Ultraviolet cutoff area and predictive ability of partial least squares regression method: A pharmaceutical case...

Area of Science:

• Analytical Chemistry
• Pharmaceutical Analysis
• Chemometrics

Background:

• The UV cutoff area (COA) is a wavelength band where solvents absorb radiation, potentially interfering with drug absorption spectra.
• While some researchers include COA due to drug-specific peaks, solvent interference can still occur even with blank experiments.

Purpose of the Study:

• To investigate the impact of including the UV cutoff area (COA) in quantitative analysis.
• To demonstrate the negative effects of COA on the predictive ability of linear chemometric methods, specifically Partial Least Squares Regression (PLSR).

Main Methods:

• A case study using pharmaceutical mixtures of Dapoxetine Hydrochloride (DAP) and Tadalafil (TAD) in pure and dosage forms.
• Comparative analysis of two datasets: one including COA and another excluding it.
• Statistical comparison of training sets, test sets, and dosage form sets using t and F statistics.

Main Results:

• Inclusion of COA significantly altered statistical parameters (t and F statistics) for dosage form analysis.
• The predictive ability of the PLSR chemometric method was negatively affected by the inclusion of COA.
• Significant differences were observed between datasets with and without COA, particularly for dosage form analysis.

Conclusions:

• The UV cutoff area (COA) should be avoided in quantitative pharmaceutical analysis, especially for routine quality control.
• Including COA can lead to inaccurate results and reduced predictive ability of chemometric models.
• Excluding COA enhances the reliability and accuracy of analytical methods for pharmaceutical mixtures.