2.2.3. Methyl Peak Profile

The peak intensity was compared using peak heights. First, the absolute peak heights of the strongest peak, Thr-d, were tabulated for five lots of each brand and five technical repeats from one lot of Lantus® (Table S3). The calculated *p* value between the five technical repeats and the five lots of Lantus® was 0.35, demonstrating the technical issues related to the spectral differences were within the inter-lot DP differences. By contrast, the Thr-d peak height in Lantus® inter-lot spectra was on average 4% higher than the peak height of the Basaglar® inter-lot spectra. The 4% difference was significant with a *p* value of 0.0061 (Table S3), which is less than the threshold value of 0.05. The 4% difference may be related to differences in assay and response *Q*-factor of the NMR probe to different formulations, usually related to electric capacity or ionic strength [42]. As a result, the comparison using absolute peak height for all methyl peaks was deemed not suitable.

However, the relative peak heights, related to the dynamics and exchange kinetics of each methyl group should still be a valid choice for comparison purposes. Here, the ratios of each peak height to the Thr-d were calculated according to Equation (3). The mean and standard deviation from both Lantus® inter-lot spectra and Basaglar® inter-lot spectra were plotted in Figure 4C. The *p* values were calculated for all 48 peaks (Table 2) and 47 *p* values were higher than 0.05 except for the Leu-t peak with *p* value of 0.0055. Ultimately, 47 out of 48 peaks were equivalent in relative peak height between the two brands, demonstrating that the HOS distribution and exchange kinetics of the insulin glargine in the two DPs were similar. The work suggested the similarity metrics for peaks that show comparable relative peak height could be at least 98% (47/48).
