Drug quality screening in developing countries: establishment of an appropriate laboratory in Swaziland.

Drug quality screening in developing countries: establishment of an appropriate laboratory in Swaziland* T.A. Kenyon,' A.S. Kenyon,2 & T. Sibiya3 A simple, low-cost, accurate thin-layer chromatography (TLC) method has been used to establish the first drug quality screening laboratory in Swaziland. For this purpose, office space at the central medical stores was first converted into a simple laboratory. Basic equipment, supplies, and materials were pur- chased, existing manpower was trained to perform accurately the TLC procedure, and a system was established for the qualitative/quantitative screening of selected drugs purchased by the Ministry of Health prior to their distribution to user facilities. The TLC method described can be used to set up simi- lar low-cost, drug quality screening laboratories in other developing countries where analytical chemistry expertise is lacking, resources are scarce, and sophisticated analytical laboratories to assess drug qual- ity are not available. Introduction Drugs form the comerstone of patient management for a wide range of illnesses that account for con- siderable worldwide morbidity and mortality, par- ticularly infectious diseases and chronic illnesses. Hundreds of millions of US$ are spent each year on pharmaceuticals, yet rarely in developing countries is the quality of drugs ever independently verified before they are administered, often as a life-saving measure, to patients. From an economic standpoint, the quality of purchased drugs may be more impor- tant now than ever. During the mid-1980s, economic constraints resulted in 20-25% cuts in drug spending by some African countries,a 80% of whose drug requirements are imported (1). The catastrophic results that may occur from counterfeit medicines are illustrated by the deaths of more than 100 Nigerian children who were given paracetamol containing an industrial solvent and by the 14 bum patients who died in India after receiving industrial rather than hospital grade glycerol (2). Clearly, countries with scarce resources must be in a position to verify the From: Project HOPE, Millwood, VA, USA. Regional Programme Director for Southern Africa, Project HOPE. Currently: Medical Director, Communicable Disease Divi- sion, Department of Health, Chicago, IL, USA. Requests for reprints should be sent to this author at the following address: 1318 Lindgate, Kirkwood, MO 63122, USA. 2 Chemist, U.S. Food and Drug Administration, Division of Drug Analysis, St. Louis, MO, USA. 3 Chief Pharmacist, Ministry of Health, Swaziland. a Jespersen E. Household responses to the economic crisis and its impact on social services in the 1980s. Unpublished UNICEF document May 1990. Reprint No. 5509 quality of the drugs they have purchased, in the interest of optimizing efforts to reduce excessive morbidity and mortality from treatable causes and to avoid needless deaths from counterfeit medicines. To the dispenser, pharmacist, or clinician the following question is vital: "Is this drug what it says it is, in the stated strength, and is it free of contami- nants?" This question can be answered by a variety of sophisticated analytical chemistry techniques. However, these procedures require considerable investment in terms of capital/recurrent costs and technical expertise, and are not widely available in developing countries. The U.S. Food and Drug Administration (FDA) has recently developed a modified thin-layer chromatography (TLC) method (3, 5) (Flinn, P.E. et al., personal communication, 1992) to screen drugs in developing countries (3). TLC is inexpensive and a trained technician requires only basic chemical and mathematical skills to per- form it. The FDA has recently demonstrated that the modified TLC method can reliably determine wheth- er or not a given drug is the same compound as a ref- erence standard (qualitative determination) and whether or not its content is within 85-115% of the standard (semiquantitative determination) (4). The U.S. Pharmacopeia generally requires individual dos- age forms of drugs to be within this range of declared content. The modified TLC procedure involves the sim- ple steps outlined below: Preparation of two solutions of the reference standard drug (85% and 115% of the expected sample concentration, resp.) and a solution of the sample (the desired concentration representing 100% of the declared quantity) in the specific solvent required for the drug being tested. Bulletin of the World Health Organization, 1994, 72 (4): 615-620 © World Health Organization 1994 615 T.A. Kenyon et al. - Spotting the sample aliquot between the spots of the two reference aliquots