As the US Food and Drug Administration (FDA) strives to streamline its regulatory process for bringing new drugs to market (see sidebar "Manufacturing data key to spurring drug development"), efforts to ensure the quality, safety and efficacy of dietary supplements and herbal medicines are gaining more attention. FDA is working with the Federal Trade Commission (FTC) to crack down on unsubstantiated health claims by supplement makers; reports of serious adverse events involving weight-loss supplements containing ephedra, a medicinal herb, have led to product seizures and FDA legal action against manufacturers. Researchers are also investigating whether herbal products with kava, another medicinal herb, cause liver toxicity. Manufacturers and regulators hope that new good manufacturing practice standards (GMPs) will help ensure the safety and quality of supplements on the market and overcome demands to regulate these products more like drugs. 	

Manufacturing data key to spurring drug development	

Under the Dietary Supplement Health and Education Act of 1994 (DSHEA), dietary supplement makers do not have to prove efficacy or even safety to go on the market, as is required for drugs. Manufacturers merely have to assure FDA that a product is "reasonably expected to be safe," shifting the burden to FDA to demonstrate that a product poses a significant or unreasonable risk to consumers before enforcement action can be taken. DSHEA permits manufacturers to make "structure/function" claims for supplements (such as "relieves stress" or "supports joint function") but not "health claims" (such as "reduces risk of cardiovascular disease"), a fuzzy distinction that has generated fierce legal battles. 				

However, the proliferation of questionable dietary product health claims - and the growth of the supplement market from $3.3 billion in 1990 to approximately $18 billion today - is prompting government regulators and enforcement officials to call for stronger oversight. Some members of Congress want to revise DSHEA to clarify government authority to remove unsafe products from the market and to require supplement manufacturers to file adverse event reports (AERs) with FDA. 	

In response, FDA's Center for Food Safety and Nutrition (CFSAN) is establishing a new AER data system for dietary supplements, food additives and cosmetics to detect more safety "signals" from the relatively few AERs filed for these products. FDA commissioner Mark McClellan has also launched a programme to clarify what health claims can be made for conventional foods, as well as supplements and other medical products, and what scientific evidence is needed to support such claims. However, a more liberal policy on health claims includes plans for stepped-up enforcement action against manufacturers that make false or misleading statements. 	

All parties agree that supplement regulation would be greatly improved by FDA's adoption of manufacturing standards. DSHEA has authorized FDA to do this, but the GMP development process has been excruciatingly slow. Manufacturers found an initial FDA GMP proposal too similar to standards for drugs. FDA tried to move the process forward by publishing a set of standards proposed by manufacturers in 1997, but then was slow to come up with its own version. A proposed rule that was headed for final publication in late 2000 was pulled by the Bush administration for further review. 				

Last October (2002), the Department of Health and Human Services (HHS) finally sent a revised proposed rule to the Office of Management and Budget (OMB) for clearance and announced in January (2003) that the proposal was coming "very soon."	

Officials at CFSAN's Office of Nutritional Products, Labeling and Dietary Supplements (ONPLDS) plan to hold public meetings to encourage broad dialogue on this long-awaited policy, but a final rule remains months, if not years, away. The debate is expected to focus on requirements for the validation of manufacturing processes and analytical methods, plus the need for expiration dating on products versus "use before" recommendations. Small supplement makers oppose strict validation policies as too costly and unnecessary, and don't want to do the extensive stability testing needed to set expiration dates. Once FDA adopts final GMPs, the agency plans to establish an inspection programme to ensure industry compliance. 	

One possibility is that FDA will modify inspection requirements for manufacturers that participate in voluntary quality audit programmes established by private organizations and laboratories. The National Nutritional Foods Association (NNFA) has a partnership with NSF International to audit and certify the quality of dietary supplements. NNFA requires members to comply with its own GMPs voluntarily. These call for firms to establish procedures and maintain records on plant maintenance, facility sanitation, equipment design, quality control operations, process controls, personnel training and distribution procedures. 								

NNFA members who want to certify GMP compliance may sign up for a third-party audit programme, a step taken by approximately 40 companies, with 60 more in the pipeline. The association has also launched a "true label" programme that conducts random tests on the ingredients in certain supplement product categories to see if what is in the product matches the label. Compliance with GMPs, explained NNFA executive director, David Seckman, at a conference on dietary supplements sponsored by the Food and Drug Law Institute (FDLI) in January, requires companies to understand, analyse and control their manufacturing and distribution processes and to demonstrate to third-party auditors that they keep required records and meet set standards. 	

NSF International has its own programme for certifying that individual supplement products meet quality standards recognized by ANSI, which involves demonstrating compliance with GMPs. Manufacturers submit products for testing and undergo an audit of production facilities. Under its agreement with NNFA, NSF will accept NNFA GMP certification and vice versa. NSF also offers GMP registration, which is particularly attractive to contract manufacturers that don't want to certify the quality of specific products. 	

The US Pharmacopeia (USP) has also established a dietary supplement verification programme (DSVP) that is open to all supplement firms. Manufacturers may display the DSVP seal on products that USP tests to ensure ingredients meet labelling strengths, have acceptable contaminant levels and meet USP manufacturing standards. USP conducts lab tests on products to evaluate composition, strength, contaminants and dissolution; reviews manufacturer quality control documentation, test methods and stability data; and inspects manufacturer facilities to ensure compliance with GMPs. 	

USP recently announced that it would award its seal to a number of products in Pharmavite's "Nature Made" line and to Kirkland products sold in Costco stores. Weider Nutrition International and Inverness Medical Innovations have products undergoing USP's testing process, and other supplement makers are signing up, according to USP vice president Srini Srinivasan. He believes that the USP audit programme complements FDA compliance efforts and that manufacturers that qualify for the USP seal will meet FDA GMP requirements. 	

Although GMPs and certification are important for ensuring that manufacturers operate high-quality production systems, they do not guarantee product purity, consistency or safety. GMPs will help bring all industry players up to a higher level of quality, observed Annette Dickinson, acting president for the Council for Responsible Nutrition (CRN) at the FDLI conference, but added that GMPs can be oversold as a guarantor of quality. Also needed are analytical methods and standards to ensure product purity and safety; her organization, for example, has issued a voluntary quality monograph on omega-3 fatty acids and is updating a publication on safe intake limit ranges of vitamins and minerals. 				

FDA, USP, National Institutes of Health (NIH) and others have launched a number of initiatives to analyse and set quality standards for supplements and botanicals, which are listed below:	

Analysis. FDA is conducting a systematic pharmacological analysis of ephedra products on the market to fully assess if they contain synthetic ephedrine alkaloids and thus would qualify as drugs and be subject to drug regulations. So far, such analysis has found that certain supplements marketed as "natural" actually contained the pharmaceutical-grade chemicals, ephedrine hydrochloride and caffeine anhydrous, which make them illegal and subject to prosecution.	

Monographs. USP is updating and expanding ingredient monographs for botanicals and dietary supplements, building on its original charge in 1820 to set standards on natural medicinal products such as chamomile, valerian and ginger. USP has published 28 monographs on botanicals and more than 100 on other supplement products based on more sophisticated analytical techniques, and dozens more are in the works. The organization has also issued reference standards for amino acids, botanicals, vitamins and minerals, and other products which can be used as standards to ensure product quality, purity and potency around the world. 	

