ARTICLE: Hernandez G, Ospina-Tascon GA, Kattan E, et. al, for The ANDROMEDA-SHOCK-2 Investigators. Personalized hemodynamic resuscitation targeting capillary refill time in early septic shock: the ANDROMEDA-SHOCK-2 Randomized Clinical Trial. JAMA. October 29, 2025. doi:10.1001/jama.2025.20402.
OBJECTIVE: Comparison of a personalized hemodynamic resuscitation (PHR) strategy targeting capillary refill time (CRT) against usual care in the improvement of patient-centered outcomes amongst patients with early septic shock1
BACKGROUND
Septic shock is a common life-threatening pathology seen in the emergency department and intensive care unit. Despite vast experience in treating septic shock, the optimal treatment strategy remains uncertain. Several studies during the past decade have examined management strategies for fluids and mean arterial pressure (MAP) targets but have failed to improve patient-centered outcomes. Today, the typical management includes source control, antibiotics, fluids, and vasopressors titrated to a MAP goal. It often involves lactate as a marker of resuscitation.
Abnormal serum lactate levels have been considered a marker of hypoperfusion, indicating anaerobic metabolism. However, studies have shown lactate to also be an end-product of endogenous (or exogenous) epinephrine, due to increasing glycolysis and pyruvate generation. Thus, an elevated serum lactate level is not specific to hypoperfused states nor an indicator of the severity of shock physiology. From a pragmatic standpoint, sepsis protocols can often result in large-volume fluid resuscitation to normalize lactate levels.
Sepsis is complex with diverse mechanisms involving hypovolemia, vasoplegia, and often cardiac dysfunction. The source of sepsis also varies, complicating the pathophysiology. Patients can present with both vasoplegia and cardiac dysfunction, making a large-volume resuscitation less beneficial (if not harmful). MAP targets are also challenging to define, as patients may have widely varying "native" blood pressures. A strict MAP target can result in large doses of vasopressors, which can also be deleterious. Given this, fluid resuscitation with serial lactates and a specific MAP target may not be the best uniform approach when treating septic shock.2,3,4
The 2019 ANDROMEDA-SHOCK study demonstrated that capillary refill time (CRT) as a non-laboratory assessment of resuscitation adequacy was non-inferior to serum lactate measurements. ANDROMEDA-SHOCK 2 looks at outcomes when applying a personalized hemodynamic resuscitation strategy. ANDROMEDA-SHOCK 2 is a large, pragmatic international randomized study in which resuscitation in early septic shock is guided by an algorithm using CRT and hemodynamic phenotyping to determine fluid, vasopressor, and inotrope use.
DESIGN
This multicenter randomized clinical trial compared a personalized hemodynamic resuscitation protocol targeting capillary refill time (CRT-PHR) vs. usual care with patients in early septic shock. Patients were enrolled from across 86 intensive care units (ICUs) in 19 countries from March 2022 to April 2025. Patients were randomized in a 1:1 ratio into the two groups. The trial was unblinded to patients and clinicians. Patients were recruited up to 4 hours after the inclusion criteria were met.
In the CRT-PHR group, resuscitation was guided by hourly CRT assessment and a multi-layered step-wise treatment algorithm, illustrated best in the JAMA article (Figure 1, which JAMA denied permission to use here). Abnormal CRT triggered the initiation of the algorithm in two tiers. First, pulse pressure was assessed and maintained at > 40mmHg with diastolic BP >50mmHg through fluids and/or norepinephrine, depending on volume responsiveness. CRT was reassessed every 30 minutes during intervention time. If CRT remained abnormal despite tier 1 intervention, then the CRT-PHR group entered tier 2 of the PHR algorithm. Tier 2 used echocardiography and further assessment of fluid responsiveness. Tier 2 also introduced an inotrope challenge and MAP push to attempt to normalize CRT. Once CRT normalized, monitoring continued hourly without additional resuscitative interventions, up to 6 hours. If CRT abnormalities persisted following all tier 2 interventions, further management exited the CRT-PHR protocol and reverted to clinician judgement.
In the usual care arm, CRT was assessed at hour 0 baseline and at hour 6 of protocol window. Their management was guided by local sepsis protocol and/or international guidelines.
CRT was obtained by counting the number of seconds it took for normal skin color to return after application of firm pressure for 10 seconds with a glass slide to the right index distal phalanx. Abnormal CRT was defined as ≥ 3 seconds.
The primary outcome was a hierarchical win ratio, a hierarchical composite. This method permits prioritization of the clinically most important outcomes, rather than treating each outcome equally. This design permits for increased statistical power compared with analyzing mortality alone (88% to show superiority in the hierarchical composite outcome for a 2-sided α of .05, per the investigators description of this sophisticated statistical analysis.
