Quality Control and Quality Assurance

QUALITY CONTROL  AND  QUALITY  ASSURANCE

It is not the aim of this section to review the entirety of quality control of sterile products or of the chemical assays and requirements of drugs and excipients prior to formulation. Only those techniques with importance either to microbiology or the confirmation of sterility of the final product are introduced here.

A)   Bioburden

It should be obvious from previous sections that a successful sterilization process is dependent on a product having a low presterilization bioburden. This will also be true of the individual ingredients, which must have low levels of microbial contamination, or else there is a danger that the contaminants will find their way into the final product or be a source of pyrogens.

Sterilization should normally be considered as the removal of the bioburden, but the high heat resistance of bacterial endotoxins means that successful steam sterilization does not necessarily guarantee that the product will pass a pharmacopoeial endotoxin test; dead bacteria are likely to remain pyrogenic.

Underestimating the level of microbial contamination prior to the terminal sterilization process will lead to a miscalculation of the sterilization dose requirements to achieve the desired SAL. The bioburden must be maintained within certain limits to justify the chosen sterilization process. When a higher number of organisms or more resistant microorganisms are encountered during manufacture of batches than was determined during the initial validation, those batches must be assumed not to be sterile. The bioburden is an estimate of the total viable count of microorganisms present before sterilization, and a knowledge of the resistance characteristics of these organisms is often an integral part of the sterility assurance calculation. To build some degree of safety into the sterilization process the sterilization conditions should be set to destroy all the bioburden by assuming that all the contaminating microorganisms are the most resistant of the species identified in that bioburden. Sterility assurance, as implied in the schemes shown in Figures 22.1 and 22.2, can only be achieved with a low bioburden and with fully validated, correctly functioning sterilizers.

B)  The Test For Sterility

The broad basis of the test for sterility is that it examines samples of the final product for the presence of microorganisms. Theoretically, the test for sterility should be applied to all products that are designated as sterile. However, the test does not examine all samples in a batch, and its results can only be considered valid if all items in a batch are treated similarly (British Pharmacopoeia, 2010). Clearly, for products which are terminally sterilized this might seem a reasonable assumption but only if there is uniform heat distribution in an autoclave or hot air oven or uniform delivery of a radiation dose. With aseptically produced products there are dangers because not all items in a batch may have been treated similarly. A successful test only shows that no microbial contamination was found in the samples examined under the test conditions. Extension of the result to a whole batch requires the assurance that every unit in the batch was manufactured in such a manner that it would also have passed the test with a high degree of probability. This highlights the weakness of the test for sterility and why the controls of sterilization processes are very important and probably of greater assurance in confirming the sterility of a batch. The test, however, remains one of few analytical methods that examine a product for sterility.

C)   Parametric Release

As there are significant limitations with the test for sterility, many authorities place considerable reliance on the validation and reliable performance of sterilizers and their sterilization cycles. Parametric release takes this reliance a step further by allowing batches of terminally sterilized products to be released without being subjected to the test for sterility. The sterilization cycle will be validated to have a SAL of 10⁻⁶ or less as the minimum safety factor. Validation studies would include heat distribution, heat penetration, bioburden, container closure and cycle lethality studies. For a product to be subject to parametric release, presterilization bioburden testing on each batch would be completed, and the comparative resistance of isolated spore-formers checked. Each cycle would include the use of chemical or biological indicators. It is hoped that these actions will provide a significantly higher level of assurance of sterility than provided by the test for sterility. This requires confirmation that each part of the manufacturing process has been satisfactorily completed, the initial presterilization bioburden is within agreed limits, that the controls for the sterilizing cycle were satisfactory and that the correct time cycles were achieved. In practice parametric release should only be used when experience has been gained on a reliably controlled and adequately validated process and where a relationship has been proved between end-product testing and in-process monitoring.

Clearly, reproducibility, regular monitoring and documentation are required. However, parametric release would imply abandoning the sterility test, an option that many manufacturers have not yet adopted, possibly because of the fear of litigation based on the premise that any sterile product would, if tested, have passed the test for sterility.

