Determination of residual dimethylsulphoxide in drug ...

14 downloads 0 Views 700KB Size Report
Feb 23, 2018 - using thermal desorber – gas chromatography ... trafluoroethylene insert which was then placed in its entirety into a thermal desorption tube.
Journal of Pharmaceutical and Biomedical Analysis 153 (2018) 193–198

Contents lists available at ScienceDirect

Journal of Pharmaceutical and Biomedical Analysis journal homepage: www.elsevier.com/locate/jpba

Determination of residual dimethylsulphoxide in drug loaded gelatin using thermal desorber – gas chromatography Adissu Alemayehu Asfaw, Kris Wolfs, Ann Van Schepdael, Erwin Adams ∗ KU Leuven – University of Leuven, Department of Pharmaceutical and Pharmacological Sciences, Pharmaceutical Analysis, Herestraat 49, O&N2, PB 923, 3000, Leuven, Belgium

a r t i c l e

i n f o

Article history: Received 14 December 2017 Received in revised form 21 February 2018 Accepted 21 February 2018 Available online 23 February 2018 Keywords: Thermal desorber – gas chromatography Gelatin Residual DMSO

a b s t r a c t Traditional headspace – gas chromatography (HS-GC) methods for the determination of residual solvents (RS) start from a homogenous sample solution. Subsequently, it is challenging to determine RS using HSGC techniques from insoluble solid samples like gelatin which is practically impossible to dissolve or distribute uniformly in water and common organic solvents. In this study, a thermal desorber combined with capillary gas chromatography and flame ionization detection/mass spectrometry (TD-GC-FID/MS) was used for quantitative determination of residual dimethylsulfoxide (DMSO) in gelatin without sample pretreatment. A sample of gelatin was sandwiched between two quartz filter double layers in a polytetrafluoroethylene insert which was then placed in its entirety into a thermal desorption tube. Factors affecting the performance of TD-GC including desorption time, desorption temperature, desorption flow and type of adsorbent were studied by applying a standard solution of DMSO in methanol on a blank gelatin bed. Validation results of the proposed method showed good linearity with an R2 -value higher than 0.999 for a wide concentration range and good sensitivity with a limit of detection and limit of quantification of 0.1 ␮g and 0.2 ␮g on tube, respectively. The proposed method shows recovery values close to 100%. In addition, a conventional HS-GC method following enzymatic degradation of gelatin was developed to verify the proposed TD-GC method. Both methods were applied for the determination of residual DMSO in gelatin that was loaded with an experimental drug. Results were comparable, but the enzyme assisted HS-GC method was more time consuming and expensive. © 2018 Elsevier B.V. All rights reserved.

1. Introduction Gelatin is derived from collagen by hydrolytic degradation which breaks crosslinks between strands and to some extent hydrolyzes the strands into fragments. Gelatin has widespread application in food as a stabilizer, thickener and texturizer. In pharmaceuticals it is mainly used to make capsules, as a binder in tablet formulations, as drug carrier and also as a coating to ease swallowing or mask unpleasant tastes [1]. Due to its biodegradability it is also often applied in the biomedical field as sealant for vascular prostheses, wound dressing and adsorbent pad for surgery [2]. If gelatin is heated, it undergoes structural and physicochemical transformations. When gelatin is treated with hot water (40–50 ◦ C), it dissolves and takes a coil conformation. Further heating results in partial or complete loss of solubility in water due to crosslinking [3]. One of the most important properties of gelatin is its gelation in

∗ Corresponding author. E-mail address: [email protected] (E. Adams). https://doi.org/10.1016/j.jpba.2018.02.047 0731-7085/© 2018 Elsevier B.V. All rights reserved.

