TOPIC:

The Buyers Club does a lot of testing. Recently we got some of the first ever Sofosbuvir+Daclatasvir tablets in and as well as splitting out the Sof and Dac with HPLC to weigh and test them we also decided to run an NMR on a sample of the complete core. Here's what it looked like (black at the bottom is the tablet spectrum and above are the pure API spectra):



And below is the commentary back and forth between a boffin called James (just to make it confusing). I was going to rename him Bruce but that was too Monty Python.

What is says in brief is that we can almost certainly test any tablet or API with a single shot NMR and validate that it contains the expected API(s) in the expected quantities quite simply (and therefore relatively inexpensively). If this had some appeal we will consider setting up a "send us your sample and we will tell you if it looks real" service. It would have to be priced up but probably likely to be under $200 a sample rather than the $5000 I've seen people get quoted.

Please note that this would be what you might call quick and dirty sanity check testing rather than the in depth certification testing we have done that left no change from tens of thousands of dollars.

Here's the conversation.....

---

Hi James

I am sending you a scout spectrum of the innards of your tablet. I made a clean slice through the coating and grabbed the powder from within the tablet and dissolved up in DMSO. There is some material that is not readily dissolved. In the attached plot the sample is on the bottom in black trace. The two spectra above it are neat Sofosbuvir and Daclatasvir. I also looked at the Ledipasvir spectrum but it looks like this tablet is using Daclatasvir. There are other detectable components as expected but crucially I believe there are clear signals from the two compounds I can measure. I also expect that they should dissolve completely in this solvent based on past observations and the small amounts of powder I am using.

My feeling on this is that HPLC might be unnecessary to validate tablet compisition as I have near baseline resolution on several well resolved signals of the two molecules. I can weigh my starting material, add a known weight of standard and determine relative proton ratios for these signals and their concentrations relative to the standard. That can then be used to calculate a weight percentage for the powder portion of the tablet.

For the purposes of inter-lab comparison I could do this for a sample also measured at NMI using their pre-separation method and compare the results for consistency. I'm inclined toward doing it my way because it doesn't risk potential losses on the HPLC column and pretty much everything that is dissolvable should be measurable in the NMR.

If you are happy for me to do so I will weigh up some material and see if I can get some quantitative information using NMR directly on the powder. Just let me know.

Cheers

James

---

Hi James,

That looks interesting. What's happened to the peaks at 8 ppm for dac?

The specific formulation should be

Sofosbuvir 400
Daclatasvir 60

So there is just under 7 x as much Sofosbuvir in there than Dac. This does rather look like tablet testing as a sanity check can be done without separation - provided you know what you're looking for vis a vis the standard tablet composition it should be impossible to have a fake tablet look anything like a real one.

Yes, if you could do further testing you think would be interesting that would be great.

Kind Regards

James

---

Hi James

The doublet from Daclatasvir at 7.3 ppm is clear and the coupling checks out. It sits between the Sofosbuvir signals in the aromatic region. I believe the hump at 7.8 ppm is the same protons which give rise to the signal at 8 ppm in the pure spectrum.

Recall that Daclatasvir shows significant changes in chemical shift of many of it's signals as a function of concentration. There is a degree of line broadening apparent for the Daclatasvir that makes me think it is not brilliantly soluble in the mixture. It might be worth my while upping the sample temperature to assist there or looking at other solvent possibilities.

If you increase the vertical scaling on the black spectrum you can also see the amide proton signals up around 14 ppm.

Cheers

James

---

Thanks James,

The near order of magnitude difference in concentration makes it a bit harder but it's great to see recognisable signals that would validate tablets quite simply without HPLC.

Kind Regards

James

---

Hi James

I just did some comparisons of mole ratios based on a few selected peaks of Daclatasvir and Sofosbuvir in that spectrum. Based on the Dac peaks around 1ppm and the doublet at 7.3 compared with three candidates in the Sofosbuvir I came up with a range of mole ratios from 7.8:1 - 11:1 Sof vs. Dac. The expected ratio based on the assumption of 400mg vs 60 mg taking in to account the molecular weights of each is 9.3:1.

So from a rough and ready perspective we are in the ballpark. It just needs to be tightened up a little.

Cheers

James

---

Hi James

The Daclatasvir is 2HCl so there is less free base than the 60 mg. Did you account for that in the calculations?

It looks in the right ballpark. I've attached the COA for it.

Kind Regards

James

---

Thanks James,

Allowing for the extra 36 g/mol the expected mole ratio is 9.8:1. In the attached plot of the tablet dissolved in d4-Methanol I have integrated the multiplet at 7.8 ppm (8 Daclatasvir protons on the aromatic rings) and the triplet at 7.4 ppm which is two protons in the phenyl ring of Sofosbuvir. They are all pretty well resolved to baseline. The calculated ratio from these signals is 10.2:1.

Lineshape is much better in MeOD. Some signals broaden in DMSO. But you lose the NH signals largely in MeOD due to replacement by the deuterons in the solvent. Still, it looks like a better solution for getting clearly resolved signals from your two compounds.

Thanks for asking Stephen of NMI for his input. That was an excellent post of his.

Cheers

J

Brilliant!

But wouldn't it be simpler if we called the rest of the team James as well?

Or maybe Brian

Very Interesting idea indeed and possibly very helpful, but maybe the details are a little hard to follow.
Staring at graphs, reading, not really 'going in' .