It is hard to pinpoint the issue as you did not follow our post guideline.
Basically, providing more compounds (as input list) does not guarantee that there will be more significant pathways or matches. Those compounds must belong to a pathway AND are significantly changed.
When you provide more compounds as reference metabolome, but it is very small - there is no much power either. The enrichment test is more effective for decent sized groups (or sets) containing at least 5 members. In your case, there is only one or two members left per pathway - the pathway coverage is too small to draw confident conclusion.