Secondary Logo

Journal Logo


The ‘propofol infusion syndrome’: the facts, their interpretation and implications for patient care

Ahlen, K.*; Buckley, C. J.; Goodale, D. B.; Pulsford, A. H.

Author Information
European Journal of Anaesthesiology: December 2006 - Volume 23 - Issue 12 - p 990-998
doi: 10.1017/S0265021506001281


Background and history

Over 10 years ago, a small number of publications reported ‘unexplained’ and ‘unexpected’ fatalities in paediatric intensive care patients undergoing sedation with propofol [{L-End} 1,{L-End} 2]. These reports mostly concerned under 4 yr olds with upper respiratory tract infections, receiving high and often increasing doses of propofol, who were reported to have developed a ‘syndrome’ consisting of some or all of the features shown in Table 1.

Table 1
Table 1:
Features of adverse event reports.

The FDA reviewed these reports and concluded in 1992 that propofol had ‘no identifiable link to cardiac adverse events’ in children, but that trials should be undertaken [{L-End} 3], since the limited information provided for these events prevented any clear determination of causality. In 1994, the manufacturer added a precautionary warning to the Diprivan prescribing information as follows:

‘…although no causal relationship has been established, serious adverse events (including fatalities) have been observed from spontaneous reports of unlicensed (ICU) use and that these events were seen most often in children with respiratory tract infections given doses in excess of those recommended for adults.’

In 1998, Bray reviewed 18 cases with an inconsistent association of clinical features, which he thought to be a ‘syndrome’. He proposed the name and suggested that muscle damage be included in the criteria [{L-End} 4].

Subsequently, Hatch proposed investigating propofol for paediatric ICU sedation but recommended: ‘children with existing sepsis or primary respiratory problems should probably be excluded’ from such studies. Additionally ‘infusion rates should be kept to those used in adults, even if other drugs have to be used to ensure adequate sedation’ [{L-End} 5].

Serious adverse events had been reported in patients with croup and epiglottitis, and the protocol for Trial 0859IL-0068 was developed while discussions surrounding the use of Diprivan for the sedation of these patients were ongoing. Children with these pathologies were excluded from clinical trials as a precautionary measure. No information is therefore available on Diprivan's safety for ICU sedation in these children, and the manufacturer has consequently contraindicated its use for such paediatric ICU patients in the prescribing information as follows:

‘Diprivan is contraindicated for the sedation of children of all ages with croup or epiglottitis receiving intensive care.’

In 2001, Cremer and colleagues [{L-End} 6] published a retrospective review of seven adult patients with head injuries, who died after having developed clinical features similar to those described by Bray. The authors presented the criteria shown in Table 2 for their proposed ‘adult propofol infusion syndrome’ scenario. Small numbers of reports, similar to those described by Cremer and colleagues, occurring in severely injured patients have been published since [{L-End} 7,{L-End} 8]. The manufacturer has documented additional reports of similar cases in its pharmacovigilance database as part of its ongoing pharmacovigilance responsibilities.

Table 2
Table 2:
Criteria for proposed ‘adult propofol infusion syndrome’.

The patients in whom this syndrome was described can be roughly divided into three groups: (i) those with severe multiple injuries, (ii) patients being treated for status epilepticus or undergoing acute opiate detoxification or (iii) critically ill septic patients with sepsis.

Historical summary – The reports presented case histories of adverse events occurring in seriously ill patients sedated with propofol, often in ‘high’ doses. In most reports on children, sepsis (usually of respiratory origin) was the reason for their ICU care. In adults and some of the children, elevated ICP refractory to treatment, or other events (as above), was the reason for their ICU care. Progression to metabolic acidosis and organ failure was described.

Paediatric ICU sedation trials

In 2002, Cornfield and colleagues reported on a trial of propofol in one hundred and nine children. Satisfactory sedation was reported with dosages up to 50 μg kg−1 min−1 (3 mg kg−1 h−1) plus bolus doses of 1 mg kg−1 not more frequently than hourly, if required. There was no metabolic or haemodynamic compromise [{L-End} 9].

In Pepperman and Macrae’s retrospective study on one hundred and six patients receiving propofol there was no unexplained metabolic acidosis and no difference in mortality. They concluded: ‘propofol compares favourably with other sedative agents’. Interestingly, they noted that of the observed incidences of metabolic acidosis, 78% were seen in children under 3 years old [{L-End} 10].

