Review of studies examining microbial contamination of vials used for preparations done with closed-system drug transfer devices

Abstract

Objectives

The main objective was to identify all studies that present data regarding microbial contamination of vials used for preparation with closed-system drug transfer devices (CSTDs). Our secondary objective was to compare the reported contamination of vials punctured with a CSTD versus no CSTD and to evaluate the quality of data reporting as defined by the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) criteria.

Methods

A literature review was conducted on 31 December 2018 on PubMed, EMBASE and Cumulative Index to Nursing and Allied Health Literature. A manual search of the archives of relevant pharmaceutical conferences was made. All studies that presented data about microbial contamination of vials punctured with a CSTD or about beyond-use date extension were included. Two researchers independently graded the articles according to the STROBE criteria.

Results

Of the 280 articles identified initially, 12 were retained for analysis. Studies evaluated microbial contamination according to different incubation times and different culture media. Nine studies did not use any comparator group. Five studies found no contamination of vials punctured with CSTDs. For the others, the contamination was between 0.3% and 27%. Three studies compared the contamination of vials punctured with a CSTD and with a conventional system and did not show a significant difference between the groups. Seven studies declared a conflict of interest. The mean number of STROBE criteria fulfilled was 12.2±4.1 out of 34 (7 not applicable) for studies, and the mean number was 5±0 out of 12 for abstracts.

Conclusions

Vials punctured in ISO5 conditions with a CSTD presented a low frequency of microbial contamination. No study showed a significant difference between vials punctured with a CSTD and with a conventional method. Centre-specific sterility testing is needed to reflect the variability of handling procedures and equipment.

Introduction

Sterile preparations must be done in controlled conditions to avoid preparation errors and microbial contamination. Contaminated sterile preparations have been linked to serious meningitis outbreaks.¹ Thus, many organisations elaborated sterile preparation standards. These standards may vary per region; for instance, there are many national standards in Europe. In Canada, the pharmacy regulatory authority from each province can adopt their own standards, but the National Association of Pharmacy Regulatory Authorities (NAPRA) provided model standards.² The US Pharmacopeia (USP) developed many standards for the USA and they also have an international influence.

According to USP <797> and NAPRA, single-dose vials used for sterile preparation that are punctured under International Organization for Standardization (ISO5) conditions must have a beyond-use date (BUD) of 6 hours in order to limit the risk of microbial contamination.³ This limit was established considering the exponential growth of bacteria after 6 hours of incubation.⁴ Microbial growth is possible even in antineoplastic solutions and it varies depending on the drug. Briceland et al ⁵ showed growth of Staphylococcus epidermidis at 35°C in doxorubicin, cytarabine for up to 21 days and in 5-fluorouracil for up to 5 days. Growth in cytarabine was dose-dependent.

Compounding in a cleanroom minimises the risk of microbial contamination. One author showed a striking difference of 16% (78/500) contaminated preparations that were done on the ward and 0% (0/500) that were done in a cleanroom.⁶ A meta-analysis of 34 studies published between 1946 and 2014 concluded that 0.5% (95% CI 0.1 to 1.6, n=6280) of doses prepared at the pharmacy showed microbial contamination,⁷ which was lower than the contamination of doses prepared in a clinical environment (3.7%; 95% CI 2.2 to 6.2, n=10 272 doses). Another meta-analysis of 16 studies published between 2000 and 2018 concluded that 0.08% (95% CI 0.00 to 0.00) of doses prepared at the pharmacy showed microbial contamination vs 7.47% (95% CI 5.16 to 9.79) when doses were prepared in a clinical environment.⁸ Indeed, four of five studies showed 0% of preparations contaminated at the pharmacy.⁸ Other authors concluded that if aseptic techniques were used and doses were prepared in a laminar airflow hood, very little contamination was observed independently of the location (traditional pharmacy room 0.30% (6/2027) vs cleanroom 0.34% (7/2030), p=1.0).⁹ Indeed, the simple introduction of isopropyl alcohol-disinfected non-sterile gloves or sterile gloves greatly reduced the contamination of preparations, showing the importance of touch contamination (bare hands or non-sterile gloves 5.2% (28/539); sterile gloves 0.34% (1/296)).¹⁰

