Data Extraction and Management

We stratified selected articles into 4 categories. 1) Features of propofol-related outbreaks worldwide (e.g., year, geographic localization, type of procedure, route of propofol contamination, type of microorganism isolated, number of cases, and number of deaths) were compiled for each outbreak during 1989–2014. Outbreak reports published more than once were occasionally encountered; only the most representative study was included to prevent data duplication. In addition, we took a conservative approach in extracting data from single cases if strong associations between propofol exposure and the infectious event were reported. 2) Laboratory-based evidence of propofol as a microbiological reservoir was retrieved regarding the frequency of contaminated propofol syringes, vials, or infusion lines used in operating rooms (ORs) or intensive care units (ICUs). 3) Epidemiologic evidence concerning the risks of infections associated with propofol was confirmed by case-control, cohort, and clinical trial studies (Technical Appendix). 4) Studies on propofol formulations were used for evaluation of the data suggestive of a reduction in microbial growth associated with specific propofol formulations.

We screened 465 abstracts and chose 53 to examine: 25 outbreak reports, 3 reports of single cases, 7 laboratory-based studies on propofol, 10 analytical studies that supported the healthcare effect of contaminated propofol in terms of the risk for infection, and 8 studies of the alternative propofol formulations (Figure 1). We retrieved an additional 5 articles by using references cited in 3 of the initial 53 articles to expand on specific points.

Worldwide Occurrence of Outbreaks Associated with Propofol-Based Anesthesia

The risk for postoperative infection depends on a variety of factors, including wound class (i.e., clean, clean-contaminated, contaminated, or dirty), the condition of the patient, type and length of surgery, use of antimicrobial drugs, and perioperative events. During the past 2 decades, several episodes of sepsis worldwide have been reported to be associated with propofol administered by syringe injection or used as a continuous infusion (2–4,6,8–15) (Table 1). These cases were reported in industrialized countries; no outbreaks have been documented in developing or low-income countries, such as those in Latin America, Africa, or Asia (Figure 2), likely as a consequence of deficiencies of surveillance programs and poor data acquisition regarding the frequency of contaminated propofol. Propofol-associated infections likely occur in developing countries with relatively higher frequency than in industrialized countries, related to the common problem of economic restraints and reduced use of universal precautions within the healthcare systems, leading to reuse of syringes and use of vials for multiple patients.

Since this anesthetic was introduced, 20 propofol-related infectious disease outbreaks have been reported worldwide, affecting 144 patients and resulting in 10 deaths; these outbreaks have lasted between 8 hours and 65 days (Table 1). However, many outbreaks related to propofol are likely undocumented and such reports do not reflect ongoing sporadic infections that are likely to be linked to propofol use.

Outbreaks have been associated with widely diverse types of procedures in both intensive care units (ICUs) and operating rooms (ORs) (Table 1). Although no specific clinical procedure has been causally related to propofol-related infectious outbreaks, endoscopic procedures have been the most frequently associated with propofol-related infections during the past 20 years.

Contrary to some healthcare perceptions, none of the reported outbreaks to date have been correlated with intrinsic batch-contamination of propofol. Nevertheless, some peer-reviewed reports of manufacturing deficiencies exist, as well as >2 outbreaks probably linked to intrinsic contamination; however, the latter were published on government Internet sites and therefore are not included in this review. A US government report traced 41 cases of infection in 2009 to 1 contaminated batch (http://www.fda.gov/Safety/Recalls/ArchiveRecalls/2009/ucm172474.htm); 9 other possible cases were traced in Australia in 2014 (https://www.tga.gov.au/alert/propofol-provive-and-sandoz-propofol-1-emulsion-injection-all-sizes-and-all-batches-update-3).

Distribution of propofal-related outbreaks has been widespread geographically (Figure 2) and temporally (Figure 3). The United States has reported 11 outbreaks, the highest number of outbreaks during the assessed period, averaging 1 every 2 years and accounting for 55% of all reported outbreaks worldwide. According to the list of healthcare-associated hepatitis B and C virus (HBV and HCV, respectively) outbreaks reported by the US Centers for Disease Control and Prevention during 2008–2014, the coincident recent exposure to propofol was considered a factor leading to the screening of >60,000 patients (http://www.cdc.gov/hepatitis/Outbreaks/PDFs/HealthcareInvestigationTable.pdf).

In a 2003 study of the literature on hospital-acquired infections worldwide, Vonberg and Gastmeier calculated a mortality rate of 13.8% related to administration of propofol (17). The data we collected indicate an estimated mortality rate in propofol–associated infections of ≈9.3% (range 0%–50%) (Table 1). This value only summarizes the current published literature describing propofol–related outbreaks, and thus it may not represent the true magnitude of the problem (2). According to an inspection of data held by AstraZeneca, 345 cases of postoperative infections or febrile syndrome occurred after propofol use in the United States during November 1989–November 2004; unfortunately, data for these cases were archived and not published (2,18).

