Direct Comparison of 19F qNMR and 1H qNMR by Characterizing Atorvastatin Calcium Content

Quantitative nuclear magnetic resonance (qNMR) is a powerful tool in measuring drug content because of its high speed, sensitivity, and precision. Most of the reports were based on proton qNMR (1H qNMR) and only a few fluorine qNMR (19F qNMR) were reported. No research has been conducted to directly compare the advantage and disadvantage between these two methods. In the present study, both 19F and 1H qNMR were performed to characterize the content of atorvastatin calcium with the same internal standard. Linearity, precision, and results from two methods were compared. Results showed that 19F qNMR has similar precision and sensitivity to 1H qNMR. Both methods generate similar results compared to mass balance method. Major advantage from 19F qNMR is that the analyte signal is with less or no interference from impurities. 19F qNMR is an excellent approach to quantify fluorine-containing analytes.


Introduction
Quantitative nuclear magnetic resonance (qNMR) has been widely utilized in pharmaceutical analysis [1][2][3], natural products characterization [4,5], and reference substances quality control [6][7][8][9]. This technique has several advantages including fast sample preparation, no necessity for reference material, and generating both structure information and quantification result in one experiment. Currently, most of the qNMR reported are proton qNMR ( 1 H qNMR). For some analytes, choosing a suitable internal standard (IS) is often challenging since the response signals in 1 H NMR are generally from 0 to 15 and signal overlapping occurs easily. When an analyte is mixed with various excipients in medicines or different metabolites in body fluids, the deployment of 1 H qNMR might be impossible due to severe signals overlapping. Some groups reported the application of heteronuclear 2D qNMR techniques which can avoid the signal overlapping mentioned previously [10,11]. But these 2D qNMR are time consuming (more than two hours) and the results are easily affected by T2 and coupling constant which are generally not a crucial parameter in 1D qNMR [10]. 19 F NMR has been utilized in characterizing fluorinecontaining pharmaceutical and metabolites in complicated matrices [12,13] since drug excipients or body fluids barely contain fluorine and do not interfere with analytes. One major advantage of 19 F NMR is that signals in 19 F NMR barely overlap due to its broad response range (ca. −200 to 100) which makes the selection of IS in 19 F qNMR much easier than in 1 H qNMR. Several groups have reported the deployment of 19 F qNMR in characterizing pharmaceutical [14,15], metabolites [16], and biooils [17].
Both 1 H and 19 F qNMR have their advantages and drawbacks. Although 1 H qNMR is potentially applicable to any analyte containing proton, the application of this method is limited when analytes are in complicated matrices. And the selection of IS should be careful to avoid interference with the analyte. On the contrary, interference from matrices or IS seldom happens in 19 There is no direct comparison of 19 F qNMR and 1 H qNMR such as signal sensitivity, linearity, and RSD. To fully understand the applicable conditions of these two methods, we chose 4,4 -difluorodiphenylmethanone which has both hydrogen and fluorine atoms as an IS to analyze the content of atorvastatin calcium with both 19 F and 1 H qNMR ( Figure 1). Directly comparing results from the same analyte and IS by different qNMR experiments can avoid the influence of sample purity. Besides, the method validation data from both 19 F and 1 H qNMR were also compared.

Materials and Analyte Preparations. 4,4 -Difluorodi
where and are the signal response of the analyte and IS, and are the number of spin atoms (fluorine in 19 F and proton in 1 H qNMR) in the analyte and IS, is the molecular weight of analyte (1155.4 g/mol), is the molecular weight of IS (218.2 g/mol), and are the mass of the analyte and IS, and is the purity of the IS. ( 1) is an essential parameter in qNMR experiments. 1 should be more than 5 times of longitudinal relaxation ( 1) to make sure more than 99% of nuclei return to their equilibrium status [18]. 1 in 19 F and 1 H qNMR experiments were determined by an inversion recovery method. 1 of the analyte signals in 19 F and 1 H experiments were found to be 0.86 s and 2.18 s, respectively. 1 of the IS in 19 F and 1 H experiments were 1.80 s and 2.97 s. So 1 in both 19 F and 1 H qNMR experiments were set as 15 s to make sure the full relaxation is achieved before next repulsion.

Optimization of Experiment Parameters. Relaxation time
Transmitter offset (O1P) and spectral width (SW) are another two important parameters in qNMR experiments. In 1 H qNMR experiments, default O1P (6.175 ppm) and SW (20 ppm) settings worked well. On the contrary, O1P and SW in 19 F qNMR experiments must be manually modified. When default O1P (−100 ppm) and SW (241 ppm) were used, the spectrum is difficult to perform phase and baseline correction. In this study, the signals of analyte and IS appeared at −111.9 and −104.4, respectively. So O1P was set at the center of two signals −108. Meanwhile, it is reported that response signals should not locate at the edge of a spectrum to avoid distortion [19]. Here, SW was set at 60 ppm to fulfill the requirement. experiment rarely occur. Generally, any fluorine-containing compound with high purity which does not react with the analyte is eligible as an IS in 19 F qNMR experiment. Here, 4,4 -difluorodiphenylmethanone was utilized (Figure 3). 19 F and 1 H qNMR methods were validated ( Table 1). The linearity of both methods was measured by using the solutions prepared by dissolving desired amount of analyte and IS in one tube. The ratio of calculated analyte mass to added analyte mass was fitted to a linear curve. The correction coefficient showed that both 19 F and 1 H methods had good linearity within 3.21-20.34 mg/mL concentration ranges with 2 > 0.99.

Precision, Repeatability, and Stability.
Precision tests were carried out by testing the same solution six times. The repeatability experiments were achieved by characterizing six independent solutions containing both analyte and IS. The RSD of precision and repeatability indicates the good accuracy of the both methods. The stability of solutions was assessed by analyzing one analyte at 1, 2, 4, 6, and 8 hours

Conclusions
For the first time, 19 F and 1 H qNMR were performed and directly compared with the same analyte and IS. Method validation data showed that 19 F qNMR has similar accuracy, sensitivity, and reproducibility to 1 H qNMR. The quantitative results from the two methods are comparable to that from mass balance measurement. 19 F qNMR is valuable in quantitatively analyzing fluorine-containing analytes which are codissolved with excipients, body fluids, or various metabolites. 19 F qNMR can be widely applied in early drug research and development as well as clinical trials.