In 1999, the manufacturer conducted a trial of Diprivan for paediatric ICU sedation in over 300 patients in the USA (Trial 0859IL-0068) to determine whether EDTA, included as a microbial growth inhibitor, caused increased zinc excretion as in adults. Secondary end-points included metabolic monitoring, safety and efficacy. Although the study was not designed to detect differences in mortality rates, an unexpected higher incidence of death was seen in the Diprivan arm during the administration period and up to 30 days after discontinuation compared to the ‘standard sedative agent’ (11% vs. 4%). This was not statistically significant and closer inspection of the data showed that no death was preceded by the constellation of events described by either Cremer or Bray. Over the first 7 days of ICU care, there was a trend to increasing base excess (BE) (Fig. 1). If propofol had impaired respiratory chain enzymes one would expect a decrease in BE and a divergence from the SSA group. Thus there was no suggestion of a time dependent development of acidosis.

Figure 1.
Figure 1.:
AstraZeneca Trial 0859IL-0068: mean arterial base excess (mEq L−1) vs. study day.

Extended licence application

A small number of authorities approved an extended licence for paediatric ICU sedation, but the FDA, who considered the trial results to be a possible safety signal, instructed the manufacturer to issue a ‘Dear Doctor’ letter indicating that Diprivan should not be used for paediatric ICU sedation in the US. Subsequently, most regulatory authorities have not approved, or have rescinded approval, of this indication. Many changed the ‘warnings’ concerning paediatric ICU sedation to a ‘contraindication’. Where the PICU indication is still licensed, continued use has not resulted in additional reports of a ‘propofol infusion syndrome’.

Safety review

In 1999, the manufacturer completed a comprehensive safety review of all adverse events in paediatric ICU patients receiving propofol reported to their database, which holds all reports of adverse events ascribed to, or possibly ascribable to, propofol use. The individual events described in the case reports and synonyms for these events were searched for in the database and reviewed. Of the few reports that were consistent with Bray’s or Cremer’s criteria for the syndrome, most also had relevant independent risk factors for the observed disease progression and for the development of the individual elements. The six patients, in whom muscle damage was reported had risk factors for its development including refractory epilepsy, a temperature of 41.5°C (possibly due to malignant hyperthermia), use of prednisolone and, in three patients, the combination of sepsis, hypoperfusion and hypoxaemia. The events occurring in six paediatric reports were consistent and remained without a satisfactory explanation. For four of these cases, the manufacturer received only minimal information despite requests for more. Since these events had all occurred in patients under 3 years old with severe viral respiratory tract infections, a precautionary contraindication was added to the Diprivan prescribing information as follows:

‘Diprivan is contraindicated for the sedation of children under the age of 3 years with serious viral respiratory tract infections receiving intensive care.’

Analysis of published reports

In a comment on the report of Parke and colleagues [{L-End} 1], Cook, who was involved in the care of two of the reported patients, highlighted omissions in the case reports [{L-End} 11]. In Case 4, propofol was discontinued 3 days before death occurred. Case 5 already had features of septicaemia and was developing renal failure when deterioration occurred, and acute worsening occurred after the first dose of ceftazidime. Cook also stated that the involved ICU and the associated university department of paediatrics found no evidence of a direct effect of propofol on any aspect of metabolism in these patients.

The manufacturer subsequently received further data on Case 5 that radically altered the interpretation of the reported events, which clearly became explicable without having to resort an unknown toxic effect of Diprivan. Three of the patients with sepsis of respiratory origin were given steroids, a fact that was not mentioned in the presentations.

The article of Bray [{L-End} 4] collated reports on eighteen patients (including those presented in Parke’s earlier publication), seven of which had already been reported to the manufacturer. Most original reporters stated or implied that Diprivan was not a likely cause of the events, since other reasonable and recognized causes existed for them. For the remainder, given the paucity of detail provided, it was often not possible to comment beyond stating that the events are recognized to occur in some patients suffering from severe presentations of those diseases and their complications.

Bray concluded that the incidence of the ‘propofol infusion syndrome’ in patients receiving propofol for more than 48 h at doses greater than 4 mg kg−1 h−1 was 33%. Given the failure to detect a similar scenario in the patients of AstraZeneca Trial 0895IL-0068 and numerous unsponsored trials, this conclusion is not supported.

