This article was downloaded by: [Universiti Putra Malaysia] On: 10 January 2012, At: 22:00 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK Journal of Liquid Chromatography & Related Technologies Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/ljlc20 Validation of HPLC Instrumentation Ioannis N. Papadoyannis a & Victoria F. Samanidou a a Department of Chemistry, Laboratory of Analytical Chemistry, Aristotle University of Thessaloniki, Thessaloniki, Greece Available online: 24 May 2011 To cite this article: Ioannis N. Papadoyannis & Victoria F. Samanidou (2005): Validation of HPLC Instrumentation, Journal of Liquid Chromatography & Related Technologies, 27:5, 753-783 To link to this article: http://dx.doi.org/10.1081/jlc-120029697 PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.
JOURNAL OF LIQUID CHROMATOGRAPHY & RELATED TECHNOLOGIES w Vol. 27, No. 5, pp. 753 783, 2004 Validation of HPLC Instrumentation # Ioannis N. Papadoyannis* and Victoria F. Samanidou Department of Chemistry, Laboratory of Analytical Chemistry, Aristotle University of Thessaloniki, Thessaloniki, Greece ABSTRACT Validation is the process to confirm that the analytical equipment, method, or system for a specific test is suitable for its intended use. A successful chromatographic analysis depends on the precise performance of the HPLC instrumentation. This article focuses on the validation of HPLC instrumentation. Definitions of each parameter are provided in conjunction with the timetable of checking their accordance towards preestablished acceptance criteria. Key Words: Validation; GMP; GCP; Qualification; System suitability. # Reprinted from the Encyclopedia of Chromatography (# 2003); Marcel Dekker, Inc.; URL: http://www.dekker.com/servlet/product/productid/e-echr. *Correspondence: Ioannis N. Papadoyannis, Department of Chemistry, Laboratory of Analytical Chemistry, Aristotle University of Thessaloniki, Thessaloniki, Greece; E-mail: papadoya@chem.auth.gr. 753 DOI: 10.1081/JLC-120029697 Copyright # 2004 by Marcel Dekker, Inc. 1082-6076 (Print); 1520-572X (Online) www.dekker.com
754 Papadoyannis and Samanidou INTRODUCTION Validation is defined as the process of providing documented evidence that an instrument, a system, a method, a product, or a procedure performs as expected within specified design parameters and requirements, so that the results obtained are reliable. Validation efforts should be broken down into separate components addressing the equipment both the instrument and the computer controlling it and the analytical method run on that equipment. After these have been verified separately, they should be checked together to confirm system s expected performance limits [system suitability testing (SST)]. Figure 1 illustrates the basic validation steps with regard to computerized HPLC equipment. The need for validation in analytical laboratories may originate from regulations such as current good manufacturing practices (cgmp), good laboratory practice (GLP), good clinical practices (GCP), quality and accreditation standards such as international standards organization (ISO) 9000 series, ISO 17025 and the European Norm (EN 45001), United States Figure 1. Basic validation steps in an analytical laboratory.
Validation of HPLC Instrumentation 755 Pharmacopoeia (USP), Food and Drug Administration (FDA), and Environmental Protection Agency (EPA). All regulations require instruments to be calibrated, well maintained, and designed to be suitable for their intended use. However, validation of equipment and analytical methods is necessary, not only due to regulations and accreditation standards, but also as prerequisite [1 5] in terms of any good analytical practice. Validation is a regular process which starts before an instrument is placed on-line and continues long after method development and transfer. It consists of at least three stages, each one being critical to the overall description of the process: Phase 1. Instrument/equipment qualification (EQ), including computer qualification. Individual modules of equipment (hardware and firmware) have to be separately validated, as well as the entire system. Validation of computer systems includes the evaluation of hardware and software. Phase 2. Analytical method validation prior to routine use and after changing method parameters. Phase 3. Analytical SST, that combines instrument, computer, and method. When the equipment and a particular method have been selected and validated, the equipment for that method goes through an SST prior to, and within, sample analyses, or practically on a day-today basis. [5] Information about validation procedures concerning HPLC instrumentation and guidelines are discussed in the following sections. EQUIPMENT VALIDATION/QUALIFICATION Validation of HPLC instrumentation, also called qualification is the procedure that ensures that the instrument is qualified, i.e., its performance complies with the method s predetermined requirements, providing reliable and valid results. Modern HPLC systems are computerized, generally consisting of analytical hardware, computer hardware, peripherals, and software. Validation of instrument hardware includes testing according to documented specifications. If HPLC instruments consist of several modules, individual modules (modular validation), as well as the entire system (holistic validation), should be validated. The latter, however, is preferred, as individual module tests should be performed as part of the diagnosis, if the system fails.
