Skip navigation

Encapsulated sumatriptan and migraine efficacy

 

Clinical bottom line

Encapsulation of ground up sumatriptan tablets in two trials of eletriptan made no significant difference to the estimation of sumatriptan's efficacy in a meta-analysis. Drug re-formulation should be examined in clinical trial reports and systematic reviews.


The conduct of clinical trials has some interesting by-ways. One potential problem brought to Bandolier's attention was that in some of the trials of eletriptan, the sumatriptan used as an active control was not formulated in the same way as sumatriptan taken by patients:

Whatever was done was done in two trials. Whatever was done could affect the results, and it would therefore be prudent to perform a sensitivity analysis to ensure that clinical results were not changed by the encapsulation of sumatriptan. Table 1 therefore shows the original results, and those with the two trials using encapsulated sumatriptan excluded. There were no large changes because of the addition of data relating to encapsulated sumatriptan.

Table 1: Sensitivity of sumatriptan 100 mg and 50 mg for encapsulation.

Original calculations

Omitting encapsulated sumatriptan

Number of

NNT
(95%CI)

Number of

NNT
(95%CI)

Outcome

Trials

Patients

Trials

Patients

Oral 100 mg            
Headache response at 2 hours

13

3522

3.3 (3.0 to 3.7)

11

3042

3.2 (2.9 to 3.6)

Pain free at 2 hours

8

2205

4.7 (4.1 to 5.5)

6

1725

4.4 (3.9 to 5.3)

Sustained headache response

4

1632

6.7 (5.0 to 10)

2

1108

10 (5.9 to 32)

Oral 50 mg
Headache response at 2 hours

6

1552

4.1 (3.4 to 5.2)

4

1081

3.6 (3.0 to 4.5)

Pain free at 2 hours

4

1083

7.8 (6.1 to 18)

2

612

7.1 (5.3 to 11)

Sustained headache response

2

539

6.0 (4.0 to 12)

no data

Comment

This is an interesting topic for two reasons.

The first is that here, the changed formulation of sumatriptan made no systematic difference to the NNTs found in the meta-analysis, and differences found for different outcomes were mostly trivial and in no case changed the result in a meaningful way. But that is a reflection of the very large amounts of information we have about non-encapsulated sumatriptan, so any large change would have been a surprise, and data needed to create such a surprise would have been so different that simple techniques like L'Abbé plots would have pointed out the difference.

The second interesting topic is that this re-formulation of existing control drugs actually goes on. Bandolier was not aware of it after 25 years of clinical trial involvement. Questioning whether any change in formulation has taken place should now become a standard question fro primary trials, and for meta-analysis. If it occurs, or is suspected to occur, then it becomes a subject for sensitivity analysis.