water which is concentration dependent and due to H-bonds linking peptide groups and hydrophobic interactions between apolar side-groups of the gelatin macromolecules. Addition of some solvents like DMSO will result in a disruption of both the H-bonds and the hydrophobic interactions in the gel network [4,5]. For the loading of drugs on carriers like gelatin, solvents are used that are removed afterwards. However, this removal is not complete and some traces will remain as residual solvents (RS). The detection and quantification of RS in pharmaceuticals is very strict and requires sensitive analytical techniques to avoid potential human health risks. Moreover, RS can influence the physicochemical characteristics of drugs such as dissolution properties and morphology [6]. GC is the most appropriate method to analyze RS from different materials [7]. It is often combined with static headspace (sHS) injection which has the advantage that the sample matrix is not introduced into the GC system [8]. Due to the fact that RS are present in very small quantities, their sensitive detection using sHS-GC is challenging and not always successful [9]. sHS-GC also suffers from difficulties to determine high boiling analytes (like DMSO) in low boiling matrices where the matrix will

194

A.A. Asfaw et al. / Journal of Pharmaceutical and Biomedical Analysis 153 (2018) 193–198

act as a dilute of the gas phase and can create overpressure in the vial. Matrix matched calibration is mandatory in sHS, but this is not always possible. The above mentioned problems can be partially solved by the full evaporation technique (FET) which is based on complete evaporation of the analyte(s) from the sample and so will overcome matrix effects [10]. However, traditional HS methods for the determination of RS start from a homogenous sample solution. As a result, it is challenging to determine high boiling RS using these techniques for solid samples (like gelatin) which are practically impossible to dissolve or distribute uniformly in water or common organic solvents. When performing HS techniques with aqueous samples, typical HS oven temperatures are only around 85 ◦ C to ensure that the HS pressure does not exceed the limitations of the instrument [11]. On the one hand this temperature is too low to volatilize DMSO (which has a boiling point of 189 ◦ C) while on the other it causes a rigid gelatin gel. To solve this problem, Li et al. performed a preceding binary solvent extraction to determine residual ethanol used to seal hard gelatin capsules [12]. However, this approach is not applicable for the determination of residual DMSO which has different physicochemical characteristics. Direct quantification by HS-GC from solid samples could be a solution for the above mentioned problems, but it is difficult to compensate for matrix effects during calibration. This is certainly true for DMSO which can interact with gelatin. Multiple headspace extraction (MHE) prior to GC has also been applied to remove matrix influences. In combination with single drop microextraction (SDME), it has been described for the determination of residual ethanol and methanol in drug products [13]. However, MHE is time consuming and the low volatility of DMSO and heat instability of gelatin remain problematic. Hashimoto et al. applied thermal desorption (TD)-GC using a pyrolyzer and on-column cryofocusing for the determination of RS from solid pharmaceuticals [14]. The method showed some limitations, mainly related to the cryofocusing including ice blockage, incomplete retention of very volatile compounds, loss of high-boiling solvents due to aerosol formation and high running costs (systems consumed up to 6 L of liquid nitrogen per hour in operation). Moreover, because the cryofocusing device was connected directly to the GC column, it was difficult to implement essential pre-desorption checks such as leak testing [15]. Nonetheless, a modified set-up of TD instruments creates interesting possibilities. In its basic form, a tube with an adsorbent used to collect samples (typically for air/environmental monitoring) is brought in the TD and heated, involving the dynamic extraction of volatile compounds in a flow of inert gas. This is known as primary desorption. The gas flow is transported to a cold trap which contains one or more sorbents and works at a lower temperature. Next, it is rapidly heated so that all retained analytes are released in a very short time (a few seconds) and transferred to the GC through a heated transfer line [16]. The simplicity of direct TD relative to conventional sHS methods is that no additional dissolution or salting-out steps are required and that it does not rely on partition coefficients or equilibria. Furthermore, complete (or nearly complete) extraction is often possible in one run [17]. In this work, TD followed by GC has been applied in an unconventional way using modified sample introduction. The developed TD-GC method allowed quantitative determination of residual DMSO in gelatin samples without any sample pretreatment. Factors affecting the performance of TD-GC including desorption time, desorption temperature, desorption flow and adsorbents were studied. The method was applied to an experimental sample of gelatin loaded with a drug using DMSO. To verify the newly developed method, results were compared with an alternative, time consuming and expensive method based on the complete enzymatic

