In Reply

Dr. Gelman summarizes very elegantly the pathophysiologic background of our norepinephrine/low-volume study¹ : general and, particularly, epidural anesthesia induce venodilation, especially of splanchnic veins (capacitance veins), which can lead to a decrease in venous return and cardiac output. These side effects can be counteracted by restoring the sympathetic tone of the plegic veins to the preoperative normal physiological status by continuous administration of low-dose norepinephrine. This avoids the need for additional unnecessary infusion of fluid that may cause interstitial edema with all its negative consequences (postoperative bowel dysfunction, pulmonary dysfunction, and cardiac arrhythmia).^2–4 

We agree with Dr. Gelman’s conclusion that the difference in results of our study between the two patient groups is quite drastic and favor administration of continuous low-dose norepinephrine instead of infusion of greater fluid volumes as in the control arm. Indeed, Dr. Gelman is right when he says that such a difference could also have been caused by overloading with too much fluid in patients in the control arm. Generous fluid administration (i.e., iatrogenic hypervolemia) can cause destruction of the endothelial glycocalix layer, the most important part of the vascular barrier that regulates homeostatic functions. This impairment of homeostasis leads to edema (third space shifting), and the fluid accumulation will not be readily mobilized in case of acute bleeding. Our patients, however, were not given up to 8 to 9 l of fluid as stated by Dr. Gelman. The 6 ml·kg⁻¹·h⁻¹ given to our control patients accords with most anesthesia guidelines recommended in reference textbooks⁵  (including a fluid bolus of 6 ml/kg after activation of the epidural catheter during induction of anesthesia) and resulted in a median overall fluid load of 4.3 l per patient for a median time of surgery of 6.5 h.¹ 

We also agree with Dr. Gelman’s concern that the use of vasoactive agents might reduce the margin of safety if they are used to compensate for intraoperative blood/fluid loss. This, however, was not the case for the patients in our study. Blood loss was monitored continuously (including weighing of all used gauze, sponges, suction contents, etc.) and replaced (as mentioned in the article) in addition to the basic fluid supplementation of 2 ml·kg⁻¹·h⁻¹ of balanced Ringer’s solution.

Whether an unexpected major blood loss would be less detrimental in an overhydrated patient with interstitial edema with decreased tissue perfusion and anesthesia/analgesia-induced venodilation and therefore unable to react adequately to an acute drop in venous pressure than it would be in a normovolemic patient under continuous low-dose norepinephrine administration with normal tissue perfusion is an often discussed but speculative question. Such controversies can only be answered by a prospective randomized trial. The results of our study speak against such speculations: No relevant impairment in tissue perfusion could be detected in the norepinephrine/low-volume group as serum lactate values and central venous oxygen saturation did not reach pathological thresholds and thus excluding a clinically relevant tissue hypoperfusion. In contrast, the number of patients with serum brain natriuretic peptide values greater than 100 pg/ml (a marker of myocardial dilation and risk for heart insufficiency) was twice that in the control arm.

Nevertheless, we fully agree with Dr. Gelman that clinical judgment and individualization of anesthesia remain the cornerstones of adequate perioperative patient management.