Application of Visible-Residue Limit for Cleaning Validation

Pharmaceutical plants must have visually clean equipment to operate according to good manufacturing practices. Formulators must visually inspect manufacturing equipment for cleanliness before formulation work begins (1). Manufacturers establish and perform visible cleanliness and analytical methods to ensure regulatory compliance. An analyst conducts a visual inspection and confirms visible cleanliness before taking swab samples for chemical analysis (2). The formulator of the subsequent batch conducts a visual inspection before manufacturing work begins. A correlation between available analytical data and visible cleanliness of manufacturing equipment over an extended period of time can expand the practice of performing visual inspections in lieu of swab sampling.

Yet, the US Food and Drug Administration's Guide to Inspection of Validation of Cleaning Processes (3) and other literature (4) discount using a visible limit as the sole acceptance criterion for cleanliness. Conversely, Mendenhall concluded that visible cleanliness criteria were more rigid than quantitative calculations and are clearly adequate for determining cleanliness (5). Other recent articles describe the justification and application of a visible-residue limit (VRL) (6, 7).

Many research teams have established quantitative VRL levels. Fourman and Mullen determined a visible limit of ~100 μg per 2 x 2-in. swab area (8) or ~4 μg/cm² . Jenkins and Vanderwielen observed residues as low as 1.0 μg/cm² with a light source (9). Forsyth et al. determined <0.4- to >10-μg/cm² VRLs for active pharmaceutical ingredients (APIs) and excipients (6).

The acceptable residue limit (ARL) for drug residue on manufacturing equipment surfaces can be determined on a health- and adulteration-based criterion (2, 9, 10). Toxicity data are the basis for a health-based ARL and cross contamination is the basis of an adulteration-based ARL. Typically the lower of the two limits is used for the ARL. If the VRL is quantitatively established and is lower than the ARL, then the VRL can be an acceptable measure of equipment cleanliness. For this study, the VRL was evaluated for several commercialized solid dosage formulations and late-phase development formulations.

Visible-residue parameters

The following VRL variables were evaluated: formulation composition, surface material, background, solvent effects, light intensity, light angle, viewing distance, viewing angle, and observer subjectivity. These factors were expanded from a previous study (6).

Stainless steel was chosen for the surface material because most manufacturing equipment surfaces are made of this material. Representative stainless steel coupons (304 grade, 2B finish) were used for spotting purposes in the laboratory setting. Larger equipment size and increased viewing distance cause reflected and ambient light that affect visibility in the manufacturing facility. To simulate these conditions, a background of stainless steel was used when viewing the spotted coupons.

The type of solvent and its solubility were also factors that affected appearance. Methanol was chosen as the solvent for the sample preparations because it leaves no residue, minimizes residual sample rings, and provides adequate solubility for most substances tested. When solubility was not achieved, samples were immediately spotted using a suspension in methanol. The common solvent allowed for a tighter control over spot sizes and concentration. The sample volume was varied for concentration differences. This procedure obviated the need for serial sample dilutions, thereby saving time and solvent consumption. A complementary volume of methanol was added to each sample to achieve a constant spot volume.

The standard light intensity in a manufacturing facility is 750 lx. Actual levels differ from room to room depending on equipment size, configuration, location, and associated shadows caused by the equipment. The larger a facility's manufacturing equipment, the greater the difference in lighting levels compared with smaller pilot-plant equipment. Large machinery without interior lighting deepens the shadows. To compensate for this lighting condition, a portable light was used for inspection as necessary. Therefore, the lighting for this study ranged from 100 lx to the portable light's intensity. For low lighting levels, ambient fluorescent light provided the same type of light as that used in manufacturing plants.

A handheld light source (Sport Shot, model VEC1 24B, Vector Machinery, Ltd., Fort Lauderdale, FL) maximized viewing conditions. By moving the light source, an observer control control the lighting conditions to optimize the incident light angle and the effect of reflected light on the formulation residue and minimize the light reflection. A light meter was used to set and verify various light-intensity levels.

The viewing distances for this study were dependent on the size of the equipment. In the pilot plant, a comfortable viewing distance of 1 ft was achievable (6). In a manufacturing facility, equipment sizes are larger and viewing distances are greater. Rather than define viewing distances for each piece of equipment, viewing distances of 5, 10, 15, and 20 ft were chosen to complement previously established data (6).

The viewing angle also is restricted by the equipment size and configuration. Therefore, residues were viewed over a range of 12–90° angles. The minimum angle resulted from a combination of a comfortable viewing angle and the viewing distance. Data at intermediate viewing angles of 30° and 45° and a perpendicular viewing angle (90° to the observer) were evaluated.

To minimize the effect of observer subjectivity, four observers viewed all samples. Sample concentration levels were spotted above and below the previously determined VRL (6) to allow for increased distances and higher intensity light, respectively. Therefore, the targeted spotting levels for the formulations were at the API's ARL (typically 4 μg/cm² , the previously determined VRL [6]) and at the ±25% VRL levels.

