Background: Neonatal exposure to general anesthetics may pose significant neurocognitive risk. Human epidemiological studies demonstrate higher rates of learning disability among children with multiple, but not single, exposures to anesthesia. The authors employ a rat model to provide a histological correlate for these population-based observations. The authors examined long-term differences in hippocampal synaptic density, mitochondrial density, and dendritic spine morphology. Methods: Twenty male rat pups (n = 5/condition) were exposed to 2.5% sevoflurane under one of four conditions: single 2-h exposure on postnatal day 7 (P7); single 6-h exposure on P7; repeated 2-h exposures on P7, P10, and P13 for a cumulative 6 h of general anesthetics; or control exposure to 30% oxygen on P7, P10, and P13. Results: Repeated exposure to general anesthetics resulted in greater synaptic loss relative to a single 2-h exposure (P < 0.001). The magnitude of synaptic loss induced by three 2-h exposures (1.977 ± 0.040 μm3 [mean ± SEM]) was more profound than that of a single 6-h exposure (2.280 ± 0.045 μm3, P = 0.022). Repeated exposures did not alter the distribution of postsynaptic density length, indicating a uniform pattern of loss across spine types. In contrast, mitochondrial toxicity was best predicted by the cumulative duration of exposure. Relative to control (0.595 ± 0.017), both repeated 2-h exposures (0.479 ± 0.015) and a single 6-h exposure (0.488 ± 0.013) were associated with equivalent reductions in the fraction of presynaptic terminals containing mitochondria (P < 0.001). Conclusion: This suggests a "threshold effect" for general anesthetic-induced neurotoxicity, whereby even brief exposures induce long-lasting alterations in neuronal circuitry and sensitize surviving synapses to subsequent loss.