SPICE 3: Does Sedation with Dexmedetomidine Increase Risk of Death at 90 Days

Study Design Summary: Multicenter trial of 3904 ICU patients who were vented for less than 12 hours at time of randomization, split roughly 50/50 into control group which used other means of sedation such as propofol, benzodiazepines or other sedatives and intervention arm which receiving mainly dexmedetomidine as sedation, with primary outcome of all cause mortality at 90 days.

Notes: In the intervention group dexmedetomidine was used as the primary sedation agent, reduce agents of predefined low dose propofol or benzodiazepines in specific situations were allowed. The control group used propofol, benzodiazepines or other sedation agents and use of dexmedetomidine was discouraged but allowed as a rescue agent if needed. Antipsychotics were allowed in both groups. The intervention group had 97.7% of patients get dexmedetomidine, and 98.3% of patients in the control group got either propofol of midazolam, which is fairly good compliance overall for the size of the study

The demographics and baseline clinical characteristics of both groups at randomization was similar. This included age, APACHE II score, rates of sepsis, type of ICU admission, median RASS score and admission diagnosis.

The primary outcome of all cause mortality at 90 days was identical between the intervention vs control group (29.1% of both groups). Secondary outcomes including all cause mortality at 180 days, median days free of coma or delirium, and median ventilator free days were also not different between the groups.

The dexmedetomidine group required more rescue medication per the authors but this data is hard to discern from the paper and it is noted that both groups received multiple medications for sedation frequently. There is also a statement in the paper that there were more episodes of bradycardia and hypotension. While comparing the intervention to the control group the rates of bradycardia were 5.1% vs 0.5%, with “serious” bradycardia as 0.7% vs 0.05% for p=0.001. For hypotension rates were 2.7% vs 0.5%, and “serious” hypotension rates of 0.5% vs 0.05% for p=0.006. Asystole was 0.7% vs. 0.1%, P=0.003. However, even given these numbers the all cause mortality rate at 90 days was the same between the two groups.

“The biologic rationale for a potential benefit of dexmedetomidine is based on experimental evidence of protective effects against neuronal, myocardial, and renal injury, along with a reduction in inflammatory mediators after cardio- pulmonary bypass and reduced mortality in animal models.”

Study Conclusion: There was not an increased risk of death at 90 days in patients receiving dexmedetomidine compared to standard sedation agents, however, there were more adverse events in the dexmedetomidine group.

Fusion Beat Bottom Line Impression: This study demonstrated that dexmedetomidine does not improve 90 day mortality in patients when used as the primary sedation agent, there were increased adverse events with use of dexmedetomidine including bradycardia, hypotension and asystole, but all cause mortality at 90 days was the same between the two groups.

MIDAS: Does Midodrine Accelerate IV Vasopressor Liberation in the ICU

Study Design Summary: Multicenter trial comparing 132 total patients who were on predefined low dose vasopressor support for at least 24 hours, randomized to receive oral midodrine or not with primary end point being time in hours to vasopressor discontinuation from randomization.

Notes: The study defined successful discontinuation of vasopressors as off pressors for 24 hours. The study was overall small in number (n=132) but they prespecificed a N of 120 to detect a 6 h difference. The groups were similar in terms of demographics, indication for ICU admission and baseline MAP.

Median time to discontinuation was 23.5 hours in midodrine group vs 22.5 hours in control group. ICU discharge readiness was the same at 5 days amongst the intervention and control group. The overall number of adverse events was reported as the same between the two groups, but a table of these events was not shown. They did, however, report that midodrine vs control group there was HTN episodes 7 vs 3, bradycardia episodes 5 vs 0, and a trial fibrillation episodes 3 vs 1; of these adverse events the only that reached statical significance was bradycardia with p=0.02.

Interestingly, there was significant decrease in time to vasopressor discontinuation in patients with epidural anesthesia receiving midodrine as compared to control group. Authors offer this explanation which I think is good, “First, the mechanism of hypotension during epidural analgesia and orthostatic hypotension, for which midodrine’s efficacy has been demonstrated in randomised trials [21, 22], is identical, namely neural vasoplegia resulting from dila- tion of both resistance and capacitance vessels in the anaesthetised area [where as it is a different cause of vasoplegia in conditions such as sepsis]. Second, it is possible that epidural analgesia may improve the absorption of oral medica- tions like midodrine. Epidural analgesia has been shown to facilitate recovery of gastrointestinal function.”

