Carryover and contamination

      Sample carryover   is a major problem that can influence the accuracy and precision of high performance liquid chromatography (HPLC), liquid chromatography-mass spectrometry (LC-MS), and liquid chromatography-tandem mass spectrometry (LC-MS/MS) bioanalysis, with the consequences being more pronounced at lower concentrations.¹ The continuous increase in sensitivity of new-generation LC-MS/MS instruments, with detection limits in the low pg/mL range and the possibility of using wider calibration ranges (>10⁴), has also drastically increased the risk of carryover during bioanalysis.² Reduction of carryover during assay development consumes time and resources and can lead to reduced productivity and delays in the drug discovery and development process.^3,4
Carryover in general is serial in nature and is caused by residual analyte from a sample analyzed earlier in the run. It does not necessarily involve only the next sample in the sequence and can affect several samples in a sequence, if many samples above the calibration ranges are analyzed. Carryover can also be random, where carryover from late-eluting residues on chromatographic columns may affect chromatograms several samples later. Carryover from analyte residues can also occur via dislodgment from a sample’s flow path through a chromatographic system and mass spectrometric detection system.
Contamination, conversely, tends to be more random, and precautions should be taken to avoid contamination during sample preparation techniques (extraction) using both manual and automated procedures. The potential for contamination and carryover is highly dependent on the calibration range selected for a given assay.
Carryover and contamination can affect both the accuracy and precision of a method and should be investigated and minimized or eliminated during method development, assessed during method validation, and monitored routinely in study samples analysis. It is critical that unexpected or random carryover and contamination not go unchecked. Unless this random carryover and contamination occurs in samples with known analyte concentrations, such as calibration standards, quality control samples, or placebo/predose samples, the contamination will go undetected and potentially erroneous results will be reported for individual samples, or an entire bioanalytical batch. When blanks or low-concentration samples follow, or are in close proximity to, high-concentration samples, there is a potential risk of contamination and carryover. This article will review the potential risks of carryover and contamination during 3 stages of a bioanalytical method (extraction, chromatography, and detection) and provide some important considerations that should be used to assess and prevent them.

Carryover and contamination from a chromatographic system can be caused by residues of a previously injected sample that are absorbed on, or trapped within, the autosampler. Carryover can also be caused by residues on columns that may randomly affect chromatograms several samples later. There are many publications that describe measures to deal with autosampler carryover, but only a few discuss column carryover. This section discusses autosampler carryover, the origins of carryover, and the means to overcome issues associated with column carryover.

The primary causes of HPLC carryover can be divided into 2 categories: autosampler carryover and column carryover. Autosampler carryover results from the residue of a previously injected sample absorbed on and/or trapped in the autosampler needle, injection port, transfer tube, sample loop, or injector valve. Typical autosampler carryover has a similar retention time to that of the analyte. This often introduces a positive bias (% relative error) and consequently has a major impact on the accuracy of quantitation, most significantly at lower analyte concentrations. Column carryover, however, can be caused by the residue of a previously injected sample on the column, both in its original form and occasionally in different forms of the analyte (eg, analyte:reagent adducts and analyte dimers)⁸ that can decompose in the ion source back to the original form of the analyte. Typical column carryover has uncertain analyte retention times and often generates random error that affects mainly the method precision.

Autosampler carryover is largely associated with the interaction of an analyte with the flow path components of the system; it has a close relationship with the chemical/physical characteristics of both the analyte and the analysis system. Analysis of extremely basic and hydrophobic compounds can be particularly problematic, because of their tendency to be present in a charged form and to adsorb to the sample path of an autosampler through ionic interaction with metallic surfaces and through hydrophobic interaction with plastic materials.⁹ Great efforts have been made by scientists and engineers to reduce carryover in 2 ways: by removing it by rinsing, and by preventing it in the first place.^9-14 Rinsing can be effective, but selection of the most effective rinsing solution, optimized for time, is no trivial matter. Rinse solution chemistry can have a huge impact and should be carefully considered to best counteract carryover. “Like dissolves like” is the primary rule to follow. Generally speaking, acetonitrile or 90% acetonitrile is an acceptable choice for rinsing/removing analytes adsorbed by hydrophobic interaction (eg, lipophilic compounds). A more protic solvent, such as methanol or 90% methanol, is an alternative for more polar lipophilic compounds. Acidified acetonitrile, alkalized acetonitrile, or methanol/isopropanol/water solution is quite efficient and universally used to dissociate analyte adsorption caused by dipole-dipole and ionic interaction (hydrophilic compounds).
Matching the pH to the organic/water or buffer ratio of the rinsing solution can dramatically reduce carryover since the pH of the rinsing solution influences the analyte charge state. For example, a basic compound exists in a positively charged state under acidic and neutral conditions and is uncharged in alkaline conditions. An acidified organic/water or an alkalized organic needle/valve wash solution is useful in removing it, but selection of an acidified organic or alkalized organic/water solution will greatly compromise the rinsing effectiveness. This effect occurs because when charged (ionized), a basic compound easily dissolves in organic/water or acidified organic/water solutions. However, in an uncharged state, it has more affinity toward pure organic or alkalized organic needle/valve wash solutions.
The pK_a of an analyte is a good indicator that should be considered when making pH adjustments to the needle/valve wash solutions. For an analyte that is hard to dissolve in common solvents (methanol, acetonitrile, or aqueous mixtures thereof), strong solvents such as tetrahydrofuran, dimethylsulfoxide, or a halohydrocarbon (eg, methylene chloride) can be used. Use of such strong solvents can, however, cause nonmetallic tubing to swell, which greatly reduces the rupture pressure of the tubing and should be avoided under ultra performance chromatography (UPLC) conditions. An ion pair reagent such as perchloric acid can be used as a rinsing solution, to reduce sample adsorption caused by ionic or coordination interactions, but the possible effect of the counterion should be considered in MS-based assays, as it may suppress ionization. Also, the introduction of any nonvolatile ion pair reagents into the MS system must be avoided.
Most modern autosamplers are equipped with 2 or more needle- and valve-wash lines, allowing multiple rinses to be performed. The first rinsing solution removes analyte residues and involves a weaker solution or mobile phase. The last rinsing solution has better compatibility with the detection system. If only 1 needle- or valve-wash for the autosampler is available, the options for selecting suitable rinsing solutions are more limited, and the compatibility of the rinsing solution with the mobile phase must be considered.