756 Papadoyannis and Samanidou The equipment used in a study should be tested on site before it is used the first time, within certain time intervals, and after repairs and it should be periodically inspected, cleaned, maintained, and calibrated according to standard operating procedures (SOP). Records of these activities should be maintained. Qualification of HPLC equipment (EQ) begins at the vendor s site. During this stage the instrument and software are developed, designed, and produced in a validated environment according to good laboratory practice (GLP), cgmp, and/or ISO 9000 standards, preferably ISO 9001, as ISO 9002 or 9003 are insufficient because they do not cover development. Qualification can be subdivided into four stages: (1) Design qualification (DQ), (2) installation qualification (IQ), (3) operation qualification (OQ), and [5 8] (4) performance qualification (PQ). Design Qualification Design qualification covers setting of user requirements, as well as functional and performance specifications. The DQ should ensure that the instrument has all the necessary functions and performance criteria that will enable it to be successfully used for the intended application. Design qualification includes: (a) description of the intended use of the equipment; (b) selection of the analysis technique, of the technical, environmental, and safety specifications, final selection of the supplier and of the equipment; and (c) development and documentation of final functional and operational specifications. Installation Qualification Installation qualification establishes that the instrument is received as designed and specified, it is properly installed in the selected environment, and that this environment is suitable for the operation and use of the instrument. During installation, one should: (a) compare equipment, as received, with purchase order; (b) check documentation for completeness or for any damage; (c) install hardware (computer, equipment, fittings, and tubings for fluid connections, columns in HPLC, power cables, and instrument control cables); (d) install software on computer following the manufacturer s recommendation and verify correct software installation; (e) make back-up copy of software; (f) configure peripherals, e.g., equipment modules; (g) switch on the instrument and check for any error messages; (h) identify and make a list with a description of all hardware; (i) list equipment manuals and SOPs; (j) inject
Validation of HPLC Instrumentation 757 and qualitative evaluate a standard; and (k) prepare an installation report describing how and by whom the instrument was installed. Operational Qualification Operational qualification or acceptance testing, is the process of demonstrating that the whole instrument or its modules will function according to its operational specification, in the selected environment, according to previously defined functional and performance specifications (acceptance criteria). This procedure can be performed to the extent of the self-diagnostics routine test or to a more detailed procedure regarding flow rate, injector precision, or wavelength accuracy. It can be done by the user or the vendor on behalf of the user. Performance Qualification Performance qualification is the process of testing and calibrating the instrument before and during routine use, verifying system performance. This test is repeatedly executed to ensure that the entire system generates valid results performing as intended, throughout representative or anticipated operating ranges. Criteria of PQ should later also be used for revalidation. After the instrument is placed on-line in the laboratory, and after a period of use, regulations require maintenance, followed by calibration and standardization. Each laboratory should have SOPs that define the period of use (usually a reasonable time interval in which the instrument operates properly). Records should be kept to track instrument maintenance. The frequency of PQ depends on the type of instrument, on the stability of the performance parameters, the specified acceptance criteria, and on the use of the equipment. The test frequency for PQ is much higher than for OQ, is always performed under similar conditions to routine sample analysis, i.e., using the same column, the same analysis conditions, and the same or similar test compounds. The PQ should be performed on a daily basis or whenever the instrument is used. For a liquid chromatograph, the most important units may be the chromatographic column or a detector s lamp condition. The test criteria and frequency should be determined during the development and validation of the analytical method. In practice, PQ can mean SST, where critical system performance characteristics are measured and compared with documented, preset limits. An example sequence for PQ includes: (a) definition of the performance criteria and test procedures; (b) selection of critical parameters: precision of the peak area/height, precision of retention times, resolution
758 Papadoyannis and Samanidou between two peaks, peak width at half height or peak tailing, limit of detection (LOD), and limit of quantitation (LOQ), wavelength accuracy of a UV/Visible detector; (c) definition of the frequency, e.g., every day, every time the system is used, before, between, and after a series of runs; and (d) specification of corrective actions in case the system does not meet the criteria. The DQ should always be done by the user, whereas IQ for large, complex and high cost instruments should be done by the vendor. In some cases, the user will even lose the warranty if he installs the system. The OQ can be done by either the user or the vendor. The decision mainly depends on the resources available at the user s site, and on the vendor s capability to offer such a service with high quality. The PQ should always be done by the user, because it is very application specific, and the vendor may not be familiar with the method. As PQ should be done on a daily basis, this practically limits this task to the user anyway. On completion of EQ, detailed documentation should be available, considering qualification checklist, procedures for testing, qualification test reports with signatures and dates, PQ test procedures, and [6 11] representative results. Computer Validation/Qualification Computer validation/qualification refers to computer hardware, peripherals, and software validation, the latter including operating software, e.g., Microsoft windows, MS-DOS, and the application HPLC software. In the hardware part, the most important unit to validate is the analog-to-digital converter (ADC) because it converts input voltage into numeric data. Concerning software, it is very difficult to validate because of its complex structure. For this reason, the vendor should check, validate, and document software according to computer aided software engineering (CASE) principles. After installation, software validation can be performed with regard to integrity of program code and data, as well as the data security, with special software packages; this can be checked automatically every time at power on. On the other hand, comprehensive validation of a data system is beyond the capability expected from many chromatography laboratories. The only parts of the software package that is created after installation are userwritten programs (macros); they should be tested, ensuring security of data with restricted access to data and regular back-up copies, documenting every change made either on the data acquisition system computer or software. Computer systems should be validated during, and at the end, of the development process and after software updates. Equipment validated under a specified set of conditions, must be revalidated whenever a change is made in those conditions. For example,
