When LC-MS/MS response falls during a batch, compare repeated controls, absolute analyte and internal-standard (IS) areas, their ratio and chromatography, then select a test that separates system response from sample-dependent loss. A declining area is an observation, not a diagnosis: matrix effects, sample instability, injection problems and instrument changes can produce overlapping patterns.
Manufacturer guidance groups signal instability into sample and material preparation, method behavior and instrument issues.1 This guide addresses a method that previously performed acceptably and now loses response during a sequence. It is a diagnostic workflow, not a replacement for the method’s established run-acceptance procedure; keep affected results under review while you establish the extent and cause of the problem. Figure 1 maps where along the path from vial to acquisition response can be lost and which comparison probes each point. For the matrix-effect background, see the LC-MS ion suppression and matrix effects guide. In this guide, QC means quality control.
LC-MS PK Peak Integrity Engine
Interrogate carryover, matrix effects, ion suppression, and integration integrity before your PK data goes to QC.
What should I record before changing the instrument?
Keep a copy of the acquisition and processing methods, sample order, original integrations and instrument logs. Identify the last acceptable control and the first suspect result. Record sample preparation time, autosampler residence time, vial position, reagent lots and any intervention. A trend through samples of different concentrations cannot by itself measure instrument drift; use repeated comparable controls.
The workflow below is summarized in Figure 3. Its control comparisons are independent tests selected by the question, not a requirement to run every test in order.
Preserve evidence: record the pattern before intervention
- Save the sequence, chromatograms and instrument logs.
- Review analyte area, internal-standard area and their ratio.
- Record retention time, peak width and pressure.
Compare controls: independent tests selected by the question
- Compare a fresh neat standard with the last acceptable control.
- Compare post-extraction-spiked matrix with a concentration-matched neat standard.
- Compare an aged aliquot with a freshly prepared aliquot.
A neat-standard comparison probes overall response, the post-extraction comparison probes matrix-dependent response and the aliquot comparison probes sample-handling loss. None uniquely identifies a failed component. Match concentration, solvent conditions and injection settings where they matter. When preparation differs, record that difference rather than attributing the entire response change to one mechanism.
Correct and verify: act on the supported cause
- Change one supported factor at a time.
- Repeat controls and representative matrix samples.
Check response, chromatography and low-level performance against the method’s established criteria before resuming reportable analysis.
Can a stable internal-standard ratio hide sensitivity loss?
Yes. This follows directly from the ratio: if analyte and internal-standard signals are multiplied by the same response factor k, that factor cancels.
Ratio = (k × Aanalyte) / (k × AIS) = Aanalyte / AIS for example, (0.5 × 80,000) / (0.5 × 100,000) = 40,000 / 50,000 = 0.80
A stable ratio therefore does not demonstrate unchanged absolute signal or adequate low-level performance: in the example both areas have halved, yet the ratio is still 0.80. Conversely, a falling ratio can reflect analyte loss, an internal-standard increase or differential response; inspect both areas before interpreting it.
An internal standard cannot be assumed to correct every failure. Its addition timing determines which preparation and storage steps it experiences. A coeluting stable-isotope-labeled standard may track shared response changes, but adequate selectivity and low-level precision still need to be demonstrated. See SIL internal standards: what they correct for the relevant assumptions.
Table 1 distinguishes useful observations from conclusions they cannot establish on their own.
| Observation | Useful comparison | What remains unresolved |
|---|---|---|
| Analyte and IS areas decline together; ratio stable | Repeated neat control and matrix control | Shared suppression, injection changes and system decline can overlap |
| Matrix controls decline; neat controls remain stable | Concentration-matched post-extraction comparison | Matrix dependence does not identify the suppressing component |
| Older aliquot responds less than fresh aliquot | Matched concentration, preparation and injection conditions | Degradation, adsorption and preparation variation still need separation |
| Height declines; area is similar; width increases | Original chromatograms and integration boundaries | Broadening may explain height loss; integration can conceal area loss |
| Retention shifts and apparent area falls | Broader diagnostic acquisition and verified valve timing | Scheduled acquisition or diversion may miss part of the peak |
| Blank response follows a high sample | Controlled high-sample/blank challenge | Carryover location is not established by sequence association alone |
How do I separate matrix effects from a system-wide decline?
Start with a fresh neat control and a representative processed matrix control. Inject them close enough in time to reduce drift as a confounder, and repeat if the result is inconsistent. A stable neat control with a depressed matrix response directs the investigation toward the matrix-containing workflow, but a pre-extraction control also includes recovery and stability losses.