Cremer and colleagues [{L-End} 6] reported events occurring in adults with severe head injuries, which are known to cause cardiovascular instability, arrhythmias, ischaemia and rapidly varying potassium levels [{L-End} 12–14]. Patient management apparently included fluid restriction with hypovolaemia and the use of inotropes and vasoconstrictors to maintain cerebral perfusion pressure (CPP) at or above 70 mmHg. The events described occurred while the average administered propofol dose exceeded the maximum recommended dose by ≥25%. The initial dosage was generally in the range recommended for sedation, but was progressively increased to two or three times the recommended dose. Such dosages were probably not solely used for sedation but also to control intracranial pressure. The infusion rates of inotropes and vasoconstrictors were increased in parallel with that of propofol, presumably reflecting the refractory nature of the intracranial hypertension. No report indicated that advanced haemodynamic monitoring had been initiated early in the ICU management; one patient did have a pulmonary artery catheter inserted shortly before death.

Reports of the unlicensed use of propofol in the treatment of status epilepticus or during acute opiate detoxification often describe very high administration rates, occasionally reaching 30–40 mg kg−1 h−1. In these reports hypotension occurred usually after the second day, and was typically treated empirically with inotropes and/or vasoconstrictors. Acidosis and hypoxaemia developed, often progressing rapidly to multiorgan failure and death. Haemodynamic parameters were not reported.

Severe sepsis – There are a number of case reports of patients with sepsis being sedated with propofol, who developed hypotension, acidosis, multiorgan failure and subsequently died. This scenario is common in the ICU and is not considered unusual unless the progression presents unusual aspects. No unusual features were described in these case reports. They generally indicated that incremental inotropes and/or vasoconstrictors were used empirically.

The main impression gained from the published reports is that the ‘syndrome’ appears to be predominantly related to a failure of tissue oxygenation secondary to haemodynamic impairment and/or distributative cardiovascular failure. In some patients with β-oxidation deficiency, propofol may theoretically impair fatty acid oxidation. Reducing the requirement for fatty acid oxidation with glucose administration may thus reduce the risk of developing the events associated with the ‘propofol infusion syndrome’.

Physiological analysis

Common features and common factors. Failure of adequate tissue oxygen uptake or utilization is the commonest cause of metabolic acidosis. The organ based elements of the reported events, i.e. rhabdomyolysis, cardiac failure and lipaemia, could be due to inadequate oxygenation of skeletal muscle, cardiac muscle and liver, respectively (see later) [{L-End} 15–17].

In all detailed reports of events resembling the ‘propofol infusion syndrome’, whether propofol was used or not (a case reported directly to AstraZeneca by J Blummer, a four-month-old patient had received lorazepam for ICU sedation), there is evidence that a failure of tissue oxygenation may have occurred. Most reports included factors suggestive of a distributive cardiovascular failure with hypoperfusion and/or hypoxaemia, such as severe sepsis (particularly of respiratory tract origin), head injuries (with fluid restriction and vasoconstrictor and inotrope therapy), multiply injured patients with inadequate resuscitation, repeated periods of sudden hypotension or hypoxia and organ dysfunction suggesting hypoperfusion or hypoxia (e.g. elevated liver enzyme concentrations diagnostic of hypoxic or hypotensive liver injury, lipaemia and renal dysfunction).

Reports that described major modifications of haemodynamic management with subsequent resolution of the symptoms of the ‘syndrome’, suggest that the haemodynamic aspects of the individual events are important, if not causal. Conversely, in patients, in whom the major therapeutic measures were solely discontinuation of propofol and empirical increase of inotrope dosage, the events usually progressed.

Clinical features

Cardiovascular failure. The primary diagnostic criterion for Bray’s ‘propofol infusion syndrome’ was ‘cardiac failure’. Of the five patients in the original article: two had evidence of fluid overload, two were reported to AstraZeneca as probably suffering from viral myocarditis by another reporter, one required 80% oxygen and was therefore likely to be suffering from significant hypoxaemia, one had pre-existing pulmonary hypertension and was hypoxic due to pneumonia and two also had cerebral oedema.

In the cases in which the events proceeded to death, there was little documentation that the cardiac function was monitored beyond the use of pulse rate, blood pressure and central venous pressure measurements. A few reports mention the use of echocardiography, the results of which were inconsistent, ranging from normal to uni- (left or right) or bi-ventricular impairment. Most stated or implied that inotropes were used empirically without sustained and/or adequate response.