Study Conclusion: Midodrine did not reduce time to vasopressor discontinuation, there were increased rates of bradycardia in the midodrine group and thus these result do not support the routine use of midodrine in critically ill patients to accelerate liberation from intravenous vasopressor support in the ICU.

Fusion Beat Bottom Line Impression: Addition of oral midodrine did not improve time to IV vasopressor discontinuation and there were slightly more adverse events in the midodrine group leading to a Fusion Beat bottom line that midodrine should not be used to accelerate vasopressor discontinuation.

WOMAN: Does TXA Reduce Mortality in Postpartum Hemorrhage

Study Design Summary: Multicenter, placebo controlled trial of adults with postpartum hemorrhage in either vaginal or cesarian section delivery with initial primary end point as composite outcome of death and need for hysterectomy, but due to well explained reasons they analyzed death secondary to bleeding as well which seems a better endpoint.

Notes: The TXA and placebo groups were identical interns of average age, type of delivery, time from delivery to randomization, primary cause of hemorrhage, blood pressure, signs of hemodynamic instability and EBL.

FOAM points: SGEM:The study was funded in part by Pfizer the maker of TXA. This should make us more skeptical of the results and how the results will be marketed by the sponsor. REBELEM: All cause mortality was unchanged (2.3% vs 2.6%, p=0.16). The NNT was large (267) with a fragility index of 0.

There is some argument that disease specific mortality is a poor outcome because it can be hard to decipher what exactly patients died of and there can be large inter-evaluator reliability. There is also the argument that patients don’t care what they died of just if they died or not. I take issue with the second argument because I’m giving TXA to stop bleeding, not prevent sepsis.

The initial primary outcome was composite of death and hysterectomy but often patients were randomized to the trial and undergoing emergent hysterectomy at the same time thus prior to any results available to the study designers they changed primary outcome top death secondary to bleeding. I think this is likely a better outcome as it is simpler with less background noise.

Death due to bleeding in the TXA group vs placebo was 1.2% vs 1.7% (CI 0.53-0.90 favoring TXA with p=0.008) when given in the first three hours and 2.6% vs 2.5% (CI 0.76-1.51) when given after 3 hours. TXA in the first 3 hours also reduced number of laparotomies to control bleeding with statistically significant numbers. This is also an important outcome that is not discussed in many of the negative FOAM posts. Risk of thromboembolic events, both venous and arterial, were similar between the two groups.

I think the study does show positive results on the results that matter but I am skeptical to trust a study funded by the drug company.

Study Conclusion: TXA reduces mortality in patients with postpartum hemorrhage when given within the first 3 hours, without a significant increase in VTE; consistent with CRASH-2 trial results.

Fusion Beat Bottom Line Impression: TXA should be given to patients within 3 hours of delivery with postpartum hemorrhage to significantly reduce risk of death.

HALT-IT: Does TXA Reduce Death in GI Bleeds

Study Design Summary: Multicenter, international placebo controlled trial of roughly 12,000 patients who were given TXA, primary outcome of death due to bleeding at 5 days.

Notes: They used a slightly higher maintenance dose of (3g vs the usual 1g over 8 hours) in patients who were deemed to have a significant risk of death from wither upper or lower GIB. Groups were well balanced in respect to average age, location of bleed (upper vs lower), blood pressure, signs of shock and comorbidities.

Primary outcome was death due to bleeding at 5 days, but they also collected data on death due to bleeding at 24 hours and 28 days as well as rebreeding at 24 hours, 5 days and 28 days. The TXA and placebo groups were identical in all aspects of these outcomes.

