Disorders of Circulatory Flow

There are three responses to mild blood loss: Stage 1) transcapillary refill, leaving an interstitial fluid deficit; Stage 2) activation of renin-angiotensin-aldosterone; Stage 3) production of erythrocytes within a few hours

Stages of Shock (Severity)

Parameter Class I Class II Class III Class IV
% Blood loss < 15% 15 – 30% 30 – 40% %> 40%
Supine BP Normal Normal Decreased Decreased
Urine output (mL/hr) > 30 20 – 30 5 – 15 < 5
Mental Status Anxious Agitated Confused Lethargic

Source: Committee on Trauma: Advanced Trauma Life Support; ACS, 1989: 57

Class I

Class I: clinical manifestations are minimal

Class II

Class II: look for tachycardia and orthostatic changes but with the knowledge that they are unreliable manifestations of blood loss [Williams TM: The Clinical Use of Orthostatic Vital Signs: 445, 1991]

Class III

Class III: there is some evidence that the tachycardia-vasoconstrictor response can be lost by this point [Am J Physiol 260: H305, 1991]. However, if blood pressure drops in supine patients, you know you’re at least Class III

Class IV

Class IV: circulatory collapse impending

Classification of Shock (Etiology)

Shock states can generally be grouped into one of four groups based on etiology

Etiologic Classification of Shock

  • Hypovolemic: hemorrhagic and non-hemorrhagic (e.g. diarrhea)
  • Distributive: sepsis, neurogenic, anaphylactic
  • Cardiogenic: MI, valvular failure structural abnormality, arrhythmia
  • Obstructive: pneumothorax, tamponade, pulmonary embolism

Monitoring Shock

Physiologic Variables

Systemic Blood Pressure

Systemic blood pressure is an insensitive measure of blood loss, but if low can be used as a reliable guide in the resuscitative effort.


Orthostatics are only slightly less sensitive and should not be relied upon [JAMA 281: 1022, 1999]. ====Filling Pressures==== Filling pressures (CVP, PCWP) do not correlate well with volume loss [Crit Care Med 12: 107, 1984], often not changing until blood loss exceeds 30%. However, placing patients in an upright position may improve the reliability of filling pressure measurements [Lancet i: 258, 1989]. Overall, experimental studies do not support the notion that central pressures are a reliable indicator of volume status [Chest 121: 2000, 2002; Crit Care Med 32: 691, 2004] – they are best used to evaluate really high or really low values, but at intermediate values are practically useless.

Oxygen Extraction

Oxygen extraction > 30% is a reasonable marker for hypovolemia, with ER > 50% or VO2 < 100 mL/min/m2 suggesting actual shock (at which point a blood lactate is helpful). When analyzing ER with regards to shock, always take into account hematocrit as it will affect the expected value.

Base Deficit

Base deficit (normally –2 to +2) is a useful measurement of acidosis and is calibrated to a pCO2 of 40, which makes it a more specific marker of non-respiratory acid-base disturbances than bicarbonate (which is measured at the patient’s given pCO2) – in an injured or bleeding patient, a base deficit is a good marker of acidosis from impaired oxygenation [J Am Col Surg 187: 384, 1998]. Base deficit is correlated with outcome and can be used to assess resuscitation [J Trauma 28: 1464, 1998]. ====Lactate==== Lactate is even better, its predictive value is not confined to the time of initial assessment and compared to lactate shows a closer correlation with blood loss [Crit Care Med 27: 154, 1999] and risk of death [Am J Surg 185: 485, 2003].

End-Expiratory CO2

End-expiratory CO2 drops as cardiac output falls and can be measured using standard nasal cannulas – it has been recommended for evaluating response to cardiopulmonary resuscitation [Crit Care Clin 8: 323, 1992] and may have a role in evaluating hypovolemia.

Hemotocrit: Do Not Use

Hematocrit should not be used to evaluate blood loss [Committee on Trauma: Advanced Trauma Life Support; ACS, 1989: 57] as it generally doesn’t change until 8 – 12 hours post-fluid loss and can be further influenced by crystalloids

Treatment of Shock

Selection of Catheters Peripheral catheters are preferred to central catheters (unless very large introducer catheters are employed) because they are much shorter. Large introducer catheters are even better, as a 9 French has a 3 mm diameter and they can be used as stand-alone infusion devices when necessary (always bypass the side infusion port). It is a common misconception that colloids flow more slowly than crystalloids – this is not true

Position and Compression

The Trendelenberg position (pelvis elevated above the head) has never been shown to be effective. Trendelenburg raises MAP, wedge, and SVR but does not affect cardiac output [Ann Emerg Med 23: 564, 1994] because venous return is unchanged (particularly in hypovolemic situations where high venous distensibility mitigates any possible pressure changes). Pneumatic compression similarly may raise blood pressure by increasing resistance, not by increasing venous return – in thoracic injuries, this can actually enhance blood loss [J Trauma 31: 846, 1991]. MAST trousers should therefore only be used for prehospitalization stabilization of trauma patients