on the TLC plate fol- lowed by a 5-minute drying period. Insertion of the spotted TLC plate into the devel- opment solution, with subsequent migration of the solvent/drug spot along the plate. Removal of the TLC plate from the development solution and allowing the plate to dry. - Since the resultant spots are not visible under white light, they are visualized either using 254-nm radiation from an ultraviolet (UV) source or by dipping them in an iodine solution for drugs whose spots do not visualize well in UV or if an electricity supply is not available (4). Interpretation of the results As shown in Fig. 1, migration of the three spots the same distance along the TLC plate indicates that the sample drug and the reference drug contain the same compound (qualitative assessment). If the intensity of the centre sample spot lies between that of the two reference drug spots, the content of the active ingre- dient in the sample drug is 85-115% of that of the reference. For example, the procedure would give reassurance that a declared 250-mg capsule of ampi- cillin is indeed this substance and that it contains 212-288 mg of ampicillin if its spot intensity lies between that of the two reference drug spots. Migra- tion of the sample spot a different distance on the TLC plate than the standard spots indicates that the sample and the standard drugs are not the same com- pound. A sample spot that migrates the same distan- Fig. 1. Appearance of spots and their location on the thin-layer chromatography (TLC) plate under various circumstances. (A) Sample meets reference; (B) sample lies below reference; and (C) sample is not the same drug as reference. A B C spot migmtion Q * * ~ visualization spot placement x x x x x x x x x and SI D S2 S1 D S2 SI D S2development - S -85%of S2= 115%of D = Drug being reference reference analysed ce as the standard spots, but is lower in intensity than the 85% standard spot would indicate that the sample contains the correct drug but that its content is inade- quate. Details of the modified TLC procedure have been described elsewhere (3, 4). Swaziland is a small country in southern Africa which relies completely on imported drugs for use in its health system. In response to WHO's recommen- dation that Member States develop a drug quality assessment capability (4), Swaziland planned to establish a drug quality control laboratory, but was unable to set up such a facility based on high tech- nology equipment and highly trained manpower. The present article reports on the establishment, in early 1992, of the first drug quality screening laboratory based on the FDA-modified TLC screening method at the Ministry of Health's central medical stores in Swaziland. Methods Office renovation In the central medical stores an office measuring approximately 3 m x 5 m was converted into a small laboratory, including the addition of electrical out- lets, a laboratory sink, shelves, cabinets, and labora- tory bench at a cost of approximately US$ 1600. Equipment, supplies, and materials The essential laboratory items and their approximate cost are shown in Table 1. The initial capital ex- penditure was approximately US$ 5000, but will of course vary depending on local prices and volume of purchases. Saturation pads and micropipettes are potentially reusable if they are carefully dried or cleaned. The total cost per test, including the TLC plate, saturation pad, micropipette, chemical solvent, and reference standard is approximately US$ 0.50. This recurrent cost, i.e., variable cost, should remain relatively constant within a given laboratory, regard- less of output. Initially, the FDA provided the drug standards used to establish the laboratory, but these may be purchased from various sources.b The FDA also supplied the simple holder assembly for the TLC plate during the development stage of the pro- cess, but similar holders can be made using simple household tools from cheap and readily available materials such as Plexiglas, thin sheet aluminium, b For example: U.S. Pharmacopeial Convention, Inc., 12601 Twinbrook Parkway, Rockville, MD 20852, USA; or WHO Collab- orating Centre for Chemical Reference Substances, APOTEKS- BOLAGET AB Centrallaboratoriet, S-105 14 Stockholm, Sweden. 