Torque is a rotational force -- it's measured/calculated by applying a given force at the end of a lever arm (the wrench) of a specific length. If you hang 1 lb. of weight at the end of a 1 foot long wrench - you're applying 1 foot-pound of torque. You see how it works? So, the problem with a cheater bar is it changes the length of the wrench/lever arm - and that throws the measurement completely off. In fact, if you look at a beam-type torque wrench, the handle has a little swivel mounting - and for proper torques, you must keep the handle/swivel right in the middle of it's travel - that assures the lever arm length is accurate.
I guess the good news is that you actually read the instructions -- something my wife assures me is damn near genetically unheard of in human males. As others said - 1) only use the torque wrench to apply torques - use a breaker bar to loosen them; 2) if it's a click-type, always reset it to 0-10 ft-lbs. before storing it away; and 3) if you drop it, you really should take it in and have it recalibrated. Click types need to be recalibrated on a regular basis anyway if you want to assure accuracy. For things like lug nuts, etc. I use a click type. For internal engine assembly (heads, main bearings, rod -bolts, etc. I prefer to use the old style beam-type torque wrench. It allows you to see the torque increasing to your set point as the needle sweeps across the face. With the proper moly lube on clean threads, it makes for nice steady, even pulls up to the specified torque. And as long as the pointer is centered at zero, you know it's accurate. With the click type, you don't know you're there until you're there, and you have no way of knowing if it's out of calibration. I only use them on big bolts/nuts where there's a wide range of acceptable torques, and when the location of the bolt in question prevents me from using the beam type - which unfortunately happens quite a bit when you're torquing things with the engine in the car.
