1) Yes - you are right, Sallen-Key topology is not well suited for high-Q filters;
2.) MFB topology is better, however - it also is not the best solution for high selectivity (by the way: You didn`t give the center frequency!)
3.) You have mentioned "state-variable filters"; I would not recommend KHN or Tow-Thomas because they suffer from phase errors (3 opamps in series); more than that, they are primarily used if all three outputs are needed;
4.) For my opinion, the best choice is a GIC-based solution (also called "Fliege filter"). However, you need two opamps. I recommend to google for "GIC filter" or "Fliege filter". One link is given below:
http://www.google.de/imgres?imgurl=http://www.schematica.com/active_filter_resources/images/fliege_bp_dual_supply_red.jpg&imgrefurl=http://www.schematica.com/active_filter_resources/op_amp_virtual_ground_circuits.html&h=299&w=395&tbnid=UkfQKtYPZXcwFM:&zoom=1&tbnh=98&tbnw=130&usg=__RMmOGMvqAIIbh83lmN3MS_k_6N0=&docid=C6RXKsxEe7_qdM&sa=X&ei=bPv_VJXjE4LmUtXHgIgE&ved=0CCMQ9QEwAA
Remark: All GIC-based filters are easy to design and have very good sensitivity (active) properties. Which means: They don`t react sensitive to opamp phase deviations because both opamps (if on the same chip) compensate each other in this respect.
GIC: Generalized Impedance Converter.