Experiment

Samples were prepared by dissolving or dispersing tablets with methanol in an appropriately sized volumetric flask to achieve the targeted API concentration. Concentrations were targeted so that similar volumes were dispensed to form the residue spots because the volume controls the spot size. The sample volume was varied for concentration differences, and a complementary volume of methanol was added to each sample to achieve a constant spot volume. Methanol evaporated rapidly under a stream of nitrogen and left no solvent residue.

Two samples were applied to each 3 x 6-in. coupon. The spots were dried under a stream of nitrogen to prevent the material from oxidizing. If these steps were not taken, then oxidation would chemically alter the material and potentially change its visual properties. The dried spots were measured to determine the amount per unit area (μg/cm² ) of each spot (see Table I).

Observers viewed the spots against a stainless steel background to more closely simulate larger manufacturing equipment and the ambient light encountered during observation. Three pieces of 27 x 34 in.-stainless steel were placed perpendicular to one another, forming a three-sided corner. The coupons were positioned at several angles to the observer within the stainless steel field (see Appendix, Figure 6).

The ambient fluorescent light source controlled the lower end of the light-intensity range. Ambient light positioned directly above the viewing surface affected residue detection, therefore indirect ambient light was used to minimize reflected-light effects. The observer held the portable light source to simulate viewing under manufacturing plant conditions. The light intensity of the portable light source on the coupons was a function of the distance from the coupon and decreased with distance (see Figure 1). Observers moved the portable light as far as an arm's length from their bodies and turned the light to adjust the angle of the incident light to the coupons. This procedure provided the best lighting conditions for each observer.

Figure 1: The effects of distance and spotlight intensity on residue detection were studied in trials with four observers. Because a portable light´s intensity is a function of the distance from a stainless steel coupon, observers adjusted the angle of the incident light to the coupons to enable the best lighting conditions.

The observers for this study were familiar with the appearance of API and formulation residues and all had 20/20 eyesight with or without corrective lenses. Observers stood perpendicular to the stainless steel background (see Figure 2, Appendix, Figure 7). The observers viewed the coupons separately and under the least favorable conditions first (i.e., at the minimum viewing angle, from the greatest distance, under the lowest light). The viewing distance, angle, and the light intensity were measured for each observation. As a final variable, the position of the observer relative to the stainless steel background was varied to ascertain the effects of the reflected ambient light and the stainless steel background (see Appendix, Figure 7).

Figure 2: Observers viewed formulation stainless steel coupons from various distances and angles.

Results and discussion

VRLs were evaluated for several marketed and late-development formulations on the basis their respective APIs' concentration. Each visible limit was designated as the lowest concentration at which all observers positively identified residue. The actual amount of material spotted in μg/cm² was based on previous VRL work conducted in a pilot-plant environment (6). The ability to detect visible residue changed compared with the previous study as a result of the wider range of parameters (i.e., light intensity range, viewing distance, and viewing angle). In addition, a sample at the ARL was spotted to address greater distances coupled with lower light intensities. Each observer viewed the spots and indicated whether they saw any residue.

Stainless steel was chosen as a representative surface for the study. The use of a VRL criteria as a result of this study is limited to the stainless steel surfaces of manufacturing equipment.

Using methanol as the spotting solvent improved the experimental results. The sample spot size and the resulting concentration levels were much tighter than in previous studies, in which various solvents were used (6). The application and drying of the spots were more controlled and more efficient. Despite the additional control, the residue spot sizes and control of the resulting spot concentrations still varied (see Table I). No noticeable effects of the solvent on the resulting residue appearance were observed.

Table I: Formulation residue concentrations.*

As expected, the overall ability to visually detect formulation residue decreased with increased viewing distance (see Table II). At 400 lx and at the minimum viewing angle (15°), observers detected the previously determined ARL and VRL for all tested formulations from 5 ft. Several VRLs were not detected from 10 ft. From 15 ft, the observers could not see most VRLs and could not consistently detect any VRLs from 20 ft. With regard to the ARLs, the observers saw most formulation residues under these viewing conditions from 10 and 15 ft. From 20 ft, the observers saw less than half of the formulation ARLs.

Table II: Effect of viewing distance on visible-residue detection.*

The ability to detect residue also diminished with decreased ambient light (see Table III). With 200-lx ambient light, VRLs were consistently detected from 15 ft and a 45° viewing angle. At 100-lx ambient light, some VRLs were not detected at 15 ft and 45°. VRLs were consistently detected from 10 ft at 100 lx, however.

Table III: Effect of light intensity on visible-residue detection.*

The ambient light source controlled light intensity at the lower end of the viewing range. The portable light source controlled the light intensity at the upper end of the viewing range. The observer moved and adjusted the portable light source to optimize individual viewing conditions within the constraints encountered for various pieces of manufacturing equipment. Therefore, the maximum intensity of the portable light source decreased with distance.

In general, the spotlight did not increase the observer's ability to detect formulation residue. The intensity of the spotlight overwhelmed the residue, and the reflecting light from the spotlight hindered the observer's ability to detect the residue. Several instances occurred in which the spotlight enabled the observer to see a previously undetected spot. Nonetheless, there were more cases in which the observer did not detect a residue with the spotlight but did detect the same residue under ambient light. In practice, the effective use of a portable light source depends on the observer and the situation.