The rate of thromboembolic events was roughly similar to previous studies, arterial thromboembolic events such as MI or stroke were the same between TXA and placebo group in this study. The TXA group had slightly higher venous thromboembolic events in this study but it was not statistically significant. This trend was demonstrated in some previous studies but never reached statistical significance in any of them. The numbers were slightly higher in this TXA group and the authors offer this explanation, “Recent research shows that acutely ill patients with cirrhosis have a mixed fibrinolytic phenotype.17 Some have increased fibrinolysis, but others have profound hypo­ fibrinolysis. The prevalence of hypofibrinolysis appears to be greatest in the most critically ill patients. Using the same clot lysis assay, reduced fibrinolysis has been shown to be associated with a small increased risk of venous thrombosis.18 In our trial, the increased risk of venous thromboembolic events with tranexamic acid appeared to be more marked in patients with liver disease, although this was an exploratory subgroup analysis and there was no strong evidence for hetero­geneity.”

“The dose of tranexamic acid used in this trial was higher and the duration of treatment was longer (4 g over 24 h) than in randomised trials of tranexamic acid in trauma (2 g over 8 h) or post­partum haemorrhage (1 g bolus with a repeat 1 g dose if bleeding continued), which did not record any increase in adverse events with tranexamic acid. Patients with gastrointestinal bleeding often rebleed after initial haemostasis, particularly within the first 24 h. Because tranexamic acid has a short half­life, we used a longer treatment duration to cover this high­risk period. Furthermore, previous trials in gastrointestinal bleeding that appeared to show a large mortality reduction with tranexamic acid used a high dose and a longer duration of treatment than trials in trauma and post­partum hemorrhage.”

Previous studies have shown that TXA is more effective if given within the first 3 hours since injury. The authors do a nice job of addressing this physiology question as to why this study did not show benefit to TXA administration, “the timing of onset is easy to determine, most patients present early, and there are well documented changes in fibrinolysis that provide a biological rationale for tranexamic acid treatment.15,16 However, in gastro­ intestinal bleeding it is difficult to determine the time of bleeding onset, presentation is often delayed (over 80% of patients presented more than 3 h after bleeding onset), and the contribution of increased fibrinolysis to bleeding is less clear.”

This was a well done study with sound logic as to why they chose a higher dose TXA. They didn’t spin their results to make it a positive study and offer sound logic as to why TXA is likely both ineffective in reducing GIB mortality and why VTE is slightly higher in TXA group. I believe the results are reliable and the you can rest your hat on these results.

Study Conclusion: TXA does not reduce the risk of death due to bleeding in GI bleed patients and should not be used outside of randomized controlled trials.

Fusion Beat Bottom Line Impression: The study demonstrates no decreased mortality to patients given TXA for GIB, and risk of VTE was slightly higher in TXA group but not statistically signifiant.

CRUSADE: Does Morphine Affect Mortality in Acute Coronary Syndrome

Study Design Summary: Multicenter observational retrospective study of NSTEMI and unstable angina patients in US split into morphine vs no morphine and morphine vs nitroglycerin groups.

Notes: Morphine vs no morphine: More patients in the morphine groups had death, MI, CHF and cariogenic shock events on admission. The groups were fairly unbalanced with the morphine group being the sicker group, but the no morphine group had larger percentage pf patients with history of CHF.

Morphine vs nitro: More death, MI, CHF, and cariogenic shock in the morphine group. The groups were fairly unbalanced with the nitroglycerin group being overall sicker, however more patients in the morphine and nitroglycerin group had history of CHF.

Quote from discussion section of paper: “In one study patients with CAD awaiting bypass surgery, morphine caused 13% reduction in coronary blood flow. In fact, in animal studies, morphine has been demonstrated quite conclusively to actually increase myocardial infarction size.”

Study Conclusion: Use of morphine either alone or in combination with nitroglycerin for patients presenting with NSTEMI and unstable angina was associated with higher mortality.

Fusion Beat Bottom Line Impression: The study demonstrates an association between morphine and mortality in patients with NSTEMI and unstable angina and needs to be evaluated further for causality.

CRASH 2: Does TXA Reduce All Cause Mortality in Trauma Patients

Study Design Summary: Multicenter study examining TAX vs placebo in 4 week all cause mortality in adult trauma patients of both blunt and penetrating trauma. Had some interesting secondary endpoints of need for surgery and blood transfusion units given.

