Determination of Genotoxic Azide Impurity in Cilostazol API by Ion Chromatography with Matrix Elimination

: Cilostazol is a commonly used active pharmaceutical ingredient (API) to treat and reduce the symptoms of intermittent claudication in peripheral vascular disease. Recently, it was found to be a potential medicine in the effective treatment of COVID-19. In addition to the positive effects of this API, genotoxic sodium azide is used in the synthesis of cilostazol that can appear in the API. In this work, a method was developed for the determination of sodium azide (as azide anion) in cilostazol API at 7.5 ppm limit level by using ion chromatography (IC) and liquid–liquid extraction (LLE) sample preparation. The liquid–liquid extraction allows the application of high sample concentrations. Because of the low limit concentration (7.5 ppm), 500 mg sample was dissolved in 5 mL solvent. By using LLE for sample preparation, the huge amount of cilostazol was omitted and column overload was avoided. The developed method was validated in accordance with the relevant guidelines. Speciﬁcity, accuracy, precision, limit of detection and limit of quantiﬁcation parameters were evaluated. The calculated limit of detection was 0.52 ppm (S/N:3) and the limit of quantiﬁcation was 1.73 ppm (S/N:10) for sodium azide. The recovery of the sodium azide was 102.4% and the prepared solutions were stable in the sample holder for 24 h.


Introduction
The cilostazol is a platelet-aggregation inhibitor and arterial vasodilator, its long-term use may prevent stroke [1]. Recently, a network-based ranking was used to prioritize drugs to treat COVID-19 symptoms [2]. Cilostazol was the fourth on this list. Several studies were published about cilostazol long-term treatment safety. It was found by W.R. Hiatt et al. that the mortality was not higher in the treatment group, than the placebo group during the examined 42 months [3]. In addition to the long-term treatment safety of cilostazol, attention should be drawn to the possible impurities because these can cause deteriorate side effect for patients in the long term. Figure 1 shows a possible synthesis pathway of cilostazol. It can be seen that sodium azide is used for forming the tetrazole ring [4]. Sodium azide is toxic and genotoxic. Its lowest lethal dose is 10 mg/kg. After poisoning, death can occur within an hour by hypotension [5].
According to the European Medicine Agency (EMA) and International Council for Harmonization (ICH), the maximum daily intake of potential genotoxic impurities for more than 12 months of exposure is 1.5 µg. The limit concentration of genotoxic impurity depends on the daily dose of API [6] that varies depending on whether it is used alone or in combination with other anti-platelet agents [1]. If it is taken alone the daily dose is 200 mg/day, if it is taken with CYP2C19 inhibitors it is 100 mg/day [7]. Considering the daily dose mentioned before, this means that the permitted sodium azide concentration is only 7.5 ppm. Due to its low limit concentration, determination of sodium azide is a problem in APIs and pharmaceutical products. Vinkovic and Drevenkar developed an ion chromatographic method for azide determination in protein samples [8]. The USP refer to a method for determining azide impurity in ibersartan API, however, the sample preparation is not applicable for other APIs in general. In this method [9], azide anions are separated by ion chromatography with 0.1 N sodium hyroxide eluent and conductivity detector. The specified column (L31) is a strong anion exchange column with quaternary amine groups [10]. This method was applied for the determination of sodium azide in a range of "sartan" drugs. The sodium azide was quantified at 15 ppm level while the retention time of azide anion was about 42 min [11]. A reversed phase liquid chromatography method was developed for sartan APIs by Gricar and Andrensek [12]. In this RP-HPLC method, UV detection was used for determination of azide at 10 ppm limit level. During the sample preparation, the APIs were precipitated and removed by filtration. In case of cilostazol the limit is lower, and the sample preparation should be carried out in different solution. In order to control this impurity at 7.5 ppm level, a large amount of the sample must be dissolved. This high cilostazol API concentration can decrease the efficiency of the analytical method by overloading the chromatographic system. Therefore, during the sample preparation, the amount of the API needs to be decreased while the quantity of sodium azide should remain the same.
The aim of this work was to develop and validate an analytical procedure for the determination of sodium azide content in cilostazol API. In our method, 500 mg cilostazol is dissolved in 5 mL of methylene chloride. Since sodium azide is hydrophilic and cilostazol is insoluble in water, aqueous phase liquid-liquid extraction (LLE) was used for the sample preparation followed by the ion chromatographic determination of the extracted azide anions. The validation was carried out in accordance with ICH Q2(R1) guideline recommendations [13].

Materials and Methods
Analytical grade dimethyl sulfoxide, methylene chloride (Fisher Scientific, Loughborough, UK) and sodium azide (Sigma-Aldrich,, Darmstadt, Germany) was used for sample preparations. Cilostazol originates from the synthesis of Egis Pharmaceuticals Plc. Water was prepared freshly using ELGA Purelab system (ELGA, Lane End, UK). Mettler Toledo analytical balances were used for weighing (Greifensee, Switzerland) and Eppendorf automatic pipettes were used for liquid handling (Hamburg, Germany).

