Introduction
Fast chromatography and green chemistry have been the most compelling trends in today’s pharmaceutical industry. Supercritical fluid chromatography (SFC) is one of the few separation technologies that can meet both ‘fast’ and ‘green’ expectations. Since it was first demonstrated in 1962, SFC has advanced remarkably [1]. The fundamentals of SFC including theory and instrumentation have been well established. By using the non-toxic, non-flammable, and relatively inexpensive carbon dioxide as eluent, SFC is a true “green chemistry technology.” In addition, due to supercritical CO2’s low viscosity and high diffusion rate, SFC has many practical advantages such as rapid equilibration, shorter cycle time, and high throughput [2]. Flow rate up to 5 mL/min is frequently used in analytical SFC and most analyses can be completed within a few minutes. Given the superior performances of SFC in the aspects aforementioned and with the enhanced and better integrated hardware and software, modern SFC instruments and methodology could be widely used and implemented as one of the major separation techniques, especially in the areas of chiral analysis and purification for pharmaceutical and other industries.
However, in the highly regulated pharmaceutical industry, analytical SFC had been limited as a research tool before it could be qualified and meet rigid GMP requirements. Historically, analytical SFC was used primarily in the high throughput screenings of drug candidates at discovery stage and pharmaceutical companies had not taken the advantages of SFC in GMP arena for release, stability and other testing of clinical and commercial products which require instrument qualification. Lack of qualified instrument also discourages SFC method development in the product development stage. For instance, in preparation of GMP product release testing, reversed, normal, or polar organic phase liquid chromatography methods were usually developed and validated even though the existing SFC methods, which were explored and finalized during preclinical developments, can offer equivalent or better performances with much shorter turnaround time. As more chiral compounds are added in product pipelines and an estimated 40% of the new chemical entities underdevelopment in the pharmaceutical industry are chiral [3], Amgen in 2009 initiated an evaluation process on a variety of SFC systems for qualification purposes. By March 2011, Amgen successfully completed the installation qualification and operation qualification on both hardware and software of a SFC system. The qualified SFC instrument will be implemented in the GMP environments for product release and stability testing. The test parameters, acceptance criteria, and test procedures of hardware qualification were intended for all SFC systems, not limited by instrument vendors or manufacturers. By using the qualification protocol, SFCs in a pharmaceutical contract manufacturer or testing laboratory can be qualified as well, which would be a prerequisite for SFC method transfer.
Qualification Strategy and Process Overview
USP general chapter <1058>(Analytical Instrument Qualification) groups instrument qualification activities into four phases: design qualification (DQ), installation qualification (IQ), operational qualification (OQ), and performance qualification (PQ). Among them, OQ testing is pivotal to demonstrate that an instrument will function according to its design in users’ environment. At Amgen, software and hardware OQ were defined by separate qualification protocols. The hardware OQ protocol can be applied to all SFC systems of different manufacturers. For software OQ, each software has its own qualification protocol as its structure, firmware, and functions including instrument control, data acquisition and processing can be very different from system to system. Hardware OQ consists of both modular and holistic tests. In holistic tests, a chosen chiral compound was analyzed by predefined SFC conditions to meet the specifications in the OQ protocol. As its testing can only be completed by the collaboration of entire instrument and software systems, holistic qualification is more efficient in qualifying both hardware and software than software qualification alone. Given the fact that it was the first qualification on a SFC instrument, obtaining manufacturers’ DQ documents was challenging, the required documents were either not available or not meeting Amgen GMP requirements. Therefore, the decision was made at Amgen to incorporate DQ tests into the ‘IQ protocol’, the installation checklist. All qualification activities including instrument moving and preventative maintenances were carried out by certified engineers. Site preparation documents, preventative maintenance and installation checklists were reviewed and approved along with IQ/OQ protocols prior to executions. PQ is scheduled periodically after IQ/OQ and major maintenances. Users will also check the instrument performance before or during their routine operations. The post-qualification documents including deviation reports were then compiled, reviewed and approved by subject matter experts, technical operation and quality assurance groups.
