Stability-Indicating HPLC for Peptides

ON THIS PAGE

1. Key findings
2. What does a stability-indicating HPLC method commit to?
3. What did the 2026 study measure?
4. Why does forced degradation matter for a certificate's purity figure?
5. What the ledger does not record
6. How to read the method line on a certificate
7. What stays unknown
8. Sources

A purity figure is worth what the method behind it can see. A stability-indicating HPLC method is the analytical chemist's answer to the hole in that sentence: a method that has been shown, by deliberately damaging the material, to separate the intact peptide from its own degradation products and to measure the intact fraction without them. The phrase is a commitment rather than a description, and a paper published this January in BMC Chemistry shows what meeting it looks like for semaglutide and tirzepatide, two peptides named in 68 of the 126 supplier records in this publication's ledger.

What does a stability-indicating HPLC method commit to?

It commits to one thing: that the method can tell the intact peptide apart from what the peptide turns into as it degrades, and can report the intact fraction with the degradation products resolved from it rather than hidden inside it. Everything else about the phrase follows from that. The commitment is not met by running a clean sample and observing a single peak, because a clean sample has nothing to hide. It is met by making a dirty one on purpose.

The paper describes the procedure as the ICH guidelines describe it, citing the Q1A(R2) stability-testing guideline for the stress conditions and Q2(R1) for validation (the paper spells the body out as the International Conference of Harmonization). In the authors' summary of that guidance, a stability-indicating assay requires forced degradation under oxidation, pH, heat, light "and others", the separation of the studied compound from its degradants "is obligatory", and the degree of degradation "should range from 10 to 20%". The window matters: too little degradation and the method has not been challenged, too much and the sample is no longer the material the method is meant to measure.

The contrast is with the figure most certificates print. As an earlier piece here explains, area-percent purity is the main peak's area divided by the total area of every peak the detector saw, under one method, on one day. That arithmetic has no term for what the method could not see. A stability-indicating method has been made to look.

What did the 2026 study measure?

The study measured how much semaglutide and tirzepatide degraded under five stress conditions, whether the method separated each intact peptide from what formed, and how the method performed on the usual validation figures. The paper was received on August 2, 2025, accepted on December 23, 2025, and published open access on January 27, 2026, as volume 20, article 31 of BMC Chemistry. Every figure below is from the full text as read at PubMed Central on 2026-09-08.

The method is ordinary reversed-phase chromatography, which is part of the point. An Inertsil ODS-3 C18 column, 4.6 by 250 millimeters with 5 micrometer particles; an isocratic mobile phase of 0.1 percent formic acid at pH 2.5 and acetonitrile in a 30 to 70 ratio; a flow rate of 1 milliliter per minute; a diode array detector reading at 220 nanometers. Under those conditions semaglutide elutes at 1.42 minutes and tirzepatide at 1.68. The authors report trying methanol first and getting broad peaks, a C8 column and getting a run of under a minute, and a 3 micrometer C18 column and getting poor peak shapes in long runs, before settling on the column and mobile phase above.

The forced degradation is the section a certificate reader should study. The authors report that strong acid or base, 1 molar of either, higher temperatures and longer exposures precipitated the peptides, "more likely due to denaturation", and that 30 percent hydrogen peroxide precipitated the solution as well, so milder conditions were chosen, and the milder conditions are what appear in the paper's Table 2:

• 0.1 molar hydrochloric acid at 25 degrees Celsius for 10 minutes: 16.90 percent degradation of tirzepatide and 12.00 percent of semaglutide, with one new peak for each.
• 0.1 molar sodium hydroxide at 25 degrees Celsius, 20 minutes for tirzepatide and 5 minutes for semaglutide: 18.97 and 13.00 percent, with two new peaks for each.
• 10 percent hydrogen peroxide at 25 degrees Celsius, 60 minutes for tirzepatide and 20 minutes for semaglutide: 15.91 and 14.80 percent.
• Direct light for 24 hours: 11.84 and 12.35 percent, with no additional degradant peak appearing for either peptide.
• Reflux at 60 degrees Celsius in water, 90 minutes for tirzepatide and 60 for semaglutide: 17.90 and 12.60 percent loss of parent peak area against a standard of the same concentration.

