Data Integrity Matters: Unknown, Unintegrated or Undetected Chromatography Peaks
In an earlier blog, I discussed the concerns about extraneous peaks that might appear in LC separations. While there can be many sources for peaks unrelated to the test substances, how do you decide which are legitimate, and which are an indication of poor quality? Which are harmless and which are dangerous? Which are expected and which are from a contamination or a change in manufacturing? This blog will consider impurities that might be in your product, yet not show up in your analysis.
Some LC Basic Science:
First let me set the stage with a couple of caveats. Most QC LC separations use a UV detector, set to a specific wavelength for detection. During method development and validation, the suitability of that wavelength setting is optimized for specificity, i.e. to be sure that the compounds of interest are detected at the setting, that the response of the analytes in question are detected at a suitably low levels, and that the wavelength chosen is a robust choice. This last is important as it preserves the methods capability to detect your compounds of interest, even with small changes in UV wavelength which can occur between detectors on different LC instruments. Does that mean that all possible impurity, contamination or degradation product will be identified in an LC UV method? Of course not! Many kinds of contaminants can be missed in any LC test.
a) Contaminants that are not extracted in the sample preparation
b) Contaminants that have no UV absorbance (for instance most polymers, lipids and even sugars are not detected by UV detection
c) Contaminants that do absorb in UV, but at a wavelength different to the one used for measuring the principle components of the sample.
It is important to note that chromatograms with no extraneous peaks is no guarantee of the absence of unexpected chemicals in the sample, just that the method did not detect anything out of the ordinary.
Searching for contaminants
However other forms of detection can assist in probing the sample for unexpected components.
For many years, using multiple different wavelengths has proven useful to identify additional substances in LC samples. These can be used for two purposes: Looking for chromatographic peaks at alternative UV wavelengths, and for examining the consistency of the UV spectra within a peak. Spectral analysis throughout a peak’s elution enables a scientist to search for additional components with different UV spectra, eluting at the same time as the main compound, something known as ‘coelution’. In its simplest form, a multi wavelength detector might monitor 4 different wavelengths. If the readings across a peak are ratioed, a spectrally homogeneous, single component peak would produce a straight-line ratio, with no spectral disturbances caused by an additional UV absorbing impurity. This technique for estimating “Peak Purity” evolved into utilizing “all wavelengths” of a photodiode array detector, allowing software to identify any spectrally different UV absorbing impurities coeluting under a measured peak.
However, Peak Purity measured by spectral homogeneity can only detect contaminants which absorb UV light, and in the case of coeluting impurities, it must have a UV spectrum sufficiently different to the main peak for it to be found. Spectral Peak Purity cannot be used to accurately quantify impurities but can be a useful indication of the presence of a non spectrally homogeneous co-eluting component at the same retention time. True quantitation requires that the impurity peak is well separated from other components, and that you can calibrate the response using standards of that specific contaminant, or standards with very similar chemistry.
There are detectors which do not rely on the UV absorbing properties of the molecule-under-test but respond to any component in the column effluent except the mobile phase. (1) These can be useful when scouting for contaminants, or during method development, to assess the completeness of a UV based method. However, the sensitivity of detectors like Refractive Index detectors, (used extensively to quantitate sugars and polymers) is almost never sufficient to detect the levels required in most pharmaceutical tests. Other universal detectors include conductivity detectors and aerosol-based detectors evaporative light scattering detectors and charged aerosol detection.
MS as a detection tool might also be described as universal, as every molecule has a mass, but not all molecules can be charged in the gaseous form that MS requires, meaning that contaminants can still be missed. In addition, MS detection is generally used in a more focused mode than UV detection.
An MS detector will be ‘tuned’ to measure product or daughter ions of a very specific mass, performing an extremely selective measurement, which is highly beneficial when measuring in complex mixtures.
This kind of focused MS detection is only useful if you know what contaminants you are looking for and configuring the MS for detecting those known contaminants. MS detection only achieves the ability to “fish” for contaminants if the analyst collects large amounts of data across ranges of masses, all of which needs to reviewed in detail.
Photodiode Array and Mass Spectrometry detection (or any other different orthogonal detector) can help development and QC laboratories scan the separated components, and search for UV or MS peaks at parameter settings different to the main method configuration. Collecting UV and MS spectra contemporaneously can help to:
a) assure the absence of contaminants with similar chemical properties to the main compounds of interest and
b) to give the chemist more identification information about extraneous peaks that appear in a chromatogram.
In the cases of peaks which come from the system, or from the sample preparation, or appear in blank injections (i.e. those we discussed in the last blog that could and should be ignored by the analyst), the UV spectra or the MS spectrum can provide confidence that the source of the peak is correctly identified. Equally, if the spectral information is unusual or doesn’t match a component we expect to see, then the structural additional information from these detectors can help to solve the identity and source of this contaminant. However, as mentioned, it is important to remember that some compounds might be invisible to even MS detectors.
Understanding the limitations of chromatographic detection is important to comprehend the concerns about impurities in drug products, specifically unknown or unexpected impurities not previously identified and characterized during process validation and formulation studies.
As detection capabilities improve, whether due to better sensitivity offered by the separation technique, or to the detection technology, the revelation of previously unseen impurities can represent a considerable challenge. New or raised levels of impurities may create a regulatory burden of additional analytical work and safety assessments, all of which needs to be communicated to Health Authorities. However, deliberately not revising and improving analytical methods, because of a fear for such discoveries, does not serve the safety of the patient. Unfortunately, in today’s regulated businesses, advancing and improving analytical methods is sometime seen as an unnecessary aspiration, especially when the existing, traditional analytical method has earlier ‘passed muster’ with pharmaceutical regulators.
However, the current plans for Method Lifecycle Management, and developing quality analytical methods for life, provides guidance for continuously evaluating method performance, and opens a regulatory framework where analytical methods, like manufacturing processes, can be improved while regulatory documentation and registration burdens from changes to validated methods are reduced. See our Method Lifecycle Management (MLCM) website for more information.
And click this link for information about Mass Spectrometry detectors suitable for QC laboratories running Empower 3.