NRPE and check_nt: wrong RAM amount in Windows

NSClient++ is used for monitoring Windows hosts both in Nagios and Icinga systems. The other day I found an interesting issue: NSClient++ agent shown wrong amount of RAM, more precisely it doubled it. I used check_nt plugin  for connecting to Windows agent. I suppose that it also counts the pagefile.sys file together with physical memory. Unfortunatelly I didn't find a way to correct it using check_nt, however it is possible with check_nrpe. NRPE allows to bypass more arguments to the agent and specify exactly what type of memory to check.
Here is an example:
First enable NRPE in agent configuration file nsclient.ini [/modules] section:
NRPEServer = 1

Next, enable NRPE arguments in [/settings/NRPE/server] section:
allow arguments = true

Restart the service. Now we can test from our Nagios/Icinga host:

check_nrpe -H 10.0.0.1 -p 5666 -c CheckMEM -a MaxWarn=80% MaxCrit=90% ShowAll type=physical

In addition, I found information on NSClient++ site, that "NRPE is the preferred way over NSClient (check_nt) and you get the most out of NSClient++ choosing this mode."

It seems strange to me that both Nagios and Icinga use check_nt in their documentation.