Skip main navigation

Vasopressin Deficiency Contributes to the Vasodilation of Septic Shock

Originally published 1997;95:1122–1125


    Background The hypotension of septic shock is due to systemic vasodilation. On the basis of a clinical observation, we investigated the possibility that a deficiency in vasopressin contributes to the vasodilation of septic shock.

    Methods and Results In 19 patients with vasodilatory septic shock (systolic arterial pressure [SAP] of 92±2 mm Hg [mean±SE], cardiac output [CO] of 6.8±0.7 L/min) who were receiving catecholamines, plasma vasopressin averaged 3.1±1.0 pg/mL. In 12 patients with cardiogenic shock (SAP, 99±7 mm Hg; CO, 3.5±0.9 L/min) who were also receiving catecholamines, it averaged 22.7±2.2 pg/mL (P<.001). A constant infusion of exogenous vasopressin to 2 patients with septic shock resulted in the expected plasma concentration, indicating that catabolism of vasopressin is not increased in this condition. Although vasopressin is a weak pressor in normal subjects, its administration at 0.04 U/min to 10 patients with septic shock who were receiving catecholamines increased arterial pressure (systolic/diastolic) from 92/52 to 146/66 mm Hg (P<.001/P<.05) due to peripheral vasoconstriction (systemic vascular resistance increased from 644 to 1187 dyne·s/cm5; P<.001). Furthermore, in 6 patients with septic shock who were receiving vasopressin as the sole pressor, vasopressin withdrawal resulted in hypotension (SAP, 83±3 mm Hg), and vasopressin administration at 0.01 U/min, which resulted in a plasma concentration (≈30 pg/mL) expected for the level of hypotension, increased SAP from 83 to 115 mm Hg (P<.01).

    Conclusions Vasopressin plasma levels are inappropriately low in vasodilatory shock, most likely because of impaired baroreflex-mediated secretion. The deficiency in vasopressin contributes to the hypotension of vasodilatory septic shock.

    Endotoxic shock is a syndrome of cardiovascular collapse and multiple organ failure in response to bacterial products.1 The central characteristic of septic shock is systemic vasodilation, the cause of which is multifactorial in view of the fact that abnormalities in vasoconstrictor and vasodilator mechanisms have been reported. Of the latter, Kilbourn et al2 found increased nitric oxide synthesis, and we3 found activation of the vascular smooth muscle K+ATP channel.

    Abnormalities in vasoconstrictor mechanisms have been less well examined, but vascular smooth muscle is poorly responsive to norepinephrine in septic shock.45 In contrast, the renin-angiotensin system appears to be appropriately activated, and its inhibition worsens the hypotension of sepsis.6 Plasma endothelin is also elevated in septic shock, but the meaning of this observation is yet to be defined.7

    Vasopressin does not play a significant role in the control of vascular smooth muscle in normal conditions8910111213 but becomes critical when blood pressure is threatened.111213 Vasopressin is markedly increased in animal models of acute sepsis,141516 but we recently found that some patients in advanced vasodilatory septic shock are exquisitely sensitive to the pressor action of exogenous hormone (D.W.L., unpublished observation, 1994). This unexpected finding raised the possibility that endogenous plasma vasopressin is inappropriately low in these patients. Thus, we examined the hypothesis that vasopressin deficiency could contribute to the vasodilation of septic shock in humans.


    Subjects were patients admitted to the intensive care units of Columbia-Presbyterian Medical Center and Allegheny Hospital. Patients were not eligible if they were younger than 18 years of age or if pregnancy was suspected, and they did not receive vasopressin if active coronary artery disease or mesenteric ischemia was present. Attending physicians identified candidate subjects, and consecutive patients meeting the entry criteria as stated below were studied. Appropriate consent was obtained, and the study protocol was approved by the institutional review board.

