The Critical Impact of Sample Management in Effective Testing

The Critical Impact of Sample Management in Effective Testing
by RichardCrowley and Jeff Stassi

Conductinga comprehensive testing program poses several challenges, among the leastpublicized of which is a well-documented sample management process. Samplemanagement is a collection of procedures from pre- through post-analysis that iscritical to the success of the project. Testing programs can range from thesimple (single product) to the extremely complex (nationwide survey of a varietyof foods). There are many factors that impact the complexity of the testingprogram (for a list, see page 68). Poor planning in any of these areas canresult in excess or inappropriate testing, unnecessarily driving up costs andpotentially threatening regulatory compliance.

Depending on the product and goal, national, regional or localsampling may be desired. If the goal is to determine the relationship betweengrowing area and nutritional content, a national sampling design is appropriate.For example, sodium levels in the soil or water vary substantially in differentparts of the country, and the levels of active components found in botanicalingredients from different geographical regions can also vary dramatically.National sampling also serves an important purpose for chain restaurants orcompanies with manufacturing plants in different geographic regions wherepreparation techniques or ingredient sources may vary. When determining theoverall nutritional value of a product, a national sampling scheme can eliminatepossible local biases and provide a statistically significant sampling.

On other occasions, regional sampling is done to confirm datagathered nationally and to detect changes in values over time. Products from individual stores may be tested to determinelocal compliance with national standards. In addition to geographic-basedsampling plans, time-based sampling may be considered when a new product ormanufacturing process is created. During a test run, samples from the productionline are taken at specific timepoints and analyzed for marker compounds toassess variability.

It is important that the number of samples tested providesdata that are both representative and statistically significant. Too few samplesmay not allow for natural variation between products and too many mayunnecessarily raise costs. In general, the Food and Drug Administration (FDA)and U.S. Department of Agriculture (USDA) specify that at least 12units from three production lots be composited for sampling if the data will beused for labeling. In a national or regional sampling scheme, a consistentnumber should be taken from each location. In cases where the study outcome isextremely critical (catastrophic consequences may occur), it may be necessary tocollect and analyze a large number of samples in order to minimize risk.Conversely, only a small number may be needed to reach a valid decision on aminor processing change.

Although the collection of samples may seem relatively benign,a prime concern is that no bias be introduced in the selection of samples. Forthis reason it is often recommended that an independent party be responsible forsample collection. To provide valid data, national samples should be collectedsimultaneously and coordinated with the laboratory to assure analysis is donepromptly to maintain sample integrity.

Good Laboratory Practices (GLP) require that the tracking ofsamples be done in such a manner to ensure chain of custody is maintained.This ensures each use of the sample is documented and can be traced to aspecific analyst, date and time.When samples arrive at the laboratory, they arecarefully unpacked, matched to the information on submission forms, and enteredinto a Laboratory Information Management System (LIMS) database for tracking andgeneration of reports. In addition, the GLPs specify that procedures are inplace for the appropriate handling of the test and control articles. Although not all studies are required to follow FDA GLPs, theprinciples of good science mandate similar processes.

Sample Preparation

Most samples need to undergo some transformation before theycan be analyzed. Obviously, you cant just stuff a ginseng root into a testtube or inject some broccoli into a gas chromatograph. Sample preparation is theprocess of transforming a sample into the form required for the type ofanalytical testing being done. This can be a multiple-step process involvingboth physical and chemical actions and is especially important when making arepresentative composite from a large number of samples.

Effective sample preparation must also ensure that nutrientssusceptible to breakdown are protected and that the process does not contaminatethe sample. For example, the determination of vitamin contentoften presents special problems due to the potential for isomerization anddegradation. Mineral analysis can be difficult due to elements found in very lowconcentrations, and testing for elements considered toxic may require limits ofdetection at trace levels. As a result, extreme care must be taken to avoidcontamination from equipment, reagents or the laboratory environment. Becausemany samples are of an unknown composition and may be limited in amount, caremust also be taken to prevent compromising or wasting the sample.

The goal of the process is to produce a homogeneous,representative sample. First, all special instructions should be followed, suchas removing the inedible portions or cooking the sample.Next, the sample must betransformed into a usable form for testing. This usually involves grinding, mixing or blending. Thephysical and chemical characteristics of the sample are a key factor in theprocess used. For example, samples low in moisture content are ground, whereassamples with high moisture content are homogenized in a blender. Care should beexercised in preparing samples with high moisture content to avoid generation ofheat and loss of moisture, which may result in undesirable changes.

Over the years, some unique methods of sample preparation havebeen developed. This is a result of the wide variety of samples analyzed, theexpanded range of components of interest, and the problems encountered inpreparing a homogeneous sample. For example, the use of cryogenics was developedfor products such as candy bars that are difficult to prepare. When usingtraditional sample preparation methods, the heat from the mixing process wouldcause the fats to separate from the other ingredients. Freezing the samples to a temperature of minus 200 degreesCelsius with liquid nitrogen stops the separation of the fat, thereby providinga better consistency.

Most testing procedures require additional sample preparationin the laboratory. In general, the need for these additional steps depends uponthe analyte of interest, matrix and instrumentation. Several techniques areroutinely used, the most common of which is extraction. Incomplete extraction isone of the primary causes of inaccurate test results. Extraction results in two phases: the extract phase, whichcontains the solvent and dissolved component, and the residual phase (i.e.,raffinate) which contains the dilutent (i.e., undissolved sample) and a smallamount of solvent. In some cases more than one extraction is done. For example,a fat-soluble analyte could be extracted with the fat and then separated using avariety of chemical methods. Separation is generally done using preparativechromatographic techniques. In chromatography, the purified sample is mixed witha mobile phase and then injected into the instrument. This mobile phase isdesigned to facilitate passage through a stationary phase, when samples areseparated.

Other steps such as digestion, concentration, derivitizationand deproteinization may be needed for some assays or matrices.

Careful documentation and handling of a sample is necessary toensure accurate results. Assay-specific techniques are continually evolving asstandards for precision and sensitivity become more demanding and a growingrange of matrices and components pose new challenges. Experience and judgmentare key factors and solid study design can eliminate hurdles to the reporting ofregulatory compliant data. The best scientists and instrumentation in the worldare of little value if the integrity of the sample is compromised beforeanalysis.

Richard Crowley and Jeff Stassi are with Covance Laboratories,a full-service analytical laboratory serving the dietary supplement andfunctional food industries. For more information, visitwww.covance.com/analytical.