Figure 1- Chromatograms of blank (A), flurbiprofen at 0.01% level (B) and 0.05% level (C). SFC conditions: 5 μm AD-H column (250 x 4.6 mm), 4 mL/min flow rate, 200 bar backpressure, CO2:MeOH 80:20 isocratic elution, 254 nm UV detection, 10 μL injection volume, 40°C column temperature
Instrument Readiness and Quality Assessments for GMP Purposes
The landscape of SFC manufacturers and their instruments has changed significantly over the years. With the new system emerging and old system phasing out, SFC instrumentation continues to evolve and advance. Driven by business needs, Amgen had conducted thorough evaluations on a number of commercial SFC systems on the market before putting them into the ‘qualifiable’ category. It was clear at the time that although manufacturers can deliver SFC systems with solid hardware suitable for the intended qualification needs, the inadequate software can disqualify a system as a potential GMP instrument. Quite a few software issues involving instrument control, data processing and reporting had been observed on separate SFC systems, which certainly will be problematic in meeting system suitability requirements of method validation and GMP analysis. In addition, low-level impurity (chiral or achiral) control is one of the top priorities of the pharmaceutical industry. Compared with HPLC, the low signal-to-noise (S/N) ratio and high detection limit has always been the weakness of SFC and limited its applications. At Amgen, a chiral SFC method was developed by using AD-H column and flurbiprofen as analyte to measure and compare the separation and detection capabilities of different SFC systems. Figure 1 shows that one of the SFC instruments can base-line separate and accurately detect flurbiprofen enantiomers at a level as low as 0.05% (10 μL injection volume, relative to 1 mg/mL concentration) with a S/N ratio = 10 and excellent linearity within the range of 0.05% to 150% of the target concentration, which is comparable to the performance of a HPLC. By excising instrument control, data acquisition, processing and reporting, and system suitability verification, this analysis provided valuable data regarding the readiness of a specific SFC instrument for GMP qualifications. Other factors such as instrument lifecycle (obsoletion and maintenance), vendor services, ease of use, training, and operation costs etc. were also included in the initial evaluation to ensure that the instrument can maintain its GMP status post-qualification.
Table 1- Summary of Critical Test Parameters and Acceptance Criteria in OQ Protocol
Setting Test Parameters and Acceptance Criteria in Hardware Qualification Protocols
For modular tests, the test parameters and acceptance criteria were set based on the instrument specifications provided by manufacturers or vendors. Most hardware components of a SFC except the backpressure regulator can be found in a HPLC and manufacturers have capitalized on this converting existing HPLC into SFC. Therefore, the existing HPLC hardware OQ protocol can be a good reference for defining SFC test parameters. For holistic tests, the design of test procedures was driven by practical needs. The historical data and evaluation results played a crucial role in setting parameters and limits. For example, the 10 μL sample loop used in the linearity and detection limit evaluations is replaced by a 5 μL loop in holistic qualification as it is commonly used in routine SFC analysis. As a consequence the detection limit has to be raised in the OQ protocol. A larger injection volume of 20 μL was initially applied as the upper limit in the autosampler injection volume precision and injection carryover tests. It was replaced by 10 μL injection volume in the final protocol to prevent peak distortions that were often observed by SFC with injection volumes over 15 μL. Since the hardware OQ protocol is intended for all SFC instruments, one detection limit is set for different systems. In principle, the detection limit varies among SFC instruments. Its true value should be determined during the method validation for a specific compound. Table 1 lists the critical parameters and their corresponding limits from the initial draft to the final version of the OQ protocol. The revisions made to the protocol were mainly based on the results of two mockup qualifications, which were heavily weighed in shaping Amgen qualification procedures. Between the protocol revisions, each changed parameter was tested in replicates to ensure the reproducibility. To our experiences, developing a suitable OQ protocol is the most important and time consuming activity in the overall qualification process.
Figure 2- Flow Rate Determination of Supercritical CO2 Fluid by Chromatographic Method at setpoints of 10 mL/min (A), 5 mL/min (B), 1 mL/min (C), and 0.5 mL/min (D). SFC conditions: 1 mL sample loop in place of a column, 200 bar backpressure, 1 μL injection volume of acetone, 215 nm – 350 nm total UV detection
Overcoming Technical Difficulties in Qualification Tests
Although SFC consists of similar components as HPLC, there are many technical challenges to overcome in qualifying a SFC instrument as it operates above the critical temperature and critical pressure (31 ºC, 73 atm for CO2) and supercritical CO2 behaves between liquid phase and gas phase. The flow meter and pressure meter used in SFC qualification must be able to sustain a pressure of 300 bar minimum, which is out of the limit for the meters used for HPLC and GC. The two meters used in our SFC qualification was custom made and the cost of flow meter alone plus its calibration was over ten thousand US dollars. An alternative procedure was developed at Amgen for measuring volumetric flow rates of supercritical CO2 fluid by applying a calibrated sample loop in place of a column between autosampler and UV detector. By injecting acetone and measuring its retention time, the flow rate was then calculated by dividing the sample loop volume with retention time. Figure 2 depicts how flow rates were determined by a 1 mL sample loop. The accuracy of this method can be improved by using appropriate sample loops – 1 mL sample loop for 0.5 and 1.0 mL/min flow rates, and 5 mL sample loop for 5 and 10 mL/min flow rates. The results obtained by this chromatographic method were comparable to those measured by the high-pressure flow meter. It is worthy of notice that due to different autosampler configurations and depending on the flow path, the flow rate determined by this procedure can deviate from the low flow rate setpoints.