Notes: Fairly large study of roughly 20,000 patients, equally balanced TXA and placebo groups. The primary end point of all cause mortality at 4 weeks was significantly lower in the TXA group (14.5% vs 16.0%), with no symptoms higher in the TXA group (14.7% vs 13.3%). There was similar risk of VTE in TXA vs placebo (1.7% vs 2.0%). The median amount of blood product transfused between the two groups was equal at 3 units.

TXA given after 3 hours didn’t reduce mortality compared to placebo and thus the 3 hour mark is the current inflection point for TXA administration.

Study Conclusion: TXA reduced risk of death in trauma patients.

Fusion Beat Bottom Line Impression: TXA should be given to bleeding trauma patients within 3 hours of injury to significantly reduce risk of death.

CRASH 3: Does TXA Reduce Head Injury Related Deaths in TBI Patients

Study Design Summary: Multinational randomized, placebo controlled study that compared giving TXA vs placebo to patients with TBI and no extracranial bleeding, primary outcome was brain injury related death at 28 days.

Notes: This study has many signs pointing towards a biased study. They have many secondary endpoints and changed their definitions of inclusion criteria (8 hours vs 3 hours from injury) and yet were still underpowered. After the study they pushed propaganda hard to make TXA standard of care for TBI patients with misleading information on videos and posters. The listed primary outcome on clinicaltrials.gov was all cause mortality which is also different from the paper.

The TXA vs placebo groups were pretty balanced in both prestudy points such as age, sex, time since injury, GCS. The TXA group had somewhat less VTE (1.8% vs 2.4%), DVT was equal in both groups, PE slightly higher in placebo group (0.3% vs 0.8%). There were very few patients lost to followup in either group, and this study is the largest RCT to date to look at TXA in TBI.

Overall the head injury related deaths in TXA vs placebo group was 18.5% vs 19.8% (CI 0.86-1.02) thus making the trial a negative trial by their pre-definition. A subanalysis of the data excluding pts with GCS of 3 or bilateral unreative pupils showed outcome was 12.5% vs 14.0% (0.80-1.00). Still a relatively negative study. Keep in mind the overall study was underpowered and thus the subanlaysis is further under powered.

The study broke down head injury related deaths based on GCS as another subanalysis. Patients with a GCS 3-8 had no difference in TXA vs placebo, patients with GCS 9-15 had difference favoring TXA (5.8% head injury related mortality in TXA group vs 7.5% in placebo group). Again, as the whole study was underpowered, these sub groups are further underpowered. They did not include a subanalysis of patients with GCS 8-12 which would have been interesting to see.

They compared TXA vs placebo group survivors in Disability Rating Scale and there was no difference.

There were less head injury related deaths in the TXA group at 24 hours, but when the endpoint was expanded to 28 days this effect was muted, some suggest this is due to increase in non-bleeding related causes of head injury related deaths. There was a subanalysis that compared TXA vs placebo in patients treated <1h, 1-3h, or >3 hours from injury and there were no differences observed. Some suggest that patients treated soon after injury likely had worse injuries and thus confounding the results.

I would be extremely weary of interpreting data from a study that had so many changes in its endpoints and after a negative study pushed propaganda hard that the study was positive.

Unpublished data apparently showed overall mortality benefit in TXA group but hard to evaluate since it was unpublished.

Study Conclusion: TXA given within 3 hours of injury reduces head injury related deaths in patients with TBI and no extracranial bleeding.

Fusion Beat Bottom Line Impression: While the study is underpowered, it demonstrated low side effect profile of TXA when considering VTE. In patients with GCS 9-15 who are less than 3 hours from injury, TXA may provide some benefit with reduction of head injury related deaths. No true definitive conclusion can be drawn as the resulting positive benefit was seen in an underpowered subgroup analysis but is worth noting as potential generator of further directed research. Giving TXA to patients in this select group that showed benefit is likely worthwhile but should not be tying up cognitive bandwidth early on in the resuscitation.

PRORATA: Does Procalcitonin Reduce Antibiotic Exposure and Mortality

Study Design Summary: Multicenter study of 630 patients split evenly into intervention and control groups, with suspected bacterial infections, looking at 28 and 60 day mortality (non-inferiority analysis) as well as antibiotic exposure between the two groups (superiority analysis).