Supplements. FDA and NIH's Office of Dietary Supplements (ODS) have awarded a contract to AOAC International (Gaithersburg, Maryland, USA) to provide accurate, validated analytical methods for 20 supplements during the next 5 years. An AOAC dietary supplement task group has drafted a priority list for evaluating ingredients based on product market share, FDA safety concerns and the availability of test methods. Expert review panels will identify best analytical methods for validating product quality and identity, starting with existing laboratory test methods for substances such as mcarotene, black cohosh, ginkgo, kava, ginseng and saw palmetto. The programme aims to train laboratory scientists to use accepted analytical methods and will disseminate official methods through an online electronic compilation of analytical methods. Recommended methods may be included in USP monographs and other standards. 	

Research. ODS at NIH supports a broad research programme to develop more information on safety and health claims for certain common supplements (for example, chromium's impact on diabetes and alternative therapies for treating benign prostate symptoms). One initiative is to create a database of research on supplement ingredients. ODS has also established several botanical research centres to identify and characterize botanicals, assess bioavailability and bioactivity, identify active ingredients and explore mechanisms of action. 	

Information. The American Botanical Council (ABC) seeks to document the value of herbal preparations by publishing more information on these products. In 1996, it issued a translation of German monographs on more than 300 herbs and combinations. ABC subsequently published a more in-depth guide to 107 popular US herbs. This year it will issue a clinical guide to 29 commonly used herbal medicines that provides information for practitioners on preparation, dosage, adverse events and herb-drug interactions.	

Safety. A joint committee formed by the Institute of Medicine's Food and Nutrition Board and the National Research Council has established a framework for evaluating the safety of dietary supplement ingredients. The panel recently issued prototype monographs and solicited public comment on six common ingredients - chaparral, shark cartilage, glucosamine, melatonin, saw palmetto and chromium picolinate. The monographs focus on ingredient safety issues, as opposed to health claims and effects, and aim to help FDA detect unsafe substances in supplements.	

Manufacturers generally support public-private initiatives to document product quality and curtail illegal claims to promote consumer confidence in supplements and herbal remedies, as well as head off more stringent government regulation. "Industry needs to step up to the plate and take responsibility for products," commented Dickinson of CRN. She considers third-party validation programmes sponsored by USP and other independent laboratories a good way to help companies demonstrate their commitment to quality. 				

At the same time, ongoing legal and regulatory battles concerning products such as ephedra have undermined public confidence in supplements overall and has prompted closer scrutiny of supplement marketing by FDA, FTC and state officials. "It's bad for industry, bad for FDA and bad for consumers for an issue like this to go on for years," Dickinson commented. Supplements may offer health benefits, but manufacturers need to be proactive about compiling evidence on product safety and benefits to retain public confidence.	

Articles

Regulars

As the US Food and Drug Administration (FDA) strives to streamline its regulatory process for bringing new drugs to market (see sidebar "Manufacturing data key to spurring drug development"), efforts to ensure the quality, safety and efficacy of dietary supplements and herbal medicines are gaining more attention.

Standards for Supplements

Although it is probably the oldest excipient used in solid dosage form formulations, lactose is still one of the most important, particularly as a diluent in tabletting. However, the inadequate compactability of a-lactose monohydrate at particle sizes that provide good flow properties of the powder mixture limits the use of crystalline a-lactose monohydrate as a filler-binder for direct tabletting. Because of the need for direct compression excipients, to accompany the progress being made on high speed rotary tablet presses, many researchers and excipient manufacturers modified crystalline a-lactose monohydrate to achieve a product exhibiting good compactability, reduced capping tendency and good flow properties.	

Two approaches were taken: the first was to process lactose by itself. For example, agglomeration of small a-lactose monohydrate crystal fractions, which display a large specific surface and have better compactability led to granules with good flowability such as Tablettose (Meggle GmbH, Wasserburg, Germany) or Pharmatose DCL 15 (DMV Pharma, Veghel, The Netherlands). The large granules fractionate at low pressure, releasing small crystals that demonstrate adequate compactability. Spray drying increased the amorphous portion of lactose in the agglomerates, resulting in improved tabletting properties compared with the granules. However, alteration of amorphous lactose to a-lactose monohydrate may change the tabletting properties of the spray dried products during storage.	

The limitations of improving the properties of lactose-only direct compression excipients led to the second approach: combinations of a-lactose monohydrate and binding and/or plastifying agents. Binary mixtures of crystalline a-lactose monohydrate with microcrystalline cellulose, povidone or starch generally increase the compressibility of the mixtures compared with pure a-lactose monohydrate. However, these mixtures do not improve flowability; hence, coprocessed materials were developed. Agglomeration of a-lactose monohydrate with povidone and cross-povidone (Ludipress, BASF AG, Ludwigshafen, Germany), for example, led to a suitable filler for direct tabletting on high speed presses. Additionally, coprocessing crystalline a-lactose monohydrate with powdered cellulose (Cellactose, Meggle) or microcrystalline cellulose (MicroceLac, Meggle) resulted in improved bonding ability and excellent flow properties. The latest material on the market is StarLac (SL), a coprocessed filler-binder consisting of 85% a-lactose monohydrate and 15% native corn starch. Exhibiting the lowest elastic recovery at high binding capacity compared with other starches, corn starch seems almost destined to form a filler-binder with a-lactose monohydrate with excellent compactability. A further advantage of starch is its bifunctionality as a binder and disintegrant. Thus, out of just two components (a-lactose monohydrate and starch) a product with good flow properties, compactibility and disintegrant functionality has been developed. The objectives of this study were, therefore, to check the tabletting properties of the pure filler-binder and its mixtures with two active ingredients (acetaminophen and vitamin C) under formulation conditions. As a comparison, all the experiments were also performed with a physical mixture of a-lactose monohydrate and corn starch.		

Table II: Composition of powder mixtures and apparent yield pressure (YP).	

SL, a coprocessed compound containing 15% (w/w) corn starch and 85% (w/w) a-lactose monohydrate (Meggle, Wasserburg, Germany and Roquette Freres, Lestrem, France), spray dried lactose (FlowLac 100 [FL], Meggle), Aerosil 200 (Degussa, Frankfurt am Main, Germany), magnesium stearate (Baerlocher, Unterschleißeim, Germany), Maisstarke extra-white corn starch (Roquette Freres), Paracetamol DC acetaminophen (Hartington, Chesterfield, UK) and C 97/Ascorbinsaure DT vitamin C (BASF AG).				

Powder characterization. Powder flowability was measured as the angle of repose and flow-through time of 100 g of powder using Pfrengle's funnel (DIN 53 916). True density was determined using a Beckman air comparison pycnometer (Model 930, Beckman Instruments, Fullerton, California, USA).			

Figure 1:  Scanning electron micrographs of StarLac (1a, 1b) and a physical mixture of 85% FlowLac and 15% corn starch (2a, 2b) at a magnification of 5003 (a) and 10003 (b).				

Bulk and tapped density were measured according to the European Pharmacopoeia (4th edition) using a 250 mL graduated cylinder and an Engelsmann tap meter (JEL ST 2, Engelsmann, Ludwigshafen, Germany).	

