GN. Predicting the need for medical intensive care monitoring in drug-overdosed patients. J Intensive Care hled 2000;15: 321328.Drug overdose (OD) is one of the most common single diagnoses admitted to medical intensive care units (XIICUs). Th;diagnosis results in relatively little morbidity or mortalit); suggesting a need to improve the methods utilized in deciding o n XIICU admission. Our objective in this study was to develop a quantitative system whereby the emergency room (ER) physician, with reasonable safety and ease, would be able to discriminate between the need for an ICU and a nonrnonitored bed. \W reviewed the charts of 216 consecutive XIICU admissions for intentional OD involving 199 patients between the years 1995 and 1998. Clinical histories, vital signs, laboratory data inclusive of toxicologic analysis, and both APACHE I1 and Glasgow coma scores (GCS) were assessed from the ER and on transfer to the AIICU 4-6 hours later. These scores, as well as individual components of the APACIIE I1 score, mere evaluated for significance. Of 216 admissions reviewed, 75 (35%) had AIICU-requiring morbidity: intutubation GI%, pneumonia 20%. arrhythmia o r ERG changes 20%, and hypotension 3%. Xlortality mas 2.7%. The remaining admissions were for ICU monitoring due to lethargy, irritability, labontory abnormalities, or simply based on the diagnosis of drug overdose. Urine examinations for drugs were positive in only 53%, with the most common agent identified being ben~bdiaze~ines (39"h). Age, Apache I1 score, and GCS were significantly different between those patients who developed XIICU-requiring morbidity and those who did not, as well as when comparing the morbidity with the mortality group. Receiver opentor control (ROC) cun-es reveal that both the MACIIE and GCS are excellent and equal predictors of morbidity, with a GCS of 5 12 having 88% sensitivity and 92% specificity in predicting XIICU-requiring morbidity. \\' e conclude that ER evaluation of GCS can be used to accurately assess and predict the need for AIICU monitoring in drug overdose. OD patients with a persistent GCS of greater than 12 or who do not demonstnte any From the Department of hledicine, Nassau County hledical Ccnter, East Aleadow, h ? : and State Universiry of New York Health Sciences Center, Stony Brook, hT.

hemodynamic, infectious, or electrocardiographic complica-GN. Predicting the need for medical intensive care monitoring tions in the ER do not require MICU admission. in drug-overdosed patients. J Intensive Care Med 2000;15: 321-328.Drug overdose (OD) is one of the most common single diagnoses admitted to medical intensive care units (MICUs). Drug overdose (OD) remains one of the three mostThe diagnosis results in relatively little morbidity or mortality, common single diagnoses admitted to medical insuggesting a need to improve the methods utilized in decidtensive care units (MICUs) in the United States, ing on MICU admission. Our objective in this study was to accounting for at least 5-13% of all admissions [1,2]. develop a quantitative system whereby the emergency roomIt is estimated that as many as 2 million poisonings (ER) physician, with reasonable safety and ease, would be able to discriminate between the need for an ICU and a occur every year of which 8% are intentional [3,4]. nonmonitored bed. We reviewed the charts of 216 consecu-Because of the potential for morbidity and mortaltive MICU admissions for intentional OD involving 199 paity, many of these intoxications result in admissions tients between the years 1995 and 1998. Clinical histories, to the MICU. Fears of later complications arise from vital signs, laboratory data inclusive of toxicologic analysis, the possibility of delayed drug actions or unidentiand both APACHE II and Glasgow coma scores (GCS) were assessed from the ER and on transfer to the MICU 4-6 hours fied ingestions, since up to half of all overdoses later. These scores, as well as individual components of the involves more than one drug [2,4,5]. Indeed there APACHE II score, were evaluated for significance. Of 216 are several reports of delayed complications from admissions reviewed, 75 (35%) had MICU-requiring morbid-overdose involving primarily tricyclic antidepresity: intubation 61%, pneumonia 20%, arrhythmia or EKG sants and lithium [1,6]. Other intoxications that also changes 20%, and hypotension 3%. Mortality was 2.7%. The remaining admissions were for ICU monitoring due to lethhave potentially dangerous delayed complications argy, irritability, laboratory abnormalities, or simply based include aspirin, methanol, and ethylene glycol. Foron the diagnosis of drug overdose. Urine examinations for tunately these events are few and their significance drugs were positive in only 53%, with the most common may be overestimated. In addition, many of these agent identified being benzodiazepines (39%). Age, Apache ingestions may be recognized as possible intoxica-II score, and GCS were significantly different between those patients who developed MICU-requiring morbidity and those tions at the time of admission by determination of who did not, as well as when comparing the morbidity with drug levels and associated characteristic electrolyte the mortality group. Receiver operator control (ROC) curves abnormalities.reveal that both the APACHE and GCS are excellent andThe majority of OD admi...

No abstract