Every figure in that list sits inside the 10 to 20 percent window the authors adopt from the guideline. Where a degradation peak appeared, the resolution between it and the parent peak "did not fall below 2", which the authors take as sufficient baseline separation, and a diode array peak purity test is reported as confirming that the parent peaks stayed homogeneous under every condition. Two of the five rows deserve a second look. Under light, both peptides lost around 12 percent of their peak area and no new peak appeared. The paper offers no account of where the lost area went; on this publication's reading, whatever formed either did not absorb at 220 nanometers or did not elute within the run, so the loss was measurable and the product was not. And under base, semaglutide reached 13.00 percent degradation in 5 minutes at room temperature, where tirzepatide took 20 minutes to reach 18.97 percent.

The validation figures are the ones a certificate usually borrows its vocabulary from. Linearity was established from 1 to 500 micrograms per milliliter for both peptides. Table 3 gives correlation coefficients of 0.9999 for tirzepatide and 0.9998 for semaglutide, and the body text says above 0.9998; the abstract says above 0.9999, which the paper's own table contradicts, so the table's figures are the ones reported here. The detection limits sit between 10 and 16 nanograms per milliliter, and here too the abstract and Table 3 disagree, assigning the two figures to the two peptides in opposite orders, so this piece cites the range rather than pairing a limit with a peptide. Small deliberate changes to the method, 2 percent in mobile phase composition, 0.2 milliliters per minute in flow rate and 2 nanometers in wavelength, moved the measured peak areas by no more than 0.64 percent relative standard deviation. Working solutions in deionized water showed no apparent degradation over 24 hours at room temperature or over a week refrigerated.

The authors describe theirs as the first validated stability-indicating HPLC method for the two peptides with degradation inside that 10 to 20 percent window, on the ground that no published method had quantified the two together in the presence of their degradation products under different stress conditions. Their own introduction cites earlier stress-degradation HPLC work on each peptide separately and faults it for degradation below the window, 0.73 to 8.73 percent for tirzepatide and under 10 percent for semaglutide. That is a priority claim, reported here as theirs and not checked against the literature. The study also quantifies both peptides in two marketed injectable products and in spiked rat plasma, and spends its second half on a set of sustainability metrics; none of that is examined here.

Why does forced degradation matter for a certificate's purity figure?

Because a method that has never been shown to separate the intact peptide from its degradation products can report a purity figure that holds steady, or climbs, as the material degrades. Area percent is a ratio of what the detector saw. If a degradation product elutes under the parent peak, it is counted as parent and the ratio does not move. If it does not absorb at the detector's wavelength, or does not elute within the run, it drops out of the denominator and the ratio can rise. The paper's own light exposure is the small-scale demonstration: around 12 percent of each peptide gone, and nothing new on the chromatogram to show for it. A method not built to look for that loss would have had no reason to report it.

That is why the phrase carries weight on a certificate and why its absence is worth noticing. A certificate stating 99 percent purity by HPLC on a stability-indicating method is saying the method was challenged with damaged material and resolved it. The same figure on a method never challenged that way is saying the main peak was 99 percent of what the detector happened to see. Both are honest statements; only one of them speaks to the material's condition. What a certificate does not tell you lists the analyses a standard purity certificate leaves out; this is a gap inside the one analysis it does include.

It also bears on time. Lyophilized peptides ship as a dry solid for the reasons the storage piece here sets out, and the paper's stress conditions push the same chemistry, hydrolysis at both ends of the pH scale, oxidation, heat and light, far harder and faster than storage would. A certificate describes the material on the day of the run. Whether the method could have seen the difference between that day and a later one is exactly what stability indicating means.

