For the Senate, it means that the numbers were just too much and the Republican won all of the close contests. In the Senate only one-third of the seats (33 or 34 states) are up for regular election. While there will be at least one and maybe three special elections for the Senate, they are all in the states that are having regular elections. Depending upon who may die or step down this year, the Republicans are only defending 8-10 seats and only one of those are in a state that Clinton won in 2016. The Democrats are defending 26 seats, many of them in states that Trump won by a large margin. So you could have a big national swing to the Democrats and still have Republicans knocking off Democratic incumbents in the Senate. For example, assume that the Democratic candidate does 5% better than Clinton, and the Republican candidate does 5% worse than Trump, the Democrats would lose 5 Senate seats that they currently have and Republicans would lose three (or 4) that they currently have -- a net of swing of two.
On the other hand, assuming that same swing in the House (which is about what current polls are estimating and is consistent with the voting in 2017 special elections), the Republicans would lose at least thirteen seats -- and probably more since some of their wins in 2016 were by incumbents who have decided to retire significantly over-performing in lean Democrat districts. So for the Republicans to gain seats in the House something significant has to change between now and November -- say bombing North Korea or Iran. (Before you say that polls can't be trusted, nationally the final polls were only 1 or 2 % off, well within the margin of error. The problem in 2016 was not with the polls, but with idiot reporters who did not think about what a 1-2% error in favor of Trump would mean for the electoral college. For the Republicans to gain house seats in 2018, there would have to be a huge problem with the polls.)