Introduction

Historically, transferring drug products from vials use a needle to pierce the vial stopper and syringes to withdraw the vial contents. The use of needles is however associated with a number of risks, mainly injuries and/or potential transfer of infections for health care providers (1). Accordingly, manufacturers have developed a number of specific devices, such as spiking devices, that can circumvent these risks during drug transfer from vials. To protect the health care practitioners, the EU issued a directive that aims at preventing injuries due to sharps, including needles, in healthcare setting and encourages elimination of sharps or use of safer alternatives (2). Also, many professional bodies, such as the Association for professionals in infection control and epidemiology (APIC) are encouraging the use of such needle-free devices (3).

Another risk for health care professionals can be related to the accidental, potential exposure to hazardous drugs (such as cytostatics) during drug transfer from vials. Closed system transfer devices (CSTDs) are a subset of spiking devices that aim at containment of hazardous drugs (4). The National Institute for Occupational safety and Health (NIOSH) defines CSTDs as “a drug transfer device that mechanically prohibits the transfer of environmental contaminants into the system and the escape of hazardous drug or vapor concentrations outside the system.” Accordingly, they should maintain integrity and product sterility by preventing microbial ingress, and simultaneously prevent exposure due to mist formation or spillage during drug transfer (5).

PhaSeal was the first FDA-approved CSTD in 1998. Since then, many other CSTD products have been developed, and currently more than 8 systems are commercially available in the US (6). Generally, CSTDs achieve their functionality through one of two mechanisms (6); either 1) a membrane to membrane system, where a needle pierces through membranes for removal of the liquid and pressure equalization, and is retraced safely once the process is finished, or 2) through a needle-free common fluid path that is activated once the pieces are correctly connected.

The interest and usage of CSTDs has significantly increased after USP <800> (7) was issued and became effective from December 2019. The monograph describes measures to be followed by entities that store, prepare, transport, or administer hazardous drugs. The later are defined as “Any drug identified by at least one of the following criteria:

  • Carcinogenicity, teratogenicity, or developmental toxicity

  • Reproductive toxicity in humans

  • Organ toxicity at low dose in humans or animals

  • Genotoxicity or new drugs that mimic existing HDs in structure or toxicity”

According to USP <800>, in case the information available is not sufficient to make an informed decision about the toxicity of a drug, it should be treated as hazardous. This is why users may be tempted using CSTDs systemically across many products and entities, possibly even more for products yet in clinical development. Thus, CSTDs seem to be also increasingly used for biotechnological drug products such as antibodies in Oncology settings, although the toxicological and safety concerns for accidental exposure of an antibody are generally different and usually much less than e.g., a small-molecule cytotoxic drug.

Among the engineering controls for containment described in USP <800>, it recommends the use of “containment supplemental engineering controls”, and names CSTDs as an example for such control measures. Interestingly, USP <800> clearly mentions that it is not known whether all CSTDs will perform adequately, simply due to the lack of a universal performance standard to evaluate CSTDs. USP <800> requests users to revert to independent peer-reviewed studies.

The majority of the studies published so far on CSTDs have focused on their main functionality; namely on the “containment” function (4). For example, Vyas et al. analyzed wipe samples from isolators taken before and after handling cytotoxic infusion preparations. They have shown that without CSTDs, the contamination with cytotoxics was significant, and that CSTDs could reduce – but not prevent – such contamination (8). The same conclusion was reached by Sessink et al. in a study involving 30 US hospital pharmacies (9). Similarly, Jorgenson et al. found that, out of 5 different CSTDs, only one system did not lead to leakage of vapors or spillage of product droplets (10). González-Haba Peña et al. also compared different systems and concluded that connectors are required to prevent spillage (11), and that contamination was observed at critical points in 2 out 3 tested systems (12). Another study by Siderov et al. also reported that the use of a closed system drug transfer device reduces surface contamination (13). Implementation of the closed system drug-transfer device for preparing and administering chemotherapy eliminated surface contamination with cytotoxic agents at the ambulatory cancer chemotherapy infusion center (14).