Mechanisms of Contamination of Propofol Formulations

Microbiological contamination of propofol lipid emulsions may occur from the environment either during manufacture (intrinsic) or after vial opening (extrinsic), the latter of which is the most frequent. The horizontal transmission of pathogens in anesthesia begins with the breach of handling precautions by anesthesia providers of devices or drugs. Frequently neglected precautions during the induction and maintenance of anesthesia include hand hygiene and protection against incidental propofol contact with the environment (19–21). In addition, surreptitious use of IV anesthetics by drug-addicted healthcare workers could raise the risk for extrinsic contamination (22). Other factors that may potentially affect the sterility of propofol in clinical use include preparation of multiple syringes for use throughout the day; re-use of vials, syringes, or infusion-pump lines on >1 patient; use of opened ampules longer than recommended by the manufacturer; and failure to wear sterile gloves during handling of propofol and to dispose of them after each contact.

The most common reservoirs associated with extrinsic contamination of propofol are syringes or micro-droppers, vials, and IV stopcock dead space. Syringes or micro-droppers have been implicated in most outbreaks (23–27). Propofol vials have been demonstrated to be a reservoir for microbes when contents are exposed to the environment (28). Delays in administration after propofol vials have been opened are a recognized risk factor; the degree of contamination of an opened vial may increase by 20%–26% after 12 hours (29). Propofol is available in vials of various volumes; a typical 20-mL vial contains 200 mg, and 50- and 100-mg vials are available. It is believed that administering a dose >200 mg to an adult in an OR will increase the probability of using an additional vial as a multi-dose vial for >1 patients. IV stopcock dead space has been shown to provide a potential route of entry for pathogenic, multidrug-resistant bacteria in infusion lines (20,27).

A number of microorganisms are associated with propofol in clinical- and laboratory-based studies at varying frequencies (Table 2). We report the specific pathogens that have been associated in several outbreaks (2–4,6,8–10,12–15,30,31), as well as all of the transmissible microorganisms that have been observed in contaminated propofol in syringes, vials, and infusion lines (23,25,27,30,32). We also describe 3 reported cases of septic shock related to propofol (33,34).

Propofol is an excellent medium not only for bacterial growth but also for fungal infections, which have been associated with propofol use particularly when poor hygienic standards are observed during the administration. Viral infections with HCV and HBV have also been demonstrated, possibly explained by the viral stability offered by propofol emulsions (9,11,13). Overall, ≈23% of the published infection outbreaks associated with propofol were caused by HCV (18.1%) and HBV (4.2%); 21.5% by Candida albicans; and 47.2% by bacteria (gram-positive 27.1%, gram-negative 20.1%). In the remaining reports, no microorganisms were identified, possibly as a consequence of concurrent antimicrobial drug therapy. The most frequent pathogens associated with propofol-related outbreaks, in order of frequency, were Staphylococcus aureus (39/144), Candida albicans (31/144), and HCV (26/144).

Laboratory-based (Table 3) and epidemiologic (online Technical Appendix Table 1) microbiological studies have demonstrated that the production of bacterial endotoxins is greatly enhanced by propofol solutions. Case reports of endotoxemia associated with the use of contaminated propofol have also been published (5,34,35).

Frequency of Contaminated Propofol Used in ICUs and ORs

Microbiological observations of opened propofol vials were reported in the first studies that identified bacterial growth in propofol, and observational studies on propofol formulations have determined the proportion of extrinsic contamination (Table 3). In 1994, Farrington et al. established an incidence rate of 6% (3/50) of contaminated propofol syringes in an ICU (31). Webb et al. retrospectively observed similar results in a different ICU setting with an incidence of 5.9% (18/302) (23). Soong observed a lower propofol contamination incidence (3.0%) in ORs and also noted an association between postoperative infections and vials from which multiple patients were medicated (24); Bach et al. found similar results (30). In 1995, McHugh and Roper reported an incidence of 6.3% (16/254) of infection when propofol was administered from vials but did not find that delays in the administration of propofol were associated in any increased likelihood of bacterial contamination (25). Cole et al. recorded the incidence of contamination as high as 17.3% (26/150) in propofol found in stopcock dead space (27).

The distribution of instructions for aseptic measures for handling propofol has shown to reduce the rate of contaminated propofol. Lorenz et al. reported that after a specific protocol for handling propofol was introduced and adhered to, a reduction in extrinsic contamination was achieved when compared with only adhering to the manufacturer’s recommendations (8.8% versus 11.3%) (26). That protocol included additional aseptic precautions, such as refilling empty syringes for use on multiple patients by using a 3-way stopcock and replacing only the infusion line to the patient. Data from a study performed in a high-complexity hospital in Cali, Colombia, showed substantial microbial growth of 6.1% (12/198) in residual propofol vials used in ORs at a tertiary hospital in Cali (32).