616 WHO Bulletin OMS. Vol 72 1994 Drug quality screening in Swaziland Table 1. Essential drug quality screening laboratory supplies, materials, and equipment for the thin-layer chromatography (TLC) procedure Approximate cost (US$) Capital expenditures Electronic analytical balance (weighing to ± 0.1 mg) Fume hood with explosion-proof fan motor Ultraviolet light source Polyethylene bag sealer Volumetric flasks (10, 25, 50 ml; 2 of each) Volumetric pipettes (1, 2, 5, 10, 25 ml) or Syringes, 1 ml, tuberculin (5, 10, 25 ml) Stoppered graduated cylinders (10, 25, 50, 100 ml) Pasteur pipettes, 2.25-cm long, with dropper bulbs Small plastic vials, with snap caps (30-mm high, 1.5 ml) Beakers (200 ml) Roll of flat polyethylene tubing (8-cm wide, 0.006 gauge) Small spatulas (18 cm with pointed tip) Recurrent expenditures 2 000-4 000 2 000 50 260 0.02 each 0.50 each 4.50 each 50.00 per roll TLC platesa 0.33 per analysis Saturation padsb 0.13 per analysis Microcapillary pipettes (3 pl) Acetone Ammonium hydroxide (500 ml) Ethyl acetate Chloroform 95% (v/v) Ethanol Ethyl acetate Formic acid (500 ml) Glacial acetic acid (500 ml) Hydrochloric acid (500 ml) Iodine crystals Methanol Methylene chloride Potassium iodide Toluene Distilled water Reference standards, primary 90.00 each a Plastic, silica-coated, fluorescent (X = 254 nm), 20 cm x 20 cm. b For example, Whatman No. 3 filters, 20 cm x 20 cm. flat polyethylene tubing, plastic report binders, wire coat-hangers, and a few screws. Alternatively, a stur- dy manufactured version of the apparatus can now be purchased for approximately US$ 50. All other items were bought through a local laboratory supply wholesaler who served other hospital or chemistry laboratories in the country. Although the method can potentially be used in areas without electricity, the electrical supply in Swaziland is reliable and an elec- tronic analytical balance measuring to ± 0.1 mg was used to weigh accurately both the reference and sample drugs. A fume hood is desirable, but we used instead a fan and an open window during initial implementation in the summer months until a hood could be purchased and delivered. Rather than buying a UV viewing cabinet, we constructed one out of black painted cardboard. Alternative analytical testing procedures for drugs could cost about US$ 50 000 for the equip- ment alone. The recurrent expenditures would also be considerably higher because of the need for more expensive high-purity solvents, use per test of approximately 20-50 times the volume of solvent required for each TLC test and of expensive dispos- able items. The salary of a more highly trained tech- nician would also need to be taken into considera- tion. In Swaziland, no additional fixed costs for personnel were necessary, and with TLC the only equipment that requires maintenance is the analytical balance. Manpower and training At the central medical stores four local staff, inclu- ding the senior pharmacist, one drug dispenser, and two trainee drug dispensers underwent a four-week period of on-the-job training in TLC procedures, with technical assistance from the FDA Division of Drug Analysis. The senior pharmacist and the chief pharmacist also participated in the training as their time permitted. Results The TLC procedure was based on FDA methodol- ogy, but was modified slightly to adapt to local con- ditions and was incorporated into the development of a local laboratory procedures manual during the training process. During the first month of training, 100 analyses were performed by the primary labora- tory technician on the 20 substances listed in Table 2. In the ensuing 6 months, the technician independent- ly performed 45 tests on consignments of 23 differ- ent drugs and developed new TLC procedures for the following additional drugs: trifluoroperazine, phe- noxymethylpenicillin, furosemide, Triple Sulfa (sul- facetamide + sulfabenzamide + sulfathiazole), pro- caine penicillin, propranolol, hydrochlorothiazide, and folic acid. The TLC assembly allows two analyses to be run concurrently and each test takes approximately 1-2 hours. In contrast, much longer times are required to complete a test using other ana- WHO Bulletin OMS. Vol 72 1994 617 T.A. Kenyon et al. Table 2. List of substances analysed using the thin- layer chromatography (TLC) procedure in Swaziland Acetylsalicylic acid Aminophylline Amoxicillin Ampicillin Chloramphenicol Chloroquine phosphate Cloxacillin Digoxin Erythromycin stearate Indomethacin Mebendazole Methyldopa Metronidazole Paracetamol Phenytoin Rifampicin Salbutamol Sulfamethoxazole + trimethoprim Tetracycline lytical methods. The policy developed at the central medical store is to analyse a randomly selected sample from each consignment received for which there is a laboratory quality screening procedure and a reference available prior to distributing the drug to user units. If a sample does not fall within 85-115% of the reference drug, as specified by the U.S. Phar- macopeia, the test is repeated. Should the sample continue to fail the screening procedure, distribution is witheld and a specimen is forwarded to the regional drug control laboratory in Zimbabwe for confirmatory analysis. Upon completion of training, the