Last period 100 70 - 90 years old died. This period 110. If they died only a fraction of the period earlier, they would have shown up in the same category anyway with some probability.

This period 5 20-40 years old died, this year 10 Normally they would have died decades later, so they would have not shown up in this category during the period.

(Of course with small numbers statisical relevance decreases. But I have understood in countries with many covid deaths in elderly care excess deaths for high age categories did not increase that much over a year.)

This would show a 10% increase in the older group, but a 50% increase in the younger group.

Of course