Medical Device Link.

DESIGN

Evaluating drug leakage

Needle-based parenteral injection devices that target the skin dermis or shallow hypodermis and deliver fluid volumes in the range of 100–200 µL are currently under clinical development for drug and vaccine delivery.^1–3 Accurate measurement of the completeness of fluid injection and the consistency of the volume of injected fluid are two of the most critical criteria when evaluating the effectiveness of a new device. Leakage of fluid from the device and/or the injection site is the root cause of variability in dose delivery by needle and needle-free injection systems.^4–8 Yet, there is a shortage of validated methods to collect and measure fluid leakage that are easy to use in a clinical trial for device performance evaluation. Furthermore, it is desirable to separate the fluid-collection step from the measurement step so that the method is easy to use in multisite clinical trials, and for batch weighing operations.

The methods compared

The capillary tube method measures the height of the collected fluid column from the tip to the 5-µL calibration mark on capillary tube wall. The measured height (mm) in 10 capillaries is correlated with the collected volume of fluid leaked from the injection sites (see Equation 1).The wicking spear gravimetric method uses the weight difference of a dry wicking spear before and after it is used to collect the leaked fluid. It correlates this weight with a collected volume result in µL, taking into account the density of the collected fluid (see Equation 2).

Equation 1

Equation 2

Study objectives

The primary objective of this study was to evaluate the two fluid leakage collection methods and determine whether they meet acceptance criteria according to fluid-volume range. The secondary objective was to determine whether the two methods are statistically different when operator effect is considered. The third objective was to validate the wicking spear method and develop a standard operating procedure for use in clinical trials that allows volume measurement in a central laboratory. The effect of intermediate storage was also investigated, where the wicking spear is placed in a glass-sample collection tube and sealed with a waterproof and airtight stopper after fluid collection and shipment from clinical site to central laboratory.

Evaluating fluid recovery

Fifteen samples were tested using both collection methods at each of three different baseline dispensed fluid set points (2, 10 and 25 µL). The dispensed fluid volumes were placed onto microscope slides, intended to simulate a known volume of saline leakage to be collected. The weight of the dispensed volume was recorded and used as the baseline “known” value for collection-volume delta and percentage recovery calculation. In the case of the 25-µL dispensed set point, three capillary tubes (Clinitubes, Radiometer, Denmark, www.radiometer.com) were required to collect this fluid volume; therefore a total fluid column height–length measurement (mm) that combined all three individual tube measurements was used to calculate the collected volume. In contrast, one wicking spear (Surgical Spears, Ultracell Medical Technology, USA, www.ultracell.com) was required for collecting 25 µL. To normalise out variability in baseline-dispensed volume between samples or groups, collection loss or volume delta (difference between as-calculated collected volume and baseline dispensed volume) and method accuracy or percentage recovery (collected volume as a percentage of baseline dispensed volume) were calculated. Those calculations were used to evaluate and compare the capabilities of the test methods.

Evaluating operator effect

The sample size was again 15 tests for each leakage collection method. However, only two different baselines of dispensed fluid volume set points (2 and 25 µL) were tested with two different operators (laboratory technicians). One operator performed all dispensing of baseline leakage volumes onto the microscope slides to prevent the introduction of additional variability at this baseline starting point. Each test method was performed by both test and measurement operators as per the procedure described above.

Validation

At day 0, one operator prepared 122 wicking spears (volumes 0, 2, 4, 8, 16, 32 and 64 µL), each stored in individual customised glass collection tubes. At day 0, 56 tubes containing one wicking spear were used to investigate test-method variability, reproducibility and repeatability, and then the tubes were stored at room temperature for 12 days. The additional 56 tubes prepared by the same operator were stored at a temperature of 4 °C to evaluate the impact of storage temperature over a period of 12 days.

At day 0 in the evaluation of variab-ility, reproducibility and repeatability, three operators were involved in weighing test tubes, re-zeroing the balance for each measurement. Each measurement was done once and repeated in triplicate. Operator repeatability was evaluated from 10 subsequent measurements of test tubes selected in a random fashion. The reproducibility of the method was calculated as the variation in the measurements made by the three operators when measuring the same tube.

One operator weighed the 122 tubes on day three, day six, day nine and day 12. The effect of the length of storage (day 0 to day 12), storage temperature, dispensed fluid volumes 0, 2, 4, 8, 16, 32, 64 µL and the collection tube on fluid leakage measurement was analysed by analysis of variance (ANOVA).

Statistical methods

Figure 1. (click to enlarge) Boxplots of percentage recovery comparing capillary and wicking spear methods.

Two sample T-tests (with significance level of 5%) were used for comparing collected fluid, collection loss and percentage recovery between the test methods at each dispensed volume. ANOVA was run at 2- and 25-µL set leakage volumes to assess method and operator effects. ANOVA was also used to assess method variability, reproducibility and repeatability and to evaluate time and temperature effect on collected wicking spears stored in customised
collection tubes.