INCLUSION CRITERIA
- Patients aged ≥ 18 years with septic shock in ICU, ED, OR, hospital wards
- Septic shock is defined as suspected or confirmed infection, plus hyperlactatemia (≥2.0 mmol/L), and requirement of norepinephrine to maintain a MAP of ≥ 65mmHg after an IV fluid of at least 1000 mL
EXCLUSION CRITERIA
- > 4 hours since septic shock diagnosis
- Anticipated surgery or dialysis procedure during the 6-hr intervention period
- Hemorrhage
- Severe ARDS
- Child B-C Cirrhosis
- DNR code status
- Refractory shock (high risk of death within 24 hours)
- Pregnancy
PRIMARY OUTCOME
Hierarchical composite outcome win ratio. The win ratio was calculated by comparing all possible patient pairs, starting with the first event in the hierarchy and stratified by median APACHE II score at admission. Components included:
- All-cause mortality within 28 days
- Duration of vital support including vasoactive agents use, mechanical ventilation, renal replacement therapy within 28 days
- Length of hospital stay within 28 days
SECONDARY OUTCOME
- All-cause mortality
- Vital-support-free days within 28 days (defined as the number of days up to 28 days in which the patient did not require cardiovascular, respiratory, and/or renal support)
- Length of hospital stay
KEY RESULTS
Final study cohorts:
- CRT-PHR group: 720 patients
- Usual care group: 747 patients
- Subjects enrolled from 86 sites in 19 countries
Among patients with early septic shock, the CRT-PHR protocol was superior to usual care for the primary composite outcome, owing primarily to a lower required duration of vital support. Overall, CRT-PHR group received less resuscitation fluids, received more dobutamine, had a lower CVP, and lower lactate levels by the end of the 6-hr period compared with usual care.
Primary Outcome
- Hierarchical composite of death, duration of vital support, and LOS (number of wins, %): 131,131 (48.9)% vs. 112,787 (42.1.%) for CRT-PHR group vs. usual care group, respectively. This yielded a stratified win ratio of 1.16 (p = 0.04)
- The total number of ties was 24,151 (9.0%)
- Of the 3 components of the win ratio (CRT-PHR vs. usual care):
- Duration of vital support yielded the highest number of wins to losses (26.4% vs. 21.1%)
- Wins for death were 19.1% vs. 17.8%
- Wins for LOS were 3.4% vs. 3.2%
Secondary Outcome
- All-cause mortality within 28 days (number of deaths, %) did not demonstrate a significant difference
- CRT-PHR: 191 (26.5%)
- Usual care: 199 (26.6%)
- Patients in the CRT-PHR group had a higher number of mean vital support-free days within 28 days (SD)
- CRT-PHR: 16.5 (11.3)
- Usual care: 15.4 (11.4)
- Proportional odds ratio 1.28 (CI 1.06-1.154)
- Mean length of hospital stay up to day 28 (SD) did not demonstrate a significant difference
- CRT-PHR: 15.3 (9.0)
- Usual care: 16.2 (9.4)
LIMITATIONS
- This was an unblinded study.
- Decisions whether to discontinue vasopressors, extubation, and initiation of RRT were made by attending intensivists per local practice, as opposed to dictated by an algorithm.
- Interrater variability in measuring CRT
- Labor-intensive protocol may not be practical in certain clinical contexts.
- Greater patient assessment frequency in the CRT-PHR randomized subjects may have contributed to the observed outcome.
- The usual care group could potentially have been unintentionally undertreated.
- Low mortality rate observed in both groups
- Usual care group may have been subject to variations in their care across study sites and countries, which could influence outcomes.
EM TAKE-AWAYS
The pathophysiologic mechanisms underlying septic shock are diverse and unique to the individual patient. ANDROMEDA-SHOCK 2 performs a personalized approach using dynamic assessments during early resuscitation of septic shock. This study challenges the current dogma of sepsis treatment and investigates the complex interactions of volume status, cardiac output, and vasodilation. ANDROMEDA-SHOCK 2 should challenge ED and ICU physicians to focus on a personalized hemodynamic resuscitation approach for treatment of patients in septic shock.
REFERENCES
- Hernandez G, Ospina-Tascon GA, Kattan E, et. al, for The ANDROMEDA-SHOCK-2 Investigators. Personalized hemodynamic resuscitation targeting capillary refill time in early septic shock: the ANDROMEDA-SHOCK-2 Randomized Clinical Trial. JAMA. October 29, 2025. doi:10.1001/jama.2025.20402.
- Muller J, Radej J, Horak J, Karvunidis T, Valesova L, Kriz M, Matejovic M. Lactate: The fallacy of oversimplification. Biomedicines. 2023;11(12):3192.
- Evans L, Rhodes A, Alhazzanni W, et. al, for the Society of Critical Care Medicine. Surviving Sepsis Campaign: Adult Guidelines. Intensive Care Med. 2021;47(11):1181-1247.
- Hernandez G, Ospina-Tascon GA, Damiani LP, et. al, for the ANDROMEDA-SHOCK investigators. Peripheral perfusion status vs serum lactate levels on 28-day mortality among patients with septic shock: The ANDROMEDA-SHOCK randomized clinical trial. JAMA. 2019;321(17):654-664.