Validation of HPLC Instrumentation 759 changes in application software may influence the correct functioning of the computer system. Revalidation refers to those items of the system that are affected by the change. [8,12] HPLC Instrumentation Qualification Instrument hardware should be validated for its intended use before and during its operation as part of the overall validation process, i.e., prior to routine use, after repair, and at regular time intervals. The main qualification tests applied to HPLC instrumentation concern: The injection system (usually an autosampler). Parameters that have to be checked are precision and accuracy. The pump. Parameters that have to be checked are flow rate precision and flow rate accuracy and stability regarding an isocratic pump. Gradient delay volume and gradient mixing volume, repeatability of the gradient, and mixing and proportioning accuracy are important when a gradient pump is used. The detector. Parameters that have to be checked are: wavelength accuracy detection, refractive index sensitivity, linearity, spectral quality (PDAD) short/long term noise, drift, and flow sensitivity. Column efficiency. Within-day and between-day reproducibility of the whole chromatographic system. The first step is to establish internal specifications and then compare results obtained towards manufacturers specifications and, finally, interpret the differences. Within-day and inter-day reproducibility is the minimum test for routine use. Chromatographic systems should be tested for suitability to the performance criteria of the method on a daily basis before and during routine application. An HPLC system should be tested before each single sample analysis or, if a series of samples are analyzed within a sequence, before each sequence. Retention time, peak height/area, resolution, and peak shape are the data used to monitor the performance of the HPLC system. Peak shape is one of the first things that experienced chromatographers notice when looking at a chromatogram. Ideal chromatographic peaks are Gaussian shaped but, in practice, most peaks show some peak tailing, which is measured using the asymmetry factor or the united states pharmacopoeia (USP) tailing factor as described in Fig. 2(A). Column plate number, N, is the most common parameter to describe the shape of the peak. Resolution R s, as described in Fig. 2(B), is the measure of separation of two peaks, as well as of the efficiency of the column.
760 Papadoyannis and Samanidou Figure 2. (A), Measurement of peak tailing using USP tailing factor and peak asymmetry factor. (B), Peak resolution (From Merck HPLC tutorial, with permission). The frequency of the control sample analysis depends upon the nature of the analysis. Successful analysis of the control samples assures that system is performing as expected under the SOP. Validation of HPLC equipment assures that valid measurements are obtained. The quality of the analytical data can be supported by keeping, in a safe place, records of the actual instrument conditions at the time the measurements were made. Backups should also be maintained. The tests can be performed using a typical C 18 analytical column 250 4 or 4.6 mm and a mobile phase consisting of water methanol or acetonitrile mixture. The analysis should concern compounds similar to those used for testing he performance of C 18 columns, such as phenol, or compounds used for
Validation of HPLC Instrumentation 761 routine analysis in the laboratory. New columns must be checked for performance, keeping records of its chromatographic parameters. Modular Calibration: Detector Performance The refractive index sensitivity affects gradient analysis and should be checked by measuring the absorbance when the cell is filled with methanol (n ¼ 1.329) and cyclohexane (n ¼ 1.427) at 270 nm. Noise and drift are measured in static (dry detector cell) and in dynamic mode at different wavelengths, e.g., 200, 254, and 390 nm. The change in the absorbance as a function of flow rate at the same wavelengths reflects flow sensitivity. Noise is expressed in AU/cm, drift in AU/hr, and flow sensitivity in AU min/ml. Some equipment units can perform calibration for accuracy automatically. For example some HPLC UV VIS detectors include holmium oxide filters for measurement and calibration of the wavelength accuracy. The characteristics of equipment may become altered over time, e.g., UV detectors lamps lose intensity, or pump piston seals abrade, or short-term noise affecting the LOD is increased due to flow cell contamination. These changes will have a direct impact on the performance of the HPLC instrument. The frequency of performance tests will be determined by experience and is based on need, type, and history of equipment performance. Intervals between the checks should be shorter than the time the instrument drifts outside acceptable limits. New instruments need to be checked more frequently and, if the instrument meets the performance specifications, the time interval can be increased. Long-term noise can be erroneously considered as late eluting peaks, while drift affects the data obtained, when the instrument is running over a long time period. Flow sensitivity affects flow rate and gradient programs using constant pressure pumps. PDA or Variable UV VIS Detector A number of reference chemicals with well-defined UV spectra have been used for detector wavelength calibration, including uracil, erbium perchlorate, holmium oxide, caffeine, etc. Pump Pump flow rate accuracy and stability can be checked at 1 ml/min with the column in place, by measuring the time required to fill a 10 ml volumetric