Table 1 Used TD-GC parameters. Parameter

Optimized setting

Range explored

230 ◦ C 180–280 ◦ C 90 min 20–120 min 120 mL/min 20–200 mL/min 4 ◦C −30 to 10 ◦ C 280 ◦ C 6 min 2–20 min Tenax TA Air Toxic/Tenax TA 40 mL/min 10–60 mL/min 80 mL/min 10–100 mL/min 4 mL/min 80 ◦ C held 1 min; 80–160 ◦ C at 40 ◦ C/ min, held 5 min; 160–250 ◦ C at 40 ◦ C/min, held 2 min GC/MS interface temperature180 ◦ C

Desorption temperature Desorption time Desorption flow Trap low temperature Trap high temperature Trap hold time Trap material Inlet split Outlet split Column flow GC oven program

digestion of gelatin combined with dehydration and FET-HS-GC analysis. 2. Experimental 2.1. Chemical reagents Active human recombinant matrix metalloproteinase-2 (MMP2, 0.1 mg/mL) was provided by Calbiochem (Merck, Darmstadt, Germany). Methanol (99.99%), N,N-dimethylformamide (DMF, 99.8%) and 2,2-dimethoxypropane (DMP, 98+ %) were obtained from Acros Organics (Geel, Belgium). Benzyl alcohol (BA, >99%), DMSO (>99.9%) and LC–MS grade water were purchased from Sigma Aldrich (St. Louis, MO, USA). Hydrochloric acid (37.5% w/w) was bought from VWR International (Fontenay-sous-Bois, France). Sodium carbonate was from Chem-Lab (Zedelgem, Belgium). 2.2. Instrumentation 2.2.1. TD-GC-FID/MS TD-GC analyses were performed on a Perkin Elmer Clarus 680 GC coupled with a Clarus SQ8T MS and equipped with a Perkin Elmer Turbomatrix ATD 350 thermal desorber (Waltham, MA, USA). Perkin Elmer empty stainless steel TDU tubes (89 mm × 6.35 mm o.d.) and polytetrafluoroethylene (PTFE) desorption tube inserts were used to place the samples in the TD tubes. Quartz filter (MN QF-10, Ø = 50 mm) from Macherey-Nagel (Düren, Germany) was used to immobilize the sample in the PTFE insert and also to apply the reference solution for method validation. Separations were carried out on an AT-Aquawax column (30 m × 0.53 mm, df = 0.50 ␮m) from Grace (Columbia, MD, USA). The TD-GC parameters are listed in Table 1. The effluent from the column was split to the FID and MS in a 3:1 ratio. A 10-microliter syringe with an 80 mm needle (Hamilton, Reno, NV, USA) was used to apply 2 ␮L of calibration solutions. Tubes with empty PTFE insert were conditioned at 250 ◦ C for 20 min with a constant helium flow (30 mL/min) prior to introduction of both standards and samples. 2.2.2. HS-GC-FID HS-GC-FID analyses were performed on a Perkin Elmer Clarus 480 equipped with a Turbomatrix 40 HS autosampler (balanced pressure system). All HS parameters are given in Table 2. HS vials and PTFE/Sil caps were purchased from Perkin Elmer as well. The column and GC oven temperature program were the same as used for TD-GC. A thermomixer “comfort” from Eppendorf (Hamburg, Germany) was used during digestion of gelatin.

A.A. Asfaw et al. / Journal of Pharmaceutical and Biomedical Analysis 153 (2018) 193–198 Table 2 Used HS-GC parameters. Parameter

Settings

Equilibration temperature (ET) Equilibration time Needle temperature Transfer line temperature Pressurization time Injection time Needle withdrawal time Injection port temperature Carrier gas pressure GC oven program

180 ◦ C 20 min ≥5 ◦ C above ET ≥10 ◦ C above ET 1.0 min 0.04 min 0.4 min ≥15 ◦ C above ET 130 kPa 80 ◦ C held 1 min; 80–160 ◦ C at 40 ◦ C/ min, held 5 min; 160–250 ◦ C at 40 ◦ C/min, held 2 min 1:5 220 ◦ C

