Events look like this:

{
  "@key": "event_one",
  "@value": 10,
  "country": "RU",
  "whatever": 5
}

{
  "@key": "event_two",
  "@value": 20,
  "country": "RU",
  "whatever": 8
}

{
  "@key": "event_one",
  "@value": 500,
  "country": "US",
  "whatever": 7
}

Histogram is using @value and US should be plotted above RU with these events (500 vs 20 + 10), but topN with count=1 will return RU, because RU occurred twice.

Colouring is affected too, here's real example:

Using terms_stats facet would solve that. #654 could help too, showing that there are other terms.