local techni- cian and the FDA technician performed equally well on five independently run test analyses (4). Only one of the drugs analysed during the train- ing period failed to pass the screening process. How- ever, the same sample met the specifications upon being analysed using high-performance liquid chro- matography (HPLC) at the FDA. An investigation indicated that the drug did not pass the screening in Swaziland because it had not been dissolved in the correct solvent; it subsequently passed the TLC screening when the proper solvent was used. All drugs tested over a 6-month period after the labora- tory was set up passed TLC screening. It should also be mentioned that the central medical stores incorpo- rated the drug quality screening laboratory into its administrative structure by creating it as an intemal department with a head (the trained dispenser) who reports to the senior pharmacist. An operating budget for the laboratory was developed and included in the submission for the coming fiscal year. Discussion The drug quality screening laboratory that we have described required minimal capital and recurrent costs, and employed readily available manpower who reliably performed qualitative and semiquantitative screening of imported drugs using the modified TLC procedure. It should be emphasized that TLC is a screening procedure for drug quality and is not in- tended to replace quality control methods that may already be in place. Nor is it intended for use as a full-scale pharmacopoeial testing procedure in a regulatory or compliance sense. However, even well- equipped laboratories could employ TLC for the rapid screening of large numbers of samples before further testing. Flinn et al. have reported that there is a 92-99% correlation between TLC spot intensity and drug concentration and have demonstrated that the eye can detect differences of 10% in drug concentration in many cases (4). At the FDA, the sensitivity and specificity of the TLC procedure have both been estimated to be 98%. The trained technician in Swa- ziland and the FDA technician agreed on all five of the test analyses run; however, it may not be accu- rate to assume that the sensitivity and specificity obtained by the FDA will necessarily be achieved under field conditions. Assuming that there was only one false-positive TLC result (sample failed to meet the reference but was actually of reference quality), we estimated that the TLC's specificity (ability to detect non-substandard drugs) in the field was 99%. The sensitivity (ability to detect substandard drugs) could not be estimated since concurrent "gold stan- dard" analyses such as HPLC were not performed. Let us assume, however, that 5% of the drugs imported into a given country are actually of sub- standard quality, and that under field conditions the sensitivity and specificity of TLC are both 98%, as reported by the FDA. As shown in Fig. 2, if 1000 drugs were analysed under these circumstances we would expect one false-negative result (drug was actually substandard but passed the TLC screening) and 19 false-positive results (drug was not substan- dard but failed the TLC screening). As a result, pos- sibly one substandard drug would have reached the population, but the drug quality screening laboratory would have had to refer 19 samples (1.9%) for more sophisticated testing (such as HPLC) when it was actually not necessary. In the absence of any drug quality control, however, it could be argued that 50 substandard drugs (5%) would have reached the pub- lic and potentially have resulted in unnecessary mor- bidity and mortality. In a situation where drug qual- ity is extremely poor, e.g., the prevalence of sub- standard drugs is 20%, the benefit of screening is WHO Bulletin OMS. Vol 72 1994618 Drug quality screening In Swaziland Fig. 2. Illustration of the performance of the thin-layer chromatography (TLC) method when used to analyse 1000 drugs under the following conditions. (A) The prevalence of substandard drugs is 5%; and (B) the preva- lence of substandard drugs is 20%. technicians' competence depends on having a mini- mum of secondary school mathematics and chemis- try and receiving practical training in TLC as well as possessing the motivation to perform specified and repetitive sequential tasks carefully. Acknowledgements We gratefully acknowledge the useful comments made by Professor W. Reinke dunng the preparation of the manuscrpt. + Fails TLC Total 50 950 + Fails TLC Total 200 800 especially great (Fig. 2). In this instance, only 16 (1.6%) drugs would have been falsely identified as being substandard, and TLC would have prevented 196 of the 200 (98%) substandard drugs from reach- ing the public. Further studies on the field