For the analysis, the Bruker ALPHA II FTIR spectrometer was used. This equipment was connected to a computer to use the “OPUS” software for analysis of the spectra.

Other equipment used is as follows: Ohaus Explorer analytical microbalance, WiseVen oven, and specac hydraulic press, where the used dies are made of 440C stainless steel with a Mohs hardness of circa 7.5 and offered in two types to prepare solid sample pellets of 13 mm diameter.

Levothyroxine sodium CRS of 89.2% purity (European Pharmacopeia Reference Standard) was used as a reference to measure peak band area values.

The pharmaceutical formulations used (Euthyrox50^®, Euthyrox100^®, Euthyrox150^®, and Eltroxin^®) were purchased from a local authorized pharmacy in Lebanon.

Sigma-Aldrich chloroform (99.8%) was used as a solvent. Himedia potassium bromide powder was used to prepare pellets for IR measurements.

A weight of 10 mg of standard LT4 was dissolved in 3 mL of chloroform. To the resultant solution, 1 g of KBr powder was added. Then, the solution was evaporated in an oven set at 70 °C, allowing the LT4 to deposit on the KBr fine solid particles, obtaining a LT4-KBr mixture having a concentration of 10 µg/mg.

Different weights from the standard LT4 powder mixture (10 µg/mg) were diluted using KBr powder to obtain five different standard powder mixtures of concentrations ranging between 0.5 and 5.5 µg/mg, where the final weight of each mixture was kept constant equal to 150 mg to be later pressed into pellets containing concentrations ranging from 25 to 800 µg/pellet. Thereafter, each standard mixture was transferred into a mechanical hydraulic die press and compressed between two metal pistons with stainless steel faces in a cylinder with a pressure value of 10 tons, enough to form a thin, translucent KBr pellet through which the beam of the spectrometer could pass.

The obtained pellets were supported on a holder and placed in the line of the IR beam of the FTIR instrument, and the spectra were recorded in the transmittance mode at 400–4000 cm^-1 by averaging 16 scans for each spectrum with a resolution of 4 cm^-1. Using the software, the % transmittance values and spectra were obtained. The standard calibration curve was constructed by plotting the peak band area (1409.27 cm^-1) against their respective concentrations.

Blank readings were recorded by scanning a 150 mg KBr pellet prepared under the same conditions to ensure that there are no interferences at the specified wavenumber.

Euthyrox (50, 100, and 150 µg) and Eltroxin (100 µg) tablets were assayed using the developed method. Ten tablets from each pharmaceutical formulation were accurately weighed and finely powdered. A weight equivalent to two tablets from each pharmaceutical formulation was dissolved in 5 mL of chloroform, stirred for 5 minutes, and then filtered. To a 3 mL volume of the filtrate, 150 mg of KBr powder were added and left to evaporate as described above. Afterwards, the 150 mg dried-up mixture was used to prepare KBr pellets, and spectral measurements were conducted as described above.

During the quantitative analysis of LT4 in pharmaceutical tablets, the main problem was the interference from the excipients, where those excipients represent more than 99.9% of the marketed tablet weight. Thus, during the FTIR determination, the peak bands from those excipients will mask all the bands relative to LT4 functional groups, hence interfering with its determination. To solve this, LT4 should be extracted from pharmaceutical tablets.

Usually, the classical liquid sampling technique by IR spectroscopy is to prepare the liquid samples using a volatile solvent and then deposit a few µls onto the KBr pellet, where the solvent evaporates, leaving a thin layer of the analyte on the pellet. However, this conventional KBr pellet sampling technique cannot be used in the determination of LT4 because, by extracting the active ingredient from the pharmaceutical tablet, the resulting solution has a very low concentration range for an NTI drug. So, taking a few µls from this prepared solution will result in the deposition of non-detectable amounts of LT4 on the KBr pellet.

Accordingly, a modification is required where extraction from the pharmaceutical tablets is performed. The extract was then mixed with KBr. Afterwards, the solvent was evaporated, leaving a detectable amount of LT4 as fine particles deposited on the KBr powder. The final step was to compress the KBr powder mixture into pellets for ease of measurement.

