Hydrostatic Gradient is Important – Blood Pressure Should be Corrected

John C. Drummond, MD, FRCPC; Alan R. Hargens, PhD; Piyush M. Patel, MD, FRCPC

To the Editor

In the Summer 2007 issue of the APSF Newsletter, Cullen and Kirby described 4 instances of catastrophic neurologic outcomes after surgical procedures performed in the beach chair position.1 They surmised that the combination of the failure to make allowance for the hydrostatic mean arterial pressure (MAP) gradient between the heart and the head with the clinicians’ acceptance of relatively low MAPs at heart level had resulted in cerebral hypoperfusion and ischemic injury. They advocated that clinicians manage patients undergoing procedures in the beach chair position on the basis of MAP measured at the level of the head or corrected to head level by imposing an arithmetic adjustment to MAPs recorded at other sites (*see footnote for illustration).

Dr. James Munis responded with a commentary in which he disputed the necessity to measure MAP at (or correct it to) head level.2 The essence of his argument (in very brief summary of a lengthy submission) is that the cerebral circulation is a closed system that functions like a siphon. While agreeing that intra-luminal pressures decrease above heart level, he argued that the same reduction in pressure occurs on both the venous and arterial sides of the circulation with, therefore, no net change in the driving (perfusion) pressure across the cerebral vascular bed and, therefore, no change in cerebral blood flow (CBF). The author of a subsequent letter to the editor of the APSF Newsletter applauded Dr. Munis’s dismissal of “the nonissues of transmural pressures, altering transducer height and ‘correction’ formulas.”3

We disagree with Dr. Munis. Furthermore, the possibility that clinicians might broadly accept this dismissal of the significance of the hydrostatic pressure gradient is of substantial concern to us. We believe that corrections for hydrostatic pressure gradients (also referred to as “gravitational pressure gradients”) in head-up positions are necessary and appropriate. One approach to convincing the readers of the APSF Newsletter of our position might be to take issue with some or all of the many arguments in favor of the closed model of the cerebral circulation that Dr. Munis has laid out in various publications.2,4,5 However, out of the concern that readers will be unconvinced (or simply confused) by such a complicated discussion, and because these arguments have been set forth in detail previously,6,7 it seems preferable to argue first that even if the cerebral circulation is a closed system in some circumstances (functioning in the manner of a siphon with balanced hydrostatic pressures in the ascending and descending limbs of closed loop), it will not be so in all situations. The closed circulation/siphon theory posits that CBF is a function of the arterial to venous pressure difference across the brain (the “perfusion pressure”) and that head elevation leads to equivalent hydrostatic pressure changes in arterial and venous pressure with, therefore, no net change in perfusion pressure or in CBF. However, the argument for parallel (and therefore compensating) changes in venous and arterial hydrostatic pressure becomes irrelevant in a circumstance in which there is direct compression of nervous tissue. In those circumstances, which arise when brain parenchyma is compressed by retractors or in the context of cervical spinal stenosis when the spinal cord is compressed by protruding discs or a hypertrophied posterior longitudinal ligament, it will be the transmural pressure, i.e., the gradient between the intraluminal and extraluminal (tissue) pressures that will be the principal determinant of flow. The siphon model would be similarly invalid when intraluminal pressure on the arterial side of the circulation has decreased more than venous pressure because of a stenosis in the arterial tree. We take the position that even if the closed (siphon) model of the circulation is in effect some, or even most of the time (which we do not accept), the “rules” are likely to be different in the context of the compression or arterial stenoses just mentioned. It would be a slim consolation to the 4 patients who sustained ischemic injuries in the beach chair position (or to their families) to know that hydrostatically reduced intraluminal pressures at the level of the head are sometimes not a matter of consequence.

Furthermore, we believe that the non-applicability of the closed circulation/siphon model goes beyond circumstances of CNS compression or vascular stenoses. In a thought experiment used by Dr. Munis to illustrate the non-importance of absolute intraluminal pressure,2,4 he asks that we imagine an intravenous (IV) infusion system with a fluid bag in its typical position some distance higher than the venous access site. Would the IV continue to flow, he asks rhetorically, if a loop of the IV tubing were raised to the level of the top of the IV bag, and what would the intraluminal pressure be at the apex of that loop of tubing? The answer to the former, from experience, is, “Yes, it would flow” and to the latter, that the pressure would be very close to atmospheric. Voila! Q.E.D! The theorem is proven. The hydrostatic gradient is unimportant! And, those who have used a siphon might accept this. Flow will continue even in the presence of the hypothetical loop.