However, when applying the 6-hour BUD, some opened vials are wasted, and this has a financial impact for healthcare centres.^{11 12} Some centres use closed-system drug transfer devices (CSTD) to prepare hazardous drugs in order to increase workers’ protection. They are ‘(…) device[s] that does not exchange unfiltered air or contaminants with the adjacent environment’.¹³ Some authors argue that by maintaining a closed environment, sterility is maintained and the BUD can be increased.^14–19 BUD extension should be supported by centre-specific sterility tests conducted with the same compounding protocol. The savings obtained by extending the BUD might offset their high acquisition cost.^{20 21}

The main objective was to identify all studies that present data regarding microbial contamination of vials used for preparation with CSTDs. Our secondary objective was to compare the reported contamination in vials punctured with a CSTD versus no CSTD and to evaluate the quality of data reporting as defined by the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) criteria.^{22 23}

Methods

Study selection and data extraction

A literature review was conducted on 31 December 2018 on PubMed, EMBASE and Cumulative Index to Nursing and Allied Health Literature. The search terms were ‘Closed-system-transfer-device’, ‘Closed system drug transfer device’, ‘CSTD’ and ‘Beyond-use date’. A manual search of the archives of relevant pharmaceutical conferences was made for relevant abstracts: Groupe d’évaluation et de recherche sur la protection en atmosphère controlée (2013–2017 period), Professional Practice Conference of the Canadian Society of Hospital Pharmacists (2013–2019), Canadian Association of Pharmacy in Oncology (2016–2018), and the mid-year clinical meeting (2012–2018) and summer clinical meeting (2013–2018) of the American Society of Health-System Pharmacists. A manual search was also conducted. Articles written in English, French or Spanish were included.

Studies were first selected based on their titles, and then their full texts were read. All studies that presented outcomes about microbial contamination of vials punctured with a CSTD or about BUD extension were included. Any study that presented their device as a CSTD was kept. All study designs were included. Studies performed by the manufacturers that were not published in an indexed or non-indexed pharmacy journal were excluded. Studies with voluntary contamination were excluded. Vials punctured with anything other than a CSTD were called ‘conventional method’. Data were extracted in an Excel spreadsheet; data extraction was independently validated by a second member of the research team. The investigators of the studies were not contacted.

Quality of reporting

For all included studies, we applied the STROBE checklist for cohort study (n=34 criteria) and the STROBE checklist for abstract conference (n=12 criteria). For each study, each criterion was reviewed (and categorised as yes, no or not applicable) independently by two members of the research team (AS, CT). Differences were resolved by consensus.

Results

Microbial contamination

A total of 280 articles were identified and 12 were included: 8 original research,^{18 24–30} 1 thesis,¹⁷ 2 poster presentation^{31 32} and 1 abstract^{33 34} (figure 1). A study published in two articles and one abstract was counted as one study.^{30 34} The English article was used for the STROBE criteria evaluation.³⁵ The studies were conducted in five different countries: USA (n=5), France (n=3), Canada (n=3) and Spain (n=1). PhaSeal was the CSTD most frequently reported (8/12). Seven studies declared a conflict of interest with a CSTD manufacturer, three did not specify, and two declared no conflict of interests.

Figure 1.

Selection of studies for inclusion in the review. BUD, beyond-use date; CINAHL, Cumulative Index to Nursing and Allied Health Literature; CSTD, closed-system drug transfer device.

All studies evaluated microbial contamination according to different incubation times, 24 hours (n=10), 48 hours (n=10), 72 hours (n=11), 7 days (n=9) and 14 days (n=5).17 18 24–33 A total of 1658 vials were tested, including 1586 vials punctured with a CSTD (between 6 and 332 vials per study, n=12 studies) and 72 vials punctured with a conventional system (12–30 vials, n=3). Nine studies did not use any comparator group (table 1). One study compared the microbial contamination with two conventional systems, and two compared both with a conventional system and another CSTD (table 2). Most (9/12) studies were conducted under ISO5 conditions.

Table 1.

