The darker the colour of the solution of (methyl) ergometrine, the greater the loss of active ingredient. This relation, which is highly significant (r=0.8487 for ergometrine and r=0.8730 for methylergometrine), may have great practical implications in the field. It may constitute a very simple method to identify, with the naked eye, those products with low levels of active ingredient.

We therefore tested the hypothesis that any ergometrine solution for which the colour is different from that of water has a level of active ingredient below 90% of the stated content. In the dilution scale of brown, 9 cannot be distinguished from clear water, but 8 can; with decreasing numbers the colour intensifies. Of the 392 samples tested, 168 were classified as colour 1-8 (that is, colour of the solution is different from clear water). Of these 168 "colour failed" samples, 24 (14%) can be classified as false-positive with over 90% active ingredient (range 90-95%). This implies a specificity of 86%: in 14% of cases the drag would fail while the level of active ingredient is still acceptable. None of the 224 "colour acceptable" samples had a level of less than 90% active ingredient implying a sensitivity of 100%: all defective samples would have been identified with this method.

Similarly impressive results were obtained for methylergometrine. Of 180 "colour failed" samples, 9 (5%) were false-positive (range 90-97% active ingredient), indicating a 95% specificity. Of 212 "colour-acceptable" samples, 20 (9.4%) actually had potencies below 90%, implying a sensitivity of 90.6%. It should however be noted that 15 of these 20 false-negatives occurred in one batch only, of a product with a 82% initial content of active ingredient only. It seems likely that no colour-producing metabolites had yet been formed in this batch. If the three batches with such low initial content are excluded from the analysis, only 5/197 of the "colour-acceptable" samples are false negatives, implying 97.5% sensitivity. Specificity is then 85%,

On the basis of these results we conclude that, both for ergometrine and methylergometrine of good initial quality, any discolouration which makes the solution different from water implies with a sensitivity of 97-100% that the product is below USP/BP standards of 90% of the stated content and should not be used. When this rule is applied, about 15% of colour failures are false-positive.

Injectable ergometrine and methylergometrine are usually presented in brown ampoules. In practice this implies therefore that the solution should be tested in a clear glass vial or test tube, comparing it with water in a similar tube under clear light against a white background.

When the solution is drawn in a syringe for injection and discolouration is visible with the naked eye, the level of active ingredient is probably very low and the product should not be used. However, this way a moderate discolouration (scale 7-8) with expected levels of 60-90% active ingredient is likely to be missed as no good comparison with clear water is possible. The method is therefore much less sensitive and would only identify the most serious cases.

One observation from this and earlier studies^7,8,39 may disturb this apparently elegant method to identify deficient products. That is that individual ampoules from the same box from the same batch kept under the same conditions, may behave differently. The most complete proof emerges from the WHO/UNICEF study⁷, in which all ampoules from the same box of ten from the same batch from the same trip were always individually tested. In one box eight ampoules measured 80-88% active ingredient while one measured 63% and one only 49%. A similar pattern, although less extreme, was observed with methylergometrine. This observation was confirmed in the present study, where sometimes two ampoules from the same batch after the same period under similar climatic exposure showed widely different results in colour and level of active ingredient.

We have no explanation for this phenomenon. If proved true, it would imply that "colour acceptable" for one or more sample ampoules from a box or batch may not necessarily mean that all ampoules in that lot are acceptable. This would of course also apply for analytical testing.