sensitivity and specificity of the TLC screening procedure should be carried out; in the meantime, similar labor- atories that use the TLC procedure should be estab- lished in areas where drug quality control analyses are currently not being performed. The TLC method described is an appropriate tech- nology for the implementation of the drug quality assessment component of the national drug policy in Swaziland. Drug quality assessment is only one aspect of the pharmaceutical management and pre- scribing process for providing the correct drug at the right dose when the patient needs it. TLC does not address other drug quality indicators, e.g., tablet dis- integration, but it could be argued that such indica- tors are of secondary importance to the most impor- tant quality issue of drug verification. Drugs that have become degraded through poor storage condi- tions or because they have long passed their expiry date and have lost potency would, however, be detected by TLC. The technique may be suited for national drug quality screening at the central level, as in Swaziland, or could be decentralized to regional stores in larger countries where greater amounts of drugs are received and placed in inventory. The Resume Contr6le de la qualit6 des m6dicaments dans les pays en d6veloppement: 6tablissement d'un laboratoire appropri6 au Swaziland Cet article d6crit le premier laboratoire de contr6le de la qualite des medicaments base sur une tech- nique simple, peu coOteuse et exacte de chroma- tographie en couche mince (TLC), qui a e 6tabli au magasin central de fournitures medicales du Ministere de la Sante du Swaziland. Les dispensa- teurs locaux, qui ont requ une instruction secondai- re, ont ete form6s a la technique et etaient compe- tents au bout d'un mois de pratique. La technique n'exige qu'un materiel de laboratoire de base (appareillage, fournitures et reactifs) et peut etre execut6e dans un espace restreint. On trouvera dans I'article une liste des appareils, des fourni- tures et des r6actifs necessaires ainsi que les four- nisseurs des substances de r6ference pour les analyses comparatives. Le coat de d6marrage estime d'un tel laboratoire, y compris la transfor- mation d'un bureau en laboratoire, est d'environ US$ 5000-10 000, et le coOt de fonctionnement par test est bien inferieur a US$ 1,00. A titre de comparaison, le materiel utilis6 pour effectuer les analyses qualitatives/quantitatives dans la plupart des laboratoires de contr6le de la qualit6 des m6dicaments est d'environ US$ 50 000 par unite, le coat de fonctionnement par test est 20 a 50 fois plus eleve qu'avec la TLC et les techniques utili- s6es exigent un personnel hautement qualifie. La TLC est une technique qui convient pour le contr6- le qualitatif et semi-quantitatif de divers medica- ments avant leur administration aux malades. En TLC, la comparaison des taches obtenues avec l'echantillon et avec la substance de r6ference indique s'il s'agit du meme compose, et l'intensite de chaque tache, visualisee en lumiere ultraviolet- te ou par coloration a l'iode, permet de savoir si la teneur de l'echantillon se situe dans les 85-115% de la valeur de reference. La sensibilit6 et la spe- WHO Bulletin OMS. Vol 72 1994 619 T.A. Kenyon et al. cificite de la TLC sont toutes deux de 98%. Les auteurs examinent la relation entre les resultats faussement positifs et la prevalence des m6dica- ments de qualite insuffisante. Une liste de 20 medicaments testes par TLC au cours de la periode de mise en service du laboratoire est donn6e. Un technicien a elabore des modes operatoires pour 8 autres medica- ments dans les 6 mois qui ont suivi. La technique de TLC decrite ici devrait sensi- blement am6liorer I'aptitude des pays en develop- pement a contr6ler la qualite des medicaments avant de les distribuer aux services utilisateurs et aux malades. 11 est propos6 de I'adopter comme methode de choix pour contr6ler la qualit6 des medicaments a tous les niveaux du systeme de soins de sante. References 1. Alubo SO. Debt crisis, health and health services in Africa. Social science and medicine, 1990, 31: 639-648. 2. Land T. Combatting counterfeit drugs. Nature, 1992, 355: 192. 3. Flinn PE et al. A simple, inexpensive thin-layer chromatography method for the analysis of theo- phylline tablets. Bulletin of the World Health Organ- ization, 1989, 67: 555-559. 4. Flinn PE et al. A simplified TLC system for qualita- tive and semi-quantitative analysis of pharmaceuti- cals. Journal of liquid chromatography, 1992, 15: 1639-1653. 5. Flinn PE et al. A portable, inexpensive system for rapid screening of drugs by thin-layer chromatogra- phy. Journal of the Association of Official Analytical Chemists, (in press). 620 WHO Bulletin OMS. Vol 72 1994