Microbial contamination of vials used for compounding with closed-system drug transfer device without comparator

References	Manipulations	Culture and detection techniques	Incubation times	Global contamination	Comments
Charbonneau et al 32	ISO5 BSC (n=4) + LAFH (n=1) CSTD (Equashield). 100 vials of TSB.	Incubation at 37°C 14 days. Macroscopic evaluation.	0, 48, hours, 7, 9, 11, 13, 14 days.	0/100 (0%) vials.	Poster. Conflict of interest: none.
Wall and Abbott31	Not fully USP <797>-compliant in ISO5 BSC + CSTD (PhaSeal). 121 vials of antineoplastic drugs from daily production.	Incubation at 35°C 14 days. CFU count.	24, 48, 72, 96 hours, 5, 6, 8 days.	5/765 (0.65%) vials.	Poster. Conflict of interest: unspecified.
Perks et al 33	ISO5 BSC + CSTD (Equashield). 192 vials of TSB.	Incubation at 37°C 14 days. Macroscopic evaluation.	1, 2, 3, 4, 5 days.	0/192 (0%) vials.	Abstract. Conflict of interest: unspecified.
Ho et al 26	According to USP <797> in ISO5 class IIB2 BSC +CSTD (PhaSeal). 12 vials of 5-fluorouracil in conditions similar to actual practice.	Incubation of 100 mL TSB intravenous bags at 35°C 14 days. Macroscopic evaluation.	0, 6, 24, 48, 72 hours, 5, 7, 14 days.	0/96 intravenous bags.	Conflict of interest: Celgene, Millennium, Seattle Genetics, Merck, Astellas and Onyx.
Barny and Blond17	Class 2 BSC + CSTD (PhaSeal). A: 160 vials of TSB. B: 30 ganciclovir stock solutions.	A: Incubation of syringes at 37°C 7 days. B: Incubation of Schaedler tubes and sheep blood plates. Macroscopic evaluation and CFU count and subsequent culture plates.	0, 72 hours, 7 days.	A: 0/480 syringes. A:19/960 plates. B: 1/65 tubes. B: 2/65 plates.	Conflict of interest: thanks to Becton Dickinson.
Sánchez-Rubio Fernández et al 30 34	ISO5 class IIB BSC + CSTD (PhaSeal). 80 vials of TSB.	Incubation of vials, intravenous bags and syringes at 25°C–35°C 14 days. Macroscopic evaluation.	0, 24, 96 hours, 7 days.	0/740 vials.	Conflict of interest: unspecified.
Rowe et al 18	According to USP <797> in ISO5 BSC + CSTD (PhaSeal). 56 vials of leftover antineoplastic drugs. 45 vials of TSB.	Incubation at room temperature 48 hours. 0.5 mL on blood agar. 0.5 mL of TSB agar. Macroscopic evaluation and CFU count.	6, 24, 48, 72 hours, 7, 14 days.	6/332 antineoplastic vials. 5/270 TSB vials 11/592 (1.8%).	Conflict of interest: Carmel Pharma.
McMichael et al 24	According to USP <797> + CSTD (PhaSeal). 332 vials of TSB.	Incubation of 2 mL syringes at 37°C 14 days. Macroscopic and microscopic evaluation and subsequent culture plates.	24, 48, 72, 96 hours, 7 days.	17/1328 syringes. 6/332 (1.8%).	Conflict of interest: Carmel Pharma.
Carey et al 25	According to USP <797> in ISO5 BSC or CACI + CSTD (PhaSeal). 332 vials of TSB.	Incubation of 100 mL intravenous bags at 35°C 14 days. Macroscopic evaluation.	24, 48, 72, 96 hours, 7 days.	1/1660 intravenous bags. 1/332 (0.3%) vials.	Conflict of interest: Carmel Pharma.

BSC, biological safety cabinet; CACI, compounding aseptic containment isolator; CFU, colony forming unit; CSTD, closed-system drug transfer device; ISO5, International Organization for Standardization 5; LAFH, laminar airflow hood; TSB, tryptic soy broth; USP, US Pharmacopeia.

Table 2.

