Tracking down pharma contamination

Despite the best efforts of pharmaceutical manufacturers to produce to the highest standards, there is always the potential for something to go wrong.

Whether through some unforeseen event during manufacture or distribution, a malicious act of a disgruntled employee or deliberate tamper by a criminally-inclined person, there are ways in which the quality of the end product can be compromised by contaminants.

Such contamination may be detected during routine quality control checks, or it may only come to light when the product is returned as a complaint, or when the extortion threat is received.

However the problem first comes to light, identifying a contaminant and finding an explanation for its presence in the product is seldom an easy matter.

To identify fully the chemical contaminant causing an 'extra' peak observed using high performance liquid chromatography (HPLC), or the obvious 'black bits' in a tablet, can require a significant amount of detective work on behalf of the scientist charged with investigating the problem.

Very often, it requires a team effort, involving the skills of chemists, microscopists, and microbiologists working together to reach a satisfactory conclusion.

Where this multi-disciplinary expertise is not available in-house, manufacturers should collaborate with a laboratory with suitable accreditation to assist in the investigation.

Chemical contaminants.

Almost by definition, a chemical contaminant will be invisible to the naked eye, but will reveal itself either through laboratory analysis during a routine quality control (QC) inspection, or through the patient reporting a problem in the product.

In the latter instance, it is usually the case that the problem is identified as a bad/wrong smell or taste, discoloration, or visible foreign material.

Such descriptions will only go a little way towards identifying the contaminant, and in any event, they depend on the user's ability to detect an anomaly in a product with which they may be unfamiliar.

Given that the patient may have an expectation that their medicine will taste bad, it is perhaps optimistic to expect that a bad taste or a bad smell will of themselves be sufficient to engender a customer complaint.

A physician or pharmacist directly involved in administering a medicine may be better placed to detect a rogue batch, though not necessarily any more able to identify the contaminant concerned.

The more common case is that routine QC checks reveal trace amounts of chemicals that should not be present in the pharmaceutical product.

The more demanding task is to say what the contaminant is, and where it has come from.

When RSSL Pharma is called on to investigate such incidents the first step is to repeat the QC procedure carried out by the manufacturer, in the expectation of repeating the finding.

This analysis will often be an HPLC or gas chromatography (GC) analysis.

If the QC analysis was by GC, then it may well be possible to identify the contaminant using GC-mass spectrometry (GC-MS).

This is a technique that involves separation of the chemical constituents of the product according to their volatility, and identifies specific compounds according to their molecular weight and the molecular weight of their ion fragments.

The transition from HPLC to LC-MS is less easy, because some HPLC solvent systems are unsuitable for LC-MS.

The most commonly used HPLC detection method involves use of UV light, ie, chemicals are detected by their absorption of UV light at a particular wavelength.

This means that non-UV-active components do not interfere with the analysis.

In all cases, however, repeating the client's analysis is necessary to give the chemist a 'hold' on the problem, and it may reveal other useful information.

For example, in a few cases investigated by our own laboratories at RSSL Pharma, it has emerged that it was the QC check itself that contaminated the product sample under test.

In these cases, RSSL Pharma had been unable to reproduce the results that the client had seen in their own laboratories.

Further investigation revealed that laboratory equipment used by the client to prepare the sample had been the source of the contamination.

While it is unusual for QC checks to cause a problem, rather than merely to reveal it, this example does illustrate the point that the sources of contamination are frequently hard to anticipate, and as a consequence, hard to eliminate.

Indeed, elimination may be impossible.

For example, it is accepted that some pharmaceutical actives will inevitably give rise to degradation products.

Provided these degradation products are well characterised, and present within the permitted levels, their presence will cause no concern.

They should nonetheless be considered as contaminants.

It is ironic that production equipment, including the very items installed to protect the product from other problems, can itself be the source of contamination.

RSSL Pharma has often come across incidents where plastic monomers or oligomers have leached out of on-line filters into pharmaceutical solutions.

Monomers, oligomers or plasticisers can also leach from plastic seals and pipes, and from plastic packaging, and it takes very little migration to produce a detectable problem.

Similarly, on some occasions, chemicals from cleaning solutions, cooling waters, and machinery lubricants, have been identified in the product.

Perhaps of greater concern, on very rare occasions, the contaminant has been shown to be an active from an earlier production run of a different pharmaceutical product.

Just as there is a potential for plastic monomers to leach from production equipment, there is a similar risk of chemicals leaching from packaging materials.

These are not common, but must be evaluated in order to be ruled-out of an investigation.

For example, a potential explanation for a yellow stain on a white tablet could be the oxidation and dimerisation product of butylated hydroxytoluene (BHT), which is an antioxidant used in plastic packaging.

Indeed, in a recent case investigated by RSSL Pharma, preliminary analysis of the solubility characteristics of a yellow chemical did at first sight indicate BHT dimer, although a more detailed chemical analysis ultimately proved the coloured material to be a degradation product of the drug itself.