D)   Pyrogens

The discovery that aqueous solutions may lead to an increase in body temperature when injected into a patient dates back to the 19th century. The agents responsible for this fever were termed ‘pyrogens’. In theory a pyrogen is any substance that, when injected into a mammal, elicits a rise in body temperature, and substances produced by some Gram-positive bacteria, mycobacteria, fungi and also viruses conform to this definition. The most common pyrogens, however, and those of major significance to the pharmaceutical industry, are produced by Gram-negative bacteria and are known as endotoxins; they are lipopolysaccharides (LPS) found in the cell envelope . The presence of pyrogens in aqueous solutions was first demonstrated by injection into rabbits whose body temperature was recorded. More sensitive methods have since been developed, mostly based on the discovery that a fraction of the horseshoe crab blood reacts with LPS as a clotting agent.

Two pharmacopoeial limit tests exist. That for pyrogens uses rabbits to assess pharmacological activity and therefore the presence of pyrogens of all kinds. The test for bacterial endotoxins uses lysed amoebocytes (blood cells) of the horseshoe crab and is therefore termed the Limulus amoebocyte lysate (LAL) test. This may be extended to many drug and device products and clearly will be developed in the future to assess the presence of endotoxins in biotechnology products.

i)  Physiological effects of pyrogens

The most characteristic effect following injection of pyrogens into humans is a rise of body temperature, but it is only one of a number of dose-dependent diverse effects. Pyrogens elevate the circulating levels of inflammatory cytokines, which may be followed by fever, blood coagulation, hypotension, lymphopenia, neutrophilia, elevated levels of plasma cortisol and acute-phase proteins. Low doses of pyrogens induce asymptomatic inflammatory reactions. Moderate doses induce fever and changes in plasma composition. Injection of high pyrogenic doses results in shock, characterized by cardiovascular dysfunction, vasodilation, vasoconstriction, endothelium dysfunction and multiple organ dysfunction or failure and death.

ii)  Characteristics of bacterial endotoxin

The release of LPS from bacteria takes place after death and lysis. Many Gram-negative bacteria, e.g. Escherichia coli and Proteus, Pseudomonas, Enterobacter and Klebsiella species produce pyrogenic LPS which is composed of two main parts: a hydrophilic polysaccharide chain with antigenic regions, and a hydrophobic lipid group termed lipid A which is responsible for many of the biological activities. The molecular size of the polysaccharide chain is very variable, and consequently the molecular weight of the LPS may vary from a few thousand to several million daltons. LPS is unusually thermostable and  in sensitive to pH changes. Molecules are able to withstand 120 °C for over 3 hours. Extremes of pH are required for rapid destruction of the LPS.

iii)              Sources

The sources of pyrogens in parenteral products include water used at the end stages of the purification and crystallization of the drug or excipients; water used during processing; packaging components; and the chemicals, raw materials or equipment used in the preparation of the product. The presence of endotoxins on devices may be attributed to water in the manufacturing process, the washing of components such as filter media to be used for the manufacture of filters, or the washing/rinsing of tubing or other plastic devices prior to their sterilization. Additionally, if the drug is biologically produced, incomplete removal of the microorganisms during purification can result in high endotoxin levels.

iv)               Measurement of pyrogens

Pyrogens have traditionally been assessed using rabbits which are stored in carefully controlled conditions and whose temperature is monitored before the administration of the test product. The British Pharmacopoeia (2010) describes a test initially based on three rabbits; the number is progressively increased if the results fall between the two values (Table 22.4). Samples of the product under test are injected into the marginal ear vein at a dose no greater than 10 ml/kg. The animals are monitored for the 3 hour period immediately after injection, at 30 minute intervals. The test assumes that the maximum rise in temperature will be detected in this 3 hour period immediately after injection. Table 22.4 describes the criteria for pass or fail as the number of rabbits used increases to the maximum of 12.