Particle size distribution was determined by laser diffraction (Mastersizer 2000, Malvern Instruments, Malvern, UK) applying the dry dispersion module Scirocco 2000 (Mastersizer 2000, Malvern Instruments) at a dispersion pressure of 0.1 bar. All values were averaged from three repetitions, and the standard deviation is given in parentheses. The properties of the excipients are listed in Table I.			

Blending. SL or a respective physical mixture (PM) of 15% corn starch and 85% spray dried lactose was blended for 15 min with the active ingredient (paracetamol [P] or ascorbic acid [AA]). Finally, 0.5% magnesium stearate was passed through a 315 mm sieve onto the mixture and mixing was continued for another 5 min. All batches were 1.5 kg and are listed in Table II. The shortened form for each mixture consists of the abbreviation of the filler (SL or PM) followed by the active ingredient (AA or P) and ends with the percentage of the active ingredient (30-70), for example SLAA30 is the shorted form of a mixture consisting of SL as a filler and 30% ascorbic acid as an active ingredient.	

Rotary tablet press. Tablets of 400630 mg were compressed on a Kilian RL-H instrumented rotary tablet press (Kilian, Koln-Niehl, Germany) at a speed of 25 rpm using a gravity feeder. Five out of 15 punch stations were equipped with 10 mm flat-face bevelled-edge tooling. Tablets were compressed at 5, 10, 15 and 20 kN and the data were recorded and processed by the Messfix software.			

Figure 2: Heckel plots of FlowLac, a physical mixture of 85% FlowLac and 15% corn starch, StarLac and corn starch.				

Single punch tablet press. Tablets were compressed on an instrumented single punch tablet press Korsch EK II (Korsch Pressen, Berlin, Germany) at a compression pressure of 200 MPa using 10 mm flat tooling. The upper punch displacement was measured using a digital incremental displacement transducer MT 2571 (Heidenhain, Traunreut, Germany). Data were acquired using the MGC Plus system including an ML 10 B voltage amplifier (HBM, Darmstadt, Germany) and the Catman software (HBM). The displacement error caused by distortion of the tooling and frame of the machine was obtained from 10 punch-to-punch compressions, which led to a linear function giving the displacement error versus the force. That function was implemented into the data acquisition software, thus correcting displacement according to the compression force.	

Compression analysis. Heckel plots were calculated using the corrected upper punch displacement data (method "in die"). All parameters were calculated for each plot and are presented as an average value of 10 single plots. The slope of the linear compression part of each Heckel plot, as well as its reciprocal (yield pressure), was determined using a stepwise linear regression from 300 data points.14 R-values represent the maximum force ratio of lower to upper punch compression force (n510). The confidence interval at a level of 95% is given in parentheses.		

Figure 3:  R-values and ejection force results from tablet mixtures containing StarLac, or a physical mixture (PM) of 85% Flowlac and 15% corn starch, together with various proportions of active ingredient (AA = ascorbic acid; P = paracetamol). Single punch tablet press at 200 MPa. P-70% was externally lubricated.	

Tablet parameters. The mass of 20 tablets was measured using an AE 200 balance (Mettler-Toledo, Gießn, Germany). The crushing strength of 10 tablets was determined using a TBH-30 tester (Erweka GmbH, Heusenstamm, Germany) at a punch velocity of 1.00 mm/s. Friability and disintegration time were both determined according to the European Pharmacopoeia (4th edition). The friability was determined in a PTF tester (Pharma Test, Hainburg, Germany) and the disintegration time was measured in demineralized water using a PTZ 1 tester (Pharma Test). All tests on the tablets were performed 24 h and 7 days after compression. 	

Powder properties. All the SL batches investigated showed excellent flow properties, because of the coarse (d[0.5]=156-195.5 Âµm) sphere-shaped particles (Figure 1) and the narrow particle size distribution (Table I), where d is the diameter. The flow-through time was, therefore, always less than 7 s for all the SL batches, which resulted in angles of repose between 26Â° and 27Â°. By contrast to the SL results, PMs of 85% spray dried lactose and 15% native corn starch did not flow at all; the entire test according to DIN 59 316 failed. Although spray dried lactose particles exhibited similar morphology (Figure 1) and flow properties (Table I) as SL, the fine starch particles (d[0.5]=13.3 Âµm) prevented the mixture of spray dried lactose and corn starch from flowing. Because a non-flowing filler was inappropriate to use in a comparison for direct compression excipients, 0.3% Aerosil 200 was added to the PM, thus a moderate flow-through time (10.5 s) for the mixture was obtained. The bimodal particle size distribution of the PM (d[0.5]=13.3 Âµm for corn starch and 132.8 Âµm for spray dried lactose) consequently led to higher bulk and tapped densities of the PM compared with SL, as a result of the ability of small starch particles to travel into the large inter-particular pores of coarse spray dried lactose.		

Compaction properties. The resistance against deformation or apparent yield pressure, which is calculated as the reciprocal of the slope (k) of the linear part of the Heckel plot15 (Figure 2), decreased in the order FL.PM.SL.corn starch (Table I). However, the low value for the yield pressure of native corn starch of 58.3 MPa needs some extra explanation, related to the Heckel equation (Equation 1). The closer the apparent density approximates the true density of the powder (the porosity gravitates towards 0), the higher the value for ln(1/[12D]). When the value of the apparent density is higher than the measured true density, the entire model becomes invalid. However, true densities are not necessarily pressure independent18 and materials such as starch may exhibit their true density even at the pressures applied during the tabletting experiment. Consequently, calculating a Heckel plot for starch employing a constant true density leads to increasingly false higher values for ln(1/[12D]) at increasing compression pressures. Thus, in the absence of a "true" true density, and considering the density of starch determined using an air comparison pycnometer is definitely not the true one, the slope of the starch Heckel plots are too steep. Nevertheless, even if the determined slope is not precise, the Heckel plot of corn starch indicates a highly plastic deforming material with an increasing elastic proportion at high compression pressure, which is in accordance with other reports.12 The lower yield pressure of SL compared with the PM (Figure 2 and Table I) shows that starch, which is bound within the SL compound (Figure 1), improved the deformation properties to a greater extent than starch, which was just physically mixed with large spray dried lactose particles. Small starch particles may travel into the pores of the spray dried lactose lattice, whereas starch particles within the compound remain in place.	

Additionally, the proportion of small starch particles in the PM decreased lubricant efficiency because of a higher specific surface area. Hence, R-values of SL compacts were higher than those of the PM (Table I) at a constant lubricant concentration of 0.5% magnesium stearate, indicating a more even stress distribution within the SL tablets. This trend is highlighted when various proportions of active ingredient (AA or P) were added (Table II, Figure 3). The PM R-values decreased with increasing proportions of active ingredient, and the ejection force increased significantly. At proportions of 70% active ingredient, the die had to be externally lubricated to reduce the ejection force on the single punch tablet press and hence led to lower values. By contrast, mixtures of SL with either AA or P in any proportion exhibited constant R-values of approximately 0.925 and ejection forces of less than half the values of the mixtures compressed with PM as a filler-binder (Figure 3).		