Notes: Patients were eligible if they were not receiving antibiotics prior to enrollment or receiving antibiotics for less than 24 hours prior to enrollment if they were included in the study less than 12 hours after admission. Prespecified algorithms advised clinicians on whether to start antibiotics and when they should be stopped. In patients who got antibiotics, procalcitonin levels were assessed daily.

Groups were well balanced in all aspects including age, SOFA score, biomarker levels such as procalcitonin and lactate at time of enrollment, infection site. Both groups had equal numbers of documented versus suspected infections as well as appropriate antimicrobial coverage.

Both groups had similar 28 day mortality, higher mortality in the procalcitonin group at 60 days (30% vs 26%), but “no patient in either group who died during days 29–60 had an infection relapse, and most deaths resulted from complications directly related to the severity of underlying disease.” Procalcitonin group had roughly 3 more days without antibiotics. There was a silightly higher relapse and superinfection rate in the procalcitonin group, with LOS in ICU and hospital same for both groups. There was not a between-group difference for the rates of emerging multidrug-resistant bacteria.

There was a significant amount of protocol deviation in both groups which makes the data hard to interpret. “53% of patients randomised to the procalcitonin group were not given algorithm-guided treatment.” Procalcitonin cutoffs between studies has been highly variable thus making interpretion of the studies together also difficult.

Study Conclusion: Procalcitonin- guided antibiotic treatment substantially lowers antibiotic exposure and is non-inferior to standard care with respect to outcomes.

Fusion Beat Bottom Line Impression: The high amount of protocol deviation makes the data very difficult to interpret and I do not think this study provides practice changing evidence that procalcitonin guided strategies reduce antibiotic exposure or affect mortality in either way. Also given the fact that studies are heterogeneous in their procalcitonin cutoffs, it makes the overall data difficult to interpret and thus not practice changing. However, I do think that adding in procalcitonin trends as another data point to evaluate whether a patient needs to start/stop antibiotics can likely be helpful.

While some may argue that the antibiotics exposure was reduced but overall mortality and resistant organism rates were not changed, with increased relapse and superinfection rates in the procalcitonin group, adding in a biomarker such as procalcitonin as part of the whole picture and reducing time on antibiotics is helpful in and of itself as prolonged exposures to unnecessary antibiotics portends itself to increase risk of unwanted sided effects.

HEAT: Does Fever Reduction Lead to More ICU Free Days

Study Design Summary: Multicenter study in which 700 patients were split to either get 1g IV acetaminophen or not; to be eligible patients had to have a documented temperature ≥38°C with known or suspected infection. Primary outcome was ICU free days from randomization to day 28. End points were discharge from ICU, death, cessation of antimicrobial therapy or defervescence.

Notes: 30% of both the intervention and control group got open label acetaminophen, but physical cooling was given as rescue therapy if temperature rose above 39.5°C, rates of physical rescue cooling and NSAID use was similar in both groups. Tylenol only decreased peak temperatures by 0.25-0.4°C, which is constant with previous studies. The number of ICU free days did not differ significantly between the groups, neither did mortality at 28 or 90 days, ICU length of stay. There were higher rates of liver dysfunction in the placebo group actually.

There was shorter ICU LOS in acetaminophen group amongst survivors (3.5 vs 4.3 days) and longer LOS in acetaminophen group amongst the non-survivors (10.4 vs 4 days). Authors comment on this, “Our observation that ICU and hospital length of stay were longer with acetaminophen than with placebo among patients who died is consistent with the finding of a study in which physi- cal cooling to normothermia delayed death in mechanically ventilated patients with septic shock.4 These observations are also consistent with a recent retrospective cohort study in which a Cox proportional-hazards analysis showed that ICU patients who received acetaminophen had a significantly longer time to death than those who did not.”

The study outcomes table showed exactly no difference in the primary outcome or any of the secondary outcomes and 28 and 90 day death rates between the two groups.

Study Conclusion: Our findings suggest that acetaminophen has a modest clinical effect as an antipyretic in ICU patients with fever and probable infection but does not reduce ICU-free days in these patients. Early administration of acetaminophen to treat fever due to probable infection did not affect the number of ICU-free days. There was no significant between-group differ- ence in 28-day mortality, 90-day mortality, or survival time to day 90.

