Peripheral Perfusion vs. Serial Lactates

Washington University Emergency Medicine Journal Club- November 2020


Ms. S is a 73-year-old female presenting to your emergency department complaining of shortness of breath and fevers that have been worsening over the last 2 days. She has a cough productive of yellow sputum and notes chest pain with coughing. Today, she developed lightheadedness, fatigue, and generalized weakness that, in conjunction with her other symptoms, led her to present for treatment. She has a history of hypertension and COPD. She adheres to her home medication regimen, has quit smoking, and has not been hospitalized for her COPD in more than a year.    

At triage, the patient’s vital signs are: T 39.3C, BP 72/30, HR 134, RR 24, and 82% oxygen saturation on room air. She is immediately brought back for evaluation and placed on supplemental oxygen. Labs are significant for a lactate of 6.3 mmol/L, WBC 16.4, and a creatinine of 3.4 mg/dL. A chest x-ray reveals an infiltrate affecting the right middle and lower lobes. You begin fluid resuscitation and broad-spectrum antibiotics. 

After the patient has received 30 ml/kg of Lactated Ringer’s, her blood pressure is 87/35. She has only had 10 ml of urine output. The patient is started on a norepinephrine infusion and a central line is placed. Bedside echo shows a mildly reduced LV EF of 40-50%, grossly normal RV function, and an IVC that is 2cm in diameter with >50% collapse with inspiration. Her skin is warm and her capillary refill is 2.5 seconds. You are undecided about giving more fluid, since you have read a lot recently about over-resuscitation worsening outcomes. You are hoping that the patient’s repeat lactate level will provide you with some additional information and keep you in compliance with the ED sepsis bundle. However, the nurses are busy attending to other patients and the point-of-care instrument is down, so there will be a significant delay in obtaining the result. There has also been a lot of focus recently on the cost of serial labs and the utility of lactate monitoring at short time intervals. You wonder if there may be another way to assess the patient’s perfusion in real time to guide resuscitation and disposition and set a phone reminder to check the literature as soon as your shift ends… 

PICO Question

Population: Adults with sepsis and septic shock

Intervention: Therapy and prognostication utilizing peripheral perfusion

Comparison: Therapy and prognostication utilizing lactate measurements

Outcome: Mortality, organ dysfunction, ICU and hospital length of stay, Wluid
administration, and mechanical ventilation-, renal replacement therapy- and
vasopressor-free days.

Article 1: Lara B, Enberg L, Ortega M, et al. Capillary reWill time during Wluid
resuscitation in patients with sepsis-related hyperlactatemia at the emergency
department is related to mortality. PLoS One. 2017 Nov 27;12(11):e0188548.
Answer Key.

Article 2: Alegrı́a L, Vera M, Dreyse J, et al. A hypoperfusion context may aid to
interpret hyperlactatemia in sepsis-3 septic shock patients: a proof-of-concept study.
Ann Intensive Care. 2017 Dec;7(1):29. Answer Key.

Article 3: Hernández G, Ospina-Tascó n GA, Damiani LP, et al. Effect of a Resuscitation
Strategy Targeting Peripheral Perfusion Status vs Serum Lactate Levels on 28-Day
Mortality Among Patients With Septic Shock: The ANDROMEDA-SHOCK Randomized
Clinical Trial. JAMA. 2019 Feb 19;321(7):654-664. Answer Key.

Article 4: Kattan E, Hernández G, Ospina-Tascó n G, et al; ANDROMEDA-SHOCK Study
Investigators and the Latin America Intensive Care Network (LIVEN). A lactate targeted resuscitation strategy may be associated with higher mortality in patients
with septic shock and normal capillary re Will time: a post hoc analysis of the
ANDROMEDA-SHOCK study. Ann Intensive Care. 2020 Aug 26;10(1):114. Answer

Bottom Line:

Resuscitation of sepsis and septic shock in the emergency department has evolved significantly over the last two decades. Since the seminal publication of Emanual River’s paper on early goal-directed therapy (EGDT) in 2001, measurement of serum lactate level has been a key component in the diagnosis of septic shock. Elevated lactate levels in such cases were felt to be a marker of hypoperfusion and tissue hypoxia leading to anaerobic metabolism. As a result, guidelines from the Surviving Sepsis Campaign recommended resuscitation be targeted to normalization of serum lactate levels. A retrospective, observational study from the ED at Washington University (Dettmer 2016) even demonstrated an inverse association between serial lactate monitoring and mortality among patients with severe sepsis and septic shock. 

However, there is increasing evidence that hyperlactatemia in sepsis is not entirely a result of hypoperfusion and anaerobic metabolism, but is also a result of increased aerobic glycolysis due to adrenergic stimulation, and may actually be an adaptive response to stressful stimuli (Garcia-Alvarez 2014). With this in mind, reliance on serial lactate levels to guide fluid resuscitation in sepsis may actually lead to over-resuscitation in some cases, and alternative measures of perfusion (such as capillary refill time [CRT], central venous oxygen saturation [ScvO2], and venous-arterial carbon dioxide gap [P(cv-a)CO2]) may be more useful. 