Tablets of pure filler-binder. SL tablets of sufficient crushing strength (50 N) could be compressed at a compaction force of 10 kN and greater (Figure 4). The 5 kN force level led to weak tablets for both SL and PM, which displayed crushing strength values of less than 20 N. The crushing strength for SL tablets increased almost linearly by increasing the compaction force and reached a value of 173 N at 25 kN. By contrast, tablets compressed with the PM started capping at compression forces exceeding 15 kN. This capping tendency was also apparent in the friability test, wherein the PM tablets compressed with a force exceeding 15 kN failed the test. However, PM tablets prepared at  compression forces of 10-20 kN exhibited acceptable friability values of just less than 0.2%. SL tablets did not tend to cap, either during the crushing strength test or the friability test, which the tablets passed with friability values of less than 0.02% for force levels between 10 and 25 kN (Figure 4). The reason for the better mechanical properties of the SL compacts was the enhanced compressibility of SL compared with the PM (Figure 2) as a result of the higher effectiveness of starch particles as subunits within the SL granules. Thus, starch particles could locate between the small a-lactose monohydrate crystals as a dry binder, increasing plastic deformation at the time the SL particles are fractioning (at an early stage of the compression). The good mechanical properties of the SL tablets did not affect disintegration. The disintegration time of SL tablets was approximately 30 s less than the tablets compressed with the PM, which varied between 2-3 min. The different batches of SL (1-3) did not show significant differences in their tablet properties.				

Tablets containing active ingredient. Mixtures of SL or PM containing ascorbic acid in proportions between 30-70% behaved poorly whilst being compressed into tablets. Together with SL, the main problem was capping, even though the type of AA used was designed for direct compression (96% ascorbic acid and 4% polyvinylpyrrolidone [PVP]). However, SL was a suitable filler for tablets containing 30-70% ascorbic acid (SLAA30-70). The only force level that could be used without capping was 10 kN. The resulting tablets displayed a tolerable crushing strength of 50 N, independent of the ascorbic acid proportion, but rather high friability (0.3-0.55%, after 24 h) (Figures 5-7), that increased at higher ascorbic acid proportions. Storing the tablets for one week resulted in reduced friability values of 0.10% for SLAA30, 0.16% for SLAA50 and 0.35% for SLAA70. Crushing strengths, however, remained unchanged. As a result of the increasing effect of hardening at higher SL proportions, the effect could be related to an alteration of amorphous lactose, which is used as a binder for the SL particles, towards a-lactose monohydrate.7 The disintegration time was always much shorter (2-6 min) than the 15 min required by the European Pharmacopoeia (Figures 5-7).	

Tabletting mixtures containing ascorbic acid and the PM were even harder to compress than the physical mixtures containing SL. Physical  mixtures that exceeded a proportion of 30% ascorbic acid immediately started sticking to the tooling and no further tabletting could be performed. Nevertheless, the PMAA30 showed decreased capping; hence force levels up to 20 kN could be applied, resulting in a crushing strength of 140 N, a friability of 0.17% and a disintegration time of 470 s (Figure 5). Repeating the tablet tests after 1 week, the same hardening effect as was seen for the SLAA tablets was displayed. Additionally, the crushing strength increased; thus values of 0.09% for the friability and 157 N for the crushing strength were determined.	

Tabletting mixtures containing the active ingredient P exhibited even more capping tendency when compressed with SL as a filler compared with mixtures containing ascorbic acid. As a result, only tablets compressed at 5 kN passed the friability test uncapped. However, they showed sufficient crushing strength (66 N) at low friability (0.1%) for SLP70 tablets. By contrast to the SL tablets with ascorbic acid, the tablets with paracetamol showed higher mechanical strength at increasing paracetamol proportions. Thus, the paracetamol for direct compression dominates the tablet properties. This result also explains the drastically increased capping compared with tablets of pure SL, as paracetamol is a well-known drug that causes capping.19 However, PM mixtures containing paracetamol had a reduced dependence of the paracetamol proportion on the compactability than SL mixtures: the PMP tablet properties were similar to the tablets of pure PM (Figures 4-7). That means shorter disintegration times, higher crushing strength and less friability compared with the SLP tablets (Figures 5-7).	

SL is a compound of excellent flow properties and enhanced compressibility compared with spray dried a-lactose monohydrate (FL). Binding small starch particles together with a-lactose monohydrate crystals into compound particles increased the starch-related effects of enhanced binding capacity and compressibility compared with a physical mixture of corn starch and spray dried lactose. Additionally, the small starch particles within the physical mixture prevented powder flow, thus 0.3% Aerosil had to be added to make a comparison of the two filler-binders possible.				

Ascorbic acid could be tabletted at all tested proportions up to 70% together with SL, whereas sticking prevented the compression of mixtures exceeding 30% AA together with the PM. 	

Paracetamol tabletted with SL proved to be a greater challenge. At a P level of 70%, tablets of good crushing strength and friability could be obtained. For P the PM had advantages of better mechanical tablet properties, particularly at P levels below 70%. Generally, all tablets hardened slightly during storage during the examined period of 1 week. None of the tabletting mixtures required the addition of a disintegrant, whereas tablets that contained SL exhibited a faster disintegration than tablets that contained the PM.	

Features

The tabletting properties of a new coprocessed excipient for direct compression were compared with a physical mixture of its components (separately and with drugs) and the individual constituents. The compaction properties were also investigated. Results indicated that the new excipient has excellent flow properties and demonstrates enhanced compressibility.

A Corn Starch/a-Lactose Monohydrate Compound as a Directly Compressible Excipient

As the introduction of new member states to the European Union edges closer, the debate in Brussels regarding healthcare for the enlarged Union becomes more heated. Health standards in new member states are predicted to increase, but the main worry for the usual protagonists is how will the European pharmaceutical industry fare?

This article is currently available in printed form only.


 To ensure access to all articles from every issue of 

It's the money that counts

The combined use of experimental fluid dynamics (EFD) and computational fluid dynamics (CFD) methods effectively identifies and addresses flow problems in the pharmaceutical industry.
Process performance an dreliability can be improved with application of this technology. This two-part article demonstrates the use of EFD/CFD with diverse examples including laboratory equipment (Part I) and production-scale equipment (Part II). Flow patterns, mixing structures and complex behaviour are all revealed with the combination of the EFD and CFD methods.

Flow problems, such as process performance and reliability, can be improved using a combination of experimental and computational methods.

Experimental and computational methods for understanding pharmaceutical flows, Part I, Laboratory scale devices

Traditional needle and syringe administration has several drawbacks. These include complexity of dose preparation, patient compliance, stability of the drug after reconstitution into its desired form and safety concerns - such as the handling and disposal of a used syringe. In response to safety considerations, several alternative drug delivery systems, such as powder injection, have been developed, together with techniques for the sustained release of drug formulations and the evaluation of new formulations for oral and injectable administration.		

It is possible to use animal experimentation to evaluate drug delivery systems; however, scientists are often uncertain how to apply the results to humans. For example, major differences between the respiratory tracts of humans and laboratory animals mean that inhaling the same amount of powder can result in dramatically different dosages. Given these issues, designers of drug delivery systems need a comprehensive design tool to enable them to evaluate different designs rapidly and economically. 	

