Abstract

Two numerical methods, Decision Analysis (DA) and Potential Problem Analysis (PPA) are presented as alternative selection methods to the logical method presented in Part I. In DA properties are weighted and outcomes are scored. The weighted scores for each candidate are totaled and final selection is based on the totals. Higher scores indicate better candidates. In PPA potential problems are assigned a seriousness factor and test outcomes are used to define the probability of occurrence. The seriousness-probability products are totaled and forms with minimal scores are preferred. DA and PPA have never been compared to the logical-elimination method. Additional data were available for two forms of McN-5707 to provide complete preformulation data for five candidate forms. Weight and seriousness factors (independent variables) were obtained from a survey of experienced formulators. Scores and probabilities (dependent variables) were provided independently by Preformulation. The rankings of the five candidate forms, best to worst, were similar for all three methods. These results validate the applicability of DA and PPA for candidate form selection. DA and PPA are particularly applicable in cases where there are many candidate forms and where each form has some degree of unfavorable properties.

Introduction

The development candidate form selection model described in Part I was a logical method that facilitated the selection of the free base form on McN-5707 (Fig. 1) from among eight forms. The eight forms were characterized with respect to critical factors for candidate selection. Forms were eliminated from further testing and consideration if a test result was unsatisfactory or problematic. The free base was selected from a final group that included the fumarate and maleate salts. The molecular weight (MW) conversion factors (CFs), i.e., ratio of MWof the salt form to the MWof the free form, were also considered in the final selection. This logical method works well when there are a limited number of candidate forms and the number can be further reduced to a reasonable number for final selection. This method does not lend itself to cases where there is difficulty in selection due to there being many candidate forms, and where there are problematic properties in all of the candidate forms.

Decision Analysis (DA) and Potential Problem Analysis (PPA) appeared to be suitable alternatives to the logical method applied in Part I. In the absence of a published comparison between DA and/or PPA and the simpler logical selection model, we decided to make a comparison of the results from these three models.

DA and PPA are scoring and weighting process models. Walkling and Appino (1) first applied DA and PPA to salt form selection. The concepts for these models are derived from models developed by Kepner and Tregoe (2). DA requires assignment of a weight to the attribute and a score for the result or outcome. It utilizes every available attribute, applied to all candidate forms. The weighted scores are summed and selection is made from among those with the higher total scores. PPA requires assignment of a seriousness factor and probability for each potential problem. PPA simplifies the analysis by reducing the number of properties considered. Potential problems related to a common property are grouped together to reduce the impact. The total sums of the probability-seriousness products are calculated for each form, and final selection is made from among the forms with the lowest scores. Normally selection is made from among the group of higher DA scores and lower PPA scores. The accuracy is not based on highest and lowest scores. DA and PPA may be applied to the same data, and the selection may be based on a consensus of scores from the two procedures. Part I generated complete data sets for the fumarate, maleate and free base forms of McN-5707. In order to create an opportunity for the comparison of the methods, the hydrochloride and phosphate hydrate forms were subjected to additional testing to complete the data set for a total of five forms.

MCN-5707 FORM SELECTION

Previously reported data (3) for five forms of McN-5707, the free base, and the hydrochloride, phosphate hydrate, fumarate and maleate salt forms were used. CFs and missing data for hydrochloride and phosphate hydrate salt forms were acquired by additional testing to create a complete the data set for five forms. The physical-chemical properties used were

1. Organoleptics, color, appearance and CF;
2. Crystallinity and melting behavior;
3. Critical solubilities, water, 0.1 N hydrochloric acid, and Simulated Intestinal Fluid (SIF) pH 7.5 at ambient room temperature (RT);
4. Hygrodynamics, percent weight change after
1. Drying at 49°C/24 hours, and
2. Storage at RT for 1 week at extremes of humidity, 11% RH and 83% RH;
5. Processability
1. Filming and sticking on compression, and
2. Rusting of a meehanite test bar after abrasion with the NCE powder followed by exposure to high RH at RT.

DECISION ANALYSIS

Eight veteran formulators, including bench level and senior managers, weight-ranked the importance (100 points total) of 13 physical-chemical properties, color and appearance, CF, melting range, crystallinity/XRD, gain/loss of moisture at RT/83% RH and at RT/11% RH, physical stability at 30°C/60% RH and at 40°C/75% RH, filmingsticking on compression, bulk density, and solubilities in 0.1 N hydrochloric acid, water, and SIF (pH 7.5). Means were calculated for each of the weight factors. The physicalchemical properties were independently reviewed and scored using a 0 to 10 grade, 10 being the highest score. The respective individual scores were multiplied by the weighting factors to obtain weighted scores and summed to obtain total weighted scores. The salt forms were then ranked highest to lowest. Selection of optimal forms was made from among the higher scores.

POTENTIAL PROBLEM ANALYSIS

The formulators also rated the seriousness (50 points total) of five potential formulation problems, filming-sticking, moisture sorption, XRD and DSC detected problem, solubilities, and salt form. A mean seriousness was calculated for each potential problem. The probabilities (0 to 1, 1 being the highest) of each potential problem were estimated independently from a review of the collected data. The respective probabilities were multiplied by the mean seriousness to obtain seriousness-probability products. The seriousnessprobability products were totaled and ranked from lowest to highest. Selection of optimal forms was made from among the lower scores.

The polling survey was intended to be all encompassing. Some of the physical-chemical properties and potential problems that were rated in the polling of the formulators were not used in the evaluations of this NCE's forms. Rusting potential was omitted from the original poll. It was assigned the same weight as filming on compression because both problems were resolved in our laboratories using the same formulation process.

RESULTS

. . .Continue to read rest of article (PDF).