Concerning the side effects, again it is true that there are individual variations, but on the average naltrexone is extremely well tolerated with very few side effects. The only proper way to measure side effects is in comparison with a placebo group - where neither the naltrexone nor the placebo patients know which medicine they are getting. The results with this type of analysis from our Finnish clinical trial (Heinälä et al., 2001) were surprisingly good.
In our two groups that got either naltrexone or placebo along with instructions to abstain, we did find significantly more side effects from naltrexone than from placebo, mostly nausea and intestinal problems, and most during the first week of treatment. Among the patients treated with the Sinclair Method, there was no significant difference between the naltrexone and placebo groups. There was no peak in the first week, and there was no time during the entire trial when the naltrexone groups showed significantly more side effects than the placebo group, or significantly more of any particular side effect.
It probably is impossible to show graphs at the forum. Pity the one here is beautiful. Plotting the mean number of side effects shown by each of the 4 groups during the week before treatment and for each week of the treatment, you see this one big peak of increased side effects, during the first week on naltrexone for the support of abstinence group. (Meaning take naltrexone and do not drink.) The naltrexone group with drinking show no peak at any time.
We speculated that the peak in the naltrexone/abstinence group was probably a mild opioid withdrawal: they had been getting endorphins from their alcohol everyday, but during the first week of treatment, the naltrexone was blocking the endorphins from binding to their receptors and causing a mild withdrawal. The specific symptoms were also similar to what might be expected from a mild opiate withdrawal.
So did the naltrexone+drinking group not show similar withdrawal symptoms? We figured this way. The usual treatment for opiate withdrawal is barbiturates or benzodiazepines. Alcohol is a dirty medicine affecting many systems, including the GABA system where barbiturates and benzodiazepines work. So the group drinking alcohol while on naltrexone are getting drug effects similar to barbiturates or benziodiazepines, which help to suppress problems from the opioid withdrawal. (Of course, a person drinking alcohol is always getting these drug effects.)

On the other hand, that was not a huge trial: only 121 patients. We have seen patients in the clinics who were showing side effects and needed attention. Since there were no placebo controls at the clinics, we cannot say for sure that naltrexone was causing the side effects. It is better, however, to be safe, so if there are side effects, see a physician.