    Septic shock was diagnosed by established criteria17 : hypotension (systolic blood pressure ≤90 mm Hg in the absence of antihypertensive agents) and low systemic vascular resistance (<800 dyne·s/cm5) before the administration of catecholamines; fever or hypothermia (temperature >101°F or <97°F); tachycardia (heart rate >90 beats/min); tachypnea (respiratory rate >20 breaths/min or the requirement of mechanical ventilation) and either a positive blood culture (63% of patients in the “Septic Shock” column of Table 1, 80 80% of patients in Table 2) or an obvious source of infection (white blood cell count >12 000/mm3 or <4000/mm3 or >10% immature [band] forms); and elevated prothrombin or partial thromboplastin time or reduction of the platelet count to less than half the baseline or <100 000 platelets/mm3.

    In all septic patients, hypotension persisted after fluid administration (pulmonary capillary wedge pressure ≥12 mm Hg) and required administration of catecholamines (norepinephrine, epinephrine, dopamine, and/or neosynephrine) to maintain systolic blood pressure >90 mm Hg for 1 to 2 days. Arterial pressure was measured by transduction through an indwelling catheter and cardiac output by the thermodilution technique using a Swan-Ganz catheter in the pulmonary artery.

    Cardiogenic shock was diagnosed by a systolic arterial pressure ≤90 mm Hg, pulmonary capillary wedge pressure >15 mm Hg, and cardiac index of ≤2 L/min before catecholamine administration18 ; patients with fever, hypothermia, or an obvious source of infection were excluded. All patients required catecholamines to maintain systolic blood pressure >90 mm Hg.

    Plasma vasopressin was measured by radioimmunoassay using published protocols.19 For this assay, the reference range (95%) for hemodynamically normal subjects is <2.2 pg/mL for serum osmolality <285 mOsm/kg. Sensitivity was 0.3 pg per tube by dilution method.

    Vasopressin (vasopressin injection USP, 8-arginine vasopressin) was administered into a central vein at 0.04 U/min (Table 1). Intravenous fluids and medications were not changed for the hour before or the first hour of vasopressin administration except that during vasopressin administration, pressor catecholamines were decreased if systolic arterial pressure exceeded 130 mm Hg. Continuous measurements of systolic arterial pressure and heart rate were averaged in 15-minute intervals. In Table 2, “Pre-AVP” values are averages of the hour preceding vasopressin; “AVP” values, from the first hour of vasopressin, are averages of the 15-minute interval in which systolic arterial pressure reached maximum value. In Fig 2, vasopressin administration (“AVP”) values are averages of the hour before or after discontinuation; the “No AVP” value is the average of the 15-minute interval during which systolic arterial pressure reached its minimum value. Data were analyzed by unpaired t test (Fig 2) and paired t test (Table 2 and Fig 1). Differences were termed significant if t>5%.


    Nineteen patients with septic shock were studied. As detailed in “Methods,” all subjects had severe hypotension that required administration of catecholamines. Despite catecholamine administration, hemodynamic data showed hypotension due to low systemic vascular resistance (Table 1). At a time that systolic pressure averaged 92 mm Hg, mean plasma vasopressin concentration was 3.1 pg/mL (Fig 1). In contrast, in a group of patients with hypotension of similar duration and severity due to low cardiac output (cardiogenic shock; Table 1), plasma vasopressin was appropriately increased, with a mean of 22.7 pg/mL (Fig 1), which is in agreement with that of other forms of hypotension of similar magnitude.2021

    To distinguish between increased metabolism of vasopressin versus decreased secretion, we measured plasma vasopressin during infusion of exogenous hormone in two patients with septic shock. At a constant infusion of 0.01 U/min, the steady-state plasma concentration increased to 27 and 34 pg/mL, values expected in subjects given this infusion rate.22 Thus, the low plasma vasopressin in patients with septic shock appears to be due to impaired hormone secretion.

    Plasma sodium concentration was normal (mean, 140 mmol/L) in the patients with septic shock (Table 1), and although the osmotically mediated secretion of vasopressin was not formally examined, no patient had clinical evidence of diabetes insipidus. This suggests that patients in septic shock have a specific inhibition of baroreflex-mediated secretion of vasopressin.