In Cremer's reports [{L-End} 6], progressively increasing inotrope, vasoconstrictor and propofol dosages were probably in response to refractory intracranial hypertension and severe fluid depletion. The systemic consequences of this regimen on haemodynamics are severe. In head injury patients, inotropes, vasoconstrictors and propofol (vasodilatory at doses occasionally used for intracranial pressure control) together with hypovolaemia over-ride the pathological intracranial vascular regulation. However their effects are not limited to the cerebral vasculature and lead to a systemic alteration of vascular auto-regulatory responses. The initial vasomotor result of this regimen at the capillary level will be a ‘hunting phenomenon’ (periodic opening of an hypoperfused, vasoconstricted capillary bed, allowing intermittent tissue perfusion dependent on central and local regulatory mechanisms). Ultimately, a resulting increase in physiological shunting will lead to patchy hypoperfusion within and among tissues [{L-End} 18], similar to that seen in the distributive cardiovascular failure of severe sepsis and described by Siegemund and colleagues [{L-End} 19].

In the myocardium, where metabolic rate and oxygen extraction ratio are high, the effect of such increased shunting would predictably be to induce ischaemia (in some reports myocardial ischaemia was suspected), arrhythmias and eventually failure.

Some publications show that head injury and its treatment can cause tissue ischaemia and necrosis [{L-End} 12,{L-End} 13,{L-End} 20]. They demonstrated ischaemia of the heart and gut, but occurrence in other tissues is likely [{L-End} 21]. In some reports of ‘propofol infusion syndrome’, ischaemic ECG changes were mentioned, but investigations following recovery revealed no disease process. Cardiac ischaemia impairs function and decreases global oxygen delivery.

Rhabdomyolysis. Whenever adequate clinical information was provided in the reports, it is possible to identify a recognized risk factor for the development of rhabdomyolysis that was independent of propofol. Muscle damage, elevated creatine phosphokinase (CPK) levels and rhabdomyolysis have been associated with both chronic and acute administration of high dose steroids [{L-End} 22,{L-End} 23]. Publications strongly suggest that high potassium flux can occur after head injury and may trigger rhabdomyolysis [{L-End} 14].

Small numbers of reports are consistent with rhabdomyolysis precipitated by steroid use or potassium flux. In other reports, risk factors included myopathies (including malignant hyperthermia), compartment syndrome and seizures.

Many others apparently suffered prolonged systemic hypoperfusion or hypoxia with or without episodes of acute hypotension and/or hypoxia. These reports gave no indication that advanced haemodynamic monitoring was implemented prior to major haemodynamic collapse. Hence, it appears unlikely that such states could have been adequately assessed. Inadequate oxygen delivery to skeletal muscle leads to anaerobic metabolism, acid production and, if sufficiently severe, increased CPK levels, muscle cell death and rhabdomyolysis. Thus it appears that acute or chronic hypoperfusion or hypoxia were risk factors for rhabdomyolysis in these seriously ill patients in the ICU.

Lipaemia. Lipaemia and fatty liver were described in some reports, and further occurrences are described in the AstraZeneca pharmacovigilance database. Unimpaired lipid regulation and metabolism in the liver require carbohydrate substrates. In their absence, phosphorylated citrate levels fall and lipid metabolism slows. Carbohydrate reserves are commonly exhausted in patients admitted to the ICU, and may not initially be adequately administered. If lipids are infused (including propofol formulations) and regulation is impaired, accumulation may begin and reach levels interpreted as ‘hyperlipaemia’ after a further 2–5 days, depending on the administration rates. The incidence and severity of hyperlipaemia in ICU patients in the AstraZeneca pharmacovigilance database correlates with higher propofol and lipid dose rates.

Anything impairing hepatic function will exacerbate hyperlipaemia. Common causes in seriously ill ICU patients include hypoperfusion, sepsis, hypermetabolic states, hypoxia, vasoconstrictors, abdominal tamponade, etc. One or more such events were commonly noted in the reported instances of hyperlipaemia. In some cases, elevated levels of hepatic enzymes indicative of hypoperfusion or hypoxaemic injury were reported. Prolonged hyperlipaemia and a poorly functioning liver appear to lead to fatty degeneration.

Some reports have indicated that patients requiring increasing dosages of propofol went on to develop features of the ‘propofol infusion syndrome’. This altered requirement was generally coincident with an increasing lipaemia. One proposed explanation for this apparent development of tolerance is that administered propofol, being highly lipid soluble, was preferentially taken up by the lipid droplets in lipaemic serum, removing it from metabolic and clinically active sites [{L-End} 24]. Such increases in dose requirement may thus be an indicator of lipaemia, which may be due to hepatic impairment.