Use a post-extraction-spiked blank matrix and a neat standard at matched final analyte concentration to examine matrix-dependent response without including extraction recovery in that comparison.2 Include an unspiked matrix control to reveal endogenous analyte or interference. Analyte-free matrix may be unavailable; in that case, document the alternate design and its limits. Do not label a lower pre-extraction response as suppression without separating recovery. The practical design is explained in post-extraction spiking.
Post-column infusion can locate retention regions where an injected extract changes an infused analyte signal. Compare a solvent injection with the matrix injection under the same gradient: a solvent-composition effect on the baseline is not automatically matrix suppression. The original experimental work demonstrated compound-dependent suppression, so one infused compound does not characterize every analyte.3
A sample-specific matrix effect and a system-wide decline can coexist. Repeat relevant neat controls after matrix exposure to test that possibility rather than treating the first comparison as permanent clearance of the instrument.
Is the apparent loss really a chromatographic or acquisition problem?
Compare area, height and width together. A broader peak can have a lower maximum height without an equivalent area decrease. Check whether integration boundaries capture the full peak and whether the same processing rules were used. Preserve the original result when investigating an integration change.
Figure 2 uses conceptual Gaussian peaks centered at zero, where A is the integrated area and σ sets the width:
y = A / (σ√(2π)) · exp(−t² / (2σ²)) so the apex height is A / (σ√(2π))
Broadening: A = 1, σ = 2 → height = 1 / (2√(2π)) = 0.5 × the reference height (A = 1, σ = 1)
Response loss: A = 0.5, σ = 1 → height = 0.5 / √(2π) = 0.5 × the reference height, with area also 0.5
Doubling σ at fixed A halves the height; halving A at fixed σ halves both area and height. Values are normalized to the reference. These are model illustrations, not measured peaks.
Verify that the acquisition window and divert-valve timing still cover the shifted peak. If appropriate for the instrument, use a separate diagnostic method with wider or unscheduled multiple reaction monitoring (MRM). Maintain sufficient sampling across the peak; changing the transition list changes cycle time and can change the measurement. Diagnostic-method results are not automatically interchangeable with validated production results. Manufacturer troubleshooting explicitly includes scheduled monitoring, retention changes and diversion timing.4
Inspect the pressure trace, leaks, mobile-phase supply and recent column or fitting changes. Pressure is contextual evidence: normal pressure does not prove correct injection volume, gradient composition or source delivery.
Could the mobile phase or source explain the decline?
A mobile-phase change deserves an immediate controlled comparison. Waters documents low response caused by impure mobile phase and notes that different analytes can be affected differently.5 Prepare replacement mobile phase from suitable fresh solvent and clean containers with the prescribed composition. Prime and equilibrate as required. Change the suspected variable rather than simultaneously replacing the column, solvents and source.
Source contamination, spray problems and autosampler faults are recognized contributors to unstable response.1 Neither visible residue nor a response trend alone establishes causality, and source contamination need not affect every analyte equally. Follow instrument-specific procedures for cleaning and maintenance; improved response after cleaning supports a source-associated contribution but does not distinguish every component altered during maintenance.
If a separate, manufacturer-approved infusion check is available, use it under documented reference conditions. Response obtained at a different flow, solvent composition or source setting is not directly comparable with an LC peak. Never compensate for an unexplained decline by changing voltages until a control passes.
How do I test autosampler residence time and sample instability?
Compare an aged aliquot with a freshly prepared aliquot at matched final concentration. Keep vial type, injection settings and preparation consistent. Where possible, prepare aliquots from a common material and alternate injections so instrument drift is less confounded with sample age. Log temperature and storage history.
Lower response from the aged aliquot supports a handling-related difference, not degradation alone. Adsorption, evaporation, precipitation and preparation error require separate controls. Check whether the IS experienced the same holding period: a freshly added IS cannot retrospectively correct losses that occurred before its addition. Checking that standards are not degraded is also a step in manufacturer troubleshooting.4
What must pass before I restart the batch?
Verify the correction under the restored production method. Table 2 lists the evidence to retain; the required limits come from your method and quality system.
| Verification | Evidence to retain | Decision |
|---|---|---|
| Repeated reference controls | Absolute areas, ratios, chromatography and variability | Demonstrate suitable and stable response |
| Representative matrix controls | Controls covering the suspected matrix or exposure condition | Confirm the problem does not recur in matrix |
| Low-level performance | Relevant low-level QC results and integrations | Demonstrate the reportable low end remains usable |
| Blank behavior | Blanks after a relevant high-concentration challenge | Verify any carryover contribution is controlled |
| Batch impact | Last acceptable control, suspect interval and interventions | Resolve affected results under the laboratory’s procedure |
Do not invent a universal allowable area loss. A control can retain an acceptable ratio while absolute response becomes inadequate. Do not automatically re-report or discard an entire batch from the response trend alone; document the investigation and apply the governing procedure to the affected interval. Escalate unexplained loss, recurrent decline after maintenance or failed manufacturer performance checks to technical support.
