The decision to discontinue venom immunotherapy requires a great deal of clinical judgement because of the potential for a life-threatening reaction to a sting. The risk of recurrence is a combination of the frequency of reaction and the severity of reaction. Early studies reported a relapse rate of 8-14% in radioallergosorbent test-negative patients when therapy was stopped after 3 years. Unfortunately, the venom skin test or radioallergosorbent test becomes negative in only 25% of patients after 5 years of treatment. An alternative criterion for stopping treatment after 5 years regardless of the skin test has been equally successful, with most post-treatment reactions being much milder than pre-treatment reactions. There was no evidence of a rebound of venom sensitivity when therapy was stopped, even when patients were stung. The level of venom-specific IgE antibodies is better suppressed by 5 than by 3 years of treatment. The risk of relapse is higher in honeybee-allergic patients and in patients who had a systemic reaction (to a sting or an injection) during therapy. The frequency of reaction may be low, but patients who had very severe pre-treatment reactions have a greater chance of the reaction being severe again, and should remain on therapy for life. Long-term observations show that the incidence of systemic reaction to a sting remains 10% for each sting that occurs, even 10-15 years after stopping treatment. Because patients may not react to one sting and then subsequently react to another sting, the cumulative frequency of sting reactions is approximately 17% after 10 years off treatment. Moreover, negative venom skin tests are not a guarantee of safety because there is almost the same 10% frequency of reaction in patients who appear to lose sensitivity. It is not yet clear whether some low-risk patients (children, mild reactors) could discontinue treatment after just 3 years.