A number of limitations of the rabbit pyrogen test are recognized. Repeated use of animals leads to endotoxin tolerance. There is low reactivity to the endotoxin produced by certain species, e.g. Legionella. There is also variability in control results when identical standardized endotoxin preparations are used, which is probably related to interlaboratory factors and variations due to seasons, rabbit species and other biological sources. Care must be taken in testing radiopharmaceuticals, and certain drugs may themselves elicit a rise in temperature on administration. The test is therefore inadequate for radiopharmaceuticals, cancer chemotherapeutic agents, hypnotics and narcotics, vitamins, steroids and some antibiotics. The presence of pyrogens may be hidden by the pharmacological activity of the product’s components. Finally the rabbit test is insufficiently sensitive to detect endotoxin in intrathecal products where only low levels of pyrogens are acceptable.

EMA (2009) is to encourage the replacement of the rabbit test with the monocyte activation test (Hoffmann, 2005) for plasma-derived medicinal products. Human monocytes from cultured cell lines mimic the human fever reaction in vitro by producing cytokines. Cytokine release can be determined, usually using enzyme-linked immunoassay (ELISA).

v)                 Measurement of bacterial endotoxins

The LAL test is considerably more sensitive than the pyrogen test. As mentioned above, although the Legionella endotoxin is not very pyrogenic to rabbits it is easily detected by the LAL test. It has been estimated that there is a 1000-fold difference in sensitivity between the two tests, but the LAL test only detects endotoxins of Gramnegative bacteria and not all pyrogens. However, the LAL test may be used for radiopharmaceuticals.

LAL test reagent comes from the American horseshoe crab Limulus polyphemus. The endotoxin-induced coagulation of its blood is based on an enzyme-mediated interaction of LAL with endotoxins. The reagents are obtained from the blood of freshly captured horseshoe crabs whose amoebocytes are concentrated, washed and lysed with endotoxin-free water. The LAL is separated from the remaining cellular debris and its activity optimized using metallic cations, pH adjustment and additives and then freeze-dried. Certain preparations interfere with the interaction between LAL and endotoxin. Chemical inhibitors may cause chelation of the divalent cations necessary for the reaction, protein denaturation or inappropriate pH changes. Physical inhibition may result from adsorption of endotoxin or be caused by viscosity of the product. Even the type of glassware may affect the test. Siliconized glassware or plastic can inhibit gel-clot formation, or prevent accurate spectrophotometric readings of the reaction end-point.

The samples of products are incubated with LAL at 37 °C. If endotoxins are present a solid gel forms, indicating the presence of endotoxins. The British Pharmacopoeia (2010) describes six separate methodologies for the test for endotoxin. These are (A) gel-clot limit test; (B) gelclot: semi quantitative; (C) turbidimetric kinetic method; (D) chromogenic kinetic method; (E) chromogenic endpoint method; and (F) turbidimetric end-point method. There are checks for interfering factors. Any validated method may be used, but the gel-clot method is the referee test in the case of dispute. Coloured products cannot be tested by turbidimetric and chromogenic methods, as precipitate formation may be mistaken for a positive response.

Kinetic LAL methods are claimed to increase the efficiency of large-scale testing, probably important when validation of depyrogenation cycles or preparation of components for aseptic processing are required. For all procedures, test validation must be conducted to rule out interference, which may be either inhibition or enhancement. Depyrogenated glassware must be used throughout.

The gel-clot method is most commonly used. The test is conducted by adding the LAL reagent to an equal volume of test solution, agitating and storing at 37 °C for 1 hour when the end-point is determined by inversion of the tubes. If a solid clot remains intact, the product is considered to contain endotoxins. Chromogenic methods utilize colorimetry but do not depend on the clottable protein. A synthetic substrate is used that contains an amino acid sequence similar to that of coagulogen, the clottable protein. The activated proclotting enzyme cleaves a p-nitroanilide chromophore from the synthetic substrate and the colour produced is proportional to the amount of endotoxin. The turbidimetric LAL method is based on the fact that an increase in endotoxin concentration will cause a proportional increase in turbidity caused by the precipitation of the clottable protein, coagulogen. The optical density is read spectrophotometrically either at a fixed time or constantly for kinetic assays as turbidity develops. The kinetic methods depend on the relationship between the logarithm of the response and the logarithm of the endotoxin concentration. The end-point methods relate endotoxin levels to the quantity of chromophore released or the amount of precipitation.

