codeine hplc high-assay issues

[peoplehi5]

we are running a codeine assay from cough syrup and are seeing an amplified codeine peak in the samples ...

we are running on an ion exchange column

mobile phase is ACN/water buffer pH 4.1 with TEA

standards are fine

system suit is fine

previous samples run back to back with high samples are fine and are prepared in same diluent as samples with amplified codeine ...

we suspect an additional ion is somewhere and co-eluting with codeine ...

any ideas???

a system flush didn't fix the problem, nor does a new column ...

thank you!!!!!

[krickos]

standards are fine

system suit is fine

previous samples run back to back with high samples are fine and are prepared in same diluent as samples with amplified codeine ...

any ideas???

If I understand you right, you managed to reproduce previously approved samples/batches, standards and system suitability test and "only" the new sample gets high (OOS?=Out Of Specification)in the same run/sequence?

If thats the case I think it is unlikely that you got a sytem error, more inclined to that you have a product error or possibly a dilution error for the high samples.

[Bryan Evans]

Specificity should've been determined during validation for this formulation. You did initiate a (documented) investigation once you observed the OOS result correct???

[peoplehi5]

we did not develop this method, it was transferred in ... yes, the investigation is underway but we are running out of ideas for causes ... no dilution involved and all other targets are normal - no elevated results - only codeine is high by >40% ... high for every sample tested ...

the problem seems to be on the samples, but this has happened before and we thought we had narrowed it down to a bad batch of solvent ... re-preps were fine ... that batch of solvent was disposed of ...

a known good sample (prepped identical to the bad samples, in the same glassware and with the same diluent) has been re-run on the system back to back with the bad samples, and that sample produced good results on this system/run ...

more fresh preps are running ... if they are good, then it is a mystery in the prep ... if they are bad, then we aren't sure what we are dealing with ...

[Bryan Evans]

Ok good - everything is being done correctly. You may want to add a vial of diluent inbetween the samples to see if you are
getting any ghost peaks (though it's usually like up to 1%, not 40% if this is the source of the problem).

Maybe check simple things - like is inj. volume the same throughout
the sequence file? Is the lamp switching to a different wavelength in the sequence file? Are all the calculations correct?

After that - PDA (or DAD) for peak (non) homogenity?

[peoplehi5]

"Ok good - everything is being done correctly. You may want to add a vial of diluent inbetween the samples to see if you are
getting any ghost peaks (though it's usually like up to 1%, not 40% if this is the source of the problem).

Maybe check simple things - like is inj. volume the same throughout
the sequence file? Is the lamp switching to a different wavelength in the sequence file? Are all the calculations correct?

After that - PDA (or DAD) for peak (non) homogenity?"

all the simple things were checked: injection volume is fine and a change would show up in the other analytes/peaks inthe run ... calculations aren't being used at this point, just comparing area counts (6 million target vs. 8.8 million integrated) ... method reloads at the beginning of each run and no changes are inthe instrument set-up in the method ...

new preps are bad, too ... thinking something happened to the diluent prior to prepping on monday ... will check one more instrument and then try a reprep with fresh diluent ... (some were already repreps and the process is LOOOONG)