Background Anaphylaxis is a systemic allergic reaction that is commonly treated in the emergency department (ED). The risk of a biphasic reaction is the rationale for observation. Objective To derive a prediction rule to stratify ED anaphylaxis patients at risk of a biphasic reaction. Methods We conducted an observational study of a cohort of patients presenting to an academic ED with signs and symptoms of anaphylaxis. We collected clinical data on biphasic reactions meeting National Institutes of Allergy and Infectious Disease/Food Allergy and Anaphylaxis Network diagnostic criteria. Logistic regression analyses were conducted to identify predictors of biphasic reactions, and odds ratios (ORs) with 95% CIs are reported. The predictive ability of the model features is summarized using the area under a receiver operating characteristics curve, or AUC. Internally validated AUCs were obtained using bootstrap resampling. Results We identified 872 anaphylaxis-related visits. Thirty-six (4.1%) visits resulted in biphasic reactions. Multivariable analysis showed that prior anaphylaxis (OR, 2.74; 95% CI, 1.33-5.63), unknown inciting trigger (OR, 2.40; 95% CI, 1.14-4.99), and first epinephrine administration more than 60 minutes after symptom onset (OR, 2.29; 95% CI, 1.09-4.79) were statistically significantly associated with biphasic reactions. The AUC of this model was 0.70 (95% CI, 0.61-0.79), with an internally validated AUC of 0.67 (95% CI, 0.59-0.76). The P value from the goodness-of-fit test was.91. Conclusions Our study demonstrated a 4.1% rate of biphasic reactions and found that prior anaphylaxis, unknown inciting trigger, and delayed epinephrine use were risk factors for biphasic reactions.