Sample Preparation with Liquid-Liquid Extraction
Because of the low concentration limit of sodium azide, large amount of cilostazol API is needed for the determination. The high concentration of API may have a negative impact on the ion chromatographic determination of azide ion. Cilostazol can precipitate in the eluent and deteriorate the column performance or overload it. A suitable sample preparation technique should be developed to avoid overloading problems. Due to the significantly different solubilities of cilostazol and sodium azide in the immiscible solutions of methylene chloride and dimethyl-sulfoxide/water, liquid-liquid extraction can be used efficiently for sample preparation. The methylene chloride solution is the lower phase and the aqueous solution is the upper phase. Then, 500 mg cilostazol is dissolved in 5 mL of methylene chloride (100 mg/mL). The sodium azide is extracted by 5 mL of dimethylsulfoxide/water. The extraction efficiency was tested on three temperatures (20, 25, and 30 • C). It was found that the extraction efficiency did not depend on the temperature. Even if the temperature affected slightly the solubility of cilostazol in the aqueous phase, it did not affected the extraction of the azide anion.

Effect of Flow Rate
The examined flow rates were over the optimum velocity. However, the results met the system suitability requirements, and the theoretical plates did not change significantly at different flow rates. Accordingly, so either flow rate could be applicable.

Effect of Eluent Concentration
The retention factor azide anion was determined at different eluent concentrations (5, 10, 15, 20, 25, and 30 mM). Figure 2 shows that as retention factors of azide anions decreased by the increasing concentration of the eluent. In ion exchange chromatography, the plot of logarithm (with base 10) of retention factor versus logarithm of eluent concentration should be linear for isocratic separations. The slope of equation (−0.989) that is fitted for the measured data points verifies that the retention behavior of azide ion is in line with equilibrium theory.

Effect of Column Temperature
The effect of column temperature on the separation of azide anions were studied in the range of 25 • C to 40 • C in six points. The results showed that the azide ion retention time did not depend significantly on the column temperature. The relative standard deviation of the six retention times (measured on six different column temperature) was only 0.18%.

Validation of the Chromatographic Method
A developed and optimized method can only be used for quantitative measurements of raw materials, intermediates or APIs if the applicability of the method was proved earlier so the method is validated. The validation of this method was performed according to ICH Q2(R1) guideline [13] for limit tests.

Solvent Preparations
The following solutions were used during chromatographic method validation: • Blank: mixture of 5 mL purified water and 5 mL methylene chloride. It was homogenized by shaking for at least one minute; • Sodium azide solution: 75 µg/mL of sodium azide in dimethyl sulfoxide; • Sodium azide reference solution: 5 mL of methylene chloride and 50 µL of the sodium azide solution was added into a HS vial. After the dissolution, 5 mL of purified water was added. It was homogenized by shaking for at least one minute. For the measurement the upper phase was used; • Test solution: 100 mg/mL of cilostazol in methylene chloride. After the dissolution 5 mL purified water was added. It was homogenized by shaking for at least one minute. For the measurement the upper phase was used; • Spiked test solution: 500 mg of cilostazol and 50 µL of the sodium azide solution was measured into 5.0 mL methylene chloride. After the dissolution 5 mL purified water was added. It was homogenized by shaking for at least one minute. For the measurement the upper phase ws used; • Limit of detection: 5 mL of methylene chloride and 50 µL of 22.5 µg/mL of sodium azide solution was added into a HS vial. After the dissolution 5 mL purified water was measured. It was homogenized by shaking for at least one minute. For the measurement the upper phase was used.

Ion Chromatographic Method for Azide Determination
As a result of the preliminary experiments the final anion chromatographic method can be seen in Table 1.

Validation Measurements
The following measurements were carried out during limit test validation: • Specificity. The specificity test has to be made for the limit validation. It can verify that the method is specific and selective for sodium azide; • Limit of quantification (LQ) and limit of detection (LD). Limit of detection (LD) and limit of quantification (LQ) were specified as the minimum concentration, at which the signal of the investigated component was at least three times (LD) and ten times (LQ) greater than the noise level; • System precision. System precision was demonstrated by calculating the repeatability of six replicate injections of the reference solution at limit level; • Accuracy and stability. In the study of accuracy, the sample was spiked with sodium azide at limit level (7.5 ppm). The concentrations were determined, and the recoveries of the spiked quantities were calculated in each case. The stability was determined by analyzing the prepared solutions over a period of 24 h in closed plastic vials in the sampler holder.
Results of validation are presented in Table 2. Figure 3. shows that no interference from blank and peak due to any impurity was observed at the retention time of sodium azide peaks. The sodium azide reference solution and sample solution stability were measured and the result was, that these were stable for 24 h in the sampler holder. The method usefulness was also proved by measuring four consecutive production batches.

Conclusions
A fast and effective high performance ion chromatographic (HPIC) method was developed for the determination of sodium azide content of cilostazol API. For the appropriate detection, a liquid-liquid extraction (LLE) step was necessary. The proposed new HPIC method developed for quantitative determination of sodium azide in cilostazol drug substance is accurate, precise, robust and selective. The solutions, which were made with LLE sample preparation, are stable at least 24 h. The method produced satisfactory validation data for the tested parameters for the appropriate ICH guidelines. The developed method is simple, cost-effective, and provides the possibility to reduce the limit concentration by up to three quarters from the current 7.5 ppm, if necessary.