Quality Control

QUALITY CONTROL

The quality control of vaccines is intended to provide assurances of both the probable efficacy and the safety of every batch of every product. It is achieved in three ways:

(1) in-process control; (2) final product control; and (3) requirements that for each product the starting materials, intermediates, final product and processing methods are consistent.

The results of all quality control tests must be recorded in detail and authorized by a qualified person as, in those countries in which the manufacture of vaccines is regulated by law, they are part of the evidence on which control authorities judge the acceptability or otherwise of each batch of each preparation.

A)  In-Process Control

In-process quality control is the control exercised over starting materials and intermediates. Its importance stems from the opportunities that it provides for the examination of a product at the stages in its manufacture at which testing is most likely to provide the most meaningful information. The WHO recommendations and national authorities stipulate many in-process controls but manufacturers often perform tests in excess of those stipulated, especially sterility tests as, by so doing, they obtain assurance that production is proceeding normally and that the final product is likely to be satisfactory. Numerous examples of in-process control exist for various types of vaccine but three demonstrate the principle.

The quality control of both diphtheria and tetanus vaccines requires that the products are tested for the presence of free toxin, i.e. for specific toxicity due to inadequate detoxification with formaldehyde, at the final product stage. By this stage, however, the toxoid concentrates used in the preparation of the vaccines have been much diluted and, as the volume of vaccine that can be inoculated into the test animals (guinea-pigs) is limited, the tests are relatively insensitive. In-process control, however, provides for tests on the undiluted concentrates and thus increases the sensitivity of the method at least 100-fold.

An example from virus vaccine manufacture is the titration, prior to inactivation, of the infectivity of the pools of live poliovirus used to make inactivated poliomyelitis vaccine. Adequate infectivity of the virus from the tissue cultures is an indicator of the adequate virus content of the starting material and, as infectivity is destroyed in the inactivation process, there is no possibility of performing such an assay after formaldehyde treatment.

A more general example from virus vaccine production is the rigorous examination of tissue cultures to exclude contamination with infectious agents from the source animal or, in the cases of human diploid cells or cells from continuous cell lines, to detect cells with abnormal characteristics. Monkey kidney cell cultures are tested for simian herpes B virus, simian virus 40, mycoplasma and tubercle bacilli. Cultures of human diploid cells and continuous line cells are subjected to detailed karyological examination (examination of chromosomes by microscopy) to ensure that the cells have not undergone any changes likely to impair the quality of a vaccine or lead to adverse effects.

B)  Final Product Control

i)  Assays

Vaccines containing killed microorganisms or their products are generally tested for potency in assays in which the amount of the vaccine that is required to protect animals from a defined challenge dose of the appropriate pathogen, or its product, is compared with the amount of a standard vaccine that is required to provide the same protection. The usual format of the test is the 3 + 3 dose quantal assay that is used to estimate the potency of whole-cell pertussis vaccine (British Pharmacopoeia, 2010). Three logarithmic serial doses of the test vaccine and 3 of the standard vaccine are made and each is used to inoculate a group of 16 mice. In the case of both the test vaccine and the standard, the middle dose is chosen on the basis of experience, so that it is sufficient to induce a protective response in about 50% of the animals to which it is given. Each lower dose may then be expected to protect less than 50% of the mice to which it is given and each higher dose to protect more than 50% of the animals. Fourteen days later all of the mice are inoculated (‘challenged’) with a suitable virulent Bordetella pertussis strain and, after a further 14 days, the number of mice surviving in each of the 6 groups is counted. The number of survivors in each group is used to calculate the potency of the test vaccine relative to the potency of the standard vaccine by the statistical method of probit analysis (Finney, 1971). The potency of the test vaccine may be expressed as a percentage of the potency of the standard vaccine. However, as the standard vaccine will have an assigned potency in international units (IU), it is more usual to express the potency of the test vaccine in similar units. Tests similar to that used to estimate the potency of pertussis vaccine are prescribed for the potency determinations of diphtheria vaccine and tetanus vaccines. In these cases the respective bacterial toxins are used as the challenge material (British Pharmacopoeia, 2010). Tests that do not involve challenge but involve titration of the antitoxin response in vitro, e.g. by ELISA (enzyme-linked immunosorbent assay), are now being adopted.

Vaccines containing live microorganisms are generally tested for potency by determining their content of viable particles. In the case of the most widely used live bacterial vaccine, BCG vaccine, dilutions of vaccine are prepared in a medium which inhibits clumping of cells, and fixed volumes are dropped on to solid media capable of supporting mycobacterial growth. After a fortnight the colonies generated by the drops are counted and the live count of the undiluted vaccine is calculated. The potency of live viral vaccines is estimated in much the same way except that a substrate of living cells is used. Dilutions of vaccine are inoculated on to tissue culture monolayers in Petri dishes or in plastic trays, and the infective particle count of the vaccine is calculated from the infectivity of the dilutions as indicated by plaque formation, cytopathic effect, haemadsorption or other effect and the dilution factor involved.

ii)                Safety tests

Because many vaccines are derived from basic materials of intense pathogenicity—the lethal dose of tetanus toxin for a mouse is estimated to be 3 × 10⁻² ng—safety testing is of paramount importance. Effective testing provides a guarantee of the safety of each batch of every product and most vaccines in the final container must pass one or more safety tests as prescribed in a pharmacopoeial monograph. This generality does not absolve a manufacturer from the need to perform in-process tests as required, but it is relaxed for those preparations that have a final formulation that makes safety tests on the final product either impractical or meaningless.