Optimal conditions were achieved by using a constant final weight of 150 mg of spectral-grade KBr powder compressed using a mechanical press at 10 tons for 5 minutes.

These conditions ensure the formation of identical translucent KBr pellets with the same path length. Thus, quantitative determination can be conducted.

In general, the solvent to be selected to dissolve or extract the drug under study should not interact with or ruin the prepared KBr pellets. In addition, the chosen solvent should be capable of exclusively extracting the active ingredients.

Several solvents were tested (water, methanol, ethanol, and chloroform). It was found that only chloroform was able to selectively extract LT4, whereas the other tested solvents were capable of dissolving excipients along with LT4, masking LT4’s IR bands and interfering with its determination.

Accordingly, the extraction of LT4 was achieved through the selective dissolution of the active ingredient using chloroform. Also, chloroform has a low boiling point (61 °C), so it can be evaporated rapidly and easily.

To ensure the efficiency of extraction with chloroform, standard LT4 powder was mixed with KBr powder without the use of chloroform, and pellets were prepared in the same concentration ranges (25–800 µg/pellet). The obtained results showed no difference between with or without extraction for the standard in terms of quantitation of LT4, yielding the same calibration. This demonstrates good performance and complete extraction using chloroform.

The obtained IR reference spectrum of LT4 can be used as a fingerprint for its identification. In addition, the intensities of the functional group bands could be used for their quantification; hence, the concentration of LT4 can be derived.

By observing the position, shape, and relative intensities of the vibrational bands in the IR spectrum of LT4 (Fig. 1), characteristic peaks corresponding to different functional groups of LT4 have been assigned and compiled in Table 1.

Figure 1. 

IR reference spectrum of standard levothyroxine sodium.

Table 1.

Download as

CSV

XLSX

IR spectral analysis of levothyroxine sodium.

Band Wavenumber (cm-1)	Functional group
517.3	C-I
1048.8	C-O (phenol)
1184.89	C-O (ether)
1241.02	C-O (acid)
1317.88	C-N
1357.16	C-N
1409.27	C=C
1538.82	N-H
1605.09	C=O
1639.26	C=C

The FTIR spectrum of LT4 (Fig. 1) reveals the presence of many characteristic peaks. Each peak represents a characteristic functional group, such as the C–I alkyl halide-stretching band, observed at 517 cm^-1, the C–O ether-stretching band, observed at 1184 cm^-1, the C=C stretch band, observed at 1409 cm^-1 and the N–H primary amine bending band, observed at 1538 cm^-1. The degree to which each band contributes to the spectrum is directly related to LT4’s concentration. The peaks at wavenumbers 517 cm^-1, 1184 cm^-1, and 1538 cm^-1 were not used in the determination since, when relating peak band area to concentration, the obtained calibration graph resulted in unacceptable values of correlation coefficients (r). The best peak chosen for the quantification of LT4 is the C=C stretch band, which was observed at 1409.27 cm^-1 (1463–1364 cm^-1), where the obtained peak is intense and did not interfere with other spectral peaks with very good correlation coefficients (r).

Under optimized experimental conditions, the developed method was validated by determining the following parameters: linearity, limits of detection (LOD) and quantification (LOQ), specificity, accuracy, and precision (repeatability and intermediate precision) according to the procedures described in ICH guidelines (ICH 2005).

For the construction of the calibration curve, the peak band area of the C=C stretch band at 1409 cm^-1 was plotted against the concentration of LT4 in the range of 75–800 µg/pellet, as stated in Table 2. Linearity parameters, including correlation coefficient, intercept, slope, and the standard deviation of the residuals for the calibration data, in addition to the variance ratio (F-values) and concentration range, are summarized in Table 2.

