ADV01-05_1-298

Figure 2

A single, modified USP procedure for separating structurally similar narcotic analgesics and

acetaminophen on an Ultra C18 column.

Figure 1

Chemical structures of narcotics and acetaminophen.

• 6 •

www.restekcorp.com

800-356-1688

Sometimes methods described in the United States

Pharmacopoeia (USP), the European

Pharmacopoeia (EP), the British Pharmacopoeia

(BP), or other compendia do not provide the

desired robustness in separation or reproducibility,

or results barely pass system suitability require-

ments. Modifications can be made to improve the

methodology, and the results compared statistically

to the original. To improve analysis efficiency and

reduce costs associated with revalidating and test-

ing, it may be desirable to create and validate a sin-

gle analytical method for a range of similar drug

products.

Many narcotics are very similar in structure, often

varying by only a single substitution. Morphine,

codeine, hydrocodone, and oxycodone are quite

similar, for example (Figure 1). Some of these

closely related compounds—all but morphine, in

fact—might be blended with other analgesics, such

as acetaminophen (APAP). USP 25 describes more

than 7 different methods to test these raw materials

and admixtures; some of the older methods do not

use HPLC as a primary test for purity.

One of the chromatographic applications in USP 25

is for the analysis of oxycodone raw material. After

reading the mobile phase section, we saw some

potential problems with the method, including:

1) The use of methanol in this analysis could lead

to high background absorption and loss of linear

range, because the analytical wavelength is 206nm,

and the UV cutoff for methanol is 235nm. In

extreme cases this also can reduce sensitivity—the

more energy the background absorbs, the less is

available to the analyte.

2) An ion-pairing agent (hexane sulfonic acid) is

introduced into the mobile phase without a buffer

to maintain pH. This could lead to widened peaks,

tailing peaks, and retention time drift.

3) Triethylamine (TEA) modifier is included in the

method. When basic compounds are analyzed on

older-type HPLC columns, TEA often is added as

competing base, to reduce the tailing caused by

acidic silanol activity. If the analytical species are

neutral, or have been “neutralized” by an ion-pair-

ing agent, TEA should have no beneficial effect.

Adding TEA, a base, to a mobile phase containing

sulfonic acids will cause acid/base neutralization,

producing a salt and water and reducing the effec-

tive concentration of the acidic ion-pairing agent.

This could lead to the formation of undesirable side

products in the mobile phase that also will absorb

in the low UV range, creating noisy baselines.

Furthermore, TEA is volatile, and its composition

might change over time if the mobile phase is

sparged.

Thus, some aspects of the method appear redundant

and some might actually compromise the separation.

In addition, some of the reagents, such as TEA, might

not be necessary for modern columns. After per-

forming the USP 25 method as written, we made

some tests to determine actual needs to achieve the

system suitability requirements as specified.

With peak shape, separation, and proper analytical

technique in mind, we attempted to eliminate some

of the perceived problems. We realized that by

using 284nm as the detection wavelength, rather

than 206nm as used in USP 25, we might not see

some impurities, but in real life the material should

be tested against some known source for potency.

(Note that with the additional reagents removed,

both Ultra C8 and Pinnacle II

™

C8 columns provid-

ed good results at the 206nm wavelength.)

Next we removed the ion pairing agent and the TEA.

We elected to keep a 20 mM phosphate buffer sys-

tem to maintain a pH of 2.5. Then we reduced the

temperature from 35°C to 27°C, to determine

whether the greater mass transfer and analyte solu-

bility in the mobile phase at 35°C had been masking