Podcast: Play in new window

Subscribe: RSS

Date: November 27, 2025 

Guest Skeptic: Dr. Justin Morgenstern is an emergency physician and the creator of the #FOAMed project called www.First10EM.com

Case: You are looking after a 65-year-old man who appears to be in septic shock. He presented after five days of fever and cough, and is now severely lethargic and hypotensive on arrival. You give him antibiotics and IV fluids immediately, but an hour later, his lactate comes back at 5, and you need to start norepinephrine to keep his MAP above 65. You put in a call to the intensive care unit (ICU) to get him transferred, and the intensivist asks you whether you have started personalized hemodynamic resuscitation targeting capillary refill time. You don’t want to sound dumb, but what the heck is personalized hemodynamic resuscitation protocol targeting capillary refill time?

Background: Sepsis is a life-threatening organ dysfunction caused by a dysregulated host response to infection. Septic shock is the most severe end of that spectrum. Patients with sepsis have persistent hypotension requiring vasopressors to maintain MAP ≥65 mm Hg and a lactate >2 mmol/L despite adequate volume resuscitation (Sepsis‑3). In high-income countries, mortality has fallen but remains substantial.  In many settings, mortality can be between 30% to 70%. High-quality ED care requires early recognition, IV antibiotics, source control, hemodynamically directed fluids, and vasopressors.

The management of septic shock has changed dramatically since the time that Ken and I started practice. We went through a period in which a very aggressive bundle of care was proposed, based on work by Dr. Emanuel Rivers, published in the NEJM in 2001. Then, we ran big trials on the components of that bundle, and found that none of them helped individually (ARISE, ProCESS & ProMISe). It was clear that these patients benefited from close attention and clinical reassessments, but aside from early antibiotics, the exact interventions needed were unclear.

For a while, many people focused on trending lactate levels. We then saw the original ANDROMEDA SHOCK study, which showed that a resuscitation strategy focused on clinical assessments of capillary refill time was at least as good as a strategy focused on trending lactates.

We have been left with the question of exactly how to improve capillary refill and which other targets are important. There has been a question about whether a higher MAP target might help (SGEM#90), especially in elderly patients with more baseline hypertension.  But the recent OPTRESS study showed worse outcomes with a higher MAP target in elderly septic shock patients.

Therefore, aside from the consensus that providing early antibiotics is a good idea, there remain many questions about the ideal initial resuscitation strategy for septic shock patients.

Clinical Question: In adult patients with septic shock, can death, duration of vital support, and/or hospital length of stay be improved by a “personalized hemodynamic resuscitation protocol targeting capillary refill time?

Reference:  Hernandez et al. Personalized Hemodynamic Resuscitation Targeting Capillary Refill Time in Early Septic Shock: The ANDROMEDA-SHOCK-2 Randomized Clinical Trial. JAMA. 2025 Oct

  • Population: Adults (≥18 y) with septic shock per Sepsis‑3 (vasopressors after ≥1 L IV fluid and lactate >2 mmol/L), within 4 hours of shock onset.
    • Key Exclusions: >4 h from shock onset; anticipated surgery or dialysis within 6 h; expected survival <90 days; refractory shock; DNAR; Child‑Pugh B/C; severe ARDS; active bleeding; pregnancy; inability to assess CRT (peripheral vascular disease, hypothermia, very dark skin tone, Raynaud phenomenon).
  • Intervention: A personalized hemodynamic resuscitation protocol targeting capillary refill time (CRT) using a 6-hour stepwise algorithm (see below).
  • Comparison: Usual care per local protocols/guidelines.
  • Outcome:
    • Primary Outcome: A hierarchical composite tested with a stratified win ratio of: (1) 28-day all-cause mortality, then (2) duration of vital support (time requiring cardiovascular, respiratory, or kidney support) through day 28, then (3) hospital length of stay through day 28.
    • Secondary Outcomes: Secondary outcomes were each of the three components of the primary outcome.
  • Trial: This is a pragmatic, multi-center, open-label, randomized controlled trial.