What the ledger does not record

This publication keeps a record on 126 suppliers, each published as its own record page and searchable through the vendor check, built from what each storefront and its certificates say. Read on 2026-09-08, the word HPLC appears in 101 of those records. The phrase stability indicating appears in none of them, in either spelling, and the phrase forced degradation appears in none. That is a statement about the ledger and it should be read as one. The ledger records what suppliers publish and what a reader can verify; it has never put the question of forced degradation to any supplier, so a record that does not mention it is the record of a question never asked rather than an answer.

The gap is specific. Validation language does appear in the ledger: on endotoxin and microbial method lines, in one supplier's statement that its certificate data came from validated in-house analytical procedures, and in a second supplier's product-page claim that a compound was validated by advanced analytical methods, with no method, figure, lot or document attached. What no record carries is a statement that the purity method behind the area-percent figure was validated against deliberately degraded material. That is the checkable question, and it is narrower than asking whether certificates mention validation at all.

How to read the method line on a certificate

Four things to do with this.

1. Find the method. A certificate that names a column, a mobile phase, a detection wavelength and a run time has given you enough to ask the next question; one that prints a percentage and the letters HPLC has not.
2. Ask whether the method is stability indicating. The question to a laboratory is plain: was this method run against forced-degradation samples of this peptide, and what resolution was measured between the parent peak and the nearest degradation peak. A laboratory that has done the work will have a figure; the paper's was at least 2 in every case where a degradant peak appeared.
3. Look at where the parent elutes. The authors' own caveat about their method is that "the proposed method has a short retention time for both TIR and SEM which may cause overlapping of the peaks with poorly retained substances due to poor interaction with stationary phase." Semaglutide at 1.42 minutes is close to the front of the run, and a parent peak that early can share its window with whatever the column did not hold. A certificate's chromatogram, where one is published, shows this at a glance.
4. Ask for the peak purity test. The paper used the diode array detector's spectrum across the parent peak to check that the peak was one substance. A certificate that reports a peak purity or purity angle figure is reporting that check; one that does not may still have run it, and where the laboratory publishes a lookup portal, its own copy of the report can be looked up and checked for it.

What stays unknown

The study is one method, in one laboratory, for two peptides, and it says nothing about any certificate issued by anyone else. Whether the method behind any given supplier's purity figure is stability indicating is a question this piece cannot answer from the ledger, because the ledger has not asked it, and cannot answer from the paper, because the paper is about its own method. The authors state a limitation in their own words, quoted above, about short retention times; they do not report testing their method against the degradation products of any other peptide, and nothing here extends their findings past the two they studied. The abstract's linearity and detection figures disagree with the paper's Table 3 in the two places described, the experimental section reverses the two reflux durations that Table 2 and the discussion agree on, and the figures above follow the tables and the discussion; a reader who checks the abstract or the experimental section alone will find different numbers. The plasma and dosage-form sections of the paper, and its sustainability scoring, were read and not evaluated. And a stability-indicating method on the day of the run says nothing about what a vial has been through since, which no certificate can report.

Sources

Khalil HA, Hassanein NA, El-Yazbi AF, Mahgoub H. A multimodal HPLC stability indicating approach for the estimation of semaglutide and tirzepatide in bulk, pharmaceutical dosage forms, and rat plasma: a six-edged sustainability appraisal. BMC Chemistry, volume 20, article 31, published January 27, 2026, doi 10.1186/s13065-025-01716-7, PubMed identifier 41588418. Full text read at PubMed Central, PMC12918740, and the identifier checked to resolve through doi.org to the publisher's page, both on 2026-09-08. Every quotation and figure attributed to the paper above is from that full text. The two ICH guidelines named are cited as the paper cites them, Q1A(R2) of 2003 for stress testing and Q2(R1) of 2005 for validation; neither was fetched for this piece. Every count of supplier records is from this publication's own vendor ledger, the file behind its 126 supplier record pages, read 2026-09-08.