762 Papadoyannis and Samanidou flask at the detector outlet. Other flow rates can be tested if analyses with narrow bore or fast LC are routinely performed. An acceptance criterion is 1.00+0.05 ml/min. Flow rate variations affect retention times. Gradient Pumps Gradient delay volume, gradient mixing volume, as well as mixing and proportioning accuracy should be evaluated. These tests are executed by replacing the column with short tubing and filling the reservoirs with water and aqueous phenol solution. Increasing and decreasing steps must be used. The above-mentioned parameters affect R t and selectivity of the separation, especially important when the gradient method has to be transferred from one system to another. Repeatability of the gradient analysis, important for the precision of the gradient method, is assessed by calculating RSD of peak area and retention times for replicate analysis of a control compound solution. Compositional accuracy is determined by making 5-min step gradients from water in one solvent line to 0.1% acetone in water in a second solvent line. The absorbance of each step against that of the 100% step is measured. A higher flow rate of 2 ml/min can be used to sharpen the step definition and increase test robustness. All four solvent lines are checked in a 1-hr solvent program. Gradient methods in HPLC depend on the dwell volume, V d, i.e., the volume between the point of mixing eluents A and B and the column inlet, including the loop of the injector. This volume differs from instrument to instrument, so that a method can be transferred only if this volume is well defined. When there is a significant difference in dwell volume between two systems, retention times and resolution are dramatically different. So one must be state in which range of V d the method is still valid. Autosampler Autosampler precision can be checked by replicate injections of a control sample, with wash injection intervals between every two sample injections. The repeatability of peak areas, mathematically expressed as RSD, is used as a criterion for autosampler precision. For example, when 10 consecutive injections of 10 ml of a solution are performed, expected RSD for peak area precision ranges from 0.5% to 1.0%. Single injections of different volumes, such as 5, 10, 50, 80 ml can also be used checking simultaneously the linearity of the injector, the detector, and the data system. Another approach to qualify the autosampler involves the gravimetric determination of the average volume of water per injection withdrawn from a tared vial after six 50 ml injections.
Validation of HPLC Instrumentation 763 The procedure takes less than 10 min and has an acceptance criterion of 50+2 ml. The same experiment can be used for the assessment of flow rate precision by evaluating the repeatability of retention times for the examined compounds, expressed again as RSD. Both terms, flow rate and injection volume precision, affect the precision of the results of the analysis. Carry over problems can be identified in the wash chromatograms at maximum detector sensitivity. They can be checked by multiple analyses of high concentration samples followed by blank injections. Column Oven A temperature accuracy test of the column oven, using a calibrated thermal probe, is used with an acceptance criterion of 35+28C. Column Two peaks are considered resolved, if the R s value is.1.5. For multicomponent mixtures, a number of R s values may have to be considered or just the resolution between the most critical peaks as an indicator of the quality of the separation. Typical values for R s range from 1.5 to 2.0 for critical resolution. Computer A computer associated with HPLC equipment may control the instrument or part of it, e.g., the photodiode array detector, or acquire data generated by the instrument, such as the peak area of analytes. Software functions: For an automated HPLC analysis, required software functions include: instrument control, data acquisition, peak integration, peak purity checks, compound identification through spectral libraries, quantitation, file storage and retrieval, and a print out of methods and data. Hardware interface: Determining true performance of an electronic interface of a chromatography data system requires special instrumentation that is not generally found in a chromatographic laboratory. Testing procedure is similar to that of calibrating any electronic instrument. The equipment hardware and computer software should be developed and validated according to a documented procedure, e.g., according to a product life cycle. The vendor should have a documented and certified quality system,
764 Papadoyannis and Samanidou e.g., ISO 9001. Quality must be designed and programmed into software prior to, and during, its development phases by following written development standards, including the use of appropriate test plans and methods. Correct functioning of software and computer systems should be verified after installation and before routine use. Operational qualification for software and computer systems is more difficult than for hardware, as (a) it is more difficult to define specifications, test procedures, and acceptance criteria; and (b) there are hardly any guidelines available on OQ of software and computer systems. Because of these problems, there is even more uncertainty for software and computer systems than for equipment hardware. The type of testing required for the qualification of software and computer systems depends very much on the type and complexity of software, i.e., if software and computer hardware are supplied by one vendor, or computer systems which are interconnected to each other and/or interfaced to analytical systems or the software are developed in the user s laboratory in addition to a vendor-supplied package (e.g., a macro). [12] ANALYTICAL METHOD VALIDATION After equipment validation, the next step is analytical method validation, which covers testing of significant method characteristics according to good analytical practice guidelines. The validation of an HPLC method is the procedure that gives the chromatographer information to determine whether the system is operating as it should for its intended use, providing accurate, precise, and reliable analytical data in a specific situation, meeting the preestablished specifications. The validity of a specific method should be demonstrated in laboratory experiments using samples or standards that are similar to the unknown samples that will be routinely analyzed. Chromatographic methods need to be validated before the first routine use. To obtain the most accurate results, all of the variables of the method should be considered, such as sampling, sample preparation, chromatographic separation, detection, and data evaluation, using the same matrix as that of the intended sample. The proposed procedure must go through a rigorous process of validation. All validation experiments should be documented in a formal report. Though the need to validate analytical methods is clear, the procedure for performing the validation is not clearly defined in terms of which validation parameters should be used, with which specific procedures to evaluate a particular parameter, and what are the appropriate acceptance criteria for a particular parameter.
