Diode array scans were taken of two of the 75-μg API samples; one with the expected recovery and one with high recovery. The
scans were not significantly different (see Figures 1b, 2b), indicating that the peaks were related, if not the same. A peak
purity profile for the sample with the expected recovery indicated that the peak was pure. The peak purity scan for the sample
with higher than expected recovery indicated that more than one species was present in the scan (see Figures 1c, 2c). Mass
spectrometry (MS) and nuclear magnetic resonance analyses of the samples were inconclusive. The sample concentrations were
too low for positive identification.
In portions of some samples, the data indicated that one metformin double bond rearranged to form a set of conjugated double
bonds (see Figure 3). The conjugated double-bond system exhibits greater ultraviolet absorbance than metformin where the double
bonds are not conjugated. This finding explains the inconsistent, high swab recoveries, the single HPLC peak, and the diode-array
impure peak composition. The manufacturer and most literature sources cite the nonconjugated version as the chemical structure
for metformin. But, at least one source (7) displayed the alternate structure in Figure 3 as the chemical structure of metformin,
supporting the conclusion deduced from the available data.
Several compounds have been tested that degraded during recovery studies. The compounds were known to be susceptible to either
by oxidation or hydrolysis. Metformin was the first example of a compound that demonstrated increased absorbance as a result
of the cleaning process. Any chemical change during recovery studies will affect recovery results as well as swab assays taken
after routine cleaning. The current study demonstrated the value of establishing the VRL along with swab recoveries for cleaning
The swab process.
The correlation between swab assay results and visible-residue limits (VRLs) for cleaning validation was examined. A review
of previously completed validation studies was inconclusive because swab results were much lower than the more-recently determined
VRLs. A current cleaning-validation study evaluated both swab testing and VRLs. Unexpectedly, high swab results led to an
investigation that showed the value of establishing the VRL in conjunction with swab recoveries for cleaning-validation programs.
Richard Forsyth* is an associate director in Global Clinical GMP Quality with Merck & Co., Inc., WP53C-307, West Point, PA 19486, tel. 215.652.7462,
fax 215.652.7106, email@example.com
Julia Roberts is a research chemist in Vaccine Pharmaceutical Research, Tara Lukievics is a global sourcing project leader in the Global Clinical Supplies Organization, and Vincent Van Nostrand is a research chemist in medicinal chemistry with Merck & Co., Inc.
*To whom all correspondence should be addressed.
Submitted: May 16, 2006. Accepted: May 25, 2006. Keywords: cleaning, validation, visible-residue limits
1. D.W. Mendenhall, "Cleaning Validation," Drug Develop. Indust. Pharm.
15 (13), 2105–2114 (1989).
2. US Food and Drug Administration, Guide to Inspection of Validation of Cleaning Processes (Rockville, MD, Office of Regulatory Affairs, 1993).
3. D.A. LeBlanc, "'Visually Clean' as a Sole Acceptance Criteria for Cleaning Validation Protocols," J. Pharm. Sci. and Technol.
56 (1) 31–36 (2002).
4. R.J. Forsyth, V. Van Nostrand, and G. Martin, "Visible Residue Limit for Cleaning Validation and its Potential Application
in a Pharmaceutical Research Facility," Pharm. Technol.
28 (10), 58–72 (2004).
5. R.J. Forsyth and V. Van Nostrand, "The Use of Visible Residue Limit for Introduction of New Compounds in a Pharmaceutical
Research Facility," Pharm. Technol.
29 (4), 134–140 (2005).
6. R.J. Forsyth and V. Van Nostrand, "Application of Visible Residue Limit for Cleaning Validation in a Pharmaceutical Manufacturing
Facility," Pharm. Technol.
29 (10), 152–161 (2005).
7. National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health,