QUALITY CONTROL

36 manufacturing chemist October 2016

A powerful tool for the detection of counterfeit drugs

Drug formulation, an integral step in theregulated pharmaceutical industry, involves thecareful blending and processing of preciselyspecified and measured components to create afinal product that can be administered to deliverthe required clinical dose. The production ofcounterfeit drugs is a major challenge to theindustry and has the potential to endangerhealth, threaten reputation and, in the long-term,reduce the funds needed to drive pharmaceuticalinnovation. Counterfeit products, although oftenpackaged convincingly, may have specificcomponents replaced or altogether removed,and/or additional components included.

Investigation of the composition of illegal drugsprovides insight into their provenance tocriminalists working in this area. However, it canbe difficult to robustly detect and identify fakeactives, contaminants and/or bulking agents,hampering efforts to investigate the associatedcrimes. Tracking down and halting counterfeitingoperations is a substantial and complex activityand calls for highly effective analytical strategies.

Counterfeiting and its associated challengesThe production of counterfeit drugs is a growingconcern for the pharmaceutical industry andaffects a wide range of pharmaceutical products,from relatively inexpensive and high-volumegenerics, through to high-value drugs still underpatent. Recipients are often duped by the near-identical appearance of the counterfeit to itsgenuine generic or innovator counterpart, whichcan make identification very difficult.

Although convincingly similar in externalappearance, the components used to formulatecounterfeits may differ substantially from thoseused to produce the original, and counterfeits are,of course, unregulated, creating a high risk ofunexpected and potentially harmful outcomes forthe individual. In this way, counterfeiting erodesconfidence in the genuine product and also in itsmanufacturer.

Criminalists are increasingly interested ininvestigating counterfeit products, with theultimate aim of identifying their source anderadicating them from the market. However, suchinvestigations can present a complex challenge.The accurate identification and quantification ofone or more individual components within acounterfeit pharmaceutical product can bedifficult and often unreliable in the absence ofappropriate technologies and methodologies. Thepharmacopoeial analyses specified to investigatefalsified products are often expensive, and there isa call for simple alternatives that offer optimal

sensitivity and specificity to maximise informationflow at reasonable cost.

During the investigation of counterfeit drugs, itis first and foremost the identification andquantification of actives, excipients and potentialcontaminants that is vital, as these studiesdetermine the performance and therapeuticprofile of the pharmaceutical product. Incounterfeit products, components are often lesscontrolled in terms of quantity and particle sizeand shape distribution. Techniques that measurethese metrics for individual components within ablend, with a high level of specificity, can thereforeprovide insight to help criminalists successfullypenetrate the superficial appearance of acounterfeit drug and determine its source.

Challenging counterfeiting with MDRSMDRS is a powerful technique that combinesautomated particle imaging with Ramanspectroscopy to provide particle size, shape andchemical identification. These capabilities makeMDRS a novel and valuable tool for theinvestigation of counterfeit pharmaceuticals andsubsequent adjudications.

Automated particle imaging generatesstatistically significant particle size and shapedistributions for a sample from images ofindividual particles, captured using advancedcamera technology. Tens of thousands of particlescan be imaged in a matter of minutes, making thetechnique substantially superior to microscopicmethods, which are far more labour intensive andpotentially subjective. In the investigation ofcounterfeit products, particle size and shape areimportant metrics as they can influenceperformance attributes such as reactivity andefficacy of delivery, and can provide clues to theprocessing provenance of a suspect material. Thatsaid, the true potential of particle size distributionmeasurements to provide valuable data in forensicscience is yet to be fully explored.

Raman spectroscopy, which is a non-destructive,rapid, and reliable technique for determining themolecular chemistry of a substance, is routinelyused in criminal investigations. It enables preciseidentification of the chemical identity of any activepharmaceutical ingredient (API) present, as wellas any excipients. Amalgamating Ramanspectroscopy with automated imaging providesenhanced insight into the physical and chemicalproperties of a blend by reliably and accuratelydifferentiating morphologically similar substances.