Validation of HPLC Instrumentation 765 Successful completion of the validation results in a method that can reliably be used to characterize real samples. Ongoing validation activities may also be necessary during the routine utilization of an analytical procedure, as well as revalidation of the analytical procedure as certain operational aspects of the method are changed during its routine and continuous application. [13 16] Validation Parameter Guidelines for Method Validation After the method has been developed, its performance must be validated with respect to several different performance parameters. To determine which operational parameters should be included in a formal validation protocol, one can look to the chemical literature to assess the practical state of the art among the practitioners of the desired methodology or, alternatively, one can examine existing guidelines published by organizations with recognized authority. Regulatory agencies and published literature provide the criteria for what constitutes a validated chromatographic method. The USP, for example, has published specific guidelines for method validation regarding pharmaceuticals, while there are no official guidelines referring to biological fluids. An attempt for harmonization was made at the International Conference on Harmonization (ICH) in 1995 and 1996 by representatives from the industry and regulatory agencies from the USA, Europe, and Japan who defined parameters, requirements and, to some extent, also methodology for analytical methods validation. In 1990, the USP 22 guideline listed eight individual parameters that must be investigated and documented in order to validate a method: (1) accuracy, (2) precision, (3) LOD, (4) LOQ, (5) selectivity, (6) range, (7) linearity, and (8) ruggedness. In 1995, USP 23 changed, to some extent, these parameters, as selectivity is now explicitly referred to as specificity, linearity, and range were combined and robustness was broken out of ruggedness. Ruggedness retains its original definition of day-to-day, instrument-to-instrument, operator-to-operator, etc., reproducibility. The robustness of an HPLC method is defined as a measure of its capacity to remain unaffected by small but deliberate variations in method parameters and provides an indication of its reliability during normal usage. So USP 23 includes: (1) accuracy, (2) precision, (3) LOD, (4) LOQ, (5) specificity, (6) linearity and range, (7) ruggedness, (8) robustness. Terminology is included in an ICH publication, but they are not part of required parameters. The definitions of these parameters and indicative acceptance criteria are discussed in the following paragraphs. Each of the above listed parameters is evaluated in in terms of validating an HPLC method. The procedures for
766 Papadoyannis and Samanidou performing the validation must be presented in a complete, well defined, practical, and understandable format. Once the validation is complete, an investigator must be able to interpret the results. Acceptance criteria allow the researcher to definitely determine, by comparing method performance data to the criteria, whether the method under evaluation is performing in a valid manner. These criteria should be universally applicable, numerically and [9 12] mathematically explicit, complete, and achievable. Accuracy and Recovery The accuracy of an analytical method is the degree of agreement of results generated by the method to the true value, or a conventional true value. Accuracy can be assessed by applying the analytical method to samples or mixtures of sample matrix components to which known amounts of the analyte have been added, above and below the normal levels expected in the samples. Method accuracy is the agreement between the difference in the measured analyte concentrations of the fortified (spiked) and unfortified samples and the known amount of analyte added to the fortified sample. Comparison of method s results can be performed using an established reference method, assuming that the latter is free from systematic errors. Secondly, accuracy can be measured by analyzing a certified reference material, and comparing the measured value with the true value as supplied with the material. If such reference material is not available, a blank sample matrix can be spiked with a known concentration that should cover the range of concern, including one concentration close to the quantitation limit. The expected recovery depends on the sample matrix, the sample processing procedure, and on the analyte concentration. Recoveries can be determined by either external or internal standard methods. Although it is desirable to attain recovery close to 100%, other values such as not less than 50%, 80%, and 90% have been used as limits. Quantification by external standard is the most straightforward approach, since the peak response of the standard is compared to the peak response of the sample. The standard solution concentration should be close to that expected in the sample solution. Precise control of the injection volume is mandatory, because it influences the accuracy. Peak response is measured as either peak height or peak area. For the internal standard method, a substance is added at the earliest possible point in the analytical scheme to compensate for sample losses during extraction, clean-up, and final chromatographic analysis. The internal standard must be completely resolved from all other peaks in the chromatogram, having chemical and physical properties as similar as possible to those of the analyte of interest, so that the detector response is similar to the solute to be quantified.
