Why interoceptive exposure is the active ingredient — not breathing exercises
The single biggest mistake in panic treatment is leading with paced breathing as a coping skill. Clients love it — and that is the problem. Paced breathing reliably becomes a safety behavior: 'if I hadn't done my breathing I would have died.' Each successful 'rescue' confirms the catastrophic prediction that the sensation was, in fact, dangerous, and prevents the new learning the brain needs. Interoceptive exposure does the opposite: the client deliberately produces the sensation they fear (racing heart, dizziness, shortness of breath), does nothing to manage it, and discovers the predicted catastrophe doesn't arrive. Repeated 30 to 50 times across sessions and home practice, this overwrites the threat association with a safety association. Breathing exercises have a place in initial stabilization for severe agoraphobia — but they should be retired before exposure begins, not used as the treatment.
Common safety behaviors that quietly maintain panic
Audit for these in session 2 and drop them deliberately across treatment. Carrying water 'in case I get dizzy.' Carrying a benzo 'I haven't taken in months but feel safer knowing it's there.' Sitting near exits. Driving only in the right lane. Avoiding caffeine, exercise, hot rooms, or anything that produces panic-like sensations. Checking pulse. Phone-checking for distraction. Calling a partner mid-attack. Each of these signals to the brain that the situation was dangerous and required management — and prevents the corrective learning. The cleanest exposure design names every safety behavior the client is using and removes them one at a time across the protocol, ending with the most defended (often the benzo in the bag).
When to coordinate with a prescriber
Two scenarios. First: severe, daily panic with marked agoraphobia where the client cannot mobilize for exposure work — short-term SSRI initiation often lowers baseline arousal enough that CBT becomes possible. Coordinate so the client understands the medication is a bridge, not the cure, and so dosing changes don't get attributed to therapy progress (or vice versa). Second: chronic benzodiazepine use. Long-term benzos undermine CBT-for-panic because every learning trial is contaminated — the client cannot tell whether they recovered because the panic passed or because the pill kicked in. Build a coordinated taper with the prescriber that begins after interoceptive habituation is in place, not before. Tapering off benzos without doing the CBT work first reliably produces rebound panic and demoralization.