I have found that caffeine is mostly metabolized over CYP1A2 (as we know), but also over CYP1A1, CYP2E1 and CYP3A4. The question is how much the individual can cope with this alternate pathways if he is a CYP1A2 poor metabolizer. 
Secondly I found in my pharmaceutics textbook that CYP2A1 activity determination is indicated due to undesirable side effects to caffeine and theophyllin.
UPDATE: 19.03.2013 12:30 GMT+1
@dd3: To clarify the second statement: I used the approach that theophyllin and caffeine are similar xanthin derivates. Since theophyllin is used as therapeutic drug in respiratory diseases (COPD etc.), I thought that scientific interest would be greater. And yes CYP1A2 plays an important role in both theophyllin tolerance and intollerance.
I found some new sources:
At first a study with CYP1A2 knock-out mice, which shows that the elimination of theophyllin takes 4-times longer in CYP1A2(-/-)mice. And they also assume that the same behaviour is also applicable to caffeine. 
Also there are two clinical studies by Japanese and Turkish scientists. The Japanese conclude, that theophyllin should be used with care in CYP1A2 poor metabolizer (even in haplotype poor metabolizer), since theophyllin has a narrow therapeutic range. But the Turkish scientist assume that according to their findings, theophyllin is also metabolized over alternative pathways such as CYP2A13, CYP1A1, CYP2E1, CYP2D6 and CYP3A4. And according to this the genetic status of CYP1A2 is not as important as expected. 
Assuming that theophyllin and caffeine are metabolized similarly, you can conclude that a mutation in CYP1A2 would lead to caffeine intolerance. In my experience people with caffeine intolerance describe severe nervousness after drinking a cup of coffee. Fatal cases I know were due to high consumption of energy drinks in children without or at least not known caffeine intolerance.