Table IV: Effect of viewing angle on visible-residue detection.

The viewing angle of the observer to the residue was a critical factor for detecting formulation residue. Under ambient light and at the minimum angle (~15°) (see Table II), the observers did not detect most VRLs at 15 ft and only detected a few at 20 ft. When the viewing angle increased to 30° (see Table IV), the observers detected more residue spots at 15 and 20 ft, but not enough to make a significant difference compared with data at 15°. As the viewing angle increased to 45° and 90°, the observers detected almost all VRLs at 15 ft (see Table IV) and most VRLs at 20 ft. The observers detected essentially all ARLs at 20 ft and at >30° viewing angles. When the observer's position changed with respect to the stainless steel background, observers detected all VRLs from 10 ft, at a 45° coupon angle, and with light intensities as low as 100 lx (see Appendix, Figure 7).

Figure 3: Distance and observer´s angle were studied to gauge their effects on visible-residue detection. According to data collected from trials with four observers, accuracy decreased with greater viewing distances and angles.

Observer variability was a factor in determining the VRL (6) for API and formulation residues. For this study, each factor examined had an effect on the observer's ability to detect the formulation residues. Observer detection was dependent on the formulation residue level, observer viewing distance, light intensity, and viewing angle. Certain observers had trouble detecting several formulation residues.

Observer variability increased with greater viewing distances, particularly for those beyond 10 ft (see Table II). This trend also was true of the observer angle factor. At 15° and 30° viewing angles, observer variability was comparable with other factors (see Table IV). At a viewing angle greater than 30°, the ability to detect residue increased significantly and observer variability decreased accordingly (see Figures 3 and 4). Residue detection was comparable using the portable light source and 400-lx ambient light (see Appendix, Figure 8) and was not a significant factor at decreasing light intensity levels until 100 lx, at which VRL detection was problematic (see Figure 5).

Conclusions

Using a visible-residue limit (VRL) to verify equipment cleanliness is a viable possibility in a manufacturing facility for formulations with a VRL that is lower than the acceptable-residue limit (ARL). The ability to detect pharmaceutical compounds down to their VRLs has been demonstrated using formulation residue level, observer viewing distance, light intensity, and viewing angle as variables.

Figure 4: The effect of distance and observers angle on acceptable-residue detection were studied. According to data collected from trials with four observers, observer variability increased with greater viewing distances and angles.

The factors that affect visible-residue detection can be determined and viewing of residues can be controlled. Under defined viewing conditions, a trained observer can detect formulation residue. The observer should be within 10 ft of the equipment surface to minimize the effect of the light intensity or the viewing angle. In addition, the observer should view the surface from a >30° angle to minimize the risk of the residue blending in with the background. Finally, the ambient light level should be at least 200 lx. Otherwise a portable light source can be used.

Figure 5: Light intensity was studied to gauge its effect on residue detection. Based on data collected from trials with four observers, using a portable light source for detection was not a significant factor at decreasing light intensity levels until 100 lx, at which detection was problematic.

Acknowledgments

The authors thanks Michael McQuade, Tara Lukievics, and Joseph Schariter for their efforts as observers during these studies.

Richard Forsyth* is an associate director in Worldwide GMP Quality with Merck & Co., Inc., WP53C-307, West Point, PA 19486, tel. 215.652. 7462, fax 215.652.7106 richard_forsyth@merck.comVincent Van Nostrand is a research chemist in Pharmaceutical R&D with Merck & Co., Inc.

*To whom all correspondence should be addressed.

Submitted: June 14, 2005. Accepted: June 29, 2005.

References

1. Code of Federal Regulations, Title 21, Food and Drugs (General Services Administration, Washington, DC, Apr. 1, 1973, Part 211.67.b.6.

2. R.J. Forsyth and D. Haynes, "Cleaning Validation in a Pharmaceutical Research Facility," Pharm. Technol. 22 (9), 104–112 (1998).

3. US Food and Drug Administration, Guide to Inspection of Validation of Cleaning Processes (FDA, Rockville, MD, 1993).

4. D.A. LeBlanc, " 'Visually Clean' as a Sole Acceptance Criteria for Cleaning Validation Protocols," J. Pharm. Sci. Technol. 56 (1), 31–36 (2002).

5. D.W. Mendenhall, "Cleaning Validation," Drug Dev. Ind. Pharm . 15 (13), 2105–2114 (1989).

6. 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).

7. 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).

8. G.L. Fourman and M.V. Mullen, "Determining Cleaning Validation Acceptance Limits for Pharmaceutical Manufacturing Operations," Pharm. Technol. 17 (4), 54–60 (1993).

9. K.M. Jenkins and A.J. Vanderwielen, "Cleaning Validation: An Overall Perspective," Pharm. Technol. 18 (4), 60–73 (1994).

10. D.A. LeBlanc, D.D. Danforth, and J.M. Smith, "Cleaning Technology for Pharmaceutical Manufacturing," Pharm. Technol. 17 (10), 118–124 (1993).

Appendix

The complete article, tables, and appendix are available on www.pharmtech.com. Coming Soon.