The efficiency of Hemoperfusion in acute poisoning cannot be clinically estimated, because: a) concomitant intestinal absorption, hepatic metabolism and urinary excretion must be taken into account. b) with supportive treatment alone, spontaneous recovery usually occurs in 98% of the intoxications in Intensive Care Units. The efficiency of hemoperfusion can only be estimated biologically. Measuring the blood level at the beginning and the end of hemoperfusion as well as measuring the clearances of the drug is misleading. A better method is to measure the amount of extracted drug, either indirectly by calculation (form hourly differences of arterio-venous measures of drug concentration multiplied by the blood flow) or directly by elution of the cartridge. In a practical way, if the blood level of drug is readily available after the patient is hospitalized, the optimum efficiency of hemoperfusion can be estimated beforehand, so that the decision to carry out the hemoperfusion can be maintained, postponed or abandoned. For the most part, the experience of toxicologists has shown hemoperfusion to be ineffective for drugs with weak extra-cellular distribution (such as Digitoxin, Tricyclic drugs, Heavy Metals, Colchicine). Its effectiveness for certain drugs, with poor in vitro dialysance (such as Paracetamol) or with small percentage of intestinal absorption (such as Paraquat) is still debatable. In the case of intoxications by hypnotic drugs, one hemoperfusion allows an average of 4 - 12% of the ingested medium and short barbiturates, 7 - 17% of the ingested Meprobamate. Whether these results can be judged satisfactory, life-saving or insignificant is largely a matter of personal standards.