In addition to the aforementioned plethora of studies on the containment function of the CSTDs, one study was recently published evaluating the compatibility of an ADC with 6 CSTD systems in a simulated in-use study (15), and another one provided the perspective of a number of pharmaceutical companies on using CSTDs with biologics (16). Other than the aforementioned literature, there seems a lack of information on other important attributes for CSTDs with potential impact on product quality. Given the increasing focus on product quality, and aspects that could negatively impact it, including stopper fragmentation and deformation, particulates, breach of container closure integrity (CCI) and thus sterility, there is a need to have these aspects evaluated (17).

In the current study, we investigated a wide range of performance aspects of CSTDs. We evaluated four typically used CSTDs from different suppliers in combination with different container closure systems (CCS), i.e. different vial sizes and vial types as well as various stoppers. The investigations involved assessing (a) integrity testing using helium leak test as a highly sensitive testing approach. Additionally, we evaluated (b) the force required to mount the vial adaptor for each CSTDs depending on the stopper material and design. We also evaluated (c) the potential for coring as well as visible and subvisible particle generation after pushing the CSTD through the rubber stopper. And finally, the (d) extractable & delivered volume was determined with solutions having different viscosity and surface tension.

Materials and Methods

CSTDs designs for 20 cc ISO vials and CSTDs for 13 cc ISO vials from four different suppliers were investigated.

  • CSTD1 (13 mm and 20 mm) consisted of two parts, a vial adapter featuring a flexible air reservoir and a syringe adapter

  • CSTD2 (13 mm and 20 mm) consisted of two pieces, a membrane ventilated vial adapter and a syringe adapter

  • CSTD3 (13 mm and 20 mm) consisted of two pieces, a vial adapter and a ventilated syringe adapter

  • CSTD4 (13 mm and 20 mm) consisted of three parts, a vial adapter featuring a flexible air reservoir, a syringe adapter and a transfer syringe

Whenever appropriate, the CSTDs performance was compared to a commonly used 2 mL disposable syringe for withdrawal, equipped with an ISO 21G hypodermic needle (without a CSTD).

Vials

ISO 2R and 6R pharmaceutical type-I glass vials were washed and autoclaved to ensure primary packaging components were particle free. The 2R and 6R vial featured a hydrophilic inner vial surface. In addition, a 2R vial featuring a hydrophobic inner vial surface was used to investigate the effect of inner vial surface properties on extractable volume.

Stoppers

13 mm serum (Serum) and lyo (Lyo) halobutyl rubber stopper featuring a proprietary coating were used to seal the 2R vials.

20 mm serum (Serum), lyo (Lyo1) and lyo (Lyo2) halobutyl rubber stopper featuring a proprietary coating were used to seal the 6R vials.

Caps

13 and 20 mm aluminum caps featuring a flip-off button were used to crimp the 2R and 6 R vials, respectively.

All container closure systems were sealed using an Integra West Capper (Genesis Packaging Technologies, Exton, PA, USA). The following capping process parameters were set for to seal the container closure systems: Compression force: 30lbs (133.45 N); capping block height: 145 (unit less machine parameter). All container closure system selections were based on prior knowledge and were used also clinically.

Placebos Surrogate Solutions

Two placebo surrogate solutions and a solution containing protein were prepared to characterize the performance of the CSTDs. The composition of the surrogate solutions were selected to mimic typical drug product formulations properties (Table I).

Table 1 Placebo Surrogate Solutions
Full size table

Penetration Force

The force required to penetrate the stopper using the CSTDs was measured using a Zwick Z010 10 kN allround table top instrument (Zwick / Roell, Ulm, Germany) featuring a 500 N load cell (force sensor) and an in-house build sample fixture. The measurement speed was set to 200 mm / min, as per ISO-8536-2 (18) The data analysis was performed with the Zwick software TestXpert version 3.7.

Integrity

Integrity of the system, including the connected CSTDs was evaluated using a Mass spectrometry based Helium leak tracer gas CCIT, as previously described by Morrical et al. (14) and Mathaes et al. (21). In brief, an in-house manufactured airtight flange for 2R or 6R vials (Fig. 1) was connected to an ASM340 Helium gas detector (Pfeifer Vacuum, Asslar, Germany). Samples were prepared by cutting an incision into the vial bottom using a diamond saw (Arnold Gruppe, Weilburg, Germany). Vials were placed into the airtight flange, the vacuum pump was turned on, and a measurement was performed while helium gas was constantly applied. Leak rates displayed on the mass spectrometer were recorded. Leak rates <10−8 mbarL/s were reported as 10−8 mbarL/s. Prior any measurement a system suitability test was performed comprising a negative (tight iron plug) and positive control sample (vial featuring an artificial leak) and a helium gas standard. In case CSTDs featured an external air vent e.g. a filter, the filters were blocked using epoxid glue to enable helium leak measurements. 3 independent measurements were performed.

