Carbon Measurement Methods for Cleaning Validation (Peer Reviewed)

August 2, 2012
Pharmaceutical Technology, Pharmaceutical Technology-08-02-2012, Volume 36, Issue 8

The authors compare direct combustion with rinse and swab sampling methods.

Cleaning validation provides assurance that the quantity of residual substances collected from equipment surfaces are within permissible limits, helping to ensure quality control and safety in pharmaceutical manufacturing facilities. Three different cleaning validation methods for measuring the carbon in residual samples of various pharmaceutical substances were compared.

The challenges of conducting cleaning validation are documented in the literature. R. Baffi et al., for example, described the diverse analytical challenges arising in validating cleaning procedures for biopharmaceutical products produced by recombinant DNA, in which a broad range of potential residual cellular components and trace levels of detergents must be quantified (1). M.A. Strege et al. described the total organic carbon (TOC) analysis of swab samples for cleaning validation of bioprocess fermentation equipment and discussed accuracy, limits of detection, limit of quantitation, linearity, and precision (2). K.M. Jenkins et al. compared the advantages and disadvantages of multiple methods for cleaning validation, including high-performance liquid chromatography (HPLC), thin-layer chromatography (TLC), spectrometry, TOC, and conductivity (3). A.J. Holmes et al. described the TOC method for measuring residual aspirin on aluminum, stainless steel, painted carbon steel, and Plexiglas (4). The latter two authors describe the swab challenge as noted in the FDA guide to inspections of cleaning surfaces.

TOC-LCPH, SHIMADZU

For cleaning validation using a TOC analyzer, the following types of sampling methods are available:

  • rinse sampling

  • swab sampling with aqueous extraction

  • swab sampling with direct combustion.

These methods were compared using a total organic carbon analyzer (TOC-LCPH, Shimadzu) to measure residual pharmaceutical products and their constituent substances.

Preparation of residue samples

Residue samples were prepared by applying various types of pharmaceutical products and their constituents to stainless steel pots. Compounds with varying levels of water solubility (i.e., soluble, insoluble, and very insoluble) were evaluated to determine how each method performed. The water-soluble substances were dissolved in water and the water-insoluble substances were dissolved in ethanol or acetone, as shown in Table I. Solution concentrations were adjusted to 2,000 mgC/L (i.e., carbon concentration of 2000 mg/L).

Table I: Substances used for residue measurements.

The carbon contents of tranexamic acid (C8H15NO2), anhydrous caffeine (C8H10N4O2), isopropylantipyrine (C14H18N2O), and nifedipine (C17H18N2O6) were estimated by molecular formula. Carbon contents of Gentashin ointment (aminoglycoside antibiotic) and Rinderon ointment (corticosteroid) were determined with the TOC analyzer by adding samples of the ointments directly into a solid-sample combustion unit (SSM-5000A, Shimadzu) since molecular formula for these compounds are unknown.

Each residue sample consisted of a 5-cm2 area on the surface of a pot to which 100 μL of each solution was applied and dried. Thus, there were 200 μg carbon in the sample at each application site.

Rinse-sampling method

In rinse sampling, the final rinse water from the cleaning of a production-equipment unit is used as the TOC measurement sample. This method is suitable for systems that cannot easily be disassembled, such as clean-in-place (CIP) equipment and narrow tubing. Sampling is considered to be difficult if the residues are not soluble in water.

To evaluate recovery of the various substances using this method, 100 mL of pure water was stirred for 15 min in the stainless steel pot that contained a patch of dried sample. TOC measurement was conducted on the rinse solution using a TOC analyzer (TOC-LCPH, Shimadzu) with a high-sensitivity catalyst. The analysis of TOC was by acidify and sparge method. The calibration curve was a 2-point curve using 0–3 mgC/L potassium hydrogen phthalate aqueous solution. A 500-μL injection volume was used. Because the carbon content in each of the residue measurement samples was 200 μg, the theoretical TOC concentration (i.e., if all carbon were to dissolve in rinse water) would be 2 mgC/L. Figure 1 shows the measured TOC concentrations for representatives of water-soluble samples (a, tranexamic acid), water-insoluble samples (b, isopropylantipyrine), and water-insoluble ointments (c, Gentashin ointment). The other samples (i.e., anhydrous caffeine, nifedipine, and Rinderon ointment) have similar profiles to the samples with corresponding solubility.

Figure 1: Total organic carbon (TOC) concentrations for (a) tranexamic acid, (b) isopropylantirine, and (c) Gentashin ointment using rinse sampling. (ALL FIGURES ARE COURTESY OF THE AUTHORS)

For the blank, measurement was conducted in the same way using water in a stainless steel pot without dried sample applied to its surface. The measured blank concentration was subtracted from each TOC concentration and divided by the theoretical value of 2 mgC/L (i.e., the theoretical concentration if all of the sample were to dissolve in the water) to determine the rate of recovery, as shown in Equation 1.

All samples were run in triplicate, and the coefficient of variation values (CV) are shown in Table II along with the TOC concentrations and the recovery rates.

Table II: Measurements using rinse sampling.

Water-soluble tranexamic acid and water-insoluble anhydrous caffeine had high recovery rates, as expected. Moreover, water-insoluble isopropylantipyrine and nifedipine had high recovery rates. However, recovery rates of Gentashin ointment and Rinderon ointment were both low, at less than 20%. Consequently, the TOC rinse method, while acceptable for some substances, is unsuitable for ointments and other similar substances.