Microbial contamination of vials used for preparation with closed-system drug transfer device compared with a conventional system

References	Manipulations	Culture and detection techniques	Incubation times	Global contamination	Comments
Whitehead et al 27	According to USP <797> in ISO5 cleanroom + 1 CSTD (brand not specified) and 2 conventional systems (1 NCDP spike and 1 syringe) 22 vials of alfaxalone without preservative.	Incubation of tubes at 36°C 18–24 hours. Macroscopic evaluation and subsequent blood agar plates.	1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14 days.	6/22 (27%) tubes (1 CSTD, 1 NCDP and 4 syringes).	CSTD unspecified. Conflict of interest: Jurox Pty Ltd.
Garrigue et al 29	Outside the isolator in an ISO7 cleanroom + classical spike and 2 CSTD (PhaSeal, Viashield/Texium). 3 media-fill tests of TSB.	Incubation at 30–35° 14 days. Macroscopic evaluation.	1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14 days.	0/30 PhaSeal. 0/30 Viashield/Texium. 0/30 Classical spike.	Conflict of interest: CSTDs provided by Becton Dickinson and CareFusion.
Savry et al 28	Outside the isolator in an ISO7 cleanroom + classical spike and 2 CSTD (PhaSeal and Spiros/Genie). 3 media-fill tests of TSB.	Incubation at 30–35° 14 days. Macroscopic evaluation.	1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14 days.	0/30 conventional. 0/30 PhaSeal. 1/40 (2.5%) Spiros/Genie.	Conflict of interest: none.

CSTD, closed-system drug transfer device; ISO5, International Organization for Standardization 5; NCDP, non-closed dispensing pin; TSB, tryptic soy broth; USP, US Pharmacopeia.

Five studies found no contamination of vials punctured with CSTDs. For the other studies, the contamination was between 0.3% and 27%. Most contaminations reported were deemed to be caused by handling errors and external contamination. The three studies that compared the contamination of vials punctured with a CSTD and with a conventional system did not show a significant difference between the groups. Two studies showed no contamination and one study showed 5 of 12 (41.7%) contamination.

Quality of reporting

Nine studies were evaluated with the STROBE checklist for cohort study and three abstracts/posters with the STROBE checklist for abstract conference. The mean number of STROBE criteria fulfilled per study was 12.2 (SD 4.1) out of 34 criteria. For abstracts, the mean number was 5 (SD 0) out of 12. There was a maximum of 17 of 34 criteria reported in the studies and 5 of 12 in the abstracts. In most studies, key elements were lacking, notably a clear study design, the setting (dates, locations), a clear explanation of the number of participants (vials), a mention of the risk of bias, an explanation of the study size, a discussion of limitations, a cautious interpretation of results and their generalisability (figure 2). In the abstracts/posters, the key elements lacking were a clear study design, a complete setting, a clear list of participants (vials) and an explanation of statistical methods (figure 3).

Figure 2.

Evaluation of study reporting according to the Strengthening the Reporting of Observational Studies in Epidemiology checklist for cohort study, with data presented as the number of studies meeting each criterion.

Figure 3.

Evaluation of study reporting according to the Strengthening the Reporting of Observational Studies in Epidemiology checklist for abstract conference, with data presented as the number of studies meeting each criterion.

Discussion

Microbial contamination

Twelve studies that presented data regarding microbial contamination of vials used for preparation with CSTDs were found. Seven showed microbial contamination of vials, but this contamination was mostly attributed to handling errors. This low contamination is used in support of extending the BUD of punctured vials up to 7 days. Most studies were done in a laboratory setting with culture broth, but some used leftover drug vials or replicated actual working conditions. Three concluded that extending the BUD would lead to savings.^{14 17 19} Most studies reported conflict of interests with CSTD manufacturers.

Sterility tests need to be centre-specific to account for the differences in handling procedures. In addition, the potential for microbial growth varies depending on the antineoplastic drug, so tests for each drug would be warranted. Different culture media should be tested to account for the different micro-organisms. Roy et al ³⁶ tested 25 antineoplastic drugs and showed that they did not behave in the same way. Eight of the studied drugs did not allow growth in any of the conditions tested. They did not notice any bacterial growth, but approximately 10% of their tests had mycological growth. There are many brands of CSTDs, but currently no universal protocol is approved to test their capacity to maintain a closed environment, so the generalisation of the results is limited.