Fig. 1
figure 1

Helium leak measurements: (A) 2 R vial configurations, (B) 6 R vial configurations.

Full size image

Sub-Visible Particle Characterization by Flow Imaging

The placebo solution and the protein solution were filtered using a sterile filter. Vials were filled and sealed under a laminar flow bench to ensure a particle free environment.

Sub-visible particles were characterized using a MFI DPA5200 instrument equipped with a silane coated SP3 flow cell (ProteinSimple, San Jose, CA).

After flushing the system with purified water, a blank measurement was performed to ensure cleanliness. Prior to sample analysis, a SST for particle size and count using a 5 μm particle size standard and 5 μm particle count standard (COUNT-CAL Count Precision Size Standard; ThermoFisherScientific, Waltham US) was conducted. We used the sample to optimize the illumination. The analyzed sample volume was set to 0.59 mL. Subvisible particles ≥5 μm were analyzed regarding the aspect ratio morphology parameter. Subvisible particles ≥5 μm were devided into particles featuring a aspect ratio ≥ 0.85 and < 0.85. The MVSS software (ProteinSimple, San Jose, CA) was used to calculate Equivalent Circle Diameter (ECD). For image analysis and morphology filter, the MFI View Analysis Suite (MVAS) (ProteinSimple, San Jose, CA) was applied.

Sub-Visible Particle Quantification by Light Obscuration (LO)

A HIAC 9703+ liquid particle counter instrument (Beckman Coulter, Brea US) equipped with a HRLD-150 detector was used to measure sub visible particles. A system calibration was performed using polystyrene beads in purified water by the supplier (Skan AG, Allschwil, Switzerland). After flushing the system with purified water, a blank measurement was performed to ensure cleanliness. Prior to sample analysis, a SST was performed for count verification using the 5 μm particle count STD (COUNT-CAL Count Precision Size Standard; ThermoFisherScientific, Waltham US). Sizing accuracy is ensured by recalibrating the instrument every 6 mo.

Visible Particles by Visual Inspection

The test samples were subject to visual inspection according to monograph EP.2.9.20. The presence of Visible particles were additionally documented using a digital camera.

Extractable Volume (Gravimetric Method)

Vials were filled with the surrogate solutions and equilibrated to room temperature (RT). The fill volume was set to 1.0 mL for 2R vials and 3.0 mL for 6R vials.

Extractable volume measurements were performed using the CSTDs or the control transfer polymer disposable syringe withdrawing the whole content of the vial while holding the vial in adequate (inverted) orientation. Extractable volume was then calculated using the extracted solution mass and the solution density. Hold up volume was calculated the difference between fill volume and extractable volume.

Sample Size

All tests were performed in triplicates. Results were plotted using Origin 2019 or JMP.

Results

Evaluation of Integrity Using CSTDs

The integrity and tightness of CSTDs in connection with the various container closure systems were evaluate using the helium leak method. Pronounced scatter in leak results for the CSTD from same vendor as well as differences among the CSTDs from different vendors were observed. For example, the two 2R vial configurations of CSTD2 and all three 6R vial configurations of CSTD1 showed helium leak rates between 2.0 × 10−6 and 1.4 × 10−4 mbarL/s, which would correspond to a single 0.1 μm – 1 μm orifrice (hypothetic pinhole without path length). The 6R vial configurations of CSTD2 showed helium leak rates of 4.7 × 10−6 and 1.5 × 10−2 mbarL/s, which would corresponds to a single 0.1 μm – 10 μm orifrice. In comparison, tightness of vials versus microbiological ingress using this method would usually be in the range of 10−7 mbarL/s.