    To examine whether vasopressin could constrict vascular smooth muscle in vasodilatory septic shock, 10 patients with this condition received vasopressin at 0.04 U/min IV. In normal subjects, significantly higher doses have little vasoconstrictor action9 and do not increase arterial pressure.910 In patients in septic shock, however, vasopressin increased systolic arterial pressure from 92 to 146 mm Hg (59% increase) due entirely to its vasoconstrictor effect. Whereas systemic vascular resistance increased 79%, cardiac output decreased 12% (Table 2). The increase in pressure occurred within minutes (<15 minutes) of the administration of hormone and frequently required decreasing or stopping the concomitantly administered catecholamine. Furthermore, in 6 of 10 patients, arterial pressure was maintained on vasopressin alone.

    Vasopressin at 0.04 U/min causes the plasma concentration to increase ≈100 pg/mL,22 which is substantially higher than the concentrations of 20 to 30 pg/mL that we found in the patients in cardiogenic shock (Fig 2) and those that others have found for this degree of hypotension.2021 To test whether this lower concentration could increase arterial pressure in septic shock, we infused the hormone at 0.01 U/min (shown above to provide a concentration of ≈30 pg/mL). Thus, in the six patients who were receiving vasopressin as the sole pressor, the hormone was stopped and within minutes, systolic arterial pressure declined from 126 to 83 mm Hg (P<.01) (Fig 2). The subsequent administration of vasopressin at 0.01 U/min resulted in a significant and sustained increase in systolic arterial pressure from 83 to 115 mm Hg (P<.01). Although systemic vascular resistance was not measured in these patients, the acute increase in arterial pressure was due to vasoconstriction because vasopressin only decreases cardiac output (see Table 2 and Reference 9).


    Plasma vasopressin was found to be inappropriately low in patients with advanced vasodilatory septic shock, indicating impaired baroreceptor-mediated vasopressin secretion. Furthermore, a dose of exogenous vasopressin that provides a plasma concentration expected for the degree of hypotension resulted in a marked pressor response in these patients. These results indicate that the low endogenous levels of hormone in septic shock contribute to the vasodilation of sepsis.

    The reason for impaired baroreflex-mediated vasopressin secretion in septic shock is unknown. First, autonomic failure is a possibility. Deficient baroreflex-mediated secretion of vasopressin is well documented in primary autonomic failure,2324 and sympathetic function appears to be impaired in septic shock.25 In support of autonomic dysfunction in our patients with septic shock, vasopressin did not cause the marked bradycardia observed under normal conditions (Table 2 ).2326

    A second potential explanation for inappropriately low vasopressin levels in human septic shock is depletion of the secretory stores of the neurohypophysis. This has been observed with strong osmotic stimuli,272829 and endotoxin is a most potent vasopressin secretagogue.16 In animal models of acute septic shock,141516 an enormous rise in plasma vasopressin during the 1 to 2 hours after endotoxin/bacterial administration (even before hypotension14 ) is followed by a rapid decline over the next few hours; no study has monitored plasma vasopressin in animals with septic shock of more than a few hours' duration.

    Needless to say, the mechanisms responsible for the profound vasodilation of septic shock are of great interest. Previous work has demonstrated abnormal activation of vasodilatory mechanisms in experimental models of septic shock.23 The findings reported herein document an abnormality of a vasoconstrictor mechanism critical for arterial pressure maintenance and provide the basis for new inquiries into the pathogenesis of the vasodilation in septic shock.

          Figure 1.

    Figure 1. Plasma vasopressin levels (AVP) of patients in septic shock and cardiogenic shock. P<.001.

          Figure 2.

    Figure 2. Systolic arterial pressure (SAP) response to vasopressin (AVP) withdrawal and readministration at 0.01 U/min (n=6). For AVP→No AVP, P<.01; for No AVP→AVP, P<.01.

    Table 1. Hemodynamic Data and Catecholamine Administration in Patients With Vasodilatory Septic Shock or Cardiogenic Shock

    Septic Shock (n=19)Cardiogenic Shock (n=12)
    Arterial pressure, mm Hg
    Cardiac output, L/min6.8±0.73.6±0.4
    Systemic vascular resistance, dyne·s/cm5837±781573±88
    Serum sodium, mmol/L140±2139±1
     Norepinephrine, μg/min5-213 (15)5-100 (35)
     Dopamine, μg·kg−1·min−15-20 (13)5-20

    Values are given as mean±SE except for catecholamine values, which are given as range (median).