Split ratio FID temperature

2.3. Standards and sample preparations 2.3.1. Standards and sample preparations for TD-GC For method optimization a 10 mg/mL solution of DMSO in methanol was prepared. For validation, a stock solution of 200 mg/mL of DMSO in methanol was prepared. Calibration solutions of seven concentrations between 0.1 and 160 mg/mL were prepared by diluting appropriate volumes of the stock solution in methanol. A 80 mg/mL solution of DMF in methanol was used as correction standard (CS) to correct the peak areas for differences in volumes applied in the TD tube. Before the calibration solutions were brought to volume (10 mL) with methanol, 2 mL of CS was added. For method optimization and validation, 10 mg of blank gelatin plugged between two double layers of quartz filter in a PTFE insert was placed in a TD tube. Finally, 2 ␮L of the reference solutions were applied on the quartz filters on the inlet side of the tube. Regarding sample analysis, 10 mg each of two experimental drug loaded gelatin samples dried under different conditions were immobilized by sandwiching them between two double layers of quartz filter in a PTFE insert which was placed in an empty desorption tube prior to TD-GC analysis. Since the amounts of residual DMSO were higher than expected (see Section 3.3), the amount of gelatin was reduced to 2 and 5 mg to avoid overloading. 2.3.2. Standards and sample preparations for FET-HS-GC 2.3.2.1. Digestion of gelatin using MMP-2 prior to HS-GC analysis. First, 50 mM HEPES buffer of pH 7.5 was prepared by dissolving 0.65 g of HEPES sodium salt in water and the pH was adjusted with concentrated hydrochloric acid before bringing to a volume of 50 mL with water. Next, 50 ␮L of MMP-2 solution (0.1 mg/mL) was diluted to 500 ␮L with HEPES buffer. The amount of enzyme and duration required to digest the gelatin were calculated on the basis of the gelatinolytic activity [18]. The following preparations were digested by continuously mixing 10 mg of sample in 1 mL of HEPES buffer with 100 ␮L of the diluted MMP-2 solution in a thermomixer at 37 ◦ C for 125 h: (i) two gelatin samples dried at different conditions after drug loading, (ii) blank gelatin for testing interference and (iii) two spiked blank gelatins for evaluation of recovery. In the latter case, the 1 mL of HEPES buffer was replaced by 1 mL of 0.1 and 1 mg/mL respectively of DMSO in HEPES buffer. When the digestion procedure was complete, each solution was diluted to 5 mL with water. 2.3.2.2. Solutions for FET-HS-GC analysis. A 20 mg/mL stock solution of DMSO in water was prepared. Calibration solutions were prepared by diluting appropriate volumes of the stock solution to obtain solutions in the concentration range from 0.05 to 2 mg/mL. To improve the sensitivity, a method developed and validated by