Table I. (click to enlarge) Comparison of the capability of the two fluid leakage collection methods.

Results

As shown in the first column of Table I, the baseline initial dispensed volume on the microscope slides for both test method groups at each volume set point are not significantly different as evaluated by two-sample T-tests. The wicking spear method has significantly lower leakage volume collection loss than the capillary tube method. Similarly, the wicking spear method collects significantly more fluid volume than the capillary tube method, regardless of initial dispensed volume. Figure 1 indicates that the wicking spear collection method has a significantly higher percentage leakage collection recovery of the initial dispensed volume than the capillary tube method. The 95% confidence interval range of percentage recovery improvement using the wicking spear method over all dispensed volume set points is between 4.52 and 7.08%. The significantly lowest method accuracy or percentage recovery performance for both tests was observed with a 2-µL set point. Evaporative loss may have an adverse impact notably on the percentage recovery at this low volume.

Table II. (click to enlarge) Test method validation based on acceptance criteria.

Figure 2. Postinjection fluid leakage measurement kit using wicking spear method. (Top to bottom) (a) Wicking spear; (b) Wicking spear in the container tube; (c) Wicking spear collecting fluid leakage on skin surface; (d) Sample collection box for kit shipment before and after fluid leakage collection.

Table II indicates that both methods satisfy the acceptance criteria for method validation as evaluated by the average accuracy calculated by the percentage of recovery at 2- and 25-µL leakagevolume set points. The ANOVA within the dispensed volume indicates an overall difference between the test methods (p = 0.034 and p < 0.0005, for average collection loss at 2 µL and 25 µL leakage volume set points, respectively); operator difference (p = 0.020 and p < 0.005 for average collection loss at 2- and 25-µL leakage volume set points, respectively); operator method interaction difference (p = 0.003 and p < 0.0005 for average collection loss at 2- and 25-µL leakage volume set points, respectively) for average accuracy (percentage of recovery) and average collection loss (volume delta) responses. When examining the 95% simultaneous confidence intervals, the overall significant operator difference results from differences in the capillary tube method only and not the wicking spear method. Nevertheless, despite this statistically significant operator difference, both methods were still able to exceed the minimum acceptance criteria.

Based on fluid-recovery percentage (fluid-collection accuracy) and the consistency of the collection method according to the operator and ease of handling of the wicking spear, the wicking spear method has been selected as the most consistent to develop a fluid-leakage measurement kit. The kit consists of a dry 7-mL glass sample-collection tube containing one wicking spear, which is to be weighted before and after fluid collection (Figure 2 ).

Table III. (click to enlarge) Variability (estimated standard deviation) of wicking spear method using dispensed fluid from a wicking spear stored in a collection tube.

The measurement of the delta volume (collected volume) using a wicking spear packaged in a customised glass sample-collection tube does not alter the measurement accuracy. Table III shows that the method variability is constant whatever the dispensed volume. An ANOVA general linear model with multiple comparisons was performed over all dispensed volume set points (0 to 64 µL) to assess the statistical significance of time and storage temperature effects. The two factors (time and temperature) as well as their interaction have a statistically significant effect on the measurement of the collected volume. After 9 and 12 days of storage, the change in measured collected volume is always below 1 µL, which is considered not to be clinically relevant. The storage at 4 °C did improve the stability of the measured volume over the period of time.

Summary

The wicking spear gravimetric method was selected as the easier and the most consistent method for fluid leakage collection. The kit uses the glass sample collection tube to store the wicking spears before and after fluid leakage collection. It has been shown to provide reliable equipment for wicking spear intermediate storage and shipment to central laboratories for weight/mass measurements. Weighting the collection tubes with wicking spear inside, in a central laboratory in a batch fashion, before and after fluid collection, contributes to minimising operator and balance errors on method accuracy. The volume detection threshold is below 2 µL and the measurement error, even after 12 days of storage at room temperature, is below 1 µL, which is considered not to be clinically relevant.

Acknowledgements

This study was supported by research grant from Becton-Dickinson’s Advance Drug Delivery Programme. The authors are thankful to Adeline Fourcot and Professor Marc Fantino (University of Bourgogne, CREABio, Dijon, France) for skillful technical help.

*Philippe E. Laurent MD, PhD is Medical Director, BD Medical, Pharmaceutical Systems, Medical Affairs, 11 rue Aristide-Bergès, 38800 Le Pont-de-Claix, France, tel +33 476 68 3501, e-mail: philippe_laurent@europe.bd.com, www.bd.com.
Ronald Pettis PhD is Senior Scientist/Group Leader, Advanced Drug Delivery at BD Technologies, Triangle Park, North Carolina, USA
William Easterbrook is Research and Development Engineer at BD Medical Surgical, Franklin Lakes, New Jersey, USA and
Julie Berube PhD is Senior Statistician at BD Corporate Statistics, Franklin Lakes, New Jersey, USA.

* To whom all correspondence should be addressed.