Validation of HPLC Instrumentation 767 However, it is sometimes difficult to find the proper internal standard, as this must be stable over the time of the measurement period, absent from real samples, not reactive with the analytes, and being eluted within a reasonable retention time. Analogs, homologs, and isomers are usually preferred. In presence of analytes with different chemical or physical properties, two or more internal standards representing these analytes should be used. Reagent blank, internal standard blank, solvent blank, and compound blank samples must also be analyzed for accurate, qualitative, and quantitative results. Procedural guidelines for accuracy determination include replicate analysis, e.g., 3 6 assays, at five levels, over the range from 80% of the lowest expected assay value to 120% of the highest expected assay value, or from 75% to 125% of label claim, six samples of drug in the matrix spanning 50 150% of the expected content. At minimum, three concentrations must be used within the analytical range (extremes and midpoint of expected or near quantitation limit, center of range and upper bound of standard curve). For trace level analyses, acceptance criteria include: 60 110% recovery for concentrations below 100 ppb, 80 100% recovery for concentrations above 100 ppb, 70 120% for concentrations below 1 ppm. In biological samples, method accuracy for discovery phase investigations should be +20% of actual, with recoveries of +10% being necessary in pre-clinical and clinical studies. Alternatively, it is recommended that the mean recovery value should be within +15% of actual except at the quantitation limit where +20% is acceptable. Precision and Reproducibility The precision of a method is the extent of agreement among individual test results, when the procedure is applied repeatedly to multiple samplings. Precision can be divided into three categories: (a) repeatability; (b) intermediate precision; and (c) reproducibility. Repeatability (intra-assay or within-day precision) is obtained when the analysis is carried out in one laboratory by one operator using one piece of equipment over a relatively short time span. It reflects the variation in replicate procedures performed within a short time period, with the same operational conditions. At least 5 or 6 determinations of three different matrices at 2 3 concentrations should be done and expressed by the RSD. The acceptance criteria for precision depend on the type of analysis. While, for compound analysis in pharmaceutical quality control, precision of better than 1% RSD is easily achieved, for biological samples, the precision may be at levels of
768 Papadoyannis and Samanidou 10 15%. For environmental and food samples, the precision is very much dependent on the sample matrix, the analyte concentration, and the analysis technique, varying between 2% and 20%. Intermediate precision is defined by ICH as the long-term variability of the measurement process and is determined by comparing the results of a method run within a single laboratory over a number of weeks. A method s intermediate precision may reflect disagreement in results obtained by different operators, from different instruments, with standards and reagents from different suppliers, with columns from different batches, or a combination of these. The objective of intermediate precision validation is to verify that, in the same laboratory, the method will provide the same results after finishing the method development. Reproducibility, as defined by ICH, represents the precision obtained between laboratories with the objective to verify that the method will provide the same results in different laboratories. The reproducibility of an analytical method is determined by analyzing aliquots from homogeneous lots in different laboratories with different analysts and by using operational and environmental conditions that may differ from, but are still within the specified, parameters of the method (inter-laboratory tests). Various parameters affect reproducibility: Differences in room environment (temperature and humidity), operators with different experience, equipment with different characteristics, e.g., delay volume of an HPLC system, variations in material and instrument conditions, e.g., in HPLC, mobile phases composition, ph, flow rate of mobile phase, Columns from different suppliers or different batches, solvents, reagents, and other material with different quality. Reproducibility is defined as long-term variability of the measurement process, which may be determined for a method run, within a single laboratory, but on different days. Reproducibility also applies to a method, either run by different operators, different instruments, or a combination of the above. The reproducibility standard deviation is typically two- to three-fold larger than that for repeatability. Precision is often expressed relative to one day as intra-day (within-day) precision or relative to a period of days, as interday (between days) precision. Reproducibility, in the sense of intra-laboratory precision, is related to the procedure being performed at two or more laboratories as in, for example, a collaborative study. Precision in retention times and peak area or height is a major criterion of a separation system. Retention time precision is important because R t is the primary means for peak identification. It is also an important performance criterion and diagnostic for an LC pump and a column. Precision of peak areas or heights is important because they are used for calculating amounts during quantification. It is also the most important