The correlation coefficient (r) obtained was high (0.999) with high values of F (low significant F), which confirmed the linearity of the calibration curves. An important statistical parameter for indicating the random error in the estimated values of y is the standard deviation of the residuals, S_y/x. Also, the importance of S_y/x originates from being used to calculate S_a and S_b, the standard deviation of the intercept (a) and the slope (b). These values showed the good linearity of the calibration graphs and their compliance with Beer’s law. The variance test for the regression lines revealed that, for equal degrees of freedom, the increase in the variance ratio (F-values) means an increase in the mean squares due to regression and a decrease in the mean squares due to residuals (i.e., the less the scatter of experimental points around the regression line). Consequently, regression lines with high F-values (low significance F) are much better than those with lower ones. In conclusion, good regression lines show high values for both r and F statistical parameters (J.N. and Mileer 2005).

The (LOD) and (LOQ) limits were calculated as (3.3 σ / S) and (10 σ / S), respectively, where σ is the intercept standard deviation and S is the calibration curve slope. The low obtained values shown in Table 2 confirm the sensitivity of the method and indicate the reliability of the method to detect and quantify LT4 by the FTIR method used.

As per the ICH guidelines, the developed method should be specific. Specificity here refers to the ability to distinguish, identify, and quantify LT4 amidst the presence of the matrix (excipients), degradation products, or impurities, namely liothyronine sodium (ICH 2005).

Figure 2. 

Overlaid IR spectra of LT4 standards.

The specificity of the method was determined by comparing the spectra of the pharmaceutical LT4 from tablets with those of the standard LT4. The results revealed overlapping spectra. In addition, to ensure the absence of excipient interference, the spectrum of an excipient mixture that is commonly employed in tablet formulations (starch, microcrystalline cellulose, magnesium stearate, and gelatin) was measured after being subjected to the same experimental conditions. Chloroform was used for extraction, followed by filtration, the addition of KBr, evaporation, and finally the pellet press. Comparing the spectra of the pharmaceutical LT4 from tablets with those of an excipient mixture showed no excipient interference, where there was no peak band at 1409 cm^-1(Fig. 3). This confirms the specificity of the developed method.

Figure 3. 

Overlaid IR spectra of the LT4 standard, Euthyrox 100^® tablet, and excipient mixture.

A common impurity that could be found in trace amounts in LT4 powder or tablets is liothyronine sodium. This could be due to degradation, manufacturing processes, or cross-contamination. Liothyronine sodium only differs from LT4 by having one less “C–I” bond (Ruggenthaler et al. 2017). The developed method for the quantitation of LT4 is based on C=C, which is the same in both, so it cannot differentiate. However, the difference between both would be detected in the fingerprint area (specifically seen in the C-I bond at 517 cm^-1). Accordingly, the complete spectrum should first be observed in order to determine whether there is the presence of LT4 or liothyronine. Also, this confirms that this FTIR method is a highly specific and selective technique for the detection of impurities or adulterations.

The accuracy of the proposed method was determined through a recovery study. Three different concentrations of LT4 standard pellets (100, 200, and 400 µg/pellet) prepared from independent stock powder were assayed three times. Good accuracy, expressed as % recovery, was obtained. The results, summarized in Table 3, show that the % recovery values do not exceed the accepted limits (98–102%), which demonstrate the accuracy of the developed method.

The precision of the method was evaluated as intra-day repeatability, inter-day precision, or intermediate precision. Repeatability studies (intra-day) were performed by analyzing three different concentrations of LT4 within the calibration range in triplicate, all on the same day, using identical working conditions. Intermediate precision (inter-day) was assessed by repeating the assay on three different days under the same experimental conditions.

Good precision, expressed by the low percentage RSD such that the values did not exceed an acceptable limit, was obtained, proving the high precision of the method. The results are summarized in Table 4.

The applicability of the proposed method was evaluated for two different pharmaceutical formulations: Euthyrox^® tablets and Eltroxin^® tablets. The results obtained from the FTIR analysis were satisfactory for the determination of LT4 in its pharmaceutical formulations and were comparable to the labeled amount expressed by high percentage recoveries with low % RSD, which indicate high accuracy and precision in the determination. Also, a statistical comparison between this method and a reference method using the student’s t-test and the variance ratio F-test was performed (Jamal et al. 2023). Since the calculated t- and F-values (ICH 2005) for each drug did not exceed the theoretical ones, this indicated that there was no significant difference between the proposed methods for the determination of the drugs in commercial tablets. The results obtained are summarized in Table 5.