Bacterial vaccines are regulated by relatively simple safety tests. Those vaccines composed of killed bacteria or bacterial products must be shown to be completely free from the living microorganisms used in the production process. Inoculation of appropriate bacteriological media with the final product provides an assurance that all organisms have been killed. Those vaccines prepared from toxins, for example, diphtheria and tetanus toxoids, require in addition, a test system capable of revealing inadequately detoxified toxins; this can be done by inoculation of guinea-pigs, which are exquisitely sensitive to both diphtheria and tetanus toxins. A test for sensitization of mice to the lethal effects of histamine is used to detect active pertussis toxin in pertussis vaccines. An improved non-lethal method is also available. The trend is to replace in vivo assays by cell culture methods where possible but these do not always emulate in vivo effects. Inoculation of guinea-pigs is also used to exclude the presence of abnormally virulent organisms in BCG vaccine. Molecular genetic methods, such as nucleic acid amplification to probe for genes specific to virulent strains, are now available but not yet in routine use for vaccine testing.

Viral vaccines can present problems of safety testing far more complex than those experienced with most bacterial vaccines. With killed viral vaccines the potential hazards are those due to incomplete virus inactivation and the consequent presence of residual live virus in the preparation. The tests used to detect such live virus consist of the inoculation of susceptible tissue cultures and of susceptible animals. The cultures are examined for cytopathic effects, and the animals for symptoms of disease and histological evidence of infection at autopsy. This test is of particular importance in inactivated poliomyelitis vaccines, the vaccine being injected intraspinally into monkeys or mice transgenic for the poliovirus receptor. At autopsy, sections of brain and spinal cord are examined microscopically for the histological lesions indicative of proliferating poliovirus.

With attenuated viral vaccines the potential hazards are those associated with reversion of the virus during production to a degree of virulence capable of causing disease in recipients. To a large extent this possibility is controlled by very careful selection of a stable seed but, especially with live attenuated poliomyelitis vaccine, it is usual to compare the neurovirulence of the vaccine with that of a vaccine known to be safe in field use. The technique involves the intraspinal inoculation of monkeys with both the reference vaccine and the test vaccine followed by comparison of the neurological lesions and symptoms, if any, that are caused. If the vaccine causes abnormalities in excess of those caused by the reference it fails the test. A modification of this test which uses transgenic mice instead of monkeys is now available. An in vitro method (MAPREC test) which relies on detecting RNA sequences specific to virulent virus has also been developed. A widespread problem with safety testing of live viral vaccines is that the host specificity of many viruses limits the availability of suitable animal models.

iii)      Tests of general application

In addition to the tests designed to estimate the potency and to exclude the hazards peculiar to each vaccine there are a number of tests of more general application. These relatively simple tests are as follows.

Sterility.

 In general, vaccines are required to be sterile. The exceptions to this requirement are smallpox vaccine made from the dermis of animals and bacterial vaccines such as BCG, Ty21A and tularaemia vaccine, which consist of living but attenuated strains. These have a bioburden limit which defines the number of permissible microorganisms but excludes pathogens. WHO recommendations and pharmacopoeial monographs stipulate, for vaccine batches of different size, the numbers of containers that must be tested and found to be sterile. The preferred method of sterility testing is membrane filtration, as this technique permits the testing of large volumes without dilution of the test media. The test system must be capable of detecting aerobic and anaerobic bacteria and fungi .

Freedom from abnormal or general toxicity.

The purpose of this simple test is to exclude the presence in a final container of a highly toxic contaminant. Five mice of 17–22 g and two guinea-pigs of 250–350 g are inoculated with one human dose or 1.0 ml, whichever is less, of the test preparation. All must survive for 7 days without signs of illness. Current pharmacopoeial monographs usually do not require this test if another in vivo test has been performed on the product.

Pyrogenicity or endotoxin content.

The pyrogenicity of a specified dose of product when administered to rabbits can be assayed by a standard pharmacopoeial method but the trend is to replace this with an in vitro assay for endotoxin . The capacity of the product to induce gelation of Limulus polyphemus amoebocyte lysate is determined against a reference endotoxin preparation and the result is expressed as IU of endotoxin. For pyrogens other than endotoxin, a monocyte stimulation test is available.

Presence of aluminium and calcium.

The quantity of aluminium in vaccines containing aluminium hydroxide or aluminium phosphate as an adjuvant is limited to 1.25 mg per dose and it is usually estimated compleximetrically. The quantity of calcium is limited to 1.3 mg per dose and is usually estimated by atomic absorption spectrometry.

Free formaldehyde.

 Inactivation of bacterial toxins with formaldehyde may lead to the presence of small amounts of free formaldehyde in the final product. The concentration, as estimated by colour development with acetylacetone, must not exceed 0.02%.

Phenol concentration.

When phenol is used to preserve a vaccine its concentration must not exceed 0.25% w/v or, in the case of some vaccines, 0.5% w/v. Phenol is usually estimated by the colour reaction with aminophenazone and hexacyanoferrate.

pH.

 The potentiometric determination of pH is made by measuring the potential difference between two appropriate electrodes immersed in the solution to be examined: one of these electrodes is sensitive to hydrogen ions and the other is the reference electrode. The pH apparatus is calibrated with the buffer solution of potassium hydrogen phthalate and one other buffer solution of different pH. The pH in the test sample should comply with the limits approved for the particular products.

Osmolality.

Osmolality is a practical means of giving an overall measure of the contribution of the various solutes present in a solution to the osmotic pressure of the solution. Osmolality is determined by measurement of the depression of freezing point of the test sample using appropriate apparatus. The osmolality of the test sample should comply with the limits approved for the particular products, for example minimum 240 mOsmol/kg.