Physiologic Effects of Head-Down Positioning

  • Cardiac: initial fluid bolus (1L), quickly offset by reflex barostimulation, vasodilation, and potentially decreased perfusion of the brain
  • Respiratory: increased atelectasis, work of breathing, movement of ETT

Resuscitation Strategies

The first priority should be to restore cardiac output, for which asanguinous fluids are generally best – PRBC have actually been shown to decrease cardiac output in resuscitation situations [JAMA 269: 3024, 1993]. Experiments suggest that colloids such as 10% Dextran-40 are 8 times more effective than crystalloids at increasing cardiac output within 1 hour [Intensive Care Med 13: 230, 1987], but there is still no data yet to show that colloids are superior to crystalloids for resuscitation in terms of outcomes [Curr Opin Crit Care 6: 395, 2000]. Generally start with 1 – 2 L crystalloid (or 6 mL/min/kg) and repeat if effective, otherwise move to colloids and blood products

Step 1) estimate blood volume (66 ml/kg men, 60 ml/kg in women, lean mass only). Adjust for obesity and old age Step 2) estimate % blood loss (see above) Step 3) volume deficit = blood volume x % loss Step 4) calculate resuscitation volume (1.0 x VD for blood, 1.5 x VD for colloid, 4 x VD for crystalloid). Blood is generally reserved for class III or IV blood loss

Common endpoints include:

  1. CI > 3 L/min/m2
  2. VO2 > 100 mL/min/m2
  3. DO2 > 500 mL/min/m2
  4. Blood lactate < 4 mM
  5. Base deficit > -2 mM
  6. Filling pressures: CVP > 15 mm Hg or PCWP 10 – 12 mm Hg (outdated???)

Base deficit (mM base needed to correct 1 L whole blood to a pH of 7.40) has been shown to correlate with volume deficits and mortality in trauma patients [Surg Gynecol Obst 173: 473, 1991; Am J Respir Crit Care Med 162: 2246, 2000], however there are those who believe that serum bicarbonate is a reliable surrogate [Am J Surg 190: 941, 2005] at least for predictive purposes

Functional criteria for transfusion might include 1) VO2 below the normal range 2) lactate > 4 mM and 3) ER > 0.5. Transfusing until VO2 no longer increases with each additional unit might be a more rational strategy than following H&H

Note that volume resuscitation to normotension may promote continued blood loss [NEJM 331: 1105, 1994], thus normal blood pressures are not appropriate endpoints in volume resuscitation

Volume Resuscitation and Fluid Selection

Normal Saline

1 L of 0.9% saline adds 275 mL to the plasma and 825 mL to the interstitial space (total of 1100 mL). One disadvantage is that its chloride content is high relative to plasma, giving rise to the possibility of hyperchloremic metabolic acidosis after significant resuscitation

Increase in Plasma Increase in Interstitial D5W 50 250 0.9% NaCl 275 825 7.5% NaCl 500 600 5% Albumin 500 500

Lactated Ringer’s

Lactated Ringer’s contains potassium and calcium and 28 mEq/L of sodium lactate (a buffer). There is no evidence that it is superior to 0.9% saline or that the lactate provides any buffering effect. Furthermore, the calcium in LR can bind certain drugs (amicar, amphotericin, blood products, defamandole, metaraminol, thiopental, and probably ampicillin, doxycyclin, and others [J Natl IV Ther Assoc 9: 480, 1986]). LR is thus contraindicated as a diluent for transfusions

Normosol and/or Plasma-Lyte

Normosol and/or Plasma-Lyte have calcium, magnesium, and different buffers (pH 7.4). The problem with magnesium is that it can lead to hypermagnesemia in renally insufficient patients, and can lead to hypotension in low flow states


Dextrose is obsolete as it is not a good volume expander and calories are generally provided through enteral or TPN routes. The excess glucose in D5 can lead to a hyperosmotic plasma and can dehydrate cells. It can also enhance CO2 production, enhance lactate production [Arch Surg 122: 765, 1987], and aggravate ischemic brain injury [Crit Care Med 20: 104, 1992]. Given the recent data on glucose control in ICU patients, Marino believes that dextrose should not be used routinely in critically-ill patients


1 L of 5% albumin adds 700 mL to the plasma and 300 mL to the interstitial space, lasting 12 – 18 hours. 25% albumin expands volume at the expense of interstitial fluid volume and thus should not be used for resuscitation in hypovolemic patients – it was designed to shift fluid in hypoproteinemic patients but even then may be questionable. A study in 1998 suggested that 1 out of every 17 patients receiving albumin died as a result [Br J Med 317: 265, 1998]– this was looked at in a large, prospective, multicenter study of 7000 patients [NEJM 350: 2247, 2004] and has not been corroborated. A recent metaanalysis attempted to suggest that if adverse events are the endpoint, albumin is safer than saline [Crit Care Med 32: 2029, 2004], however this study included patients with ascites (known to benefit from albumin) in their analysis – this subgroup overwhelmed the others, none of whom showed any benefit