Figure 3- Determination of Gradient Composition Accuracy by Chromatographic Method. SFC conditions: 4 mL/min flow rate, 200 bar backpressure, 265 nm UV detection, co-solvent = 0.5% acetone in methanol, 40 °C zero-volume union temperature, 0 μL injection volume
Chromatographic method was also applied in validating gradient composition. Methanol with 0.5% acetone was used as co-solvent and its composition increased stepwise by 20% increments from 0% to 100% in a gradient program. The composition at each step was held for 3 minutes. Triggered by a blank injection, a chromatogram, as shown in Figure 3, was acquired with a zero-volume union. The gradient composition was then calculated by comparing the height of each step with that of 100% step. In addition, the linearity of composition measurements and the stability at each composition step were calculated and verified against the specifications in OQ protocol.
In HPLC qualification, the aqueous solution of caffeine was often used in calibrating the wavelength accuracy of UV detector. As for SFC, a normal phase chromatography, injecting aqueous caffeine is not preferred and methanol as the co-solvent/sample diluent can shift the three characteristic wavelengths of caffeine (205, 245, and 273 nm) and fail the test. To prevent the wavelength shifts, caffeine in water was forced into an off-line flow cell and the flow cell was then put back into the SFC system to acquire UV spectrum. This procedure could be cumbersome for some SFC instruments, but it is simple and effective.
Achieving satisfactory S/N by SFC can be challenging. It was found that in general, SFC has higher levels of baseline noise than HPLC. High backpressure, e.g. 200 bar and steady temperature control of the mobile phase through UV detector reduced baseline noise significantly and improved S/N ratio. Our results demonstrated that the S/N ratio criterion (not less than 10) could only be met by optimizing both backpressure and temperature. In fact, during the qualification process not all tests passed first time. When there were failures and repeated tests such as in injection carryover test and S/N ratio determination, deviation reports were generated to capture root causes and corrective actions and to justify the final results. The failures actually demonstrated the rigidness of qualification protocol.
Conclusions
The IQ and OQ of both hardware and software were successfully completed. Through the qualification process, evidence was collected demonstrating that the SFC system performs suitably for its intended purpose and meets the GMP requirements of supporting pharmaceutical product developments. The critical quality attributes of a SFC system must be closely examined and assessed prior to aqualification. The test devices, including highpressure flow meter and pressure meter, need to be refined and standardized. As the SFC instrumentation, especially software continues to improve, the test procedures and qualification protocol should be retuned accordingly to meet new expectations. The success of SFC qualification will further extend its applications across many fronts and establish SFC as one of the essential separation technologies in pharmaceutical industry.
Acknowledgements
The author would like to thank David Semin, Hue Lu, Janet Cheetham, Donna Norton, and Joey Ling of Amgen for their support and contributions to this successful SFC intrument qualification.
References
- Berger, T. and Berger, B. “A Review of Column Developments for Supercritical Fluid Chromatography” (2010) LCGC North America May.
- Larry, T. “Supercritical Fluid Chromatography” (2010) Anal. Chem. 82, 4925-4935.
- Akin, A., Antosz, F.J., Ausec, J.L., Greve, K.F., Johnson, R.L., Magnusson, L., Ramstad, T., Secreast, S.L., Seibert, D.S., and Webster, G.K. “An Orthogonal Approach to Chiral Method Development Screening” (2007) Curr. Pharm. Anal. 3, 53-70.
Author Biography
Jingshun Sun is a Senior Scientist at Amgen in Thousand Oaks, CA where he leads SFC qualification and a number of other technology developments. Previously, he was a scientist at Emisphere Technologies, where he led analytical supports in improving macromolecule drug delivery. He also worked in Barr Pharmaceuticals supporting formulation developments. Jingshun received his doctorate degree from University of Florida, Gainesville in pharmaceutical sciences-medicinal chemistry.
This article was printed in the May/June 2011 issue of American Pharmaceutical Review - Volume 14, Issue 4. Copyright rests with the publisher. For more information about American Pharmaceutical Review and to read similar articles, visit www.americanpharmaceuticalreview.com and subscribe for free.