Swab-sampling with water-extraction method

Swab sampling with water extraction consists of wiping the inside surface of the production apparatus with a fibrous swab material, extracting the adhering material with water, and conducting TOC measurement of the extract solution. Since the residue is physically wiped off from a fixed area of the surface, sampling efficiency is high. Residues that are insoluble in water, however, are difficult to extract with water. Accordingly, evaluating water-insoluble residues with this method may present similar difficulties to the rinse-sampling method.

Figure 2: Total organic carbon (TOC) concentrations for (a) tranexamic acid, (b)isopropylantirine, and (c) Gentashin ointment using swab sampling with water extraction.

To evaluate the recovery of the various substances using swab sampling with water extraction, the sample applied to the stainless-steel pot was wiped off with a 5-cm2 piece of fibrous swab material, which was placed in a glass jar containing 100 mL of pure water. The fibrous swab material (Texwipe Alpha 10 swab washed in pure water and dried) consists of polyester so that very little organic material is extracted from the swab itself. The residue was extracted by stirring for 1 h, and TOC measurement was conducted using the same equipment and conditions used for the rinse-sampling method. Three replicates of each sample were run. As in the rinse-sampling method, because the carbon content in each of the residue measurement samples is 200 μg, the TOC concentration (i.e., theoretical TOC) in the extraction solution would be 2 mgC/L if all of the sample were wiped off. Representative data are shown in Figure 2. For the blank, measurement was conducted in the same way by wiping the stainless pot, which had no sample applied before conducting extraction. Recovery rate was determined using Equation 1. The results are shown in Table III.

Table III: Measurements using swab sampling with water extraction.

Water-soluble tranexamic acid and anhydrous caffeine had high recovery rates as expected. Moreover, water-insoluble isopropylantipyrine and nifedipine had high recovery rates of approximately 90%. However, recovery rates of Gentashin ointment and Rinderon ointment were both low, at less than 10%. These results show that the TOC water-extraction rinse method is reliable and accurate for some substances, but unsuitable for ointments and perhaps other such substances due to the low recovery rates.

Swab-sampling with direct-combustion method

Swab sampling with direct combustion consists of wiping the inside surface of the production apparatus with a piece of quartz filter-paper swab material, and then conducting measurement using a direct-combustion carbon-measurement system. The swab material with adhering residue is measured directly (i.e., without first extracting with water) in a TOC analyzer using a connected solid-sample combustion unit or module (SSM).

To evaluate the rate of recovery of the different types of substances using this method, paper swab material (45-mm diameter Advantec quartz glass paper QR-100, heat treated at 600 °C for 15 min) was used to wipe the sample adhering to the stainless steel pot and placed in the sample boat, which is then placed in the SSM (SSM-5000A, Shimadzu) connected to the TOC analyzer (TOC-LCPH, Shimadzu). Three replicates of each sample were run. The SSM uses 400 mL/min oxygen as a carrier gas. The calibration curve is a 1-point calibration using 1% C glucose aqueous solution. The total carbon (TC) content on the swab was measured directly by the TOC analyzer. Selected measurement data are shown in Figure 3.

Figure 3: Total organic carbon (TOC) concentrations for (a) tranexamic acid, (b) isopropylantirine, and (c) Gentashin ointment using swab sampling with direct combustion.

Since the carbon content in each of the residue measurement samples is 200 μg, the TC value would be 200 μg if all of the sample were wiped off. For the blank, measurement was conducted in the same way by wiping the stainless pot, which had no sample applied. The measured blank value was subtracted from each TC value, and then divided by the theoretical value of 200 μg using Equation 1 to determine the rate of recovery. The results are shown in Table IV. A high recovery rate of about 100% was obtained for all the substances, regardless of whether they were water soluble or water insoluble.

Table IV: Measurements using swab sampling with direct combustion.

Conclusion

The measurement methods used here and their respective recovery rates are summarized in Table V. When using the rinse- and swab-sampling methods, some of the water-insoluble substances had high recovery rates while others had low recovery rates. It is thought that this may be due to differences in the affinity with which the substances adhere to the stainless steel pot. Accordingly, it is possible that residue evaluation using these methods would be difficult for substances with low recovery rates.

Table V: Summary of measurement results.

In contrast, high recovery rates were obtained for all the substances when using the swab sampling with direct-combustion method, regardless of whether the substances were water soluble or water insoluble. Therefore, this method is considered to be the most versatile measurement method for conducting cleaning validation, especially when multiple compounds are being manufactured in the same vat, if the compounds are unknown, or if there is a possibility the known compounds will decompose into other compounds.

Robert Clifford*, PhD, is industrial business unit manager at Shimadzu Scientific Instruments, 7102 Riverwood Drive, Columbia, MD 21046, tel. 800.447.1227, rhclifford@shimadzu.com. Minako Tanaka is a scientist at Shimadzu Applications Development Center, Kyoto, Japan, minako@shimadzu.com.

*To whom all correspondence should be addressed.

Submitted: Nov. 17, 2011. Accepted: March 12, 2012.

References

1. R. Baffi et al., J. Parenter. Sci. Technol. 45 (1), 7–12 (1991).

2. K.M. Jenkins et al., PDA J. Pharm. Sci. Technol. 50 (1), 6–15 (1996).

3. M.A. Strege et al., BioPharm Intl. 9 (4), 42(1996).

4. A.J. Holmes et al., PDA J. Pharm. Sci. Technol. 51 (4), 149–152 (1997).