Authors’ Conclusions: “Among patients with early septic shock, a personalized hemodynamic resuscitation protocol targeting capillary refill time was superior to usual care for the primary composite outcome, primarily due to a lower duration of vital support.”

Quality Checklist for Randomized Clinical Trials:

  1. The study population included or focused on those in the emergency department. No
  2. The patients were adequately randomized. Yes
  3. The randomization process was concealed. Yes
  4. The patients were analyzed in the groups to which they were randomized. Yes
  5. The study patients were recruited consecutively (i.e. no selection bias). Yes
  6. The patients in both groups were similar with respect to prognostic factors. Yes
  7. All participants (patients, clinicians, outcome assessors) were unaware of group allocation. No
  8. All groups were treated equally except for the intervention. Unsure
  9. Follow-up was complete (i.e. at least 80% for both groups). Yes
  10. All patient-important outcomes were considered. Unsure
  11. The treatment effect was sufficiently large and precise to be clinically significant. Unsure
  12. Financial conflicts of interest. The researchers at each study site had to get their own fund, so this was truly researcher-driven. Although a few of these researchers have ties to industry, they do not seem to have any obvious financial conflicts of interest that would be relevant to this protocol.

Results: They randomized 1,501 patients and were able to analyze 1,467. The median age was ~66 years, and 43% were female. The severity of sepsis was an APACHE II of ~18 and a SOFA of 8. Sources of infection were ~48% abdominal, ~20% respiratory, and ~20% urinary. Baseline supports included invasive ventilation ~48%, norepinephrine 100% (median ~0.21 to 0.23 µg/kg/min). Lactate median ~3.6 mmol/L. Median time from shock criteria to randomization was 2 hours (IQR 1 to 3).

Key Results: Compared with usual care, CRT‑PHR produced a statistically significant advantage on the hierarchical composite, driven mainly by shorter duration of vital support, while 28‑day mortality was nearly identical between groups.

  • Primary Outcome: Based on their stratified win ratio analysis of their composite outcome, there were 131,131 wins for the CRT-PHR group and 112,787 wins for the usual care group, resulting in a win ratio of 1.16 (95% CI 1.01-1.33, p=0.04)
  • Secondary Outcomes: 
    • No statistical difference in mortality (26.5% vs 26.6%, p=0.91).
    • There was a statistical decrease of 1 day in mean vital support free days (16.5 vs 15.4). Vital sign support free days is also a composite outcome. They don’t report the specific components of this outcome, but based on the supplementary material, there doesn’t seem to be any real difference in the use of mechanical ventilation or vasopressors, and so the entire difference in this trial might come down to a 1-day difference in renal replacement therapy.
    • There was no statistical difference in hospital length of stay (15.3 vs 16.2).

1. Unmasked Trial with Subjective Outcome Changes: Open-label trials increase the risk of performance bias. Mortality is an objective outcome and was completely unchanged here. The part of the component outcome that changed was “vital support free days”. However, treatment decisions are highly subjective and can be influenced by treatments the patient has already received. Therefore, this outcome is at high risk of bias in an unmasked trial.

2. Composite Outcomes: Composite outcomes have been discussed on the SGEM several times. They are endpoints in clinical research that combine multiple individual events. In this trial, they combined all-cause mortality, vital support and length of stay into a single measure to capture the overall effect of the intervention.

Composite outcomes can be useful when individual events are too rare to provide adequate statistical power, allowing researchers to detect treatment effects with smaller sample sizes or shorter follow-up. However, composite outcomes must be interpreted cautiously because the components may vary greatly in clinical importance or frequency [1,2,3].  In addition, the trial’s reported benefit may be driven largely by less important or more common components rather than the outcomes that matter most to patients.

 In this ANDROMEDA SHOCK-2 trial, there was no 28-day mortality benefit (26.5% vs 26.6%; HR 0.99), but there was a statistical difference in fewer days of vital support. Performing a hierarchical composite analysis gave a win ratio was 1.16 (95% CI 1.02–1.33; P=0.04). This means a net advantage for the CRT‑personalized resuscitation strategy.