Evaluation of Force Required to Mount the Vial Adaptor

The 4 tested CSTDs were mounted on the different container closure configurations with the force for assembling the system being evaluated using the Zwick stress-strain instrument. Results are presented in Fig. 2. Pronounced differences in the required force for assembly were observed. For the 2 R configuration, the type CSTD being used had a strong impact on the penetration force. The lowest puncture force was needed to mount CSTD3 on a serum rubber stopper (25.6 N – 28.7 N). In contrast, the combination of CSTD2 and a lyo rubber stopper required 248.1 N – 284.9 N. The type of CSTD used had a major influence on the required force. Interestingly, the variability of the rubber stopper puncture force measurements for CSTD3, CSTD4 and CSTD1 was low, whereas the results for CSTD2 showed an increased variability. The rubber stopper design only had a marginal impact on penetration force.

Fig. 2
figure 2

Penetration force required to mount the vial adaptor for (A) 2 R vial configuration, (B) 6 R vial configuration.

Full size image

Similarly, for the 6R vial configuration, the CSTD type had a strong impact on the required rubber stopper penetration force. The lowest force was needed for CSTD3 and CSTD4. (31.8 N - 49.9 N). In contrast, CSTD1 required 143.6 N – 152.5 N. The rubber stopper design had only a marginal effect on required penetration forces. Compared to the 2 R vial configuration, the variability of the measurement was lower for the 6R vial configuration (Fig. 3).

Fig. 3
figure 3

Visible rubber stopper fragments.

Full size image

Evaluation of Visible and Subvisible Particles when Using CSTDs

Visible Particles

Visible particles (VP) were readily observed in the placebo solution after penetration of the rubber stopper using the Zwick instrument with any of the CSTD devices. The VP displayed a characteristic morphology of rubber stopper fragments. Particles were observed in any combination, independent of the used rubber stopper and CSTD combination.

Sub-Visible Particles (Light Obscuration and Flow Imaging)

The CSTDs were mounted on different CCS configurations filled with a placebo solution (0.02% polysorbate 20 in purified water). All CSTDs introduced sub-visible particles (SbVP) into the placebo solution, though to a variable degree (Fig. 4).Pronounced differences were observed between the different CSTD types: CSTD1 introduced the lowest amount of SbVP into the placebo solution. CSTD2 and CSTD4 could be ranked in between CSTD1 and CSTD3. CSTD 3 consistently introduced the highest amount of SbVP into the product solution. For example, the measured SbVP particle counts ≥10 μm ranged between 2761 and 13,639 particles/mL. The measured number of sub-visible particles ≥25 μm were between 181 and 812 particles/mL. The SbVP results were consistent across all SbVP size bins, and independent of the used vial size configuration and the used rubber stopper. The ranking between the CSTDs regarding SbVP count results for CSTDs 1,2 and 4 were consitent across SbVP size bins the used vial configuration and the rubber stopper.

Fig. 4
figure 4

Sub visible particles (light obscuration) of 2 R vial configuration and 6 R vial configuration.

Full size image

In comparison, the conventional method of solution withdrawal using a polymer transfer syringe with a 21G needle, introduced only a negligible amount of SbVP into the placebo solution.

Flow imaging was used to further characterize the type of sub-visible particles. Sub-visible particles were characterized by morphology. The results show that the large majority of particles introduced by CSTDs are likely attributed to silicone oil (Fig. 5).

Fig. 5
figure 5

(A) Morphology characterization of Sub visible particles ≥5 μm as analyzed by FIM, (B) example images of CSTD3 6R.

Full size image

Hold-Up Volume

The hold-up volume, i.e. the volume that could not be extracted from the vial, was evaluated using 3 different solutions with viscosities of 1, 15 and 19 mPa.s, representing a common range of viscosity for many parenteral formulations. Figures 6 and 7 shows the delivered volume of the different vial/CSTD configurations in combination with a specific placebo solution. These are compared to 2 control measurements using a more routinely used needle and syringe setup, one with an inverted vial and another in an upright vial position. The results show that, in comparison to the controls, the cumulated hold-up volumes in the vials and the transfer device can be significant, reaching up to approximately 1 mL for CSTD4 and the highest viscosity solution tested in 6R vials. Factors influencing vial hold-up volume were: 1) the CSTD system, where significant differences among the systems were observed. CSTD3 consistently contributed to the least losses in delivered volume among the 4 tested systems in 6R vials, while CSTD1 and CSTD2 consistently contributed highest volume losses, 2) the solution viscosity, where increasing the viscosity tends to increase the volume losses, 3) vial size, where the hold up volume with 6R vials was higher than 2R vials. The stopper type (whether serum or lyo stopper) seemed to have low to negligible effect on the vial hold-up volume.