Computational fluid dynamics (CFD) uses numerical methods to solve the equations that govern fluid flow. The basic premise of CFD is to divide the domain to be analysed into small volumes or elements (Figure 1). For each of these elements, a set of partial differential equations is solved; these approximate a solution for the flow to achieve the conservation of mass, momentum and energy for each volume or element. The three basic equations are solved simultaneously with any additional equations implemented in a particular model to obtain flow velocities, pressure and, hence, any derived quantities such as shear stress.				

The conservation of momentum, with F representing the convective flux and G representing the diffusive flux, is given in Equation 1. The conservation of mass is given in Equation 2.	

In addition to these model equations, the equations of turbulence are solved if they are applicable to the particular flow regime being analysed. From the flow domain, each cell is then analysed in relation to its neighbours in an iterative process until errors in the total conservation equation have been reduced to an acceptable level across the whole of the area being studied. 		

quick evaluation of various design and process parameters 

savings in work hours and other operating costs 

ideal conditions to be simulated and greater control of the physical process 

specific phenomena to be isolated for study.

Experiments only permit data to be extracted at a limited number of locations in the system by using measuring instruments, such as pressure and temperature probes, heat flux gauges and laser doppler velocity meters. CFD allows the analyst to examine a large number of locations in the region of interest and yields a comprehensive set of flow parameters for examination. The resulting design and parametric data, such as dose concentration and variation, particle dispersion process, as well as three-dimensional visualization of design performance, is invaluable. 

Cost savings can range from 10–30% depending on the type of device, its application, the complexity of the physical models and the target area (disease) for the medicine. Some of the current and possible applications of CFD in the development of drug delivery systems are reviewed below. 	

Inhalation technology is extensively used for treating lung diseases such as asthma, cystic fibrosis and emphysema. This method offers rapid and easy administration of drugs and the ability to administer low dosages. Most drug adsorption in the lung occurs across the alveolar epithelium.1 The rate of drug deposition depends on particle size and rate of inspiration.2 Dry powder inhalers and metered dose inhalers (MDIs) are two commonly available inhalers, with MDIs containing chlorofluorocarbon (CFC) gases being the most prescribed inhalation systems during the past few years. 	

Figure 1:  Fluid region of pipe flow discretized into a finite set of control volumes (mesh).				

However, MDIs have two main weaknesses: the actuation of the aerosol and the low dose concentration (typically 10% of the normal dose). Pharmaceutical companies started researching inhaler designs to improve ease of use and the efficiency of drug deposition (current efficiency is 30–35%3), and the reproducibility of the dosage. The research was accelerated because of the global drive towards reducing the level of CFC gas emissions. 	

Drug delivery systems normally involve the transportation of a secondary-phase material to the area being dosed. These secondary materials may be medicated particulates, liquid drops, a gaseous species or a mixture of all three. A variety of established multiphase modelling methodologies is available to the CFD practitioner to characterize a device for dose concentration, dose variation and the particle dispersion process. They can also be used to identify reasons for particle loss and review potential design improvements. 		

Figure 2:  Carrier particle trajectories and residence time in a dry powder inhaler.(4) 	

Figure 2 illustrates the residence time and trajectories for the medication particles in a transient simulation of an inhaler. Figure 3 illustrates the air flow in the inhaler.4 The simulation results were used to identify regions of high velocities, shear forces and circulation loops, particle residence times for larger and smaller diameter particles, and the separation of medication particles from carrier particles.	

CFD can also be used to evaluate other drug delivery systems such as needleless liquid and powder administration for the treatment of chronic problems and neurological diseases.5 A concentration profile of drug degradation rate can be evaluated using the values from diffusivity and reaction kinetics. In addition, these data are pertinent for regulatory authorities that seek information regarding bioavailability and bioequivalence for orally administered drug products.(6) 			

Figure 3:  Flow field across a dry powder inhaler.(4) 				

Complex oral cavity and oesophagus shapes can be discretized into a fine grid to model drug absorption and dissolution processes. Limitations of these approaches are the necessary simplification of the complex and varying geometry present in real life, the boundary conditions for the fluid and the porous nature of the oral cavity. 				

Drugs intended to be delivered via the respiratory tract are frequently micronized to create microparticles in the respirable range. These particles have a high surface area and high charges forming an unstable cohesive system. This means the particles could adhere to any surface on the way to the targeted region of the respiratory system. 			

Figure 4:  Mesh over nasal passages. Courtesy of TGS, Inc.	

CFD can be used to trace the trajectories of a medication particle through the nasal passage and lung. Figure 4 illustrates the nasal passage, to identify the region of high air velocity. CFD can also be used in model die design, extrusion, blow moulding, device filling and debubbling, and other elements of device manufacturing.	

CFD can be used for a range of applications, including the design of drug delivery systems and drug absorption. Accuracy of the CFD simulations will depend largely on the geometry, the physical models that are employed and the boundary conditions used in the simulation. Although CFD does not eliminate the necessity of experimental work, it provides quick and cost-effective evaluations of new designs, new molecules, the delivery process and the performance sensitivity of the drug delivery systems to changes in the parameters.	

Interest in more advanced drug delivery systems has increased, with an acceleration in the discovery and development of novel therapeutic macromolecules for targeted applications. Computational fluid dynamics is a design tool that allows producers of these and other products to evaluate different models rapidly and cost-effectively.

Design, Development and Optimization Using Computational Fluid Dynamics

Accepted practices and operating procedures have been developed in the bio/pharmaceutical industry for the design, production and installation of process piping so that process systems operate in a clean and sterile manner. It is also understood that the materials of construction must not affect the purity of the product. An ideal process line will be designed to give the smoothest possible flow of product. To achieve this, the internal bore of the weld should be consistent throughout the entire system. To avoid potential contamination areas, known as 'bug traps,' the bore of the system should be smooth and free from marks and crevices. Bug traps are small voids in the material surface where product can stick and deteriorate during a period of time (Figure 1). By designing a system with predominantly welded joints, the number of problem areas is minimized; mechanical joints, by their nature, have areas that cause potential bug traps and should, therefore, be kept to a minimum.	

If the line is not installed to a satisfactory standard, it will not last long and may contaminate the product travelling through it. The most important consideration when installing a stainless steel process line is to ensure that the steel will not degrade during installation, particularly during the welding operation.	

Figure 1:  A potential bug trap on a welded tube.				

Stainless steels contain chromium and nickel, which give the material its stainless properties. If the steel is not adequately protected from oxygen during the welding process, it will dramatically degrade. Another serious risk to the product comes from pipelines that have not been thoroughly cleaned prior to use; any debris left in the system will contaminate the product flowing through it.		

Figure 2:  Egmo high quality fittings (Aalco, Bio-Pharm Division).	

The main points to consider when creating a clean process pipeline are as follows:				

post-installation cleaning and treatment of pipeline.

To achieve a clean process pipeline, it is important to select high quality tubes and fittings from a reliable source that can provide mill certification, chemical analysis and guarantee the quality of the tube finish (Figure 2). Tubes should be supplied from a controlled warehouse in protective packaging to eliminate any chance of tube contamination prior to installation. The following should be considered when selecting tubes and fittings:				

Sulfur control. Stainless steels contain a trace amount of sulfur, which is not often considered a problem. In practice, however, sulfur content can have a major effect on the ability to weld the material. A difference of 0.001% in sulfur content can cause a significant difference to the weld profile; this is because of the Maragoni effect, which changes the surface tension temperature coefficient; that is, the heat movement in materials changes with different sulfur content.		