Validation of HPLC Instrumentation 769 performance criterion for an LC injection system. Precision should be determined using a minimum of five replicate chromatograms. For bioanalytical samples, precision is studied at least at low, medium, and high concentrations. This is repeated on separate days to calculate intra-day and inter-day precision. Precision reflects a procedure s ability to reproduce the same, but not necessarily the correct or expected, result each time it is correctly performed. Precision is assessed by repetitively injecting a number of samples and statistically evaluating the resulting data. Important issues related to the precision determination include the number of replicates required and the type of sample to be tested. For the determination of repeatability, recommendations include: (1) 5 10 replicates for release or stability assays; (2) duplicate measurements made on 10 samples at each of three different analyte levels; (3) five replicates at three levels (LOQ, mid-range and upper calibration bound); and (4) replicate samples at analyte levels of 80 120% of expected for dosage forms and drug substance tests. For intermediate precision, the repeatability experiments should be performed on 2, 3 5, or at least 10 separate days. To assess reproducibility, the experiments have to be performed in at least two laboratories. System precision should be 1 2% RSD (or higher for low level impurities). The repeatability is generally 1/2 1/3 of the reproducibility. However, for biological samples, an RSD of 10 15% should be acceptable as the minimum precision. Systematic errors result from sources traced to the methodology, the instrument, or the operator and affect both accuracy and precision of the measurement. Random errors affect the precision and are difficult to eliminate, as they result from random fluctuations in the measured signal due to noise and other factors. The imprecision of the entire procedure is often dominated by the random errors of the most imprecise step. The difference between an accurate and a precise method is illustrated in Fig. 3. Selectivity Specificity The terms selectivity and specificity are often used interchangeably. The term specificity generally refers to a method that produces a response for a single analyte only, while the term selective refers to a method which provides responses for a number of chemical entities that may or may not be distinguished from each other. Since there are very few methods that respond to only one analyte, the term selectivity is usually more appropriate. The USP defines selectivity of an analytical method as its ability to measure accurately an analyte in the presence of interferences, such as synthetic precursors,
770 Papadoyannis and Samanidou Figure 3. Comparison of precision and accuracy parameters of an analytical method. excipients, enantiomers, and known (or likely) degradation products that may be expected to be present in the sample matrix. Selectivity in HPLC is obtained by setting optimal chromatographic conditions, such as mobile phase composition, column temperature, and detector wavelength. There is a variety of ways to validate selectivity. One approach is to demonstrate a lack of response in the blank biological matrix. A second approach is to check whether the intercept of the calibration curve is significantly different from zero. Peak purity is one important parameter in chromatography to determine whether the peaks within a sample chromatogram are pure or consist of more than one compound. To determine the homogeneity of a chromatographic analyte response the use of UV/visible diode-array detector (DAD) and massspectrometers is recommended, acquiring spectra during the entire chromatogram. The spectra acquired are normalized and overlaid for graphical presentation; in case they are different, the peak consists of at least two compounds. Usually the simplest and most frequently used peak purity algorithm compares three spectra across a peak: on the upslope, at the peak apex and on the downslope. Specificity relates to the ability of a method to measure only what it is intended to be measured with a requisite level of accuracy and precision, even though the sample may contain other, related compounds. The most commonly cited specificity evaluation procedure is the analysis of a placebo, wherein the sample matrix without the analyte is analyzed and the resulting
Validation of HPLC Instrumentation 771 system response is examined for the presence of responses, which interfere or overlap with that of the analyte of interest. Other procedures for specificity include: (1) peak re-analysis, wherein the peak of interest is collected and reanalyzed by different chromatographic conditions or with methods that are sensitive to analyte structure; (2) collaboration in which the sample is quantitatively analyzed using two or more detection/separation strategies and the results compared; and (3) use of detectors (e.g., MS or PDAD to assess peak purity). Linearity and Range The linearity of an analytical method is its ability to extract test results that are directly, or by means of well-defined mathematical transformations, proportional to the concentration of analytes in samples within a given range. In chromatography, peak parameters are related to analyte concentration via standardization procedures. This relationship is then used to convert a sample s peak parameter to its analyte concentration. Linearity is determined by a series of 3 6 injections of five or more standards, whose concentrations span over the range of expected concentrations, or 80% of the lowest expected level to 120% of the highest expected level, or 50 150% or 10 200% of the expected working range, or 25 125% of the target range specified. The response should be directly or, by means of a well-defined mathematical calculation, proportional to the concentrations of the analytes. A linear regression equation applied to the results should have an intercept not significantly different from zero; otherwise, it should be demonstrated that there is no affect on the accuracy of the method. Frequently, the linearity is evaluated graphically in addition to, or alternatively to, mathematical evaluation. The correlation coefficient of the best linear least squares regression model should be between 0.98 and 1.00 or greater than 0.999 with the slope and intercept reported. Though there is no rule stating that the relationship between instrumental response and analyte concentration must be directly linear for a procedure to be valid. The desire to have a linear relationship reflects the practical consideration that a linear relationship can be accurately described with fewer standards than a nonlinear relationship, and the subjective expectation that a linear relationship is more rugged than a more complicated, nonlinear one. A method s range is linked to its linearity. The range is the interval between the lower and upper analyte concentration, for which it has been demonstrated that the analytical procedure has a suitable level of accuracy, precision, and linearity. The range is expressed in the same units as analytical method s results (e.g., ng/ml, ppb, %, etc).