Lipaemia itself can impair mitochondrial oxygen uptake, and may thus exacerbate the oxygen utilization problems promoting the development of aspects of the ‘syndrome’ and may contribute to its refractory nature.

Lipaemia occurring in seriously ill patients receiving propofol is predictable, manageable and not due to some unknown ‘propofol toxicity’. The Diprivan prescribing information indicates that infusion rates of propofol and other solutions with lipids should be reduced in patients with inadequate clearance of fat. In most reports of ‘hyperlipaemia’ documented in the AstraZeneca pharmacovigilance database, whether occurring in association with the development of a ‘propofol infusion syndrome’ or in isolation, these recommendations appear not to have been followed.

Possible metabolic causes

Laboratory investigations have shown that inhibition of mitochondrial respiratory chain enzymes only occurs at concentrations of propofol not likely to be seen clinically [{L-End} 25,{L-End} 26]. These investigations did not take into account the extensive protein binding of plasma propofol that would in vivo decrease mitochondrial exposure to as little as a tenth of the investigated conditions. In addition, the effects that such inhibition would predictably cause are not seen in practice. The ‘syndrome’ and/or acidosis do not appear to occur with anaesthesia, where higher dose rates are administered; propofol effects diminish rapidly after discontinuation, but discontinuation has not been demonstrated to result in recovery from the events comprising the ‘syndrome’. (A long-lasting metabolite has been postulated to cause the ‘scenario, but none have yet been identified.) Propofol would decrease oxygen consumption; this it does, but to the same extent as midazolam [{L-End} 27].

A case study of a child who developed clinical features of the ‘syndrome’ reported successful treatment with haemofiltration. Before filtration, high concentrations of malonylcarnitine (3.3 μmol L−1) and C5-acylcarnitine (8.4 μmol L−1) were detected that returned to normal after recovery. The abnormalities resembled those seen with impaired mitochondrial function. Hyperlipaemia, which is recognized to cause disturbances of acylcarnitine metabolism, was seen in this patient. Further reports of altered fatty acid oxidation have confirmed the coexistence of abnormal acetylcarnitine metabolism with features of at least a subgroup of patients with the ‘syndrome’ [{L-End} 28–30].

In an editorial, Wolf and Potter [{L-End} 30] indicated ‘Children with inherited defects in β-oxidation are often asymptomatic until such time as they are stressed by, for example, starvation (low glucose) or infection’. Because their usual source of energy, glucose, has become inadequate they turn to fat metabolism as their sole energy source. Wolf and Potter further indicate that ‘under these conditions they may develop life-threatening rhabdomyolysis, cardiac and hepatic insufficiency associated with hypoglycaemia’. Children with β-oxidation deficiency suffer acute conditions caused both by failure of the respiratory chain, resulting in cellular hypoxia, and direct toxicity of accumulated long chain fatty acids [{L-End} 31]. The accumulation of long chain acylcarnitines is toxic for heart and skeletal muscle [{L-End} 32,{L-End} 33]. Any propofol effects on mitochondria will be related to exposure, and the syndrome is almost exclusively seen where ‘high’ dosages have been given. Therefore, irrespective of the causality, prudence suggests limiting dosage to the prescribing information recommendations and appropriate carbohydrate supplementation [{L-End} 34].

Clinical implications and improving patient care

No evidence of an effect of propofol on the acid–base status nor any incidence of ‘propofol infusion syndrome’ was detected in controlled studies. However, the failure to see either the events consistent with the ‘propofol infusion syndrome’ or metabolic effects in clinical trials cannot exclude the possibility of such rare events. It seems clear that its manifestation requires a seriously ill patient with tissue hypoxia. ‘High’ dose propofol administration may exacerbate tissue hypoxia by its cardiovascular effects, the associated lipid load and theoretically by impairing mitochondrial function. The analysis of the available information suggests principles for optimizing the care of ICU patients and identifying those, who may be susceptible to developing of syndrome or its individual elements, irrespective of the sedative agents used.