3. Lipstick on a Pig? One of the criticisms of composite outcomes is that each outcome is valued the same. Changing the analysis into a hierarchical model with pre-specified outcomes of importance attempts to address this limitation (death first, then major morbidity and then resource utilization).

Analysis typically compares each patient in the intervention arm with each patient in the control arm (or matched pairs) and determines a “win,” “loss,” or “tie” based on the highest‑priority endpoint on which the pair differs. Only if tied at that level does the comparison proceed to the next level in the hierarchy.

The net effect is summarized as a win ratio (or a “stratified win ratio,” if stratified by baseline risk). The number of wins for the intervention divided by the number of wins for control (values >1 favour the intervention). Foundational methods include the Finkelstein–Schoenfeld global rank (“win–loss”) test [4], Pocock’s win‑ratio framework [5], Buyse’s generalized pairwise comparisons [6], and the Wei–Lachin family of multivariate rank tests.

This definitely makes some sense, but the question is whether the methodology truly helps clinicians understand what is going on. One problem with the hierarchy is that the primary outcome is still presented with a single number and still discussed as a single outcome in their conclusion and abstract, and so the hierarchy isn’t actually represented. Instead of discussing the outcomes separately, they still seem to get lumped together, making it somewhat confusing to determine what happened to these patients. (I had to read the paper multiple times and dig into the abstract to get a sense of what was changed here.

The hierarchical methodology is not such a big problem if done well because it is a reasonable approach to improve the standard composite outcome. Hierarchies should be registered up front to avoid HARKing [7], and results for each component should be reported side‑by‑side to avoid selective emphasis.

In this and other statistical methods, we need to avoid the “spin”. The results are the results and how we interpret and communicate the clinical meaning of the results is important. I would start with patient-oriented outcomes (POO). Mortality, functional status, and major morbidity are likely what patients care about most. Composites must not hide neutral or adverse effects on top-tier outcomes.

It is important to be transparent about what drove the “win. In ANDROMEDA‑SHOCK‑2, the hierarchical primary favoured CRT‑personalized resuscitation (win ratio 1.16), driven largely by fewer days of vital support, while mortality was unchanged. This information should be presented plainly so clinicians and patients can judge whether the trade-off is worthwhile in their context.

4. Bundle of Care: Studies that look at bundles of care can be harder to interpret. Which parts of the bundle matter? Are they all equal? When lumped together, it is even possible that some parts of the bundle are hurting patients, but the harms are just outweighed by the benefits of other components. This is especially important when the bundle is very intensive and will be difficult to replicate in many clinical environments.

We have done this before in sepsis. We had the surviving sepsis bundle by Dr. Rivers. Over time, we found that none of the components of the bundle were improving patient outcomes. It seems like just extra attention was the key. We should probably learn from that history.

I think this question is especially important given that the OPTRESS trial was published while this trial was being run and showed harm from targeting higher MAPS in elderly patients with septic shock. The protocols and patients were not the same, but it is a reminder that this bundle could include components that cause harm.

5. Opportunity Cost: This intervention, as described in the manuscript, will take a lot of clinician time. Obviously, that time is worth it if we are saving lives. However, we must remember that every minute spent with this patient is a minute taken away from other undifferentiated patients in the department waiting for our services.

This is a problem seen in emergency medicine. Any time we throw resources at a problem (code MI, code stroke, code sepsis, code paronychia), we can show improvement in performance metrics. Given the questionable outcomes being changed here, we have to think very carefully about the question: Is this worth it?

Comment on the Authors’ Conclusion Compared to the SGEM Conclusion: We would quibble slightly over the language. They conclude that this personalized resuscitation protocol was superior to usual care based on their composite primary outcome, primarily due to a lower duration of vital support. Although technically true, that seems to imply that there were other benefits as well.

We would conclude that the trial found no change in mortality, but this resuscitation protocol might decrease total days of vital support, primarily in the form of renal replacement therapy.

SGEM Bottom Line:  This is interesting research, and warrants follow-up, but considering its open-label design and lack of mortality change, we do not believe that this protocol is ready for routine practice.