195

van Boxtel et al. was applied, based on the use of an acetone acetal as water scavenger prior to FET-GC for the analysis of high boiling analytes in aqueous samples [11]. So, 4 mL of each of the digested solutions (from Section 2.3.2.1) or the calibration solution were transferred into a volumetric flask of 50 mL and 40 mL of DMP was added to react with water. Reactions were catalysed by adding 10 and 20 ␮L of hydrochloric acid (37.5% w/w) to the volumetric flask for the calibration solution and the digested solution, respectively. The double volume of hydrochloric acid added to the digested solution was to compensate for the buffer used during digestion. The solution cools down significantly during the hydrolysis reaction which was considered to be complete when the flask was again at room temperature. Before completing to volume with methanol, 1 mL of the appropriate internal standard solution (benzyl alcohol in methanol, 4 mg/mL) was added. Samples were neutralised afterwards by addition of a spoonful (∼ 5 g) of sodium carbonate to obtain a saturated solution. Since sodium carbonate does nearly not dissolve in the mixture, solutions were stirred for 10 min with a magnetic stir bar to improve contact. Supernatant was decanted in a centrifuge tube and centrifuged at 4500g for 10 min. Next, 1 mL of centrifuged solution was transferred into a HS vial and the open vials were placed in a vacuum oven (≤ 0.1 kPa) at room temperature for 5.5 min to remove the excess of methanol, acetone and DMP. After evaporation, the vials were capped with a PTFE/Sil cap to be analysed with FET-HS-GC [11]. 2.4. Method validation 2.4.1. Method validation for TD-GC Selectivity was monitored by evaluating the separation of analytes and the peak purity through MS, as well as by checking the possible interference of compounds extracted from the blank quartz filter, blank gelatin or the loaded drug. The linearity was tested by analyzing calibration solutions with the described TD-GC method at seven concentrations in a range from 0.1 to 160 mg/mL, each injected in triplicate. After plotting the peak areas vs. the concentrations, the regression coefficient was calculated. The limit of detection (LOD) and limit of quantification (LOQ) were determined using the signal-to-noise (S/N) ratio: S/N = 3 for LOD and S/N = 10 for LOQ. To exclude any interference from contaminants, blank gelatin was analysed as described above. Repeatability was tested by analyzing six tubes of 20 mg/mL standard solution in 1 day and expressed as percentage relative standard deviation (RSD). Before loading the 2 ␮L of reference solution, the sample tubes (containing blank gelatin sandwiched between the quartz filters) were conditioned for 20 min at 250 ◦ C under a helium flow of 30 mL/min to ensure that they were free from volatiles. The accuracy of the method was determined by comparing the detector response after spiking 2 ␮L of standard solutions at three concentration levels of 1, 20 and 160 mg/mL (corresponding to 2, 40 and 320 ␮g respectively) on quartz filter in the presence and absence of 10 mg of blank gelatin. The recovery without the blank gelatin was considered to be 100%. Complete extraction of the analyte during the first run was confirmed by performing a second run on the same tube. 2.4.2. Method validation for FET-HS-GC Linearity was assessed by the regression coefficient of a five point calibration curve in the range from 0.05 to 2 mg/mL. Repeatability was tested by analyzing the 1 mg/mL concentration six times. The recovery was evaluated by spiking blank gelatin with two concentrations (10 mg of blank gelatin spiked with 0.1 and 1 mg of DMSO) and then treated with the procedures described in Section 2.3.2.1. Finally, all the calibration and recovery solutions were treated with the procedure mentioned in Section 2.3.2.2 and then analysed by the FET-HS-GC method.

196

A.A. Asfaw et al. / Journal of Pharmaceutical and Biomedical Analysis 153 (2018) 193–198

Fig. 1. Sample introduction to TD tube.

3. Results and discussion 3.1. TD-GC method 3.1.1. TD-GC method development Quartz filter paper was used as support material to apply the reference solution and to immobilize the gelatin because it is inert, has a high thermal stability (above 900 ◦ C) and is easy to cut to fit into the thermal desorption tube [19]. A sample or blank gelatin was brought into a PFTE insert with a double layer of quartz filter at both ends (see Fig. 1) before it was put in the desorption tube for TD analysis. This was done to ensure that molten sample does not move out of the tube and can easily be removed after analysis. Spiking was performed on the inlet side of the tube so that the spiked compounds of interest should pass through the gelatin bed during the primary desorption. This imitates adsorption/desorption of real samples as closely as possible. An important consideration during the primary desorption is whether the volatile components released from the sample are being properly retained on the cold trap sorbent bed. With a single sorbent any breakthrough will result in the loss of compounds, the extent of which will be unknown. If the sorbent is too weak, the very volatile compounds will break through, while the less volatile ones are retained. On the other hand, if the chosen sorbent is too strong, the very volatile compounds will now be retained, but it will be more difficult later to desorb the less volatile ones. For this reason most traps are composed of two beds, a weaker and a stronger [20]. In this study, first an Air Toxic trap was tried which has two types of adsorbents, carbon black as weak adsorbent and carbon molecular sieve as strong adsorbent. However, the peak for DMSO was broad and showed tailing due to the problematic desorption of DMSO from the strong adsorbent. Therefore, a Tenax TA trap which is weak, inert and hydrophobic, was used and the obtained chromatograms showed sharp peaks and no interference. 3.1.2. Optimization of TD parameters To investigate the influence of the TD parameters, a set of experiments was performed in which the FID response was monitored in function of a single parameter. An overview of the ranges explored and final settings is given in Table 1. For each experiment, 2 ␮L of standard solution was applied on a blank gelatin plug between quartz filter inside a PTFE insert. The resulting data for the optimization were visualized in graphs expressing the relative peak area in function of the respective parameter whereby the lowest area in each dataset was arbitrarily chosen as reference point (100%). Desorption at high temperature may result in decomposition of analyte, while a low temperature could lead to incomplete desorption of the residual solvent. Fig. 2(a) depicts that desorption of DMSO reached a plateau between 220 and 240 ◦ C. So, 230 ◦ C was selected as optimum primary desorption temperature. No decrease in the FID peak area of DMSO or no extra peak in the MS chromatogram was noticed till the temperature reached 240 ◦ C.