Haemodynamic and vascular function – Prevention or treatment of the ‘syndrome’ depends on restoring peripheral oxygen delivery and uptake by treating the underlying causes. In most reports, impaired haemodynamics appear the primary cause of oxygen delivery failure. Although many intensivists accept that monitoring haemodynamics and oxygen delivery and uptake is required to assess cardiovascular function in seriously ill patients, only five patients (adult and paediatric) suffering fatal events with any resemblance to the ‘syndrome’ were documented to have been invasively monitored in the entire AstraZeneca pharmacovigilance database. In all five reports, adequate correction of grossly abnormal haemodynamics was not successful or not undertaken or not reported. In some non-fatal cases, appropriate use of monitoring with therapeutic response resulted in recovery.

Despite technical problems in determining the mentioned parameters, particularly in very small children, such monitoring should be considered, particularly if the patients are very ill and their condition is deteriorating with acidosis or cardiac failure. For smaller patients, use of ultrasound aortic flow, computerized pulse waveform analysis or impedance cardiac output monitors together with echocardiography, is more practical but may not give the full range or accuracy of invasive measurements.

Monitoring may reveal unexpected findings, such as congenital heart defects, pulmonary hypertension and uni-ventricular failure, whose recognition may prove vital to successful management. Physicians not familiar with the acquisition and interpretation of these measurements and derived parameters may require additional training [{L-End} 35].

Steroids – A high proportion of the children developing ‘the syndrome’ were given steroids, hence the need for steroids should be carefully reviewed.

Lipaemia – The earlier described physiological analysis of this situation suggests likely methods of prevention and management. The possibility that lipaemia is due to hepatic functional impairment and that the impairment may be due to a failure of haemodynamics or oxygen delivery, should be considered. The co-administration of carbohydrate allows continued lipid metabolism, given adequate hepatic function. Limitation of total lipid load may be needed, lower lipid administration rates will prevent onset of ‘hyperlipaemia’ or reduce to tolerable levels plasma lipid increases. Lower propofol doses are appropriate where plasma lipids are increasing, possibly necessitating adjuncts if such doses are insufficient. In some markets, Diprivan 2% formulation can halve associated lipid dosage.

Factors adversely affecting hepatic function should be treated. In most instances, treatment of the primary disease process is the correct approach and includes treatment of hypoperfusion states, sepsis, hypermetabolic states, hypoxia, etc. Some treatments may adversely affect hepatic function (vasoconstrictor use, high pressure ventilation, etc.) and should be used with caution.

Head injured patients. Fluid restriction necessitating the use of vasopressors to maintain CPP increases the risk of tissue hypoperfusion and impaired hepatic function. Zornow and Prough provide evidence that severe reductions of circulating volume, and associated need to use vasoconstrictors, does not improve outcomes in this patient group [{L-End} 36]. Manley and colleagues state that factors vital to good outcomes in these patients are the maintenance of an adequate CPP and the avoidance of systemic hypotension [{L-End} 37].

Bingham demonstrates that serum osmolarities exceeding 378 mOsm kg−1 in one group of head injured patients correlated with death and recommends an upper limit of 340 mOsm kg−1 for these patients [{L-End} 38].

Luerssen and Wolfla [{L-End} 39] describe a balanced approach to head injury care, with three levels of intensity of treatments and monitoring for increasing refractoriness of raised intracranial pressures; normovolaemia or a slightly raised intravascular volume is recommended for all.

Whenever patients develop signs of cardiac failure, metabolic acidosis, etc. during aggressive treatment with fluid depletion and vasoactive agents, we recommend that management should be guided by haemodynamic parameters and aimed at optimizing oxygen delivery and uptake. Most will probably require optimization of intravascular filling, reductions in vasoactive agent dosages and reduction in the propofol or barbiturate dosage.

Seriously injured patients

Seriously injured patients who developed a situation similar to the ‘propofol infusion syndrome’, all developed this condition after prolonged phases consistent with poor perfusion, often with episodes of acute hypoxia or hypoperfusion and supervening infection. None reported the use of haemodynamic or oxygen delivery monitoring prior to major cardiovascular collapse. Appropriate monitoring and therapeutic responses might have prevented occurrence of such circumstances in some cases.

Prescribing information

In response to the reports discussed the manufacturer analysed the situation and amended the package inserts to include the warnings described earlier. The manufacturer’s prescribing information for Diprivan now also includes a recommendation to optimize haemodynamic and oxygen delivery parameters in sedated ICU patients. European Regulatory Authorities will be requiring statements in the prescribing information for all propofol preparations recommending that the administration of propofol be reduced or stopped if metabolic acidosis, rhabdomyolysis, hyperkalaemia and/or rapid progressive heart failure should occur, despite the lack of evidence for the efficacy of this as an isolated manoeuvre.