Comparison with conventional methods

Only three studies compared CSTDs and conventional methods, and none showed a significant difference between both groups. In three studies, between 6 and 10 vials were punctured in each group. This small sample size is not sufficient to show a statistically significant difference. If 0.5% is used as the baseline level of microbial contamination of doses prepared at the pharmacy,⁷ in order to prove the superiority of CSTDs in reducing this contamination to 0.1%, a study with 2311 vials in each group would be needed (alpha 0.05, power 0.8). Even though many studies showed no microbial contamination of vials punctured with a CSTD under ISO5 conditions, no study was able to demonstrate that adding a CSTD improved this outcome (and no study was powered to do so).

Considering the low level of microbial contamination of vials punctured under ISO5 conditions, it is our opinion that a convincing study should compare vials punctured with CSTDs and with a conventional method, be adequately powered, and free from conflict of interests.

Perhaps CSTDs could offer an added benefit of maintaining sterility when drugs are prepared outside the pharmacy, as preparations done in these areas are known to be more prone to microbial contamination.^{6 7} However, CSTDs are not a replacement to using biological safety cabinets.³ A group evaluated the transfer of micro-organisms in sterile preparations after voluntarily introducing contamination either on the rubber stopper of the vial or at the top of the inlet of the CSTDs.³⁷ None of the devices evaluated maintained sterility (ie, Chemoprotect Spike, Clave Connector, PhaSeal, Securmix and the conventional system), and the contamination risk increased with the number of couplings. The authors note that the low inoculum probably represented a higher contamination risk than real-life contamination (the low inoculum magnitude was estimated per their tests representing potential contamination scenarios). Vials punctured with PhaSeal were less contaminated, but it should be noted that even the addition of a two-step decontamination of the vials and the use of a PhaSeal CSTD did not succeed in reducing the bacteria count to zero viable cells for all strains tested, thus highlighting the need for excellent handling practices in a controlled environment.

Quality of reporting

The majority of the STROBE criteria were not reported in the studies. Key elements for data interpretation were missing, such as study design, setting, date of study and risk of bias. The quantity of vials at each step of the study was often hard to grasp. The discussions lacked a limitation section, a cautious interpretation of results and an adequate statement about the generalisability of results.

Strengths and limitations

This is the first literature review about the microbial contamination of vials used for preparation with CSTDs. Considering the emerging literature on this topic, we considered three published abstracts from two Canadian meetings to have a better literature scan. Further review could eventually consider only published studies in indexed journal. Few studies compared the use of CSTDs with conventional methods, limiting the possibility of concluding on the added benefit. The methods used in each study were variable, which limited the interpretation of data. Only three different CSTD brands were represented in the studies.

Conclusion

Following the implementation of high compounding standards, the majority of vials punctured in ISO5 conditions with an aseptic technique were not contaminated. Vials punctured under the same conditions with a CSTD also presented a low frequency of microbial contamination. No study showed a significant difference between vials punctured with a CSTD and with a conventional method. Centre-specific sterility testing is needed to reflect the variability of handling procedures and equipment.

What this paper adds.

What is already known on this subject

• Single-dose vials used for sterile preparations that are punctured under International Organization for Standardization 5 (ISO5) conditions must have a beyond-use date (BUD) of 6 hours in order to limit the risk of microbial contamination.

• When applying the 6-hour BUD, some opened vials are wasted, and this has a financial impact for healthcare centres.

• In the literature, some authors argue that closed-system drug transfer devices (CSTDs) maintain the sterility of vials, which potentially increases their BUD and leads to cost-savings.

What this study adds

• This is the first review of the literature on the microbial contamination of vials used for preparation with CSTDs.

• Vials punctured under ISO5 conditions with a CSTD presented a low frequency of microbial contamination, but no study showed a significant difference between vials punctured with a CSTD and those with a conventional method.