The point is that all potential explanations must at least be considered.

In a different incident, an intense red colour was spotted in a pale pink tablet.

Initial questions about what bright red chemicals existed that might have contaminated the product were put aside when analysis showed that the intense colour was due to 'wetting' of the pink tablet with tiny drops of the lubricant used on the production line.

The redness was not an attribute of the lubricant, merely a consequently darkening of the pink colours already present in the tablet.

Experience essential.

To reach any conclusion about the identity and source of the contaminant is no easy matter, and for the investigating team it is always useful to have a full production history of the rogue sample, and a sample of a perfect product (a control sample) against which to compare it.

Armed with this information, the experienced chemist can usually target his/her efforts more precisely against an unknown chemical that often first appears as a tiny, unwanted 'blip' on the HPLC analysis.

This chromatogram will, of course, reveal some information about the contaminant.

The size, shape and position (retention time) of the contaminant peaks relative to the expected peaks will at least indicate whether it is polar or non-polar.

In addition, most routine QC HPLC methods use UV detection at a defined wavelength, whereas RSSL Pharma will generally choose the more powerful diode array detection (DAD).

With DAD, the UV spectrum of the impurity peak can be reconstituted, provided that it is present at sufficient concentration.

The UV spectrum provides some information on the nature of the contaminant (e.g whether or not it is a substance related to the active).

Sometimes, it is possible to use the retention time and UV data to produce a 'shortlist' of possible identities for the contaminant.

The candidate compounds can then be sourced and their HPLC retention times checked against that of the contaminant.

In some cases, LC-MS may be used.

Alternatively, the chemist may employ preparative HPLC, thin layer chromatography (TLC) or even (in some cases) solvent extraction to isolate the offending substance, and spectroscopic techniques such as MS, Fourier transform infra-red spectrometry (FT-IR) and nuclear magnetic resonance (NMR) spectrometry to identify its formula and structure.

Once the chemical has been identified, it is common practice to use it to spike a sample of the perfect product and perform the original QC analysis again, to verify that the same result emerges.

This is actually more easily said than done, since it is not necessarily the case that the contaminating chemical is commercially available.

For example, the dimer from nylon 6,6, (which could leach from a nylon filter) is not a chemical that is produced as anything other than an incomplete by-product of polymerisation.

It cannot be bought 'off the shelf' for the purpose of spiking a pharmaceutical product.

Nonetheless, once the contaminant's identity has been established, it is usually possible to identify the source, except perhaps, when there is more than one potential source of the same contaminant.

All of the techniques that might be applied to the investigation of a problem brought to light by routine QC checks are also applicable to the investigation of contamination identified by a patient or threatened by a criminal.

In the latter case, of course, the threat may not actually involve contamination of any kind.

The role of the analytical laboratory may be to determine whether the threat is genuine and/or achievable.

For example, an extortion threat might be received that sounds plausible in theory, but in practice, results in such damage to the original product that it is rendered unusable.

The consumer/patient may be presented with no risk simply because they cannot use a product that has been contaminated in the specified manner.

On the other hand, analysis may well be necessary to verify that a suspect product has indeed been tampered with in the way that has been threatened.

Clearly, where the criminal has identified the chemical he/she intends to use to 'poison' the product, it is a simpler task to determine the presence (or absence) of the offending chemical.

Vague threats are less easy to verify, although it should always be possible to show the presence of a contaminant, even if it is less easy to say what it is.

Foreign bodies.

What is true for contamination with trace amounts of chemicals is also true for the 'bulk' contamination that manifests itself as a foreign body.

Once again, production equipment, laboratory equipment, packaging materials, production errors and deliberate acts of tamper are all potential sources of contamination.

In a sense, the only difference is that the foreign body can be seen by the naked eye.

Microscopy is always the first line of attack in investigating foreign body incidents.

There is much that can be learned even from examining the size, shape and surfaces of a foreign body with simple light microscopy.

However some of the most powerful analytical tools are the scanning electron microscope, linked to X-ray microanalysis equipment and also X-ray microfluorescence spectrometry.

Both can be used to reveal the elemental composition and elemental distribution of a sample and are especially valuable in the analysis and comparison of glass and metallic fragments.

FT-IR microscopy is similarly powerful in being able to identify small particles of an organic nature, such as plastics.

Glass fragments are a potential foreign body hazard in any pharmaceutical product that uses glass packaging.

Any breakage on the filling line can easily contaminate other bottles and phials, but it is not necessarily the case, of course, that the glass fragment found in glass packaging has come from the packaging itself.

There are many other sources that need to be investigated.

Microscopy of any original surfaces of a glass fragment can provide important information on its mode of manufacture, whether, for example, the original glass article was moulded (e.g a bottle) or spun manufactured (e.g light bulb).