It is not currently possible to clearly state how propofol contributes to the development of the ‘propofol infusion syndrome’ or to its individual components.

Most events comprising the ‘syndrome’ appear explicable by recognized processes and may be rationally treated. Haemodynamic compromise and/or poor oxygen delivery should be assessed and managed. Whenever it occurs, lipaemia and its underlying cause should be addressed. The possibility of a mitochondrial effect at high doses should encourage the use of doses within, or at least close to, the recommended range. Glucose administration may theoretically assist in the management of lipaemia or where mitochondrial defects are present.

The authors hope that the information provided by this review will clarify the issues and physiology surrounding the so-called ‘propofol infusion syndrome’ and its individual component features and provide both an understanding of these events and a basis for physicians to prevent or treat them for the benefit of patients.

AstraZeneca regrets that it cannot disclose confidential information on individual patients held in its pharmacovigilance database. The discussion of this complex subject has been abbreviated for the purposes of publication; an opportunity to respond can be found at


This manuscript was prepared with editorial assistance from Mrs Evelyn Frearson, medical writing consultant (with financial support from AstraZeneca) and Dr Philip Wood (AstraZeneca). All authors are currently employees of AstraZeneca and based in Sweden, UK and USA, respectively. Final editing was performed by the editorial staff of the European Journal of Anaesthesiology.


