A high-performance liquid chromatographic (HPLC) method with ultraviolet (UV) detector had been developed for simultaneous quantification of danshensu, protocatechuic aldehyde, caffeic acid, salvianolic acid D, rosmarinic acid, salvianolic acid B and salvianolic acid A in Danshen injection. According to the UV spectra of these components, three detection wavelengths have been selected as follows: 280 nm for danshensu and protocatechuic aldehyde, 326 nm for caffeic acid, salvianolic acid D and rosmarinic acid, 286 nm for salvianolic acid B and salvianolic acid A. The limit of detection (LOD) was improved to be in the range of 0.008~0.160 μg/ml. Moreover, excellent linear behavior over the investigated concentration range was observed, with R>0.999 for all the analytes.

DSIs were supplied by a Chinese pharmaceutical manufacturer (Qingchunbao, Zhejiang, China). Standard substances including DSS, PA, CA, RA and SaB were purchased from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China). SaD and SaA were isolated from aqueous extract of DS. These isolated salvianolic acids were identified on the basis of mass spectrum (MS) data, purities of which were all over 97% by HPLC analysis. The HPLC-grade acetonitrile and formic acid were purchased from Merck (Darmstadt, Germany) and the Tedia Company (Fairfield, USA), respectively. Deionized water used throughout the experiments was produced using a Mill-Q academic water purification system (Milford, MA, USA).

The HPLC system HP 1100 series (Agilent Technologies, Waldbronn, Germany) was equipped with Chemstation Software (Agilent Technologies) and was comprised of a quaternary pump, an online vacuum degasser, an autosampler and a thermostated column compartment. A UV detector was used for the chromatographic analysis and a DAD detector was used to obtain UV spectra of seven standard substances. All separations were carried out on a Tigerkin C₁₈ column (200 mm×4.6 mm i.d., 5.0 µm particle size) from Dalian Sipore Co., Ltd. (Dalian, China). A linear gradient elution of Eluents A (0.5% (v/v) aqueous formic acid) and B (0.5% (v/v) formic acid in acetonitrile) was used to run the separation. The elution programme was well optimized and conducted as follows: the first linear gradient was 5%~20% Eluent B in the range of 0~10 min, the second one was 20%~25% Eluent B in the range of 10~17 min, and the last one was 25%~55% Eluent B in the range of 17~35 min. Then the system was restored to initial conditions after 5 min. The solvent flow rate was 1.0 ml/min, the injection volume was 20 µl, and the column temperature was maintained at 30 °C. The chromatograms were recorded at 280 nm in the range of 0~13 min, 326 nm in the range of 13~23.5 min, and 286 nm in the range of 23.5~35 min.

The standard stock solutions of DSS (4.05 mg/ml), PA (5.10 mg/ml), CA (1.0 mg/ml), SaD (1.15 mg/ml), RA (2.70 mg/ml), SaB (1.00 mg/ml) and SaA (1.10 mg/ml) were prepared in phosphoric acid-methanol-water (0.5/80/19.5, v/v/v) and stored away from light at 4 °C. Working solutions of the lower concentration were prepared by an appropriate dilution of the stock solution.

DSI of 200 µl was diluted to 1 ml with deionized water and the dilution was injected into the HPLC system.

An aqueous acetonitrile solvent system was often used in the HPLC analysis of phenolic acids in DS (Wasser et al., 1998; Liu et al., 2000; Zhang et al., 2005; Zhou et al., 2006; Ma et al., 2007) and the modifiers were formic acid, acetic acid, or phosphoric acid. Water (0.5% (v/v) formic acid) and acetonitrile (0.5% (v/v) formic acid) were selected as the mobile phases in the experiment. From the UV spectra of these seven phenolic acids (Fig.2), 280 nm for DSS and PA, 326 nm for CA, RA and SaD, and 286 nm for SaB and SaA, were selected as detective wavelengths. Five column temperatures (15, 20, 25, 30 and 35 °C) were investigated and the separation of CA, SaD and SaB was found to be improved when the temperature stepped up. However, base-line separation of DSS was obtained at 30 °C (Fig.3). As a result, the column temperature was determined as 30 °C. The most suitable flow rate was found to be 1.0 ml/min. Finally the chromatograms, after optimization of the mixture of standard solution and the real sample solution, were described as in Fig.4.

Integrated chromatographic peak areas (Y) were plotted against the corresponding concentrations (X, μg/ml) of the seven constituents in the standard solutions to obtain calibration curves based on linear regression analysis. The regression curves were obtained from six concentration levels and all the analytes had good linearity (R>0.999) in the investigated ranges (Table 1). Limit of quantification (LOQ) was defined as the lowest concentration level resulting in a peak height of 10 times the baseline noise (the signal-to-noise ratio (S/N) is 10). The minimum concentration, which could be calculated at S/N=3, was considered to be the LOD. The LOD values for seven target components were in the range of 0.008~0.160 µg/ml (Table 1), which were better than those (0.04~0.43 µg/ml) in a previous report (Wasser et al., 1998).

The analysis repeatability was examined by using six duplicate samples from the same batch of DSI, treated according to the sample preparation procedure and analyzed with the established HPLC method. The stabilities of these seven compounds were tested at six time points (0, 2, 4, 8, 16 and 24 h) during 24 h. The result showed that DSS, PA, CA, SaD, RA and SaB were stable in the testing period, except for SaA with a stability of just 4 h (Table 2). According to this decomposing speed, it was almost certain that SaA in DSI became unstable just after opening the bottle of injections. After 4 h, the chromatographic peak of SaA in DSI had almost disappeared, while SaA in stock solution was found to be stable. It was thought to relate to their different pH values (about pH 7 in DSI and pH 3 in stock solution). The instability of SaA was thought to indicate that its relative standard deviation (RSD) value of repeatability analysis was more than 5% (Table 2).

The intra- and inter-day instrument precisions were determined by analyzing three different concentration solutions (low, medium and high) of authentic standards. The intra- and inter-day experiments were tested by injecting three times in a single day and three times on six successive days, respectively. The results are shown in Table 3. Due to its instability, the maximum RSD values in the precision test were obtained from SaA.

In the experiment, three different quantities (low, medium and high) of authentic standards were added to sample solutions. Then the mixture solutions were analyzed using the developed HPLC-UV method mentioned above and the quantity of each component was subsequently calculated from the corresponding calibration curve. The result (Table 4) showed that the recovery of these six phenolic acids, except SaA, ranged from 94.32%~106.05%. In previous experiments, SaA was found to be unstable in DSI and relatively stable in phosphoric acid-methanol-water (stock solution). Due to this fact, the recoveries of SaA that were obtained were much higher than 100% at low concentration and lower than 100% at medium and high concentrations. These recoveries were tested from low concentration to high concentration.

The developed method was applied to the simultaneous quantification of DSS, PA, CA, SaD, RA, SaB and SaA in DSIs from different batches. The result (Table 5) showed that as the unique target component for quality control of DSI, the content of PA approximated to 250 µg/ml in all of the samples, while contents of other components differed greatly among batches from the same company. The variety may result from the low level of quality standards. It is well known that the preparations of traditional Chinese medicines generally include multiple bioactive constituents while only one component (e.g. PA in DSI) was considered as the target for quality control. It meant that only the target compound was well monitored during the manufacturing process and other components were ignored.