Furthermore, because MDRS is applied toindividual particles, it offers heightenedsensitivity when compared with bulk Raman

Introducing morphologicallydirected Raman spectroscopy

In this article, we look at howMorphologically-DirectedRaman Spectroscopy (MDRS) isbeing used to reliably identifycounterfeit drugs. Case studydata show how thecomprehensive and detailedinformation it provides canhelp researchers to gain abetter understanding of themixtures encountered duringforensic investigations, andidentify the source of theoperation.

QUALITY CONTROL

analysis. MDRS can identify small quantities of anAPI, even when a high volume of bulking agent ispresent. Comparing samples analysed usingMDRS to a database of other known samples canoffer insight into the provenance of the product.The following case studies illustrate theinformation that MDRS can efficiently generate;all measurements were made using a MorphologiG3-ID from Malvern Instruments.

Case study 1: comparing over-the-counter (OTC)aspirin productsTwo OTC aspirin products were analysed usingMDRS: one was a reference-labelled drug, theother a generic alternative. Samples of bothproducts were dry dispersed using the automatedoperating procedures of the instrument and thencharacterised in terms of their particle size andshape distribution. More than 1200 individualparticles were uniformly selected across thesedistributions for analysis by Raman spectroscopy.Comparing individual particle spectra with therespective reference spectra for the active (aspirin)and the binder (starch) allowed the chemicalidentification of specific particle populations, andthe generation of a particle size distributionspecifically for the active in each product (Figure 1).

In this instance, the particle size distributions ofthe active in the branded and generic productswere found to be very similar; if the brand werecompared with a counterfeit pharmaceutical, thedistributions may be expected to be much lessclosely aligned. Individual images of the binderand aspirin particles show that the componentswould be difficult to reliably differentiate on thebasis of particle morphology alone, as they are ofsimilar size and shape.

This example therefore highlights the value ofthe Raman data, which enables robustdifferentiation between the two components andthe consequent generation of a particle sizedistribution specifically for the active. Should thebranded product be compared with a counterfeit,

October 2016 manufacturing chemist 37

Fig. 1 Overlay of the particle size distribution for aspirin confirms close similarity between thebranded and generic products; particle images (inset) show that aspirin and binder cannot bedifferentiated on the basis of particle morphology alone, highlighting the value of MDRS forgenerating component-specific data

Fig. 2 MDRS is sufficiently sensitive to detect and characterise baking soda present in DXM in aratio of 1:999, enabling the generation of a particle size distribution for each component and anestimate of relative quantities present

these same spectra would enable the identificationof any active substances and/or fillers present,with particle shape and size data providingcomplementary information to aid identification ofthe source.

Case study 2: detecting an active at lowconcentrationsIn a second study, a powder-based drug,dextromethorphan hydrobromide (DXM), wascontaminated at 999:1 (w:w) with baking soda tochallenge the sensitivity of MDRS analysis.Traditional bulk Raman analysis of this mixtureproduced spectra of the DXM alone, and was thusunable to detect the baking soda ‘contaminant.’The results of analysis by MDRS show that theparticles are not distinguishable on the basis ofmorphology but can be robustly quantified withreference to library Raman spectra. The particlesof the baking soda ‘contaminant’ werereproducibly chemically identified, even at thisvery low level (Figure 2), enabling the generationof a particle size distribution specifically for bothDXM and baking soda, and an estimation of theamount of each present.

Case study 3: using morphological classificationto cut analysis timesIn a final study, two commercially availableartificial sweetener blends with differentsweetening agents (saccharin and sucralose) butthe same bulking agent (dextrose) was analysed.Sweetening agents are several times sweeter thana comparable amount of table sugar. Consequently,artificial sweeteners require a bulking agent orfiller such as dextrose to increase the volume ofthe blend and to mitigate any undesirable tasteresulting from the presence of the sweeteningagent. Identification of the sweetening componentin artificial sweeteners cannot be accomplishedwith bulk spectroscopic analysis because of its lowconcentration relative to the bulking agent.