1. Parke TJ, Stevens JE, Rice ASC et al. Metabolic acidosis and fatal myocardial failure after propofol infusion in children: five case reports. BMJ 1992; 305: 613–616.
2. Anonymous. Side-effects of propofol (Diprivan). Ugeskr. Laeger 1990; 152(16): 1176.
3. FDA Reports 1992; 54: 14.
4. Bray RJ. Propofol infusion syndrome in children. Paediatr Anaesth 1998; 8: 491–499.
5. Hatch DJ. Propofol-infusion syndrome in children. Lancet 1999; 353: 1117–1118.
6. Cremer OL, Moons KGM, Bouman EAC et al. Long-term propofol infusion and cardiac failure in adult head-injured patients. Lancet 2001; 357: 117–118.
7. Friedman JA, Manno E, Fulgham JR. Propofol. J Neurosurg 2002; 96: 1161–1162.
8. Prasad A, Worrall B, Bertram E et al. Propofol and midazolam in the treatment of refractory status epilepticus. Epilepsia 2001; 42: 380–386.
9. Cornfield DN, Tegtmeyer K, Nelson MD, Milla CE, Sweeney M. Continuous propofol infusion in 142 critically ill children. Pediatrics 2002; 110: 1177–1181.
10. Pepperman ML, Macrae D. A comparison of propofol and other sedative use in paediatric intensive care in the United Kingdom. Paediatr Anaesth 1997; 7: 143–153.
11. Cook S. Propofol infusion in children. BMJ 1992; 305: 952.
12. Cruickshank JM. Reduction of stress/catecholamine-induced cardiac necrosis by beta1 selective blockade. Lancet 1987; 2: 585–589.
13. Dixit S, Castle M, Velu RP et al. Cardiac involvement in patients with acute neurological disease. Confirmation with cardiac Troponin 1. Arch Intern Med 2000; 160: 3153–3158.
14. Tse H-F, Yeung C-K. From profound hypokalaemia to fatal rhabdomyolysis after severe head injury. Am J Med 2000; 109: 599–600.
15. Feinberg MS, Di Segni E, Freimark D et al. Transient left ventricular outflow tract obstruction and pseudohypertrophy due to severe hypovolaemia and catecholamine administration in a potential donor for heart transplantation. J Cardiovasc Diagn Proced 1999; 16: 125–129.
16. Perkin RM, Anas N, Lubinsky P. Myocardial ischaemia and disparate ventricular function after paediatric head injury. Crit Care Med 1991; 19: 587–588.
17. Schranz D, Stopfkuchen H, Jungst BK. Hämorrhagisches Lungenödem und Herzkreislaufinsuffizienz nach isoliertem Schadelhirntrauma. Monatsschr Kinderheilkd 1981; 129: 248–250.
18. Gustafsson U, Wardell K, Nilsson GE et al. Vasomotion in rat skeletal muscle induced by hemorrhage as recorded by laser-Doppler flowmetry. Microvasc Res 1991; 42: 224–228.
19. Siegemund M, van Bommel J, Racovitza I et al. Sepsis, microcirculation and vasodilators. Nederlandse Vereniging voor Intensive Care 2002; 6(5); 12–16.
20. Balasubramanian V, Townsend S, Boots RJ. Does splanchnic ischaemia occur in isolated neurotrauma? A prospective observational study. Crit Care Med 1999; 27: 1175–1180.
21. Vitale GC, Larson GM, Davidson PR et al. Analysis of hyperamylasemia in patients with severe head injury. J Surg Res 1987; 43: 226–233.
22. Gayan-Ramirez G, Decramer M. The effect of corticotherapy on respiratory muscles. Revue des Maladies Respiratoires 1998; 15: 33–41.
23. Reichhart MD, Lobrinus JA, Kuntzer T et al. Acquired neuromuscular disorders in critically ill patients. (Les complications neuromusculaires acquises de reanimation.) Med Hyg 2003; 61: 970–976.
24. Knibbe CAJ, Zuideveld KP, Dejongh J, Kuks PFM, Aarts LPHJ, Danhof M. Population pharmacokinetic and pharmacodynamic modeling of propofol for long-term sedation in critically ill patients: a comparison between propofol 6% and propofol 1%. Clin Pharmacol Therap 2002; 72(6); 670–684.
25. Rigoulet M, Devin A, Averet N et al. Mechanisms of inhibition and uncoupling of respiration in isolated rat liver mitochondria by the general anesthetic 2,6-diisopropylphenol. Eur J Biochem 1996; 241: 280–285.
26. Schenkman KA, Yan S. Propofol impairment of mitochondrial respiration in isolated perfused guinea pig hearts determined by reflectance spectroscopy. Crit Care Med 2000; 28: 172–177.
27. Kress JP, O'Connor MF, Pohlman AS et al. Sedation of critically ill patients during mechanical ventilation: a comparison of propofol and midazolam. Am J Resp Crit Care Med 1996; 153: 1012–1018.
28. Wolf A, Weir P, Segar P, Stone J, Shield J. Impaired fatty acid oxidation in propofol infusion syndrome. Lancet 2001; 357(9256): 606–607.
29. Withington DE, Decell MK, Ayed TA. A case of propofol toxicity, further evidence for a causal mechanism. Pediatr Anesth 2004; 14: 505–508.
30. Wolf AR, Potter F. Propofol infusion in children: when does an anesthetic tool become an intensive care liability? Pediatr Anesth 2004; 14: 435–438.
31. Vasile B, Rasulo F, Candiani A, Latronico N. The pathophysiology of propofol infusion syndrome: a simple name for a complex syndrome. Intens Care Med 2003; 29: 1417–1425.
32. Corr PB, Creer MH, Yamada KA, Saffitz JE, Sobel BE. Prophylaxis of early ventricular fibrillation by inhibition of acylcarnitine accumulation. J Clin Invest 1989; 83: 927–936.
33. Bonnefont JP, Djouadi F, Prip-Buus C, Gobin S, Munnich A, Bastin J. Carnitine palmitoyltransferases 1 and 2: biochemical, molecular and medical aspects. Mol Aspects Med 2004; 25: 495–520.
34. Roth MS, Martin AB, Katz JA. Nutritional implications of prolonged propofol use. Am J Health-Syst Pharm 1997; 54: 694–695.
35. Wasserberger J, Ordog GJ, Shoemaker WC. Inadequate critical care curriculum in pediatric emergency medicine fellowships. Acad Emerg Med 1994; 1: 503.
36. Zornow MH, Prough DS. Fluid management in patients with traumatic brain injury. New Hori Sci Pract Acute Med 1995; 3: 488–498.
37. Manley G, Knudson MM, Morabito D et al. Hypotension, hypoxia and head injury – frequency, duration and consequences. Arch Surg 2001; 136: 1118–1123.
38. Bingham WF. The limits of cerebral dehydration in the treatment of head injury. Surg Neurol 1986; 25: 340–345.
39. Luerssen TG, Wolfla CE. Pathophysiology and management of increased intracranial pressure in children (Chapter 4). In: Andrews BT, Hammer GB, eds. Pediatric Neurosurgical Intensive Care. Neurosurgical Topics Series. AANS Publications Committee. The American Association of Neurological Surgeons. New York: Thieme Medical Publishers, 1997. ISBN-10: 1879284456/3131324813.


© 2006 European Society of Anaesthesiology