Qualitative Methods Presence or Absence

For qualitative or presence or absence methods, the validation criteria suggested are:

• Limit of detection

• Ruggedness

• Robustness

Standard methods for detection of microorganisms are in themselves variable, particularly when low levels are being enumerated, making critical comparison of standard versus rapid methods particularly problematic.

The first problem is the preparation of a homogenous suspension from which successive representative samples can be drawn. This can be influenced by the type of organism, incubation conditions, length of incubation, media, etc., all of which may cause problems such as clumping of bacteria, chain elongation, or the formation of "gummy" exudates.

Before a comparison is attempted, time should be spent on determining the most consistent method of preparing a homogenous suspension. This may require different techniques for different microorganisms. In our experience, preparing such suspensions for fungi is even more problematic and may require aggressive homogenization techniques.

For existing methods, where presence or absence is based on turbidity, problems include the ability of bacteria to multiply to such an extent, that detection of turbidity may well be a function of the process by which they are removed from the substrate of interest (e.g., a cream); or the way in which any interference from the substrate is overcome. Additional or totally different manipulations needed for a rapid method may well confuse the equivalency.

Accuracy

Bearing in mind some of the issues identified, accuracy can probably only be performed by preparing very low concentrations of the target organisms (1 to 5 cfu/test unit) and inoculating these into a number of containers of the chosen media sufficient to obtain, after incubation, both positive and negative results. Accuracy of the rapid method is then based on providing at least the same degree of recovery, i.e., a similar relative proportion of both positive and negative results.

Precision

Precision can be determined by repeating the exercise on different lots of the same product. Due to the critical nature of presence or absence detection methods, e.g., sterility tests, we suggest that the validation be split into two phases. The first phase would be a comprehensive comparison across a wide range of target organisms, both pharmacopoeial and recent isolates with a range of products (if applicable). The second phase would be an ongoing comparison for a prolonged period or until a predetermined number of comparisons performed side by side during routine usage had been reached. It is difficult to set a time or number against this, but for sterility testing it is suggested that 12 months or 100 separate tests may be appropriate.

Specificity

In this case specificity can be determined either by demonstrating that the rapid method can detect growth in the presence of the test article over a wide range of organisms (suitable for turbidimetric methods), or by demonstrating that the method does not erroneously detect the presence of extraneous matter from the test article and generate a positive result.

Limit of Detection

As for specificity the limit of detection comparison is performed by preparing low level inocula (1 to 5 cfu per test article) and demonstrating that the rapid method is as capable as the conventional method. At such low levels, a number of the replicates should show negative growth. We suggest that a range of organisms should be used, including those from the pharmacopoeia, plus the normal isolates from the test articles under comparison. Considerable replication is required to make this meaningful; not less than 10 replicates need to be performed for each organism used.

Ruggedness

The supplier of the rapid method should be able to supply data on the impact of using different instruments, different analysts, etc. However, this does not preclude the necessity of performing some measure of ruggedness in-house. Once again, the ability to prepare uniform samples significantly impacts on the discriminatory value of the test. The key test variables, e.g., analysts, reagents, time or temperature (where ranges are quoted), could all be challenged to show that under normal conditions such operational variability does not impact on the test's ability to correctly identify the presence (or absence) of a range of microorganisms.

Robustness

As for ruggedness the impact of small but deliberate variations in the method parameters, and their subsequent impact on the comparability of the methods, is probably best left to the manufacturer. As a guide the following would be expected to have been covered:

• Instrument performance over time

• Effects of different mixing times or techniques

• Effects of different incubation time or temperature variations within accepted limits

• Effect of ambient temperatures

• Effect of variability of any dispensing devices used

• Effect of lot-to-lot variation of reagents

• Effect of using reagents at end of shelf life

Much of this data has not been routinely generated for most conventional methods.

Quantitative Methods

It is for these techniques that the vagaries of microbiology, particularly in homogenous test sample preparation, really begin to bite! In order to demonstrate equivalency with the conventional method, the following parameters need to be evaluated:

• Specificity

• Limit of quantification

• Ruggedness or robustness

• Equivalence

Accuracy

Before evaluating this parameter, it is worth reviewing the difficulties associated with determining the closeness of a test result to the true value in a microbiological context. Table 7.1 lists the number of replicates needed to claim a 90% probability that the results between two methods differ from between 10 to 100% for a range of bacterial conditions.

Table 7.1 Number of Replicate Determinations Needed for Enumeration of Bacteria at Different Concentrations

% Difference

1 cfu

10 cfu

30 cfu

50 cfu

100 cfu

300 cfu

10

1887

189

(63)

38

19

6

20

517

52

14

10

5

2

25

345

35

12

7

3

1

50

106

11

4

2

1

1

100

37

4

1

1

1

If a technician wished to claim with a 90% probability, that two methods used to enumerate a suspension containing 30 cfu with an acceptance criteria of no more than a 10% difference in results, he would need to perform 63 replicates! There is little regulatory guidance as to what constitutes agreement, although Ph Eur 2002 (2.6.12) states that where a limit of 102 is given then results up to 5 x 102 may be considered compliant. USP 24/NF 19 (1231) provides data on plate-count enumeration, stating that the error on a count of 3 cfu on a plate from a 10-1 dilution is 58%.

To determine accuracy, cultures of a number of organisms are prepared providing a range of dilutions, i.e., 100; 75; 50; 25; 10% of the original suspension, counts in the order of 30 to 300 cfu over the dilution range. Assuming the lowest count seen is around 30 and with an acceptance criteria of no more than a 25% difference in results, then 12 replicates of each dilution using the conventional and rapid method could be used.

Statistical methods of comparing the two data sets such as students t-test or analysis of variance could be used.

Was this article helpful?

0 0
Lower Your Cholesterol In Just 33 Days

Lower Your Cholesterol In Just 33 Days

Discover secrets, myths, truths, lies and strategies for dealing effectively with cholesterol, now and forever! Uncover techniques, remedies and alternative for lowering your cholesterol quickly and significantly in just ONE MONTH! Find insights into the screenings, meanings and numbers involved in lowering cholesterol and the implications, consideration it has for your lifestyle and future!

Get My Free Ebook


Post a comment