Surface interferometry gives information on the curvature of a fragment, distinguishing between flat glass (e.g window) or curved glass (e.g tumbler, milk bottle).

Using this technique it is possible to estimate the radius of curvature of a minute glass fragment and thus form a conclusion about the diameter of the region of the item from which it originated.

More information is available from the use of X-ray microfluorescence spectrometry.

This technique relies on the basic principle that different elements have characteristic fluorescence spectra when irradiated with X-rays of defined energy.

Detection of the emitted radiation, and subsequent data processing, reveals the elemental composition of the glass fragment and allows it to be compared with reference samples, either from the factory, or from our own database of over 500 glass types.

Using this technique, it is possible to differentiate between sheet glass, lighting glass, containers (bottles and jars), lead glass, borosilicate (ie, heat resistant glass) and domestic glass (tumblers, dishes etc).

Metal swarf from machinery is another common foreign body contaminant, again arising from wear and tear on factory machinery.

Lubricants are relatively frequently found as co-contaminants in this kind of case.

However, like glass, metal fragments and objects may arise from a variety of other sources such as packaging (laminated foil), even dental fillings.

Most factories do have metal detection facilities on-line which helps limit the problem of metal fragments reaching the patient but none of this equipment guarantees an end to the issue.

The origin of a tiny metal fragment or dust can only be determined once its elemental composition is known.

As with glass, this is can be achieved by X-ray microfluorescence spectrometry or by X-ray microanalysis, which allow distinctions to be made between different base metals, steels and other alloys.

The same analysis can be used to determine any match between samples and reference materials supplied from the factory.

In a similar way, plastic fibres can also arise from the use of plastic packaging, or from wear and tear on plastic fixtures in the process line.

Polymers and plastics have replaced metals and glass in many applications.

As a result, plastics present an increasingly common foreign body problem.

A combination of microscopy and spectroscopy techniques can be used to identify these materials.

Fourier transform infra-red (FT-IR) spectroscopy or microspectroscopy can be used to characterise the chemical structure of the sample.

The spectrum obtained by FT-IR can be compared with reference spectra from RSSL's own plastics database, or the extensive polymer library also available at RSSL.

A second technique, known as differential scanning calorimetry, can also be used to characterise plastics in terms of their melting point, degree of crystallinity and glass transition temperature.

In some cases this will differentiate between different forms of the same polymer.

Whilst glass, metal and plastic fragments from plant and packaging constitute the majority of foreign body cases, they are by no means the only cases.

RSSL Pharma has investigated a wide variety of incidents over the years and identified a wide variety of insects, human hairs, pieces of paper, drops of silicone and 'black bits' (often consisting of thermally degraded product).

In conclusion, one last class of foreign body is worthy of mention; sub-visible particles in injectable solutions.

USP<788 and EP 2.9.19 set out limits on the numbers of sub-visible particles (<10um and <25um) that are acceptable in injectables, and it is clear why injectables must be tested to ensure that they conform with the USP and EP as appropriate.

However, sub-visible particles can also be a problem in inhalers and foams, whether for reasons of efficacy or product quality.

There is no pharmacoepial test for this last group of products, but RSSL Pharma has developed its own validated methods for quantifying the sub-visible particles, and for differentiating between particles and bubbles.

The actual measurement and counting is performed on a light obscuration particle counter, which has a particle size range from two to 125um.

(though particles above 50um would be considered within the visible range).

It is also worth mentioning in respect of sub-visible particles that sample preparation is key to obtaining accurate and meaningful results.

In the same way that chemical contamination during the client's own QC testing procedures has sometimes been implicated in cases that RSSL Pharma has investigated, so mistakes/oversights during sampling have sometimes introduced sub-visible particles into the products under test.

This may merely be a case of using a poorly prepared sample bottle, but of course, it is an expensive mistake to make, given that a whole production batch will have to be withheld on the strength of the negative results.

Far better is to send a complete production container to the testing laboratory in order that the sampling can be done by the laboratory itself.

Microbiology.

Pharmaceuticals are a relatively inhospitable habitat for bacteria, and certainly, those that are produced in a sterile environment are unlikely to suffer from microbial contamination.

However, the water used in the production of non-sterile products may be a source of microbial contamination, and spore forming moulds, such as penicillium and aspergillus do occasionally arise in over-the-counter medicines.

These organisms can survive in the high temperature of the production environment, and may be able to proliferate if the product comes into contact with moisture.

Though seldom dangerous, clearly such contamination is likely to be distressing to consumers and is always worthy of investigation.

Always investigate.

While it may not be possible to prevent every contamination incident, it is always in the interests of manufacturers to investigate every incident.

The information acquired by a successful investigation will also offer an explanation as to how it occurred, and should go a long way to helping the manufacturer to decide what measures need to be introduced to prevent a repeat occurrence.