    Table 2. Hemodynamic Data and Catecholamine Doses in 10 Patients With Vasodilatory Septic Shock During an Observation Period and During Administration of Vasopressin

    Arterial pressure, mm Hg
    Cardiac output, L/min6.6±1.05.8±0.3<.01
    Systemic vascular resistance, dyne·s/cm5664±551187±8<.001
    Heart rate, bpm111±9111±12NS
     Norepinephrine (n=5), μg/min4-213 (32)0-213 (0)
     Epinephrine (n=2), μg/min5, 113, 11
     Dopamine (n=2), μg·kg−1·min−112, 200
     Neosynephrine (n=3), μg/min40-149 (125)0-78 (53)

    Pre-AVP indicates an observation period before administration of vasopressin; AVP, administration of vasopressin. Values are given as mean±SE except for catecholamine values, which are given as range (median).


    Correspondence to Donald W. Landry, MD, PhD, Columbia University, Department of Medicine, 630 W 168th St, New York, NY 10032.


    • 1 Parrillo JE, Parker MM, Natanson C, Suffredini AF, Danner RL, Cunnion RE, Ognibene FP. Septic shock in humans: advances in the understanding of pathogenesis, cardiovascular dysfunction, and therapy. Ann Intern Med..1990; 113:227-242.CrossrefMedlineGoogle Scholar
    • 2 Kilbourn RG, Gross SS, Jurbran A, Adams J, Griffith OW, Levi R, Lodatao RF. NG-methyl-l-arginine inhibits tumor necrosis factor-induced hypotension: implications for the involvement of nitric oxide. Proc Natl Acad Sci U S A..1990; 87:3629-3632.CrossrefMedlineGoogle Scholar
    • 3 Landry DW, Oliver JA. The ATP-sensitive K+ channel mediates hypotension in endotoxemia and hypoxic lactic acidosis in dog. J Clin Invest..1992; 89:2071-2074.CrossrefMedlineGoogle Scholar
    • 4 Meadows D, Edwards JD, Wilkins RG, Nightingale P. Reversal of intractable septic shock with norepinephrine therapy. Crit Care Med.1988; 16:663-666.CrossrefMedlineGoogle Scholar
    • 5 Chernow B, Roth BL. Pharmacologic manipulation of the peripheral vasculature in shock: clinical and experimental approaches. Circ Shock..1986; 18:141-155.MedlineGoogle Scholar
    • 6 Baker CH, Sutton ET, Dietz JR. Endotoxic alteration of muscle microvascular renin-angiotensin responses. Circ Shock..1991; 36:224-230.Google Scholar
    • 7 Weitzberg E, Lundberg JM, Rudehill A. Elevated plasma levels of endothelin in patients with sepsis syndrome. Circ Shock..1991; 33:222-227.MedlineGoogle Scholar
    • 8 Grollman A, Geiling EMK. The cardiovascular and metabolic reactions of man to the intramuscular injection of posterior pituitary liquid (Pituitrin), Pitressin and Pitocin. J Pharmacol Exp Ther..1932; 46:447-460.Google Scholar
    • 9 Graybiel A, Glendy RE. Circulatory effects following the intravenous administration of Pitressin in normal persons and in patients with hypertension and angina pectoris. Am Heart J..1941; 21:481-489.CrossrefGoogle Scholar
    • 10 Wagner HN Jr, Braunwald E. The pressor effect of the antidiuretic principle of the posterior pituitary in orthostatic hypotension. J Clin Invest..1956; 35:1412-1418.CrossrefMedlineGoogle Scholar
    • 11 Aisenbrey GA, Handelman WA, Arnold P, Manning M, Schrier RW. Vascular effects of arginine vasopressin during fluid deprivation in the rat. J Clin Invest..1981; 67:961-968.CrossrefMedlineGoogle Scholar
    • 12 Schwartz J, Reid IA. Effect of vasopressin blockade on blood pressure regulation during hemorrhage in conscious dogs. Endocrinology..1981; 108:1778-1780.Google Scholar
    • 13 Schwartz J, Keil LC, Maselli J, Reid IA. Role of vasopressin in blood pressure regulation during adrenal insufficiency. Endocrinology..1983; 112:234-238.CrossrefMedlineGoogle Scholar