Case Resolution: Thankfully, you always keep a close eye on your critically ill patients, and so you can tell the intensivist that the patient’s cap refill is improving since the fluid bolus, but has not yet normalized. She asks you to start one more fluid bolus, and then she will reassess the patient clinically when they arrive in the ICU.

Dr. Justin Morgenstern

Clinical Application: This paper reminds us that septic shock patients still have a high mortality, and they may require individualized plans to deal with their shock. Although the evidence might not be strong enough to recommend this specific protocol, it is a good reminder that the answer to low blood pressure is not just reflexive IV fluids. A more holistic approach to the patient, including clinical assessment of tissue perfusion and adjuncts like echocardiography, may help guide your resuscitation.

What Do I Tell the Patient?  Most critically ill septic shock patients are not awake enough to have this conversation, but I will tell them or their family that they have a life-threatening infection, and that researchers around the world are doing their best to find ways to improve outcomes, and for now, I will just keep a very close eye on them and adjust their treatment as needed.

Keener Kontest: Last week’s winner was Dr. Olof Larén-Sandgren, an attending emergency physician Växjö, Sweden. He knew Batman’s one rule is that he doesn’t kill. 

Listen to this week’s episode to hear the keener question. If you think you know the answer, send an email to TheSGEM@gmail.com with “keener” in the subject line. The first correct answer will receive a shoutout on the next SGEM episode. 

Other FOAMed:

Remember to be skeptical of anything you learn, even if you heard it on the Skeptics’ Guide to Emergency Medicine.

References

  1. Montori VM, Permanyer-Miralda G, Ferreira-González I, Busse JW, Pacheco-Huergo V, Bryant D, Alonso J, Akl EA, Domingo-Salvany A, Mills E, Wu P, Schünemann HJ, Jaeschke R, Guyatt GH. Validity of composite end points in clinical trials. BMJ. 2005 Mar 12;330(7491):594-6. doi: 10.1136/bmj.330.7491.594. PMID: 15761002; PMCID: PMC554039.
  2. Ferreira-González I, Busse JW, Heels-Ansdell D, Montori VM, Akl EA, Bryant DM, Alonso-Coello P, Alonso J, Worster A, Upadhye S, Jaeschke R, Schünemann HJ, Permanyer-Miralda G, Pacheco-Huergo V, Domingo-Salvany A, Wu P, Mills EJ, Guyatt GH. Problems with use of composite end points in cardiovascular trials: systematic review of randomised controlled trials. BMJ. 2007 Apr 14;334(7597):786. doi: 10.1136/bmj.39136.682083.AE. Epub 2007 Apr 2. PMID: 17403713; PMCID: PMC1852019.
  3. Cordoba G, Schwartz L, Woloshin S, Bae H, Gøtzsche PC. Definition, reporting, and interpretation of composite outcomes in clinical trials: systematic review. BMJ. 2010 Aug 18;341:c3920. doi: 10.1136/bmj.c3920. PMID: 20719825; PMCID: PMC2923692.
  4. Finkelstein DM, Schoenfeld DA. Combining mortality and longitudinal measures in clinical trials. Stat Med. 1999 Jun 15;18(11):1341-54. doi: 10.1002/(sici)1097-0258(19990615)18:11<1341::aid-sim129>3.0.co;2-7. PMID: 10399200.
  5. Stuart J. Pocock, Cono A. Ariti, Timothy J. Collier, Duolao Wang, The win ratio: a new approach to the analysis of composite endpoints in clinical trials based on clinical priorities, European Heart Journal, Volume 33, Issue 2, January 2012, Pages 176-182, https://doi.org/10.1093/eurheartj/ehr352
  6. Buyse M. Generalized pairwise comparisons of prioritized outcomes in the two-sample problem. Stat Med. 2010 Dec 30;29(30):3245-57. doi: 10.1002/sim.3923. PMID: 21170918.
  7. Kerr NL. HARKing: hypothesizing after the results are known. Pers Soc Psychol Rev. 1998;2(3):196-217. doi: 10.1207/s15327957pspr0203_4. PMID: 15647155.