In order to minimize analysis time and optimize sample throughput, the optimum desorption time was investigated. As indicated in Fig. 2(b), around 80 min complete desorption was obtained, but after 100 min a decrease was observed. This breakthrough in function of time is caused by the increasing gas volume (=time × flow) and might be influenced by the fact that the lower temperature could not be maintained in the centre of the trap because of the hot gas flowing through it for a long time. Since only the compartment around the trap is temperature controlled, the low temperature of the trap itself may not always be guaranteed during desorption. So, 90 min was selected as desorption time. Carrier gas flow rates may be increased to enhance the desorption process from the tube and the trap, and to facilitate transfer to the cold trap. Implementation of sample splitting often benefits the thermal desorption process by allowing higher tube and/or trap desorption flows than (and independent of) conventional GC column flows. Increasing the flow can be very useful to lower the temperature and to decrease the time required for desorption. As shown in Fig. 2(c), the peak area of DMSO increases with the tube desorption flow till 120 mL/min. Above this value a small decline in response was observed. This might be due to the increased gas volume as explained above, resulting in loss of analyte before injection.

3.1.3. TD-GC method validation Analysis of blank gelatin revealed the presence of a few compounds. Some could originate from the degradation of gelatin itself. Anyhow, they did not interfere with the compound of interest. To confirm complete extraction of DMSO and DMF during the first run, a second run was performed on the same tube. As a result, no DMSO or DMF peak was observed in the chromatogram of the second run (Fig. 3). The calibration curve was constructed by repetitive injections (n = 3) of each calibration point. The R2 -value was found to be higher than 0.999, indicating that a good linear relationship was obtained. The LOD and LOQ of the method were 0.05 and 0.1 mg/mL respectively (corresponding to 0.1 and 0.2 ␮g on tube). The repeatability of the method was good with RSD values not exceeding 3%. Recoveries were found to be 99.8%, 99.7% and 99.8% for spiking with 2, 40 and 320 ␮g, respectively.

3.2. FET-HS-GC method 3.2.1. FET-HS-GC method development An attempt was made to prepare a solution of a gelatin sample for conventional HS injection to obtain comparative quantitative data for the DMSO peak obtained by TD-GC. However, the sample was insoluble in most of the common organic solvents available in the laboratory such as tetrahydrofuran, DMF, acetone and methylene chloride. On the other hand, it would have been convenient to have gelatin dissolved in a suitable solvent to perform conventional HS-GC. An aqueous gelatin solution was obtained after digesting the gelatin according to the procedure described in Section 2.3.2.1. Completeness of the digestion procedure was confirmed by leaving the solution for more than 24 h. No coagulation of the gelatin was observed. Sensitive analyses of high boiling polar compounds like DMSO in aqueous samples is not evident [21]. The use of FET-HS-GC can be beneficial, but still a pressure limit has to be respected, meaning that the introduced sample volume and therefore the sensitivity is limited for aqueous samples [22]. To overcome the abovementioned problems, DMP was used as water scavenger prior to FET-HS-GC analysis enabling sample enrichment leading to 10fold improvement in sensitivity [11].

A.A. Asfaw et al. / Journal of Pharmaceutical and Biomedical Analysis 153 (2018) 193–198

197

Response %

800

0 150

200

250

300

600 400 200 0 0

T (°C)

50 100 Time (min)

c) Desorption flow 250 200 150 100 50 0 0

100

200

300

Fig. 2. Charts illustrating the influence on the detector response for DMSO of (a) the TD primary desorption temperature, (b) the TD primary desorption time and (c) the TD primary desorption flow.

Fig. 3. Chromatogram of first desorption (upper chromatogram) and second desorption on the same tube (lower chromatogram). Degradation products are related to gelatin.