By collating the incidences of contaminated propofol in containers or devices, the data suggest that it is more common to encounter contaminated previously used vials of propofol (5.6%, 31/552) than in used syringes (5.4%, 48/896) and infusion systems (4.4%, 31/702). Similarly, analyzed by hospital location, the percentage of contaminated propofol is greater in ORs (7.3%, 103/1,405) than in ICUs (4.0%, 36/894), correlating well with the expected prevalence of opened vials and syringes used for bolus injections of propofol for multiple patients in the OR compared with those in the ICUs.

Risk for Infection Data Derived from Analytical Studies

After we reviewed the initial microbiological studies, our interest increased in ascertaining potential links between the use of propofol infusions and the incidence of infections and sepsis after surgery. We analyzed 10 epidemiologic studies (Table 4; Technical Appendix Table 1); of these, 4 assessed the associations with infection when practitioners followed manufacturers’ instructions for propofol handling, 5 assessed the association with infection when practitioners did not follow instructions, and 1 did not report on this issue. Studies following manufacturers’ precautions stated such within the articles, but the degree of compliance of such precautions was not documented (Technical Appendix Table 2). In 4 of 5 studies (80%) during which practitioners did not follow handling precautions, high infection risk was noted. In 4 studies in which precautions were followed, 2 (50%) scored above the significant risk threshold (Technical Appendix Table 1). These findings underscore the controversy surrounding the utility of existing handling protocols and demonstrates the continued high potential of propofol as a causal factor of iatrogenic infection. Other authors have explain the susceptibility to infection because of the attenuation of the immunological activity caused by propofol infusions. More studies of specific handling protocols are required before a significant risk reduction is clearly observed. We developed an algorithmic approach that shows certain crucial measures to prevent future propofol-related outbreaks of infections (Figure 4); this approach was based on our analyses and summaries of the epidemiologic and clinical data selected.

Proposed Propofol Formulations without Risk for Infection

There are currently no propofol formulations without infection risk; however, several investigators have advocated the use of antimicrobial additives. As required by the FDA, the efficacy of such additives must retard microorganism growth to <10-fold at 24 hours after extrinsic contamination of propofol (1); many suggested antimicrobial agents have been rejected because of their poor efficacy, additional side effects, or higher costs (1). The fact that propofol is a lipid emulsion poses problems when additives are considered because admixture with other substances, especially charged species with differing partition coefficients, can interact with either the lipid phase or the emulsifying agents, resulting in emulsion destabilization. Older and newer advances on the use of additives are described in Table 5, which includes the scientific progress related to patients, clinical trials, original articles, and brief reports.

Propofol with EDTA

As an antimicrobial ion chelator, EDTA exerts its effect by removing divalent and trivalent metal cations, causing rupture of the microbial cell membrane by loss of control of osmotic pressure gradients. This combination is approved by the FDA; clinical trials demonstrated antimicrobial efficacy and safety in humans (18,36). For manufacturers, EDTA is the most broadly incorporated agent in combination with propofol formulations. Despite its widespread use, some controversy remains over its selection as the optimal additive (37). Moreover, in several developing countries, cost remains a considerable limiting factor for the use of EDTA-containing propofol formulations (A. Zorrilla-Vaca et al., unpub. data).

Fospropofol Disodium

A newly introduced agent, fospropofol disodium, is a water-soluble pro-drug of propofol that currently has a small evidence base for its use of 3 published clinical trials in the literature; in these studies, fospropofol was assessed for use in sedation for colonoscopies, bronchoscopies, and coronary artery bypass graft surgery, and showed an acceptable safety profile (38). The advantage of this drug in reducing the risk for infectious events is that it does not have a lipophilic formulation that would support bacterial growth. Nonetheless, the drug has some disadvantages that could discourage its use, such as transient paresthesias and pruritis in the perineal and perianal regions (38).

Anesthetic Mixtures

Pain on injection is a common side effect of propofol. Anesthesiologists use a variety of strategies to reduce this, such as the addition of 1–2 mL of lidocaine to the propofol before injection. Lidocaine, like other amide local anesthetics, has bacteriostatic properties, which could theoretically reduce the chances of infection (29). It is, however, not currently known whether lidocaine has sufficient antimicrobial effect to make a clinical difference in infection rates.

Other Additives

Benzyl alcohol at concentrations of <2% has been used as a preservative agent in propofol formulations (39). Despite its bacteriostatic activity, benzyl alcohol used in propofol formulations is limited by its toxicity and instability in the combination. Other additives, such as phenylmercuric nitrate, phenylmercuric acetate, chlorobutanol, and phenol have been studied experimentally with propofol; however, all of these agents were rejected because of their potential toxicity. The sodium metabisulfite–containing formulation, created originally with the aim of reducing the pain of propofol injection, has been shown to possess preservative properties. Unfortunately, it has a labeled pH from 4.5 to 6.4, which is different from the FDA requirement of a pH of 6–8.5 for propofol (1). The nonlipid nanoemulsion and EmulSiv filter are the most recent alternative propofol formulations (40). In recent studies, these formulations of propofol have attained a level of antimicrobial activity above that observed with propofol with EDTA (40). Costs are currently a limiting factor for their use, but these 2 options seem to provide some promise for the future if production costs decline.