772 Papadoyannis and Samanidou Limits of Sensitivity: Limit of Detection and Limit of Quantitation Sensitivity is the ability of a method to reliably respond in a consistently recognizable manner to decreasingly smaller amounts of analyte. Frequently utilized measures of sensitivity are the LOD and LOQ. For chromatographic procedures, LOD is the lowest concentration of analyte that can be detected above the baseline detector noise at the most sensitive instrument setting, but not necessarily quantified. In practice, in chromatography, the detection limit is the injected amount, which results in a peak with a height twice or three times as high as the baseline noise. The LOD can be determined either directly or from other validation data. Its direct measurement involves an analysis of the method s peak-to-peak baseline noise or an analysis of the variation in the method s blank response. In either case, LOD is calculated as either 2 or 3 times the variation in measured response, where the factors are associated with the 95 and 99% confidence intervals for a normal distribution. Practically, LOD can be measured by the serial dilution of samples until the peak can no longer be observed. The LOD can be estimated as the value of the linear calibration curve s y-intercept. The LOQ is the minimum injected amount that gives precise measurements, in chromatography typically requiring peak heights 10 20 times higher than baseline noise. Another approach is the EURACHEM: a number of samples with decreasing amounts of the analyte are injected six times. The calculated RSD% of the precision is plotted against the analyte amount. The amount that corresponds to the previously defined required precision is equal to the LOQ. Alternatively, LOQ is the lowest injected amount of analyte, which results in a reproducible measurement of peak areas and can be reproducibly quantitated above the baseline noise. Peak heights are typically required to be about 10- to 20-times higher than baseline noise. Quantitation implies that the measurement possess a specified accuracy and precision. In some applications, LOQ is defined as the smallest concentration included in the standard curve. The baseline response method for estimating LOQ involves the following procedure: the chromatogram resulting from a blank injection is examined over a range of 20 peak widths and the noise is measured as either the largest peak to peak fluctuation or as the largest deviation (positive or negative) from the mean response. The LOQ is then calculated as the product of ten times the measured deviation and the calibration curve slope. The LOQ can also be determined as the lowest analyte concentration for which duplicate injections results in a % RSD 2%. In routine applications, it has been recommended that LOQ should be within the working linear concentration range and that a
Validation of HPLC Instrumentation 773 specification limit should be no lower than twice the LOQ. For clinical applications, the LOQ should be at least 10% of the minimum effective concentration. Ruggedness It is generally expected that an analytical method will perform in an acceptable manner each time it is used. While a consideration of method ruggedness is a necessary part of any method s validation, it is a critical issue for compendial methods because of their widespread use in many different laboratories. Ruggedness establishes a method s ability to perform effectively in the face of variations on operational and environmental conditions, which can reasonably be expected to occur whenever the method is applied. More specifically, ruggedness is the reproducibility of test results obtained by the analysis of samples under a variety of normal test conditions such as different laboratories, analysts, instruments, reagent lots, elapsed assay times, temperatures, etc. Thus, ruggedness addresses unintentional variation in the method introduced by its application, at different times, by different people, at different locations, using different instrumentation and materials. A rugged method will be able to withstand minor operating or performance changes and has built in buffers against typical procedural abuses, such as, differences in care, technique, equipment, and conditions. For the determination of methods ruggedness within a laboratory, a number of chromatographic parameters, for example, flow rate, column temperature, detection wavelength, or mobile phase composition is varied within a realistic range, and the quantitative influence of the variables is determined. If the influence of the parameter is within a previously, specified tolerance range, the parameter is said to be within the method s ruggedness range. Clearly, ruggedness is assessed by analyzing aliquots from homogeneous sample lots using operational and environmental conditions that differ, but are still within the method s specified operating range. The ruggedness test should be performed at several values of each operational parameter that affects method performance, e.g., mobile phase flow rate and composition, (ph, buffer concentration, ion pairing reagent concentration, percent organic phase, column temperature, injection volume, gradient dwell time, column lots or column manufacturers, different room temperature and humidity in separate laboratories, detection wavelength, analysts with different experience, instruments from various vendors, reagents from different suppliers, columns from different batches, sample and standard preparation procedures, and operating temperature. For operator-related ruggedness, a