In Situ Dissolution Testing Using
a UV Fiber Optic Probe Dissolution System

Kevin C. Bynum*, Erik Kraft, John Pocreva, Emil W. Ciurczak, Philip Palermo
Purdue Pharm, L.P., International R&D, Ardsley, NY

Purdue Pharm, L.P., International R&D, Ardsley, NY

* e-mail address for correspondence: kevin.bynum@pharma.com

Introduction
Dissolution testing is a time-consuming and labor-intensive test procedure. The technique requires a sample of dissolution medium be removed from each of the test vessels at a specified time point(s). This sampling procedure can be handled either manually or by means of an automated sipper mechanism. Each method has its own inherent errors associated with it. The manual technique is subject to analyst variability. Two of the most common errors associated with the manual sampling technique are improper sampling within the vessel and timing errors in sampling. Additionally, this is the most labor-intensive technique and is especially taxing for controlled-release formulations, which often require 24-hour testing periods. Automatic sampling techniques can save analyst labor, but still require a good deal of maintenance and set-up time. Also, automated sipping mechanisms are complex and contain many moving parts that can malfunction, resulting in sampling errors. Sipper mechanisms are prone to sample contamination, carryover, bubbles, dilution, filter clogging, and tubing leaks.

We have developed a simple solution to these problems. We have decided to eliminate this sipping step and conduct an in-situ spectroscopic measurement in the dissolution vessels. We pipe the light (not the liquid), to a series of 12 dip-type transflectance probes (C-Technologies, Cedar Knolls, NJ) that reside in the 12 vessels throughout the dissolution. A pair of deuterium lamps is used as light sources. After interacting with sample, the light is piped to a series of 12 Photo-Diode Array (PDA) spectrophotometers. The spectrometers are Zeiss MMS UV (190-385nm) modules obtained from Hellma (Forest Hills, NY). At the start of each run, the system Figure 1 collects "dark" spectra, 100% transmittance spectra, and standard scans. It also runs a fully automated system suitability procedure. Once this procedure passes the specified limits, the probes are placed in the vessels Figure 2, the dosage units are dropped into the vessels, and the system acquires dissolution data without further analyst intervention. A full absorbance scan is collected in real time from each vessel at time intervals ranging from minutes (for a 24-hour dissolution) to seconds (for immediate-release dissolution). The system is controlled by an in-house developed and validated software package. Once the full spectrum is collected and stored on our computer network, the resultant "percent dissolved" is calculated in real-time and displayed for the user on the monitor. When the run is complete, the final report is prepared by simply pushing a button; no post run calculations are required.
To demonstrate the performance of the system, three dissolution experiments were conducted. The first consisted of a number of immediate release formulations, where the amount of excipent was adjusted to control the immediate release characteristics. The second experiment was a twelve-hour dissolution, conducted on three samples of a controlled release formulation. The third experiment was a 24-hour dissolution conducted on twelve samples of a once-a-day controled release formulation.

Immediate Release Experiment
Four separate samples were used to study the ability of the system to profile immediate release formulations. The system was set up to acquire a full UV spectrum from each of the four vessels every 10 seconds. A Hanson SR-8 dissolution apparatus was fitted with paddles (USP Apparatus II) rotating at 100 RPM and was equilibrated at 37° C for one hour. The fiber optic probe system was set up by scanning dark, 100% transmittance, and standard scans on all four probes. The probes were then placed into the vessels, and the samples inserted. The system acquired an absorbance scan every 10 seconds for the duration of the experiment. A single-point baseline correction was utilized to eliminate small baseline offsets. The resultant "percent dissolved" was calculated by the software in real time and the dissolution curve displayed to the analyst as the experiment proceeded. At the end of the experiment, the results were immediately ready for the analyst to print and review.