    • 14 Wilson MF, Brackett DJ, Hinshaw LB, Tompkins P, Archer LT, Benjamin BA. Vasopressin release during sepsis and septic shock in baboons and dogs. Surg Gynecol Obstet..1981; 153:869-872.MedlineGoogle Scholar
    • 15 Wilson MF, Brackett DJ, Tompkins P, Benjamin B, Archer LT, Hinshaw LB. Elevated plasma vasopressin concentrations during endotoxin and E. coli shock. Adv Shock Res..1981; 6:15-26.MedlineGoogle Scholar
    • 16 Brackett DJ, Schaefer CF, Tompkins P, Fagraeus L, Peters LJ, Wilson MF. Evaluation of cardiac output, total peripheral vascular resistance, and plasma concentrations of vasopressin in the conscious, unrestrained rat during endotoxemia. Circ Shock..1985; 17:273-284.MedlineGoogle Scholar
    • 17 Bone RC, Balk RA, Cerra FB, Delling RP, Fein AM, Knaus WA, Schein RM, Sibbald WJ. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Chest..1992; 101:1644-1653.CrossrefMedlineGoogle Scholar
    • 18 Califf RM, Bengtson JR. Cardiogenic shock. N Engl J Med..1994; 330:1724-1730.CrossrefMedlineGoogle Scholar
    • 19 Sakurai K. A simple and highly sensitive radioimmunoassay for 8-arginine vasopressin in human plasma using a reversed-phase C18 silica column. Folia Endocrinol..1985; 61:724-736.CrossrefGoogle Scholar
    • 20 Minaker KL, Meneilly GS, Young JB, Landsberg L, Stoff JS, Robertson GL, Rowe JW. Blood pressure, pulse and neurohumoral responses to nitroprusside-induced hypotension in normotensive aging men. J Gerontol Med Sci..1991; 46:M151-M154.Google Scholar
    • 21 Robertson GL. The regulation of vasopressin function in health and disease. Recent Prog Horm Res..1977; 33:333-386.Google Scholar
    • 22 Mohring J, Glanzer K, Maciel JA Jr, Dusing R, Kramer HJ, Arbogast R, Koch-Weser J. Greatly enhanced pressor response to antidiuretic hormone in patients with impaired cardiovascular reflexes due to idiopathic orthostatic hypotension. J Cardiovasc Pharmacol..1980; 2:367-376.CrossrefMedlineGoogle Scholar
    • 23 Zerbe RL, Henry DP, Robertson GL. Vasopressin response to orthostatic hypotension: etiologic and clinical implications. Am J Med..1983; 74:265-271.CrossrefMedlineGoogle Scholar
    • 24 Kaufmann H, Oribe E, Oliver JA. Plasma endothelin during upright tilt: relevance for orthostatic hypotension? Lancet..1991; 338:1542-1545.CrossrefMedlineGoogle Scholar
    • 25 Garrard CS, Kontoyannis DA, Piepoli M. Spectral analysis of heart rate variability in the sepsis syndrome. Clin Auton Res..1993; 3:5-13.CrossrefMedlineGoogle Scholar
    • 26 Cowley AW Jr, Monos E, Guyton AC. Interaction of vasopressin and the baroreceptor reflex system in the regulation of arterial blood pressure in the dog. Circ Res..1974; 34:505-514.CrossrefMedlineGoogle Scholar
    • 27 Cooke CR, Wall BM, Jones GV, Presley DN, Share L. Reversible vasopressin deficiency in severe hypernatremia. Am J Kidney Dis..1993; 22:44-52.CrossrefMedlineGoogle Scholar
    • 28 Negro-Vilar A, Samson WK. Dehydration-induced changes in immunoreactive vasopressin levels in specific hypothalamic structures. Brain Res..1979; 169:585-589.CrossrefMedlineGoogle Scholar
    • 29 Jones CW, Pickering BT. Comparison of the effects of water deprivation and sodium chloride inhibition on the hormone content of the neurohypophysis of the rat. J. Physiol..1969; 203:449-458.CrossrefMedlineGoogle Scholar


    eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

    Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.