Figure 3: Orbital cutting machine (Otto Tool, Inc.).	

Surface finish. It is important that tubes and fittings have a smooth internal surface and that the supplier can guarantee this, particularly in welded tubes where the internal surface will be affected in the welded area.	

Manufacturer specifications. There are many specifications to which tubing can be manufactured, but the most important regulation is ASME-BPE 1997. It is important that tubes should be manufactured with a statistical process control (SPC) production line to give guaranteed quality tubes. 		

Figure 4:  An orbital welding system (Hewes Engineering).	

Batch control. All tubes and fittings should be ordered by lot to ensure they are from the same batch of material. This ensures that all materials have the same sulfur content and will eliminate any problems with weld quality.	

Traceability. All tubes and fittings should be traceable to their original source. This allows any problems encountered during fabrication, installation or service to be tracked back, and problems quickly identified and rectified.		

Figure 5:  Manually welded process tube (photo courtesy of Delstron Systems).	

During fabrication, it is important to follow recommended fabrication techniques to minimize contamination of the tube bore and material degradation. Each stage of pipeline  fabrication has the potential to cause problems in the final process line. The main fabrication techniques used are cutting, tacking and welding.					

Cutting. Lengths of tube will need to be cut to length during every stage of installation. To maintain the stainless properties, it is important to control the heat input during the cutting process; this would usually require an orbital cutting machine (Figure 3) with a well conditioned cutting blade. The use of other methods, or indeed an orbital cutting machine with a blunt blade, allows heat to build up during the cutting process and can degrade the material being cut. Overheating during cutting can be identified by the formation of oxides on the surface of the tube near to the cutting area.	

Tacking. When tubes are cut to length, they will need to be tacked to fittings or other pipes to produce the required pipe run. The atmosphere in the internal bore should have an oxygen content of less than 500 ppm during tacking and welding so that the material does not degrade. The tack should also be small, clean and must not fully penetrate the wall of the material. Dirty tacks, which will be dark blue or black and in extreme cases may have a crusty surface, will also affect the quality of the final weld.		

Figure 6:  Orbital welded process tube (Delstron Systems).	

Welding. Welding has the greatest degrading effect on the properties of the material and is one of the most complex and critical operations in the fabrication process.	

The most suitable process for welding stainless steel tubes is tungsten inert gas (TIG) welding. This process uses a non-consumable tungsten electrode that is shielded from the atmosphere by an inert gas, usually argon. This process provides a controlled heat source, which can be enhanced by a pulsed arc. 		

Figure 7:  Fully enclosed weld head located on a process tube.				

To attain high quality weld profiles and to minimize material degradation, a consistent weld speed must be used. The most effective way to control weld speed is to use an automatic orbital welding system (Figure 4). Figures 5 and 6 show the difference between a weld made by a manual pipe welder and a weld made with an orbital welding system. The manual weld is significantly more oxidized and less consistent than the automatic weld.	

It is now considered standard practice to use orbital welding equipment with fully enclosed weld heads for pharmaceutical process line installation. Fully enclosed weld heads envelop the entire external surface of the weld area (Figure 7), which is then filled with an inert gas, usually argon, before the weld cycle starts. This guarantees a clean outer surface and should the joint open during the welding cycle, it is inert gas and not air that will be drawn into the internal bore.		

Figure 8:  Welded process tube with no back purge (Delstron Systems).	

As well as protecting the external surface of the material, it is also essential that the internal surface is protected. If stainless steel is not protected from oxygen during the welding process, then the chromium and carbon in the material will be oxidized and form chromium carbides in a process known as coking (Figure 8). Once the chromium has been oxidized, the steel loses its corrosion resistance.	

Oxidation can be avoided by filling the internal bore of the tube with an inert gas; this method is called back purging and a range of systems is available that localize the area at which the gas is applied, thus avoiding filling large sections of tube. These systems use a low volume of gas and reduce the time taken to get the oxygen level of the back purge to a suitably low level (Figure 9).	

All welding should be performed to the highest standard and each weld must be tested and certified to that standard. If orbital welding equipment is used, the equipment and operators will also have to be certified. An approved weld procedure should be submitted with the certification and this weld procedure should be adhered to at all times. Modern welding equipment has the facility to store the weld procedure, limiting operators' ability to change the parameters. Test coupons should be submitted and approved prior to the welder commencing work on the installation. Further test coupons should be submitted to ensure that quality is continually being maintained.	

Figure 9:  Principle of an internal purge system (courtesy of Huntingdon Fusion Techniques).				

All welds should be logged on the system diagram with a record of when welds were done and by whom. This helps if problems occur on a particular weld. Ideally, a data log will be kept for each weld; modern welding equipment can print or store this information and each log file will include date, time, weld number, welding current, travel speed and average arc voltage, and will record any deviations from programmed parameters.	

The internal bore of the tube should also be filled with a suitable inert gas because the oxygen level should be less than 500 ppm before welding commences. Modern welding equipment can be connected to an oxygen analyser to prevent the weld operation beginning until the the oxygen has reached the required minimum level.	

Post-installation cleaning and pipeline treatment

Because of fabrication, debris will remain in the system and this will have to be flushed out. In addition, all weld areas should be passivated as the metal crystalline configuration at the surface changes during cooling, resulting in depletion of the chromium oxide layer.				

For post-installation treatment, the system is flushed with water to remove debris and a detergent is then circulated to remove oil, grease and organic residues. The system is then flushed again with water to remove all traces of the detergent. Valve bodies, diaphragms and other areas where debris may have accumulated are swabbed and then inspected under UV and white light to ensure that organic residues have been removed.	

Nitric acid is then circulated to dissolve iron oxide and any other contaminants that have adhered to the steel. This is particularly important at weld areas where the metal is relatively rough and there is a possibility of iron oxide residue. Nitric acid is an oxyacid, which oxidizes the chromium layer at the metal surface to form resistant chromium oxide.	

Even with electropolished biotubes, it is essential that the system is passivated to enrich the chromium oxide layer at the metal surface of the weld. Where nitric acid is unavailable or must be avoided because of safety or environmental reasons, other chemicals, such as mixtures of citric acids and organic chelants, can be used.The system is finally flushed with demineralized water to remove all traces of nitric acid. 	

When planning the installation of a new stainless steel process pipeline, it is important to understand the significance of using the best possible materials and techniques to ensure it is clean when installed. This article examines the factors that must be considered when planning and constructing a new process line.

Installing Clean Process Lines

Process validation has been widely discussed and criticized by the pharmaceutical industry during the past 20-30 years. Regulatory guidelines in the US and Europe have slowly been modernized, and revisions will certainly continue; in autumn 2001 for example, new guidelines for process validation came into force in Europe.(1,2) The unofficial debate regarding process validation has always been quite lively, although ostensibly negative. The advantages and disadvantages of process validation have never been systematically evaluated, and validation is frequently performed without a real understanding of the work involved. The challenge for the pharmaceutical industry is to streamline and/or simplify validation without sacrificing product quality.(3) 	

To successfully fulfil the challenge, those practising validation need to be aware of the best way to perform validations and the real aim of these procedures. This review examines how pharmaceutical process validation has evolved, the attitudes towards it and how it has been accepted by the industry. 	