Hetastarch as a synthetic colloid made of amylopectin molecules and is significantly less expensive than albumin. MW is similar but half-life is much longer, although this is misleading as its osmotic effects disappear within 24 hours (cleaved by amylase within 24 hours, cleared by kidneys weeks later). Hetastarch can sometimes elevate PTT but as of 1992 had never been proven to affect bleeding rates [Crit Care Clin 8: 235, 1992]. It now appears that hetastarch can increase bleeding tendency, but usually only if given in doses > 1500 cc over 24 hours [Intensive Care Med 25: 258, 1999; Crit Care Med 29: 1261, 2001] or in patients with an underlying coagulopathy [Anesthesiology 101: 560, 2004]


Dextrans are glucose polymers made by Leuconostoc bacteria. Dextran-70 lasts longer and is thus preferred. Anaphylaxis occurs in 0.032%. Dextrose can coat RBC and interfere our ability to cross-match blood products, so red cell preparations have to be washed. Some have proposed that dextrans may cause renal failure but this is unproven – animal studies in sheep [Shock 5: 289, 1996] show that Dextran does not affect renal function and in fact can improve GFR in hypovolemic animals

Data On and Composition Of Fluids

Composition of IV Fluids

Na Cl K Ca Mg Buffers PH Osmolality Plasma 141 103 4-5 5 2 Bicarb 7.4 289 0.9% NaCl 154 154 0 0 0 None 5.7 308 7.5% NaCl 1283 1283 0 0 0 None 5.7 2567 Lactated Ring. 130 109 4 3 0 Lactate 6.4 273 Normosol 140 98 5 0 3 Acetate 7.4 295 Plasma-Lyte 140 98 5 0 3 Gluconate 7.4 295

Volume Expansion and Half-Life of Various Colloids

Plasma Volume Expansion Serum Half-Life 5% Albumin 0.7 – 1.3 16 hr 6% Hetastarch 1.0 – 1.3 17 days* 10% Pentastarch 1.5 10 hr 10% Dextran-40 1.0 – 1.5 6 hr 6% Dextran-70 0.8 12 hr

Data on Colloids

Despite is superiority in filling the vascular space, colloid has not been shown to improve survival rates [Crit Care Clin 9: 313, 1993; Crit Care Med 24: S12, 1996; Crit Care Med 27: 200, 1999; AIM 135: 149, 2001; NEJM 350: 2247, 2004]. The problem with these studies, according to Marino, is that they classify all hypotensive patients regardless of etiology (ex. dehydration vs. blood loss)

Hypertonic Resuscitation

Hypertonic resuscitation has recently generated renewed interest in certain mechanisms of injury, including head injury [J Trauma 42: S61, 1997], dehydration [Shock 6 (suppl): 32, 1996], and penetrating injury [Ann Surg 213: 482, 1991]. Some studies have shown that small-volume (4 mL/kg) infusion of hypertonic saline can restore mean arterial pressure and microvascular perfusion [J Trauma 54(suppl): 89, 2003; Surgery 122:609, 1997]. Interestingly, a recent RCT in patients with head injury did not show any difference in outcomes between HS and LR [JAMA 291: 1350, 2004]

Effects of Fluids on Immune Function

Recent evidence suggests that isotonic crystalloid solutions may actually aggravate the immune dysfunction [Crit Care Med. 2000;28:74–78]. By contrast, a number of experimental studies reveal that hypertonic solutions have favorable immunomodulatory effects on hemorrhage/resuscitation-induced leukocyte activation [Shock 8: 235, 1997; J Trauma 49: 580, 2000]

Predictive Measures: lactate, acidemia

Regarding the composition of the resuscitation fluid, despite encouraging results with new preparations of stroma-free hemoglobin and hypertonic salt solutions with colloid, searches for the optimal combination of oxygen-carrying blood substitute, colloid, and electrolyte solution for limited fluid resuscitation with the smallest volume should continue. For titrating treatment of shock, blood lactate concentrations are of questionable value although metabolic acidemia seems helpful for prognostication [Crit Care Med 24: S12, 1996]

Correcting Anemia: Is Hematocrit Useful?

Hematocrit is not an accurate indicator of plasma volume or oxygen carrying capacity, especially in patients who recently received or lost fluids. A more rational transfusion trigger is extraction ratio [J Trauma 32: 769, 1992], which is defined as VO2/DO2 or as (SaO2 – ScvO2) / (SaO2) – this calculation requires a venous blood sample.