The impact of hepatic glucose concentration on the sympathetic response to progressive hypoglycemia was examined in chronically cannulated conscious male dogs (n = 6). Graded hypoglycemia was induced via peripheral insulin infusion (30 pmol kg-1-min-1) with either peripheral (PER) or portal (POR) glucose infusion. Over the 260-min experimental period, arterial glycemia was adjusted from 5.2 ± 0.1 to 2.5 ± 0.1 mM in decrements of %0.5 mM every 40 min. Arterial glycemias were not sigfntly different between PER and POR at any measured level. However, hepatic glycemia was signiicantly elevated at all times during POR (8.4 ± 0.8 to 3.4 ± 0.2 mM) when compared to PER (5.2 ± 0.2 to 2.5 ± 0.1 mM). Plasma epinephrine values were significantly greater during PER vs. POR at all arterial glycemias below 4.0 mM. At the lowest level of arterial glycemia studied (2.5 ± 0.2 mM) the epinephrine response above basal was 3-fold greater for PER (8.7 ± 1.7 nM) when compared to POR (2.6 ± 0.6 nM) (P < 0.01). Plasma norepinephrine results were similar for the two protocols, with PER demonstrating a 3-fold greater response above basal when compared to POR at 2.5 mM arterial glycemia (P < 0.05). While the sympathetic response was markedly different between protocols when expressed as a function of arterial glycemia, when expressed as a function of hepatic glycemia this discrepancy was largely eliminated. This latter observation supports the liver as the primary locus for glycemic detection relevant to the sympathoadrenal response when hypoglycemia develops slowly-i.e., over a period of 2-3 h. A comparison of the current findings with our previous observations suggests that the hepatic glucosensors may play a greater role in hypoglycemic counterregulation as the rate of fall in glycemia is less.In 1924 Walter B. Cannon and coworkers (1) provided the first convincing evidence that insulin-induced hypoglycemia resulted in increased sympathetic output. At that time, they proposed that the enhanced sympathetic activity was likely due to glucopenia "local" to the autonomic nervous system. This position was supported by the earlier work of Claude Bernard (2) and others (3, 4), demonstrating that lesions to specific aspects ofthe brain had a profound impact on glucose metabolism. The existence of specific "glucoreceptors" within the hypothalamus was later proposed by Mayer and Marshall (5). Subsequent studies in which direct microinjections were used have now clearly identified glucosensitive neurons within the ventromedial and lateral hypothalamus (6). In addition, substantial evidence has accumulated over the years delineating the efferent capacity of the central nervous system (CNS) to impact upon glucoregulation (6,7). This has led to the prevailing concept that the brain "senses" ambient glycemia and effects the requisite glucoregulatory mechanisms.Evidence for the role of the CNS in glycemic detection has relied largely on nonphysiological stimuli, such as brain lesions, electrical stimulation, glucose analogues, and direct (in th...