High-Quality Glass Improves Confidence in Analytes Measurement

Published on: 

Choosing low-adsorption glass vials helps ensure detection accuracy.

In drug development, production, and clinical use, it is vital to have a complete picture of the array of molecules in a sample, from the APIs and excipients to their degradative by-products and unanticipated contaminants. However, the concentrations of these analytes of interest can vary greatly in a sample.

For example, in a product stability test, a drug developer may find the parent or target molecules in great abundance and yet only trace amounts of degradation products. Similarly, in metabolism studies, the percentage of the parent compounds transformed to bioactive forms (or metabolized for rapid clearance from the body) or into potentially toxic side-products can vary dramatically.

Because even trace amounts of some analytes can signal problems with product stability or impact clinical efficacy and safety, the accurate detection and quantitation of these analytes are critical to the success of a drug development program. One of the most important, but often overlooked, considerations in maintaining such accuracy is the choice of autosampler glass vials. Both the type and quality of glass can significantly impact the quality of your results.

The need to understand glass quality

The glass used in autosampler vials—typically, a first hydrolytic class borosilicate glass—is not completely inert and can react with analytes via free silanol groups on the vial wall, causing adsorption (1,2). The degree of glass reactivity can be described by its expansion coefficient. Glass used for these vials will typically have expansion coefficients ranging from basic-grade 70 type all the way up to high-grade 33 type, the number describing the percentage of reactive free silanol groups present on the glass surface.

Another complicating factor for analyte adsorption is the increased presence of imperfections in lower-quality glass, such as irregular surface structures, scratches, and holes (see Figure 1). Such imperfections can lead to higher surface activity because of increased surface area and, thus, more exposed silanol groups, increasing the risk of adsorption of susceptible analytes.

However, not all analytes are equally susceptible to reaction with free silanol groups. Although systematic investigations of susceptibility factors are few, researchers have identified several common features of molecules that are more easily adsorbed, including the presence of trisubstituted N-atoms or tertiary amines (3–5). Other molecules prone to reacting with glass surfaces are those with strong water bridge binding tendencies, such as compounds with sulfur atoms, multiple acid or base groups, or multiple alcohol functions.

Furthermore, and most importantly, when susceptible analytes are low in abundance, the percentage of analyte adsorbed by the glass surface is greater, relative to that adsorbed when analytes are higher in abundance. This larger percentage of free analyte loss significantly increases the potential for and size of measurement errors.

Analyte sample conditions can impact glass-analyte interactions

Advertisement

Ultimately, for a drug to serve its purpose, it must be able to function within complex biological matrices, such as serum, plasma, or other bodily fluids. This means that to fully characterize a drug, developers must evaluate the compound and its associated analytes in the same matrix or artificial analogue. Unfortunately, components of these matrices can significantly influence how glass surfaces adsorb analytes (6). A matrix can mask molecules and alter their reactivity simply by changing the viscosity of the sample, blocking the analytes’ functional groups, or by altering molecular folding (e.g., in the case of protein therapeutics) (7,8). Drug developers may also find that components within a given matrix can increase the background noise in a mass spectrograph, making it more difficult to detect signals of any analytes-of-interest.

These potential issues are further compounded by the practical realities of an analytical lab, where the time between sample preparation and liquid chromatography-mass spectrometry (LC-MS) analysis can vary significantly for any number of reasons. For example, a sample may be prepared for testing only to be set aside while the lab shifts its efforts to more critical or time-sensitive samples. The longer a sample remains in a glass vial, the more time analytes have to interact with the glass surface. This can complicate efforts to accurately quantify the analyte and increase the divergence of that value from the target or true value.

Case study: impact of glass quality on tricyclic antidepressants analysis

To demonstrate the possible scale of the problem, researchers quantified a mixture of tricyclic antidepressants (TCAs) in glass vials of different quality and sources (9).

The researchers spiked real matrix (serum) with a series of TCAs, including Doxepin, Clomipramine, and Bromoperidol, storing the samples in vials made of 33-type glass, two sources of 51-type glass, and another 51-type glass that they silanized to inactivate the surface. They then performed LC-MS analysis on samples in five different vials of each lot over a time course of 0, 2, 4, 12, and 24 hours to determine recovery and reproducibility.

As shown in Figure 2, the 33-type glass vials offered the greatest reproducibility both from vial to vial and over the entire time course. Over the course of 25 runs of five vials, the recovery values deviated only 13% from the target. Furthermore, the researchers noted perfect results in the chromatogram for Doxepin, with peaks showing the same height and symmetry across all samples.

The reproducibility for both 51-type glass vials was inferior to that seen with the 33-type glass. In one case, the researchers determined the overall variance in recovery was 35% from the maximum, and reproducibility between vials within the batch was also poorer. In the second case, recovery was even less consistent, with the researchers finding a variance of 86% and significant shifts in Doxepin concentration over time. As with the other 51-type vial, they found low reproducibility from vial to vial.

Interestingly, as shown in Figure 3, where the silanized 51-type vials should have provided excellent recovery and reproducibility due to inactivation of the surface silanol groups, the researchers discovered significant variance from vial to vial (up to 86%). Digging further, they determined that the variance was largely due to the silanization reagent causing perturbation of the exact m/z.

In sum, these results highlight the significant variability that can be introduced through the choice of the glass vial.

Quality glass ensures quality results

Across the pharmaceutical pipeline, from discovery and development through to manufacturing and clinical use, the need for accurate and reproducible results is critical. This can be particularly challenging in the detection and quantitation of analytes, where low-abundance contaminants, degradation products, or metabolic by-products can significantly impact a drug’s viability, and the erroneous analysis of a molecule can result in significant risks to both product and patients.

To ensure the depth of understanding of compound behavior and the certainty of analytical results, drug developers need to do everything they can to reduce sources of variability and error. Central to that effort in sample handling is the use of high-quality glass vials.

References

  1. T. Mizutani. J. Pharm. Sci. 70 (5) 493-496 (1981).
  2. J. Jednacak-Bisĉan and V. Pravdic. J. Coll. Inter. Sci. 90 (1) 44-50 (1982).
  3. T. Mizutani and A. Mikutani. J. Pharm. Sci. 67 (8) 1102-1105 (1978).
  4. J. R. Shallenberger, et al., Surface Interface Anal. 35 (8) 667-672 (2003).
  5. W. J. Lokar and W. A. Ducker. Langmuir. 20 (2) 378-388 (2004).
  6. C. H. Suelter and M. DeLuca. Anal. Biochem. 135 (1) 112-119 (1983).
  7. P. Van Dulm and W. Norde. J. Coll. Inter. Sci. 91 (1) 248-255 (1983).
  8. C. J. Burke, et al., Int. J. Pharma. 86 (1) 89-93 (1992).
  9. Schaaf, L., et al., Thermo Fisher Scientific white paper 21833. (2018).

About the author

Detlev Lennartz is sample handling product manager at Thermo Fisher Scientific.