Fisher's exact test

Hello, I'm trying to calculate odds ratio and CI of odds ratio for my thesis data using 2-sided Fisher's exact test. I'm using SPSS and everything was ok till i found that one variable equals zero. The contingency table looks like that:
resistant sensitive
dead 3 31
alive 0 76
SPSS couldn't complete the calculation because of the zero value. I read in the statistical book i'm using as a reference that in cases when one variable equals zero, i'm supposed to add 0.5 to all the 4 values in the contingency tables. Soi provided SPSS with the values
resistant sensitive
dead 3.5 31.5
alive 0.5 76.5
SPSS calculated the odds ratio to be 9.6 instead of 17. When i looked at the results, i found that it approximated all values to the closest integer,i.e. it increased all values by 0.5 and calculated the results using 4, 32, 1, 77 as values. My question is: is it right to use those SPSS results or should i search for another software/manual formula to claculate the odds ratio when one variable equals zero? If so, please tell me where i can find it.
Thank you very much in advance.