Frequently asked questions
Should I clean the source first?
Not as the first step, unless an instrument fault requires an immediate stop. Preserve the sequence, chromatograms and logs and run the relevant control comparisons first. Cleaning may restore response, but it also removes the evidence needed to locate the original problem, and an improvement after cleaning cannot tell you whether the source was the only contributor. If the comparisons point to the source, follow the instrument-specific cleaning procedure, then repeat the same controls so the before-and-after evidence is directly comparable.
Does a passing tune prove the assay is working?
No. A tune measures instrument performance under its own conditions, usually with a clean infused standard. It does not establish extraction recovery, chromatographic delivery, matrix behavior or the production assay’s low-level performance. A system can pass its tune while matrix controls are suppressed, while an aged aliquot has lost analyte or while a shifted peak falls outside a scheduled acquisition window. Treat a passing tune as evidence about one part of the system and keep the assay controls as the decision basis.
Can carryover cause this problem?
Carryover is residual material reaching a later run, and it usually adds unwanted response rather than removing it. Contamination and matrix loading can also impair response, but a declining area alone does not demonstrate carryover. If blanks after high samples show analyte-associated peaks, run a controlled high-sample/blank challenge with fresh blanks and localize the contribution as described in the companion guide, LC-MS/MS carryover localization.
Is a stable ratio enough to release results?
No. A stable analyte-to-IS ratio shows that the two signals changed in proportion, not that the absolute response, signal-to-noise or low-level performance is still adequate. Review the method’s full acceptance requirements, including relevant QC results and low-level performance, rather than using the ratio alone. Where absolute response has fallen markedly, check the lowest calibrator and low-level QC integrations before deciding whether results in the affected interval can be reported under the laboratory’s procedure.
Should I raise source voltages or gas flows to recover the signal?
Not while the cause is unexplained. Changing source settings until a control passes can mask the original problem, alters a validated method and leaves matrix, sample and chromatographic causes untested. First run the control comparisons, then correct one supported factor at a time. If an approved infusion check is used to assess the source, run it under documented reference conditions, because response at a different flow, solvent composition or source setting is not directly comparable with the LC peak.
The takeaway
A falling response during an LC-MS/MS batch is an observation that several mechanisms can produce. Preserve the evidence, look at absolute analyte and IS areas as well as their ratio, compare area, height and width together, then choose independent control comparisons by the question: a fresh neat standard for overall response, a post-extraction spike for matrix dependence and a fresh aliquot for sample handling. Correct one supported factor at a time, then verify response, chromatography and low-level performance against the method’s own criteria before restarting.
References
- SCIEX, “Diagnosing signal instability when processing data in Analyst, MultiQuant, or SCIEX OS software,” knowledge base article, sections on sample and material preparation, LC-MS method instability and instrument issues (accessed October 2026).
- SCIEX, “What is matrix effect and how is it quantified?,” knowledge base article (2026), answer section: post-extraction spiked matrix compared with a neat standard at the same concentration.
- R. Bonfiglio, R. C. King, T. V. Olah, K. Merkle, “The effects of sample preparation methods on the variability of the electrospray ionization response for model drug compounds,” Rapid Commun. Mass Spectrom. 13(12), 1175–1185 (1999).
- SCIEX, “Reference and troubleshooting guide,” page 2, low-sensitivity, scheduled MRM and sample-delivery branches (accessed October 2026).
- Waters, “Low response for one or more analytes in an LC-MS application,” WKB260060, cause and additional information sections.
Further reading
- LabVeda, Ion suppression and matrix effects in LC-MS/MS – the cluster pillar.
- LabVeda, Post-extraction spike: matrix factor and recovery.
- LabVeda, LC-MS/MS carryover localization.
Reviewed against primary and manufacturer sources. Every manufacturer statement on this page is checked against the cited source; the comparison sequence and its interpretation are LabVeda’s diagnostic synthesis, and no pattern is presented as a uniquely validated diagnosis. Numerical examples are illustrative calculations from the equations stated and are not acceptance criteria. For validated or compendial methods, the applicable procedure and regulatory framework take precedence over the general rules of thumb given here. Evidence review: October 2026.