Fig. 6
figure 6figure 6

Hold-up volumes, i.e. volume that could not be extracted from the 2R vials for A) type I glass + Lyo stopper, B) type I glass + Serum stopper, C) type I glass hydrophobic inner vial surface + Lyo stopper, D) type I glass hydrophobic inner vial surface + Serum stopper.

Full size image
Fig. 7
figure 7

Hold-up volumes, i.e. volume that could not be extracted from the 6R vials for A) type I glass + Lyo stopper, B) type I glass + Serum stopper.

Full size image

Discussion

USP <800> suggests the use of Closed-System-Transfer Devices (CSTDs) for hazardous drugs, in order to protect healthcare professionals (HCP) from accidental exposure to potentially hazardous drugs.

The first question to be answered, is for which drugs and products this procedure would apply and should be followed, hence, which drugs may be hazardous for health care professionals. Secondly, it must be assessed, if and how CSTDs impact product quality and/or patient safety to ensure there is no adverse impact.

Addressing the first question, the primary intention of USP <800> is for drugs that are known to have a safety concern due to their carcinogenicity, teratogenicity, developmental toxicity, reproductive toxicity, organ toxicity or genotoxicity. USP <800> references the NIOSH list of hazardous drugs (19), with the great majority of the listed molecules belonging to the class of small molecules. Out of large list molecules, the NIOSH list includes only 7 biologic molecules; namely 4 antibody drug conjugates (ADCs), one mAb and 2 Fc fusion proteins. However, the NIOSH list considers the hazards only, and does not consider the likelihood of (systemic) exposure.

For investigational drugs, where no sufficient data is available to exclude the above mentioned safety concerns, USP <800> recommends treating them as hazardous drugs. This can lead to clinical sites being tempted to broadly use CSTDs with biotechnologically produced products, despite their significantly low risks for healthcare professionals upon unintended exposure (20), simply due to the low risk for systemic exposure via airborne particles / inhalative route. Towards this direction, USP <800> allows the use of risk assessments to waiver engineering controls if there would be little hazards and/or low likelihood of exposure based on the route of administration, dose or nature of the drug, leading to. However, given the final responsibility to such risk assessments lies in the hand of the site handling the investigational drug and not the sponsor, we expect an interesting debate. There is thus a clear need to increase knowledge and awareness among practitioners for those new classes of biologic therapies (21).

To address the impact of CSTDs on product quality, our study aimed to evaluate commercially available CSTDs in typical use and how they impacted (a) integrity and tightness, (b) forces required to mount the CSTD and penetrate the rubber stopper (usability), (c) contamination with visible and sub-visible particles that relate to the CSTD and (d) impact of the CSTD on delivered (extractable) volume.

