System suitability failure and a growing degradate in a fixed-dose combination assay
An RP-HPLC assay for a fixed-dose combination tablet (drug A 500 mg, drug B 50 mg) uses a 250 × 4.6 mm, 5 µm C18 column, methanol : 0.02 M phosphate buffer pH 3.0 (55:45, v/v), 1.0 mL/min, 240 nm, 20 µL injection. On a new column lot the system suitability run gives resolution between A and B of 1.2 (limit not less than 2.0), tailing factor for B of 2.4 (limit not more than 2.0), %RSD of five replicate injections 0.8% (limit not more than 2.0%) and plate count 2,100 (limit not less than 2,000). The analyst raised the buffer pH to 3.4 to 'improve the tailing' and wants to proceed. In a 24-month stability sample, a new peak at relative retention 0.85 has grown from 0.04% to 0.6%, the assay of B has fallen from 99.2% to 94.1%, and the assay of A remains 99.5%. The peak at RRT 0.85 has the same UV spectrum as B.
1. Judge the analyst's pH adjustment. What changes are permissible within method-adjustment rules and what would require revalidation? What is the likely cause of B's tailing and poor resolution?
Read the worked answer
The analyst's change is not a permitted adjustment and would not have been allowed even if it had worked. Method-adjustment rules such as those in USP <621> allow small, defined changes - typically pH of the aqueous component by about ±0.2 units, buffer concentration by ±10% and the organic component by up to ±10% relative - without revalidation, provided system suitability still passes. A 0.4-unit pH shift is outside that window and changes ionisation, retention and possibly analyte stability, so it needs revalidation of specificity, linearity, accuracy, precision and robustness.
More importantly, the run that failed system suitability is invalid: resolution 1.2 against a limit of not less than 2.0 and tailing 2.4 against not more than 2.0 mean no result from that run can be reported. The most likely cause is column variability, not pH. Tailing of a basic analyte at low pH is the classic signature of interaction with residual silanols, and a new column lot with different end-capping or carbon load changes both tailing and selectivity. Before blaming the column, check buffer preparation and pH meter calibration, equilibration volume (at least 10-20 column volumes), column temperature and sample solvent strength. Then try a low-silanol, end-capped column of the same nominal chemistry; changing the column chemistry is a method change requiring revalidation, whereas replacing a lot of the same chemistry is not.
2. Is the 24-month result a real loss of B or an artefact? Explain what peak purity and mass balance would show, and how you would demonstrate that the method is stability-indicating.
Read the worked answer
The peak at RRT 0.85 has the same UV spectrum as B, so it is almost certainly a B-related degradate, and the fall in B from 99.2% to 94.1% is consistent with real degradation. It does not add up on its own, however: 0.6% of a single degradate cannot explain a 5.1% loss. Possible explanations are other degradates elsewhere in the chromatogram, a volatile or unretained degradate, losses to the column or filter, a response-factor mismatch, or a standard or integration error.
Mass balance is the test: the sum of assay plus all detected degradates should lie within about 95-105% of the initial value, and anything outside that range means degradates are being missed. Peak purity by photodiode array on both the B peak and the RRT 0.85 peak establishes whether either is composite. If the degradate co-eluted with B, the assay of B would be falsely high, not low, so a low assay argues for genuine loss.
Demonstrating that the method is stability-indicating requires forced degradation - acid, base, oxidation, thermal, photolytic and humidity stress - showing that every degradant formed is resolved from both main peaks and that peak purity is maintained, plus confirmation that the degradate's response factor is known or bounded.
3. Set out the forced-degradation experiments and the specification decisions you would make for the RRT 0.85 peak.
Read the worked answer
Forced degradation should aim for 5-20% degradation of each drug, with a placebo and an unstressed control: 0.1 M HCl and 0.1 M NaOH at 60 °C for a defined time, 3% hydrogen peroxide, dry heat at 105 °C, ICH Q1B photostability (option 2) and humidity stress. Each chromatogram must show resolution of every degradant from A, B and from each other, and the degradate at RRT 0.85 should be isolated or characterised if it exceeds the identification threshold.
The specification decision rests on ICH Q3B. For drug B at 50 mg per day the reporting threshold is 0.1%, the identification threshold 0.2% and the qualification threshold 0.5% or 5 mg total daily intake, whichever is lower. At 50 mg per day, 0.5% corresponds to 0.25 mg, so the percentage limit governs. A degradate growing to 0.6% therefore crosses the qualification threshold and cannot be accepted as an ordinary unspecified impurity; it must either be qualified by toxicology or limited below 0.5% with a shelf-life justification supported by degradation kinetics.
The method must then be revalidated for the degradate (linearity, LOD/LOQ, accuracy, precision), the specification and stability protocol updated, and the change filed with the regulator as a post-approval variation. Releasing the 24-month data on the failed suitability run would be indefensible.