3.2.2. FET-HS-GC method validation When blank gelatin was analysed, no peaks from the sample matrix were noticed in the chromatogram. A calibration curve with 5 different concentration levels of DMSO was constructed in the range from 0.05 to 2 mg/mL and an R2 -value of 0.998 was obtained. The results for recovery were found to be acceptable (100.5% and 99.2% for blank gelatin spiked with 0.1 and 1 mg of DMSO, respectively) and revealed that there was no effect of the

digestion procedure or the matrix. RSD values (n = 6) were below 1.5%. 3.3. Application The two methods were applied to analyze residual DMSO in experimental drug loaded gelatin samples to investigate the difference between two drying approaches (freeze drying alone versus freeze drying followed by vacuum oven). The samples were treated

198

A.A. Asfaw et al. / Journal of Pharmaceutical and Biomedical Analysis 153 (2018) 193–198

Table 3 Residual amount of DMSO in drug loaded gelatin samples using FET-HS-GC and TD-GC. Method

FET-HS-GC TD-GC

Residual DMSO (RSD, n = 6) in drug loaded gelatin Freeze dried at 60 ◦ C

Freeze dried at 60 ◦ C + vacuum oven at 80 ◦ C

13.7% (0.7%) 13.6% (2.2%)

3.6% (0.2%) 3.7% (1.4%)

according to the protocols mentioned in Section 2.3.1 for TD-GC and Section 2.3.2 for FET-HS-GC. The results are shown in Table 3. When using 10 mg of sample for the TD-GC method, results fell out of range because they were higher than expected (>3%). To avoid overloading, the amount of sample was reduced to 2 and 5 mg for the “freeze dried” and “freeze and oven dried” sample, respectively. For both samples the amount of DMSO is above the ICH acceptance limit (0.5%) [6] so that further improvement of the drying process is mandatory. A statistical comparison of the results obtained by FET-HS-GC and TD-GC proved that there is no significant difference between the two methods (p = 0.5 at 95% confidence interval). In case of samples with a low residual amount of DMSO, the sample size in the desorber tube can easily be increased to obtain a more sensitive determination. 4. Conclusions In this study, the application of TD-GC for the determination of residual DMSO in solid gelatin samples was investigated. The use of enzymatic digestion of gelatin prior to more conventional FET-HS-GC was also applied for comparison, but it is less obvious from a practical point of view since the TD-GC method does not require long lasting sample pretreatment and also no effort to select a suitable solvent. As a result, the TD-GC chromatogram shows no interference from a large solvent peak nor from impurities present in the solvent. In addition, since larger sample sizes can be brought in the TD tube, it is possible to further increase the sensitivity if necessary. Quantitative results of the proposed method showed good agreement with the conventional HS-GC method. The results also indicated that TD-GC has a great potential for the quantitative determination of RS directly from solid products due to its ease of operation, sensitivity and reliability. Acknowledgements Prof. Guy Van den Mooter (Drug Delivery and Disposition, KU Leuven) is kindly acknowledged for the provision of the drug loaded gelatin samples. The authors would like to thank the Interfaculty Council for Development Co-operation (IRO, KU Leuven) for support.