As mentioned earlier, the most widely evaluated functionality aspect for CSTDs is their goal on ensuring containment (lack of exposure to liquid and/or vapor). However, the majority of the studies described so far use rather qualitative methods. The lack of a harmonized quantitative method is also described in USP <800>, as well as in NIOSH’s continued effort to develop an acceptable quantitative method (22). In an effort to develop a quantitative method for evaluating the containment functionality of CSTDs, we employed the helium leak container-closure integrity test, being currently one of the most sensitive methods for evaluating container closure integrity. The results from the current study showed significant variability among the same CSTDs from a single vendor, as well as among different CSTDs. The variability of the measurement was very high. The high variability may be associated to variability in sample preparation as well as dimensional variability of CSTDs. The rubber stopper design and the vial configuration had only a marginal effect – the smaller differences were challenging to delineate. It is worth noting that the helium leak method holistically measures the leaking helium gas stream, and cannot delineate whether the obtained leak is derived from a single large orifice or multiple smaller ones. In conclusion, low leak rates (e.g. 10−6 – 10−4 mbarL/s) seen for the 2R configurations of CSTD2 and CSTD4 and 6R configuration of CSTD1 and may indicate tight CSTDs, whereas CSTDs with higher leak rates require additional characterization. In comparison to acceptance criteria for typical parenteral container closure systems, in the range of 10−7 mbarL/s, all data points suggest that the used CSTDs did not provide sufficient tightness and containment to prevent potential microbiological ingress. Obviously, when using CSTDs, parenteral preparations cannot be considered sterile any more and microbiological hold time considerations – similar to when using blut needles for product removal from vials – should apply. Depending on leak size, also tightness to prevent potential product leakage may need to be specifically assessed.Another aspect that was evaluated in our study was the penetration force when using the CSTDs in a simulated-use study. Currently no international harmonized rubber stopper puncture force method including an established maximum allowable puncture force exists. We evaluated the force in alignment to ISO-8536-2 (18), which is used to evaluate the puncture force and depth. Nevertheless, one should mention that the force measured during the current study represents more of the force required for “assembling” the CSTD rather than pure puncture force, since there is a component required for fitting the wings of the CSTD on the edges of the vial cap. Zhao et al. evaluated the puncture force for five chemo spikes, either already vented with CSTDs or can be vented with CSTDs (23). They found out that the penetration force depends on the crimping parameters. Our results show that CSTD3 and CSTD4 had the lowest penetration forces for both of the 2R and 6R configurations. In contrast, CSTD1 and CSTD2 required higher assembly forces, also differing among the 2R and 6R configurations. The forces measured were up to 300 N. Such high forces likely represent a significant challenge for the user for adequate handling. In summary, the CSTD type had the highest impact on the required assembly force, including the rubber stopper puncture force. Interestingly, no correlation between CSTD type and the vial size was observed. This means, a CSTD type of a specific vendor may perform adequately with the 2R vial configuration whereas it may not perform with the 6R vial configuration. Of note, the orientation of the CSTD and the rubber stopper puncture speed may significantly impact the measured puncture force, which can lead to pronounced differences between different test setups/methods.

The potential contamination of the product formulation with (visible and sub-visible) particulates is another aspect that was also evaluated in this study. For most cases, we observed visible particles, which were determined to be rubber stopper fragments. The occurrence of rubber stopper fragments is expected to be impacted by the used rubber stopper, dimensions and sharpness of the CSTD spike, the orientation of the CSTD spike and the rubber stopper puncture speed. Stopper coring and fragmentation is known and described in the public domain, and has been mentioned elsewhere to depend on the type of needle and type (formulation and dimensions) of stopper used. However, the topic is hardly systematically investigated among different CSTDs. In one publication, Katherine et al. compared two types of PhaSeal CSTDs, namely P50 and P55, where the former has a metal needle, while the later has a plastic needle. They showed that using the P55 protector reduced the incidence of particulate contamination from 5.3% in case pf P50 protector to 0.72%. (24). To note, rubber stopper formulations and products are also likely developed versus current standards and requirements that use needles instead of spikes (e.g., Ph Eur 3.2.9, ISO 7864, and USP <381>).

Sub-visible particles were evaluated by both light obscuration (LO) and flow imaging. LO results showed that the number of particles differed among the CSTD systems, with CSTD3 showing a significant number of sub-visible particles. Flow imaging suggested that the majority of the observed particles are related to silicone oil droplets. This is also in line with the results from Petoskey et al. (15). The fact that such large number of silicone oil particles can be shed from some CSTDs into the product solution may lead to failing the established limits for sub-visible particles in parenteral preparations, if considering also for the simulated-use conditions. Additionally, there is a level of concern in industry that silicone oil may lead to aggregation of some protein molecules (25).

As can be further seen from the results, several factors affect the delivered (extractable) volume. Different CSTDs show different impact on volume losses keeping all the other parameters constant. One of the more important reasons for volume losses is the CSTD design itself, particularly the length of the needle/spike and position of the opening. Since for volume extraction, all CSTDs need to be in the inverted position, the deeper the needle position into the vial, the more volume that cannot be extracted out of the vial. For comparison, the inverted volume extraction using a needle and syringe was applied as well as the upright position. One can consistently see that the inverted volume extraction leads to similar volume losses or less, when compared to CSTDs. This is due to the fact that the needle can be moved to extract as much as possible out of the vial, while the CSTDs are usually fixed in position. Another factor was the vial size, particularly the opening diameter. It is clear that, because of the inverted solution extraction position, some liquid in the vial neck would not be extractable, and thus with larger vial openings, this lost volume would be higher. Also the viscosity of the fluid impacted the losses in delivered volume.