Controlled Release Experiment-12 hour release
Three samples of controlled-release analgesic formulations were used to demonstrate the ability of the system to monitor long-term dissolution experiments. The experiment was run on a Hanson SR-8 dissolution apparatus configured with baskets (USP Apparatus I) operating at 100 RPM. The system was equilibrated at 37° C for one hour. The fiber-optic probe system was set up by scanning dark, 100% transmittance, and standard scans on the 3 probes. The probes were placed into the vessels and the samples added. The system acquired an absorbance scan from all of the vessels at ten-minute intervals for 12 hours. A novel (proprietary) algorithm was used to correct for the particulate matter that was present from the controlled release matrix. The percent dissolved curve was calculated in real time and was compared with data obtained from the HPLC method that the fiber optic technique is replacing.

Controlled Release Experiment-24 hour release
Twelve samples of a once-a-day controlled-release analgesic formulation were used to demonstrate the utility of the system for routine stability testing. This experiment was conducted on two, six vessel Hanson SR-6 dissolution apparatus configured with paddles operating at 100 RPM. The system was equilibrated at 37° C for one hour. The fiber-optic probe system was set up by scanning dark, 100% transmittance, and standard scans on all 12 probes. The probes were placed into the vessels and the samples added. The system acquired an absorbance scan from all of the vessels at ten-minute intervals for 24 hours. A single-point baseline correction was utilized to eliminate small baseline offsets. The percent dissolved curve was calculated in real time and was compared with data obtained from the HPLC method that the fiber optic technique is replacing.

Results
The results of the immediate release experiments are presented in Figure 3. It can be seen that the data-rich graphs show both the gross and subtle differences between formulation dissolution profiles.
The 12-hour controlled-release experiment in Figure 4 shows the versatility of the method in that specifications for percent dissolved at nearly any time point may now be set. In most dissolution trials, fewer time points are sampled, making "off-time" specifications difficult. That is, if the samples were drawn at four, eight, and 12 hours, a percent dissolved figure for 10 hours would have to be extrapolated in current systems. With the fiber optic system, the software would simply print out the 10-hour result.
In addition, the close agreement of the in-situ probe data with the HPLC data indicates that the resident probes do not effect the release of the active component from this formulation.

The 24-hour controlled-release experiment in Figure 5 shows that the system works well under routine stability conditions. The close correlation of the data with the HPLC results demonstrates the accuracy of this technique.
Discussion

As Figure 3 illustrates, the in-situ method also allows for rapid data acquisition. Our system has been able to acquire scans on all 12 channels once every 12 seconds. This rapid-scanning procedure has allowed us to profile the release characteristics of immediate and explosive release formulations.
As is shown in Figure 4, the UV Fiber Optic Probe Dissolution System generates a continuous dissolution curve consistent with the external assay performed by HPLC. This accuracy is also demonstrated in Figure 5, which demonstrates the application of the system for the analysis of 6 or 12 sample lots over a 24 hour period.
The increase in "data density" during the dissolution experiment increases the statistical significance of the data collected. This can be very useful when using this data to demonstrate an In-Vivo In-Vitro Correlation. The increase in data allows for predictive modeling of dissolution curves.

The system is much easier to set up than systems with sippers. Once the system passes the system suitability portion of the experiment, the system will run continuously for the specified dissolution time. This ruggedness is due to the solid state nature of the system. Since there are no moving parts, breakdowns are rare.

The absence of a chromatographic step makes the setup quick and easy while eliminating any error associated with HPLC. It takes the average analyst about two hours to set up a dissolution experiment (including all reagent preparation). The system also performs all calculations in real time; which eliminates the need for post-dissolution paperwork, while allowing for post-run review of the data.

Conclusion
A UV Fiber Optic Probe Dissolution System has been developed for the analysis of solid dosage forms. The system uses 12 dip-type fiber-optic probes coupled to 12 separate PDA spectrophotometers to acquire continuous dissolution curves in real time. The system is applicable to the analysis of both immediate and controlled-release formulations. The system is accurate, quicker, and easier to set up when compared with conventional HPLC or UV-sipper systems. The data is acquired, calculated, and secured using an in-house developed and validated software package. The software package uses a novel scatter-correction algorithm to generate accurate dissolution curves in a turbid medium. The data that the system generates is as accurate as the HPLC methods that it is replacing.