QUALITY CONTROL

38 manufacturing chemist October 2016

To detect the active ingredient, manualmicroscopic particle picking must be done prior toRaman spectroscopic analysis, which is timeconsuming, subjective and requires significantskill. Thus, different brands of sweeteners cannotbe rapidly differentiated by traditional non-destructive analytical techniques.

In this study, morphological data was collectedfor approximately 150,000 particles of sample A(saccharin/dextrose) and 100,000 particles ofsample B (sucralose/dextrose). Following thisanalysis, particles with a circle equivalentdiameter (CED) greater than 7.0 µm were taggedfor chemical identification using the instrumentclassification software. Particles with a 7.0 µmCED were chosen for analysis for each sample fortwo reasons: (1) smaller than 7.0 µm includedaggregate particles of both mixture componentsthat produced spectral mixtures that were notidentifiable and (2) the particle size distributionsdid not have a significant number of particlesbelow 7.0 µm CED.

Of these tagged particles, 3000 per sampleanalysed were targeted for Raman spectroscopy, togive a morphological and chemical analysis time ofapproximately 18 hours in total. Point spectra foreach of the three components present (saccharin,sucralose and dextrose) were generated using puresamples of each. Comparing the gathered particlespectra against these reference spectra enabledthe robust chemical classification of individualparticles and the generation of particle sizedistributions for the specific componentpopulations within each blend.

Scatterplots of the Raman spectra correlationscores for each sample show the two distinctpopulations present. A correlation valueapproaching 1 is indicative of a close match to thereference spectra for each component. Mid-rangecorrelations therefore suggest the presence of acertain amount of aggregated material (Figure 3).

Particle size distribution data for the twosamples shows that the particle size of thesaccharin (data not shown) and sucralose (seeFigure 4) is quite different from that of the bulk.

In addition, as before, the specificity afforded bythe MDRS technique enables the detection,differentiation and quantification of a very lowlevel of sweetening agent in a high-volume bulkingmaterial in both samples (Figure 5), in a way thattraditional bulk Raman analysis would not be ableto achieve.

Looking aheadThe growing prevalence and sophistication ofcounterfeit pharmaceuticals presents a growingchallenge to the pharmaceutical industry, posing adirect threat to reputation, health andsustainability. Robustly identifying counterfeitscan be difficult and accessing the informationneeded to trace the product back to its origin isoften even more complex. Analytical techniquesthat deliver the necessary insight in a cost-efficient manner are therefore essential.

The combination of automated particle imagingand Raman spectroscopy makes MDRS a reliableand efficient tool for the investigation of

counterfeit pharmaceuticals, and one that offersenhanced sensitivity to low level contaminantsrelative to bulk Raman spectroscopy.

MDRS allows criminalists to penetrate thesuperficial appearance of counterfeit drugs toreliably identify individual components within theproduct and gather both chemical andmorphological data to establish their provenance.In this way, MDRS can provide the evidenceneeded to successfully prosecute counterfeitingoperations and safeguard both industry andsociety against the threat to health and innovationthat counterfeiting poses.

Fig. 5 The high specificity of MDRS makes it possible to detect and quantify very low levels ofsweetener in a blend

Fig. 4 Particle size distribution data for individual components in the blend shows that sucraloseis present as a much finer powder than the bulk dextrose; the dextrose distribution (green) is ontop of the overall distribution (red), which reflects the fact that the majority of the mixture iscomposed of dextrose

FOR MOREINFORMATIONDeborah Huck-Jones, PhD,Product Manager, AnalyticalImaging, MalvernInstruments(www.malvern.com), AndrewKoutrakos, PhD, student atUniversity of Verona, andBrooke Kammrath, PhD,Assistant Professor ofForensic Science, University ofNew Haven.

Fig. 3 Scatter plots of the particle correlation values for samples A and B highlight the ability ofRaman to reliably differentiate the two components present in the sweeteners