The three most often referred to definitions of pharmaceutical process validation are those presented by the European Agency for the Evaluation of Medicinal Products (EMEA), the US Food and Drug Administration (FDA) and the Pharmaceutical Inspection Co-operation Scheme (PIC/S). The latest versions of their definitions are described in the sidebar "Current definitions of pharmaceutical process validation."	

Current definitions of pharmaceutical process validation				

The three definitions are very similar; the only difference is that FDA expresses a minor uncertainty of the concept, despite the efforts of validation, by stating that process validation only provides a high degree of (not absolute) assurance that the process will produce the intended product.	

This amendment by FDA is, however, essential. Even when approaching process validation as scientifically as possible, by incorporating elements of validation during each stage of product development(6–8) and evaluating the influence of different process parameters on the final product with statistical principles,(3,9) the possibility of erroneous results still remains.	

The definition and meaning of pharmaceutical process validation has never really been clear. The expression "pharmaceutical process validation" is said to have first been used approximately 30 years ago by the US authorities,(10) and since then, people have demanded more precise and understandable definitions. But the more the authorities have tried to explain their meanings and provide guidelines, the greater the ensuing confusion and misunderstanding.11-14 John R. Sharp,(14) who at the time was the principal medicines inspector at the Department of Health in the UK, said that such definitions only confused those who already had an idea of what pharmaceutical process validation is and would not enlighten those who did not. He said that process validation is nothing more than common sense - it is simply proving that a process does what it is designed to do.(14,15)				

There is also much confusion as to what constitutes process validation and what does not. Pharmaceutical process validation has always been understood in one of two ways - either as the total validation activity in a pharmaceutical manufacturing site from development qualification of the equipment to the final validation of three consequent batches of the final product; or as the final production-scale validation of a pharmaceutical preparation only. This discrepancy is not surprising because interpretations vary between the authorities(1,4) and the textbooks.(10,16) 	

Concept. The idea of process validation is not new and is common in many different fields of life; one can find the need for process validation in almost any kind of process, for example, building a spaceship or treating illnesses. It is important to be sure that these processes do what they are expected to do, otherwise the consequences could be serious.				

Sharp also interpreted pharmaceutical process validation simply as a step in developing the maintenance of the quality of manufactured medicines.14 Process validation has been included in the first interpretations of good manufacturing practice (GMP) to ensure that medicines are safe and have the identity and strength they are supposed to have. 		

Table I:  Regulatory history of process validation in the US and Europe.	

US regulations. Bernard T. Loftus,(10)  a former director of FDA, previously described how the principles of process validation evolved in the US from the first current good manufacturing practice (cGMP) in 1963 to the first Guideline on General Principles of Process Validation in 1987.4 Prior to 1963, the only way for FDA to prove that a process had not done what it was designed to do was to take samples from the final product, analyse them and show deviations from the specifications.	

From 1963, the law stated that a pharmaceutical manufacturer had to follow cGMP regulations whilst FDA received authorization to inspect manufacturing facilities. This was a direct consequence of a series of accidents in which people were injured and even killed. 	

These incidents led to the evaluation of manufacturing processes, but it still took a long time before the authorities could point out clear and serious production faults and demand better procedures and processes.	

Things began to change during the late 1960s and early 1970s when new types of incidents, such as poorly mixed, highly potent tablets and insufficient sterilization procedures for large volume parenterals caused serious patient disorders.(10,15) Many speeches pointing out the need for process validation were made by US authorities and the expression "validated manufacturing process" was finally defined in the Drug Process Inspections Compliance Program in 1978. The more precise definition and adjustment of the concept for process validation was published in the Guideline on General Principles of Process Validation in 1987 and, since then, exhaustive process inspections have been routinely performed by FDA. It took a long time before process validation was directly named in US cGMP regulations. To this day, process validation is not clearly defined, although a new version of the US cGMP rules, including more precise definitions of process validation, has been in preparation for a couple of years.(17) 	

European regulations. At the same time, a similar process was taking place in Europe (Table I), particularly in the UK. In 1968, new guidelines were introduced in the UK governing the sale and distribution of medicinal products, and in 1971 the first edition of Orange Guide, a guide to GMP, was published by the Pharmaceutical Inspection Convention (PIC). This guide already mentioned validation as being beneficial, but it was only in the 1977 and 1983 editions that validation practices in sterilization, aseptic processing and non-sterile processes were included. 	

In 1989, the first edition of the European Guide to GMP superseded all national guidelines within the European Union (EU).(16) This guide from the European Commission, together with the EMEA guidelines, has served as a model for all European countries regardless of whether or not they belong to the EU. As in the US, European process validation for sterile and aseptic processes was accepted and put into practice soon after the first regulatory guidelines demanded it. Yet, up until recently, process validation for other non-sterile processes has been introduced much more slowly and in limited amounts. More precise guidelines and definitions were also demanded for the European medicinal industry before the systematic validation of all pharmaceutical manufacturing processes was regarded as necessary. 	

The preparation of more detailed instructions went on between 1997-2001,18 and in September 2001, Annex 15 to the EU Guide to GMP came into operation titled Qualification and Validation2 and the Note for Guidance on Process Validation for marketing authorization applicants was introduced.1 It is worth mentioning that these instructions were issued at the same time as the GMP Annex and Note for Guidance on Parametric Release.(19,20) Parametric release is important because it adds more value to process validation.	

At the beginning of the 1980s, US authorities published several guidance speeches relating to process validation.(12,13,15,21) Additionally, Loftus defined process validation as nothing more than proving that a process works; that is, common sense. He also underlined that FDA emphasized "what to do" and left the "how to do it" to the manufacturer.(15) 				

He stated that the regulatory body should insist on documented validation and was preventing initiatives and progress with meaningless regulatory controls.	

Ted E. Byers, a consultant and former FDA inspector, explained how to best organize validations and how to maximize their benefits.(21) He introduced the team approach as a model for process validation organization and asked for better co-operation between FDA and the industry when planning individual validation cases. In conjunction with the preparation of the US Guideline on General Principles of Process Validation, Edmund M. Fry, director of the Parenteral Drug Association (PDA), attempted to clarify some points of the draft guideline that had raised questions from the industry.(12,13) His main message was that guidelines are not mandatory, but simply explain one approach to accomplishing the objective of the law. However, FDA comments suggested that it never wanted to define too many details of process validation, but that the industry itself asked for them. 	

In both the US and Europe, the authorities underlined how beneficial it would be if the industry made an effort to provide constructive comments on the guidelines before they were published.(12,18) As Loftus from FDA stated: "There is clearly a public policy need that we regulate the industry, but not that we manage it."(15) 				

James P. Agalloco was more positive about process validation, highlighting the business aspects and listing the numerous benefits to the manufacturer.(22) However, there was one person who attacked FDA's principles of process validation - John Sharp. He initiated a discussion on process validation just before the US guideline came into operation,(14) although the majority of the discussion took place between 1993-1995.(23-29) Sharp believed that validation in the US was in danger of becoming a cult-like activity where the original purpose had already been forgotten by spiralling onwards and upwards on a stairway of increasingly demanding standards. Not only did he blame FDA for this, but also found the pharmaceutical industry guilty. Sharp claimed that the need for validation should be inversely proportional to the adequacy of product design, raw material control and other quality control aspects, and directly proportional to the hazard to the consumer that a process failure might present.	

