Collection of a few drops of blood on filter paper from the heel of newborns has been the specimen of choice in neonatal screening laboratories for nearly 50 years.   Robert Guthrie pioneered the use of this sample collection device for detection of phenylketonuria (PKU).  The filter paper, which is prepared from cotton (cellulose) linter, is attached to a form for recording patient information.  The card has historically been called a Guthrie Card or PKU card.   The filter paper is usually marked with dashed circles that define zones for the nursery to apply blood from the heel of newborns.   The volume of a blood-filled circle is around 50-75 µL.   Each card usually has between 4 and 6 circles such that the total volume of blood collected from a newborn is approximately 300-500 µL.  Filter paper for use in newborn screening is an FDA-registered Class II medical device with specifications regarding the volume of blood at 55% hematocrit that will be present in a punched disc of a specific diameter (i.e., specific surface area).  The most common sample sizes obtained for analyses are either a 1/8th- or 3/16th-inch punches, which correspond to blood volumes of 3.4 and 7.7 µL, respectively for a 55% hematocrit standard.

Typically, newborn screening labs are not considered diagnostic laboratories although some diagnostic laboratories do perform newborn testing of dried blood spots.  I have often written on the topic “Screening versus Diagnostics” and it relates to perceived and often real differences between obtaining samples as a fluid in a vial, tube or other container versus a dried specimen.   The issue is not quite that simple so the following discussion is required.  First, the volume of blood per unit area in a dried blood sample (DBS) will vary between newborns due to variations in hematocrit.   Higher hematocrits are associated with higher blood volumes per unit area of the blood spot and are presumably due to less spreading on application.  In addition, lower volumes of blood per punch are present in samples that do not fill the circles and are significantly less than 25 µL. This is presumably due to lack of saturation of the card.  There are other problems associated with the collection of specimens but it is worth noting however that this variation is usually 15% or less, a value often acceptable in many laboratories.  Second, the volume of blood or plasma and urine is very small (<10 µL) as compared to a typical clinical chemistry analysis that utilizes a few hundred microliters.  Therefore methods have to exhibit sufficiently low limits of detection to be able to measure the concentrations of these metabolites.  

What I presented above are some of the limitations often cited for the use of filter paper.  But I do believe they can be overcome for many assays.   The analytical sensitivity issue is overcome by methods with a broad range of sensitivities.   Tandem Mass Spectrometry has worked well in this regard.  The issue of quantification of analytes can be improved by using stable isotopes although it is important to note that these standards are often added to the extraction step of a DBS and not as good as adding stable isotopes to liquid specimens (known as isotope dilution mass spectrometry, IDMS).   One approach that has worked well in newborn screening is to calculate concentrations relative to other metabolites. For example, the ratio of Phenylalanine to Tyrosine (Phe/Tyr) is perhaps a better indicator of PKU and its subsequent treatment monitoring that the concentration of Phenylalanine alone.  It also has less variability due to blood volume.  These ratios, which I call molar ratios, require a non-classical approach for issues such as LODs LOQs etc. There are other advantages to filter paper that are well known such as easier sample storage, reduced shipping cost, improved sample stability, lower collection volumes, and reduced risk of infection that could take several more paragraphs to describe.

Finally, it is worth noting that a few major pharmaceutical research laboratories have begun to investigate the use of dried blood spots and there have been more presentations on the subject recently.  It is likely there will be many new solutions to improve this method of sample collection. The question perhaps should be asked: when, rather than if, clinical laboratories will use dried blood specimens routinely.