Determination of the extractable volume is important to ensure correct dosing and hence, product efficacy in clinical use. The FDA guidance for industry on “Allowable Excess Volume and Labeled Vial Fill Size in Injectable Drug and Biological Products” aims at ensuring the excess volumes as sufficient for accurate delivery of the nominal volume, but also avoiding having too much excess volume to prevent drug misuse. Recommended excess volumes can be based according to or derived from USP general chapter <1151>.

With no additional overfill to compensate for CSTD-related volume losses, the losses during withdrawal are expected to exceed typical overfills, hence potentially compromising the administration of the desired drug dose and potentially leading to systematic underdosing. Hence, in case where CSTDs must be used due to the hazardous nature of the drug, and a significant likelihood of systemic exposure to manifest the hazard, we recommend to study the impact on delivered (extractable) volume with specific CSTDs. After selection of a CSTD, manufacturers then need to define their mitigation strategies. This may e.g. require to increase excess volumes (overfills) to the vials to account for losses during dose extraction and preparation for administration when using this particular CSTD. If pursuing such strategy, it has to be considered that any added overfill to account for one CSTD may be leading to overdosing/underdosing when using another CSTD if it would lead to different volume losses. In summary, hold-up volumes due to CSTDs are significant, and vary across products. While a number of pharmaceutical companies have published an opinion about the challenges the industry is facing with CSTDs, their position was mainly requesting feedback from the regulatory authorities (16). Until this urgently needed health authority guidance is available, pharmaceutical manufacturers need to come up with specific approaches or strategies when aiming to allow or even require the use of CSTDs.

Conclusions

CSTDs are intended to reduce the exposure of health care professionals in clinical settings to hazardous drugs. Interest in CSTDs has increased significantly after the issuance of USP <800>, which recommends their use as supplemental engineering controls for containing the exposure to hazardous drugs. While there is still a need for more regulatory guidance on using CSTDs, as well as harmonized testing procedure, the literature so far focused mainly on the containment feature of CSTDs, and hardly on other important functionality attributes and impact on product quality. In this study, we studied (a) the containment and integrity of CSTDs in simulated use, (b) the penetration forces required to mount the CSTDs, (c) contaminations of product solution by visible and subvisible particles and (d) delivered (extractable) volume. The results show that, as previously shown by other authors and methods, CSTDs may not be completely tight, and could lead to leakage. As such, they are valuable tools for reducing, but not completely eliminating exposure to hazardous drugs. CSTDs are for sure not ensuring microbiological tightness and microbiological in-use hold times must be considered also when using these devices. Particle evaluation showed that CSTDs can potentially lead to significant amount of particles contaminating the product solution, e.g. due to stopper coring and shedding of rubber particles. In addition, significant amounts of subvisible particles, particularly silicone oil, contaminated the product to be potentially administered to patients, with the particle level depending on the type of CSTD used. If considering product requirements for particulates in parenteral preparations, these solutions would be incompliant with typical regulatory requirements. Finally and very importantly, volumes up to almost 1 mL could not be extracted from the container closure system, depending on the vial size, CSTD type and solution viscosity. This significantly impacts the dose delivered and hence can lead to systematic underdosing and hence adversely impact efficacy.

In conclusion, our study results show that CSTDs we evaluated in combination with typical container closure systems can lead to adverse impacts on product quality. We recommend pharmaceutical manufacturers to carefully evaluate and advise users in which cases CSTDs are required, and to test product quality impact for these products with a specific CSTD. Any impact on product quality or dosing accuracy may need to be addressed by adequate means. These may for example include adapting (increasing) the overfill to compensate for losses in product volume and/or by adding additional in-line filters during administration to mitigate for particulates potentially contaminating the product formulation due to CSTDs. The assessment of protein stability itself was out of scope of this manuscript and is subject to further studies. In case considering the use of CSTDs, specific in-use studies that include both extractable volume, particulates, penetration forces, tightness as well as content and purity of the active ingredient are highly recommended.

Finally, and in light of the high risks on product quality versus the little value they can bring when employed with the great majority of investigational biologics, we believe that there is a need for an aligned action between the regulatory bodies, pharmaceutical manufacturers and clinical sites to give a clear guidance on using CSTDs for biologics, limiting their use where they bring real value to health care professionals with no risks to patients.