Here are five things you could look at to start with.

1) Is there a circuit that is establishing a voltage bias on one or more IC pins before the IC has a power voltage applied to it? This could happen with the phone line becasue of the -48V feed from the central office.

2) Is there a circuit that is keeping a voltage bias on one or more IC pins after power is removed from the ICs? This could happen due to voltage stored in some capacitor or due to the phone line which has the aforementioned central office feed on the line.

3) Is there something about the design of your circuit that nominally places an I/O pin at a voltage level that is outside of its normal operating range as specified in the part data sheet?

4) Is some circuit connected with a capacitor in such as way as to push a pin of an IC outside its normal voltage range when a transient event occurs on the other side of the capacitor? This could happen for instance if a capacitor connected to the phone line had a large bias across it when the line was on-hook and then suddenly the line goes offhook.

5) Is there a connection in the circuit where some inductive component is producing a ring back spike when the signal driving the inductance changes suddenly? This could happen for instance with a relay coil that suddenly went from ON to OFF and produces a large voltage spike that can couple into IC pins and lead to latchup.

Good Luck
Michael Karas