But the shortcoming of this analogy is that the siphons that the readers have used were invariably composed of relatively rigid tubing. What if some portion of that loop at the apex of the system was composed of collapsible tubing with a consistency similar to that of a Penrose drain? As previously argued in detail,6 the tubing would collapse; flow would cease; and the siphon would not function. The cerebral vessels are not all rigid like the tubing in the IV analogy. At least some of the cerebral veins, venules, and capillaries are non-rigid. Munis and Lozado acknowledge this, but argue that at least some vessels within the nervous system will be patent at any given moment in time and that at least a portion of flow will continue.4 Perhaps, but we cannot take reassurance from the notion that at any given time “some” of the brain is not ischemic. Again, it would be a slim consolation to the 4 devastated patients (or to their families) to know that blood flow had continued to some portions of their nervous systems while disabling damage was evolving in others.

We cannot assert that Munis and associates’ belief in a closed model of the cerebral circulation in which hydrostatic pressure gradients are of no physiologic consequence has yet been refuted definitively. However, we take the position that such a model is highly improbable and at best unproven. We are concerned that there is a substantial potential for the occurrence of additional neurologic injuries if clinicians accept Kleinman’s opinion that “transmural pressures, altering transducer height and ‘correction’ formulas” are “nonissues.”3 We hold the view that clinicians managing patients in significantly head-up postures should continue to measure blood pressure at (or correct it to) head level. We think that if unbiased observers with no prior knowledge of the elements of this debate were to visit and examine the existing body of information, they would be intrigued by the intricacies of the physiologic discussions. But we suspect that they would conclude, in the absence of definitive proof that hydrostatic pressure gradients were never of consequence in determining blood flow to the nervous system, that blood pressure should always be measured at, or corrected to, head level.

* As blood flows vertically from the heart, there will be a reduction in arterial pressure that is related to the weight of a column of blood. That reduction will be approximately 2 mm Hg for each inch (2.54 cm) of vertical displacement. For illustration, consider a patient in a semi-recumbent position such that the external auditory canal (EAC) is 12 inches above the mid-point of a blood pressure cuff on the upper arm. If MAP as measured by the cuff were 65 mmHg, the MAP at the EAC would not be greater than 41 mm Hg.

John C. Drummond, MD, FRCPC
Professor of Anesthesiology
University of California, San Diego
Staff Anesthesiologist, VAMC San Diego

Alan R. Hargens, PhD
Professor of Orthopaedic Surgery
University of California, San Diego

Piyush M. Patel, MD, FRCPC
Professor of Anesthesiology
University of California, San Diego
Staff Anesthesiologist, VAMC San Diego


  1. Cullen DJ, Kirby RR, Beach chair position may decrease cerebral perfusion; catastrophic outcomes have occurred. APSF Newsletter 2007;22(2):25,27.
  2. Munis J. The problems of posture, pressure, and perfusion: cerebral perfusion pressure defined. APSF Newsletter 2008;22(4):82-3.
  3. Kleinman B. Reader calls attention to change from baseline pressure. APSF Newsletter 2008;23(1):19.
  4. Munis JR, Lozado, LJ. Giraffes, siphons, and Starling resistors: cerebral perfusion pressure revisited [Editorial]. J Neurosurg Anesthesiol 2000;12:290-6.
  5. Hicks JW, Munis JR. The siphon controversy counterpoint: the brain need not be “baffling.” Am J Physiol-Regul Integr Comp Physiol 2005;289:R629-632.
  6. Seymour RS, Hargens AR, Pedley TJ. The heart works against gravity. Am J Physiol-Regul Integr Comp Physiol 1993;265:R715–R720.
  7. Seymour RS. Model analogues in the study of cephalic circulation. Comp Biochem Physiol A Mol Integr Physiol. 2000;125:517-24.