Not really.
You have two options though. One is simple to merge all of the capture files and the use follow TCP stream then over that merged capture. But of course the reason you have separate files might well be for size reasons, so joining them might not be practical.
The second is for you to identify what makes a stream and use that. For a stream (in a general sense) is identified by a tuple (a set) of the source and destination IP addresses and TCP ports. In older versions of wireshark when you did "Follow TCP stream" it would prepare a display filter with this tuple and display it, as below
(ip.addr eq 192.168.0.118 and ip.addr eq 212.58.253.70) and (tcp.port eq 43047 and tcp.port eq 80)
Now if you then copied this display filter, and then opened a different capture file it, it could then be used as a display filter to show the same stream.
Now in newer versions of Wireshark (I'm not sure exactly when this appeared), when you "Follow TCP Stream", you now get a display filter something like:-
tcp.stream eq 54
Now for most purposes this makes it easier to select and remember different streams, the "tcp.stream" is a generated field that only has relevance to the capture file loaded. Basically for every new TCP stream that wireshark decodes (based on each unique source and dest IP and TCP port tuple) it generates a new TCP stream. And of course each capture file will almost certainly have different streams and probably appearing in different order. The only easy way that I can see to create a display filter that references a particular stream across different captures (in the newer Wireshark versions) is unfortunately going to need to be done manually and would follow the first format I mentioned.
Regards, Martin
MartinVisser99@xxxxxxxxx