References [1] M. Marques, Enzymes in the dissolution testing of gelatin capsules, AAPS PharmSciTech 15 (2014) 1410–1416. [2] A. Bigi, S. Panzavolta, K. Rubini, Relationship between triple-helix content and mechanical properties of gelatin films, Biomaterials 25 (2004) 5675. [3] M. Djabourov, J. Leblond, P. Papon, Gelation of aqueous gelatin solutions I. Structural investigation, J. Phys. 49 (1988) 319–332. [4] V.A. Grigor’eva, L.Z. Rogovina, G.L. Slonimskii, The rheological properties of solutions and gels of gelatin in different solvents, Polym. Sci. U.S.S.R. 17 (1975) 165–171. [5] P.V. Kozlov, G.I. Burdygina, The structure and properties of solid gelatin and the principles of their modification, Polymer 24 (1983) 651–665. [6] Guidelines for Residual Solvents, International Council on Harmonization (ICH), Geneva, Switzerland, 2009. [7] Identification and Control of Residual Solvents (Section 2.4.24), European Pharmacopoeia, 9th ed., European Department for the Quality of Medicines, Council of Europe Strasbourg, 2017, 2018. [8] Y. Sitaramaraju, A. Van Hul, K. Wolfs, A. Van Schepdael, J. Hoogmartens, E. Adams, Static headspace gas chromatography of (semi-)volatile drugs in pharmaceuticals for topical use, J. Pharm. Biomed. Anal. 47 (2008) 834–840. [9] Y. Sitaramaraju, A. Riadi, W. D’Autry, K. Wolfs, J. Hoogmartens, A. Van Schepdael, E. Adams, Evaluation of the European Pharmacopoeia method for control of residual solvents in some antibiotics, J. Pharm. Biomed. Anal. 48 (2008) 113–119. [10] D.M. Kialengila, K. Wolfs, J. Bugalama, A. Van Schepdael, E. Adams, Full evaporation headspace gas chromatography for sensitive determination of high boiling point volatile organic compounds in low boiling matrices, J. Chromatogr. A 1315 (2013) 167. [11] N. van Boxtel, K. Wolfs, A. Van Schepdael, E. Adams, Application of acetone acetals as water scavengers and derivatization agents prior to the gas chromatographic analysis of polar residual solvents in aqueous samples, J. Chromatogr. A 1425 (2015) 62–72. [12] J. Li, S. Shao, M. Solorzano, G.J. Allmaier, P.T. Kurtulik, Determination of the residual ethanol in hydroalcoholic sealed hard gelatin capsules by static headspace gas chromatography with immiscible binary solvents, J. Chromatogr. A 1216 (2009) 3328–3336. [13] Y. Yu, B. Chen, C. Shen, Y. Cai, M. Xie, W. Zhou, Y. Chen, Y. Li, G. Duan, Multiple headspace single-drop microextraction coupled with gas chromatography for direct determination of residual solvents in solid drug product, J. Chromatogr. A 1217 (2010) 5158–5164. [14] K. Hashimoto, K. Urakami, Y. Fujiwara, S. Terada, C. Watanabe, Determination of residual solvents in pharmaceuticals by thermal desorption-GC/MS, Anal. Sci. 17 (2001) 645–648. [15] E. Woolfenden, Thermal desorption for gas chromatography, in: C.F. Poole (Ed.), Gas Chromatography, Elsevier Inc., Waltham, 2012, pp. 245–247. [16] N. Ramírez, A. Cuadras, E. Rovira, F. Borrull, R.M. Marcé, Comparative study of solvent extraction and thermal desorption methods for determining a wide range of volatile organic compounds in ambient air, Talanta 82 (2010) 719–727. [17] E. Woolfenden, Optimising analytical performance and extending the application range of thermal desorption for indoor air monitoring, Indoor Built Environ. 10 (2001) 222–231. [18] J.L. Seltzer, K.T. Akers, H. Weingarten, G.A. Grant, D.W. McCourt, A.Z. Eisen, Cleavage specificity of human skin type IV collagenase (Gelatinase): Identification of cleavage sites in type I gelatin with confirmation using synthetic peptides, J. Biol. Chem. 265 (1990) 20409–20413. [19] A.A. Asfaw, K. Wolfs, A. Van Schepdael, E. Adams, Thermal desorption-gas chromatographic methodology for the determination of residual solvents in mesoporous silica, J. Chromatogr. A 1500 (2017) 160–166. [20] E. Woolfenden, Monitoring VOCs in air using sorbent tubes followed by thermal desorption-capillary GC analysis: summary of data and practical guidelines, J. Air Waste Manage. Assoc. 47 (1997) 20–36. [21] K. Ui, M. Abo, A. Okubo, S. Yamazaki, An improved method for the analysis of dimethyl sulfoxide in water samples, Anal. Sci. 20 (2004) 223. [22] H.C. Hu, X.S. Chai, C.H. Wei, D. Barnes, Increasing the sensitivity of headspace analysis of low volatility solutes through water removal by hydrate formation, J. Chromatogr. A 1343 (2014) 42.