as general anaesthetic agents, Liu 1983), apnoea and associated oxygen desaturations can recur. Attempts have been made to determine which infants are most at risk (Cote 1995; Spear 1992; Welborn 1994). Factors such as younger postmenstrual age and past history of apnoea seem important, although such risk scoring does not detect all infants who develop postanaesthesia problems (Cote 1995).

Methylxanthines such as caffeine are thought to stimulate breathing efforts and have been used in clinical practice to reduce apnoea in preterm infants since the late 1970's. Caffeine is better than theophylline in other attempts to reduce apnoea.

The mechanism of their action is not certain. Possibilities include increased chemoreceptor responsiveness (based on increased breathing responses to CO2), enhanced respiratory muscle performance and generalised central nervous system excitation. Other reviews on the effects of methylxanthines such as caffeine on apnoea, published in The Cochrane Library, include Henderson-Smart a; Henderson-Smart b; Henderson-Smart c and Henderson-Smart d. A systematic review has also been carried out on the effects of regional versus general anaesthesia on postoperative apnoea (Craven 2003).

If prolonged, apnoea can lead to hypoxaemia and reflex bradycardia which may require active resuscitative efforts to reverse. There are clinical concerns that these episodes might be harmful to the developing brain or cause dysfunction of the gastrointestinal tract or other organs, although there are no data to support this. Frequent episodes may be accompanied by respiratory failure of sufficient severity as to lead to intubation and the use of intermittent positive pressure ventilation (IPPV).

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Electronic searches

The standard methods of the Cochrane Collaboration and its Neonatal Review Group were used to select trials, evaluate the quality of the trials and to extract data.

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Assessment of heterogeneity

Data synthesis

Subgroup analysis and investigation of heterogeneity

Three eligible trials were found. Details are given in the table 'Characteristics of included studies'. In general, infants were born at 30 to 32 weeks gestation and studied at 40 to 44 weeks postmenstrual age. All three trials evaluated caffeine given intravenously in a single dose during general anaesthesia. The dosage varied from 5 mg/kg (Welborn 1988) to 10 mg/kg (LeBard 1989; Welborn 1989). In all three trials apnoea was defined as any apnoeic episodes of > 15 seconds duration, or shorter ones associated with bradycardia. In two trials, oxygen desaturations (< 90%) on pulse oximetry were recorded (LeBard 1989; Welborn 1989).

This is detailed in the table 'Characteristics of included studies'.

Concealment at randomisation: all used formal randomisation, although the steps taken to fully conceal allocation assignment were not clear.

Blinding of treatment: in each trial a saline placebo was used.

Blinding of outcome assessment: in two trials apnoea/bradycardia was assessed blindly from a polygraphic recording. Methodology in the third trial (LeBard 1989) needs clarification as this has only been published in abstract form.

Completeness of follow-up: not clearly stated.

Three eligible trials were found. In each trial apnoea/bradycardia occurred in fewer infants treated with caffeine. The typical estimate for relative risk is 0.09 (95% CI 0.02 to 0.34). Absolute risk difference is -0.58 (95% CI -0.74 to -0.43) indicating that fewer than two infants have to be treated with caffeine to expect to prevent one with postoperative apnoea. In two trials (LeBard 1989; Welborn 1989) continuous recordings of oxygen saturation detected hypoxaemic episodes (< 90%) in fewer treatment than control infants [RR 0.13 (95% CI 0.03 to 0.63]. No infant in any trial required intubation and mechanical ventilation. No adverse effects were reported.

Caffeine reduces the occurrence of apnoea, bradycardia and oxygen desaturation after general anaesthesia, although it has yet to be determined whether these episodes are clinically important. If such episodes are detected by monitors, responded to promptly by nurses and do not lead to a need for mechanical ventilation, then these episodes are of uncertain significance. Apnea and hypoxaemia might be more important in infants at particularly high risk of respiratory failure, such as those with chronic lung disease. This has not been evaluated.

No adverse effects of caffeine were reported, although the number of infants studied in these three trials is too small to exclude less common ones.

One trial (LeBard 1989) have many unclear methodology and was only published as an abstract. The other two trials had good methodology.

Two randomised trials (Krane 1995; Welborn 1990) have suggested that apnoea is more common following general anaesthesia than following spinal anaesthesia without premedication in infants having inguinal hernia repair, one of the most common operative procedures in ex-preterm infants. A Cochrane review (Craven 2003) did not find evidence to support that hypothesis.

Variations in the type of general anaesthetic agents might affect postoperative ventilatory function. A trial (Sale 2006) compared two general anaesthetic agents (desflurane and sevoflurane) in terms of postoperative ventilatory events or in the number of apneas in the postoperative period and found no difference

After general anaesthesia caffeine can be used to prevent postoperative apnoea/bradycardia and episodes of oxygen desaturation in preterm infants if this is deemed clinically necessary. In view of the small numbers of infants studied in these trials and uncertainty concerning the clinical significance of the episodes, caution is warranted in applying these results to routine clinical practice.

There is a need to determine which infants might benefit most by this treatment. Studies confined to those most at risk of apnoea (prior history, younger postmenstrual age) and those that might require mechanical ventilation (chronic lung disease) would be of value.

Risk of bias table

Risk of bias table

Risk of bias table

LeBard SE, Kurth CD, Spitzer AR, Downes JJ. Preventing postoperative apnea by neuromodulator antagonists. Anesthesiology 1989;71:A1026.

Welborn LG, De Soto H, Hannallah RS, Fink R, Ruttimann UE, Boeckx R. The use of caffeine in the control of post-anesthetic apnea in former premature infants. Anesthesiology 1988;68:796-8.

Welborn LG, Hannallah RS, Fink R, Ruttimann UE, Hicks JM. High-dose caffeine suppresses postoperative apnea in former preterm infants. Anesthesiology 1989;71:347-9.