vi)               Endotoxins in parenteral Pharmaceuticals

The limits for endotoxin are based on the dose of the product. Put simply, the endotoxin limit, EL, which represents the maximum amount of endotoxin that is allowed in a specific dose, is inversely related to the dose of the drug; it may be assessed from the following equation (United States Pharmacopeia, 2010):

EL = K/M                                                                     (2)

where K is the threshold human pyrogenic dose of endotoxin per kg body weight and M is the maximum human dose of the product in kg body weight that would be administered in a single 1 hour period. M recognizes that the pharmacological effects of endotoxin are dosedependent. The endotoxin limit is the level at which a product is adjudged pyrogenic or non-pyrogenic. Gelclot reagent sensitivities are generally in the range 0.015–0.5 EU/ml. As examples of endotoxin limits, the United States Pharmacopeia (2010) states limits of no more than 0.5 EU/ml for Dextrose Infusion, no more than 5 EU/mg promethazine in Promethazine Injection USP, no more than 10 EU/mg of mitomycin in Mitomycin for Injection USP and no more than 24 EU/mg warfarin sodium in Warfarin Sodium for Injection. The British Pharmacopoeia (2010) has a limit of 0.25 IU/ml in Glucose Intravenous Infusion; this value is similar for many BP intravenous infusions. As another example, insulin should contain no more than 10 IU/mg of endotoxin. The endotoxin limit for drugs gaining access to the cerebrospinal fluid is reduced to 0.2 EU/kg because the intrathecal route is the most toxic route for endotoxins.

vii)              Depyrogenation and the Production of Apyrogenic Products

Pyrogens and endotoxins are difficult to remove from products once present and it is easier to keep components relatively endotoxin-free rather than to remove them from the final product. Rinsing or dilution is one way of eliminating pyrogenic activity provided that the rinsing fluid is apyrogenic. Closures and vials should be washed with pyrogen-free water before sterilization. Pyrogens in vials or glass components may be destroyed by dry heat sterilization at high temperatures. A recommended condition for depyrogenation of glassware and equipment is heating at 250 °C for 45 minutes. Pyrogens are also destroyed at 650 °C in 1 minute or at 180 °C in 4 hours. The British Pharmacopoeia (2010) states that dry heat at temperatures above 220 °C may be used for the depyrogenation of glassware. Sterilizing tunnels are designed not only to sterilize at 250–300 °C but also to remove pyrogens. These processes equate to incineration, although removal by washing, also termed dilution, may be used. Filtration, irradiation or ethylene oxide treatment have limited value in reducing pyrogen or endotoxin loads.

The removal of pyrogens from Water for Injections may be effected by distillation or reverse osmosis. Distillation is the most reliable method for removing endotoxin. Care has to be taken to avoid splashing in the still as pyrogens have been carried over in droplets. Another source of endotoxins is the Water for Injection system. Generally, circulating hot water at temperatures above 75 °C provides an environment that is not conducive to microbial growth and thus the formation of endotoxin. Circulating water at approximately 60 °C causes some concern as some Gram-negative organisms, e.g. Legionella pneumophila, will survive and grow at 57 °C. The water-producing systems may be sanitized by circulating water at 75–80 °C.

Pyrogen-free water can be produced using an ultrafiltration membrane with a nominal molecular weight limit that is low enough to ensure the removal of endotoxins under all conditions. Hollow fibre polysulphone membranes can be sanitized with sodium hydroxide, which efficiently destroys pyrogens. A nominal molecular weight limit of 5000 Da should efficiently remove endotoxins. However, many endotoxin-producing microorganisms multiply in ambient temperature Water for Injection systems, especially reverse osmosis (RO) systems, in which the filters are not absolute and may be used in series in order to manufacture pyrogenfree water.