CRT is easily measured at the bedside, requiring no additional equipment beyond a chronometer (i.e. a clock or the stopwatch on your phone), and yields immediate results. A prospective, observational study out of Chile (Lara 2017) found that while baseline CRT was not associated with increased risk of adverse outcomes (a composite of ICU length of stay ≥ 72 hours, need for mechanical ventilation ≥ 48 hours, need for renal replacement therapy, and in-hospital mortality), an abnormal CRT after initial fluid resuscitation was (RR 4.4, 95% CI 2.7-7.4). There was also a significantly increased risk of 28-day mortality among those with an abnormal CRT after initial resuscitation (RR 6.7, 95% CI 2.9-16). This study was unfortunately limited by a rather small sample size (n of 95) and a failure to adhere to the STROBE guidelines for reporting in observational studies

Another study out of this same hospital in Chile (Alegría 2017) used a more broad approach to identify septic patients with hypoperfusion. In this retrospective study, patients with septic shock and a “hypoperfusion context,” defined as an ScvO2 < 70%, a P(cv-a)CO2 ≥ 6 mmHg, or a CRT ≥ 4 seconds, generally required more aggressive treatment. This subset of patients required more norepinephrine (0.19 vs. 0.09 mcg/kg/min), were more likely to receive dobutamine (31% vs. 5%), and received more IV fluids (6732 ± 2524 vs. 5940 ± 2756 mL) than those without a “hypoperfusion context.” There was a trend toward increased mortality (RR 2.86, 95% CI 0.39-21) but this did not achieve statistical significance, owing in large part to the small sample size (n = 90, of whom 70 had a “perfusion context”). 

While both of these studies suggest that alternative measures of hypoperfusion can be used to identify patients at higher risk of poor outcomes, neither demonstrated how or whether such measures can be used to guide resuscitation. More specifically, it remains to be seen whether such measures are superior to serial lactate measurements in guiding resuscitation of patients with septic shock. The ANDROMEDA-SHOCK study out of South America attempted to do just that. This multicenter, randomized controlled trial compared a resuscitation strategy guided by every half-hour CRT measurements to a strategy guided by every 2-hour lactate measurements in patients meeting Sepsis-3 criteria for septic shock: sepsis plus a lactate ≥ 2.0 mmol/L and need for vasopressors following fluid resuscitation to maintain a mean arterial pressure (MAP) ≥ 65 mmHg. While there was a trend toward a decrease in the primary outcome (28-day mortality) in the CRT-guided group, this did not quite achieve statistical significance (34.9% mortality vs. 43.4%; hazard ratio 0.75, 95% CI 0.55 to 1.02). There was also less organ dysfunction at 72 hours, but no difference in 90-day mortality, mechanical ventilator-, renal replacement-, or vasopressor-free days within 28 days, or ICU or hospital length of stay. 

While it would be easy to dismiss ANDROMEDA-SHOCK as a negative study, it should be noted that at the very least a CRT-guided resuscitation strategy was not inferior to serial lactate measurements. Although this was not designed as a non-inferiority study, one could argue that it should have been given the nature of the intervention. In situations where serial lab measurements are not feasible, a simple bedside test appears to be at least as good, and could easily serve as a surrogate. More academically, there has been a call among some scientists to rethink (or even entirely abandon) the concept of statistical significance, and while the 95% confidence interval for the hazard ratio for mortality in this study crossed 1, it barely did so. As there is no harm associated with CRT measurement, it could easily be argued that this is evidence enough to begin guiding resuscitation based on CRT rather than serial lactate measurement. 

Earlier this year, a secondary, post hoc analysis of the ANDROMEDA-SHOCK study was published. In this analysis, the authors chose to look only at those patients whose CRT had normalized by hour 2 of the resuscitation period. For this subgroup of patients (n = 184), those undergoing further resuscitation based on serial lactate measurements had a significantly higher 28-day mortality compared with those whose resuscitation was guided by CRT measurement (40% vs. 23%; RR 1.8, 95% CI 1.1 to 2.8). Patients in the lactate-guided group received more supportive therapies and resuscitative interventions (including fluid administration, vasopressor use, and inodilator tests), and it is possible that these additional interventions, while well-intentioned, led to over-resuscitation based on a poor measure of perfusion status.  

While interesting, these results should be viewed with caution. This was a retrospective, secondary analysis of data and is at a different level of evidence than the original randomized controlled trial. In particular, this analysis excluded patients whose CRT did not normalize at hour 2. By excluding a subset of patients, the previous prognostic balance achieved by randomization was destroyed. While certainly thought-provoking, these results should be hypothesis-generating rather than practice changing. 

As previously noted, serum lactate may remain elevated in patients with septic shock for reasons other than ongoing hypoperfusion, and hence may be a poor measure of perfusion status throughout the resuscitation process. Overall, the data we reviewed supports the notion that hypoperfusion in sepsis can be measured by methods other than serum lactate levels, including simple bedside CRT measurement. In addition, CRT may be used to guide fluid resuscitation in these patients, with a suggestion that its use is superior to serial lactate measurement. At the very least, these papers suggest that factors other than serial lactate should be monitored, and it may well be that more frequent bedside assessment alone may improved resuscitation in these patients.