The present study assessed the long-term(i.e., 24 months) efficacy of the ketogenic diet (KD) as an add-on therapy in children with refractory epilepsy, with focus on seizure frequency, seizure severity, and tolerability. Most patients were treated with the MCT-diet. At one and two years, 33% and 23%, respectively, of the 48 included patients were still on the KD. After three months, one year, and two years of treatment, 16.7% of the patients were responders. The highest responder rate (i.e., 22.9%) was seen at six and nine months of treatment. Of the fifteen patients with seizure clusters during baseline, 60% were responders after three months when looking at cluster reduction and most of them were not responders for the total seizure frequency. From three months of treatment onwards, most of the patients had a relevant decrease in seizure severity which was mainly related to the most severe seizure type. Gastrointestinal dysfunction was often reported, especially in the first six weeks of treatment. Growth deceleration was present in 30% of the patients, and weight reduction in 15%. Improved arousal was mentioned in 30% of patients. No patients developed ECG abnormalities or kidney stones. Increase in lipid profile was rare. The KD is an effective therapy for children with therapy-resistant epilepsy. Effectiveness is reflected in the reduction of seizure frequency as well as in the reduction of seizure severity. After 6 months of treatment, it is obvious which patients are responders and tolerate the treatment well. Most of these patients will continue to benefit from the KD for a longer time. Long-term use of the diet was well tolerated.