Representing European authorities, Sharp found the European method of validating more reasonable than the US approach. He also emphasized the need for an improved level of interpretive consistency among US authorities. According to Sharp, this consistency was an area where the British Medicines Inspectorate had invested a lot of time.	

Michael H. Anisfeld,(27) who worked at Interpharm Consulting (Buffalo Grove, Illinois, USA), agreed with Sharp and emphasized the importance of the patient perspective. He asked: "How much validation can the patient and the world afford?" and made comparisons of the expenditures on validations and their influence on prices of medicinal products with annual incomes of people in developing countries in need of those medicines. He also highlighted the exaggeration of US regulations and their consequences; some companies were not manufacturing products aimed at the US market in accordance with regulations because they did not want to go through costly FDA validation programmes for new facilities.	

James Akers28 (Akers, Kennedy & Associates, Cary, North Carolina, USA) agreed with both Sharp and Anisfeld. According to Akers, validation was becoming a negative force within the pharmaceutical industry and the time had come for some fundamental change in the way the industry approached the subject. He wanted to go back to the basics of validation and forget all the bureaucratic exercises with irrelevant definitions and terms. Furthermore, Akers observed that process validation served as an excuse to do nothing new and, therefore, hindered the creativity of scientists and engineers.	

R.A. Nash, associate professor of St John's University, Jamaica, did not agree with Sharp. Nash had been working in the field of validation since the early 1980s(23,31) and found Sharp's opinions to be personal attacks; he consequently wrote a heated response.26 However, in 1996,3 Nash also asked for process validation to be streamlined and simplified.				

Although Gene K. Estes et al.32 had established that process validation emerged primarily from the actions of the pharmaceutical industry, Agalloco suggested the idea first came from a regulatory body,(29,30) and was adopted as a regulatory burden rather than benefit. Agalloco found that validation could be defined as an application of total quality management (TQM), stating that it was the quality of the validation that is important and not the quantity.	

In 1995, Sharp criticized FDA-type validations and went as far as to suggest the term "validation" be abandoned altogether.(33,34) All that inspectors would then need to ask, he said, would be: "Prove to me that this process works, in that it achieves what it is intended to achieve."	

During the 1990s, Robert G. Kieffer, a consultant at RGK Consulting (Yonkers, New York, USA), challenged the need for risk-benefit analysis and the need for remembering to take into account the human element when planning process validations,(35,36) suggesting that the steps in a process involving human intervention would present the highest risks. Additionally, the sophisticated technology being used made more demands on people's knowledge and skills. Kieffer also anticipated new trends in pharmaceutical manufacturing - the pressure to do more with less, improve quality with lower costs, the globalization of the pharmaceutical industry and an increase in contract manufacturing. These trends are currently guiding the work of industry and also the work of regulators.(37) Kieffer also suggested that the basic ideas of GMP should be based not on quality control but on total quality, which meant the empowerment of all staff. Another consultant, Keith Powell-Evans, stressed the need to educate individuals involved in the validation process.(38)	

In Europe, the preparation of the new guidelines on process validation began in the late 1990s. The European authorities explained the new proposals,(18,39,40) consultants and pharmaceutical schools introduced their theories about how to best survive the validation effort41-43 and positive experiences from the field were presented by the industry.(44,45) 				

S. Roman44 (Laboratoire Glaxo Wellcome, France) found validation to be one of the main tools for quality assurance, resulting in reduced production and control costs. M.J. Girault45 (Marion Merrell Bourgoin SA, France) presented calculations showing that the cost of validation compared with the total costs of the plant was approximately 1.2%, and compared with the total costs of the quality department was 7.8%. 	

D. Caubel from the French Agency for Medicines evaluated the meaning of the guidelines and stated that they should serve mainly as a framework for process validation.39 They should give guidance for reasonable documentation practice but they should only give detailed instructions in particularly sensitive cases, such as cross-contamination prevention and aseptic preparations. 	

Many recent comments go into specific details of validation and/or overall quality assurance, warning regulators of giving too precise instructions in their guidelines.(46,47) 	

A more open approach from the regulatory side is required and a need to aim ultimately for parametric release for all dosage forms is emphasized.(37,48) Interesting additions to the discussion are comments from the less-developed countries where the concept of validation is new.(49)	

The debate regarding pharmaceutical process validation began in the US during the late 1970s and the first definitions of process validation were introduced at the beginning of the 1980s.(11-15, 21, 31) The discussion continued throughout the 1980s and during the preparation of US guidelines.(12-14, 22) However, the most critical comments were only published after the guidelines had come into operation.(23-28) 				

Comments on process validation emerged in European literature later than in the US and most of them were positive and accepting in their nature, apart from those by Sharp. This remained the case between 1997-2001 during the preparation of European guidelines on process validation.(38-45)	

The authorities, the pharmaceutical industry, consultants and pharmaceutical schools have all contributed to the discussion. Many of them have been critical of some aspects of process validation but none have been totally against it. In fact, many constructive and economic ideas regarding process validation have been suggested.(29,30,33-38)	

A clear evolution of the discussion can be recognized. The first comments were from US authorities who tried to explain the demands from a regulatory point of view,(11-13,15,21) referring to comments that they had received from the industry.11-13,15 It is important to note that in these early regulatory speeches, the regulators often stressed that they would not like to write too precise instructions about process validation, because they found that the practice varies so much from case to case.12,13,15 But, according to their references, it was the industry that wanted to have these detailed instructions and then started pointing out defects in them.	

FDA had gone into too much detail in its guidelines and that encouraged misunderstandings, negative attitudes and a substantial amount of unnecessary work in the industry. The European guidelines, although often one step behind, were better accepted. 	

Since the early 1990s, there has been a greater concentration on the risk-benefit analysis, statistical analysis and prioritizing work of process validation.(28,35-37) It appears that the pharmaceutical industry has now accepted that process validation is here to stay and that it is not simply a tool for quality assurance.(50,51) 	

Process validation has become part of strategic quality management and performance improvement;(52) all its benefits (to patients as well as manufacturers) have been recognized and from now on, the challenge is to find the right level of cost-effective validation without any risks to consumers.	

According to literature published during the last 30 years, pharmaceutical process validation has evolved. For most people in the industry, it originally appeared as a regulatory requirement and, hence, was often regarded as a burden.				

Despite originally being criticized, process validation is well accepted and regarded as part of total quality and process management. The US has been more critical of process validation; however, the lack of European literature on the subject does not reveal the true European attitude to process validation.	

1.	“Note for Guidance on Process Validation. CPMP/QWP/848/96; EMEA/CVMP/598/99” (EMEA, 7 Westferry Circus, Canary Wharf, London E14 4HB, UK, 2001).

2.	“Annex 15 to the EU Guide to Good Manufacturing Practice — Qualification and Validation” (European Commission, Rue de la Loi, Wetstraat 200, B-1049 Brussels, Belgium, 2001).

The renewal of European regulations regarding pharmaceutical process validation in autumn 2001 has been the subject of much discussion. The purpose of this article is to summarize and review the development of pharmaceutical process validation, and demonstrate how industry opinion regarding the concept has shifted from that of a regulatory burden to something driving total quality management and cost benefits.