Fusion Beat Bottom Line Impression: Tylenol for fever offers no benefit in the ICU. The observation of length of stay prolongation in non-survivors would need to be addressed specifically in a separate study that also takes into account time on antimicrobial therapy, which this study did not do, to provide specific recommendations.

I do think this observation of prolonged time to death is important and shouldn’t be overlooked but without data on time on antimicrobials it is a hypothesis generating observation only.

This study used 39.5°C as cut off for beginning rescue therapies which is appropriate based on previous studies, above this there is known deleterious effects of the fever.

PROSEVA: Does Prone Positioning Improve Survival in ARDS

Study Design Summary: 466 patients with severe ARDS were split roughly 50/50 into prone versus stand care groups. Prone group spent at least 16 consecutive hours per day in the prone position. To be included patients needed to be intubated and mechanically vented for ARDS for less than 36 hours. Primary end point was 28 day all cause mortality.

Notes: Severe ARDS here is P:F <150 mm Hg. The most common cause of ARDS in this study was pneumonia, which is probably consistent with what most ICUs see. More patients in the supine group needed rescue therapies; the groups were very well balanced and equal amongst factors such as pH, PPlat, TV, PEEP, PaO2, PaCO2, FiO2, SOFA score, coexisting conditions and such. Both groups received low-tidal ventilation and therapeutic paralysis. The prone group had more adverse events such as unexpected extubation but this was a smaller difference than the survival benefit conferred.

The prone group had a decreased rate of mortality at both 28 days and 90 days as well as shorter time to successful extubation and greater proportion of patients extubated at 90 days. The prone group had shorter ICU stays and more ventilator free days.

All patients in the study had standard ICU beds with proning protocol, no proning beds.

Study Conclusion: Prone positioning for at least 16 consecutive hours per day reduces 28 day all cause mortality in patients with severe ARDS.

Fusion Beat Bottom Line Impression: Prone positioning provides mortality, LOS and extubation benefit to patients with severe ARDS (P:F <150mmHg).

REALITY-AHF: Does Early Lasix Improve Mortality in Acute Heart Failure

Study Design Summary: Observational multicenter study of patients presenting to ED for acute heart failure. No protocol on which medications/treatment to give or timing of medications since it was observational. Primary outcome was all cause in-hospital mortality. 1,291 total patients, 481 in early treatment group, 810 in nonearly group. Early treatment was defined as lasix within 60 minutes of ED arrival, nonearly was defined as >60 minutes but less than 24 hours. Median time overall was 90 minutes. for all 1,291 patients.

Notes: The early treatment group were more likely to arrive by ambulance, onset of symptoms <6 hours prior to arrival, higher HR and BP, as well as more peripheral and pulmonary edema, JVD, orthopnea by a large margin. There was decreased in-hospital mortality in the early treatment group. This maybe was because the disease was recognized earlier and thus treated earlier. Similar to the increased mortality rate of low symptomatic brain mets.

There was also an inflection point in the mortality of the groups. If lasix was given within the first hour there was an association with decreased mortality. Patients given lasix just after 60 minutes had increased mortality which then dropped off quickly, but slowly increased back again the further out from 1 hour from admission they got. This phenomenon may point more towards a disease process driving mortality rather than the lasix since the body and the drug do not respect time in that fashion.

The early treatment group had worse symptoms but improved mortality. Similar NYHA III and IV between the two groups. More patients in non-sinus rhythm in nonearly treatment group.

Hard to attribute in-hospital mortality to lasix within the first 60 minutes of hospital presentation based on face validity. PPV and inodilators would have more face validity as patients who need these are likely on the cusp of death, but as long as ABCs/pulm edema controlled, lasix to continue the fluid management after initial fluid management is important but maybe not in-hospital mortality level important.

The study mentioned that more patients had history of HF in nonearly group but then the tables seemed to contradict that later. So unsure how to interpret that.

Study Conclusion: IV lasix given within 60 minutes of hospital arrival was independently associated with better in-hospital mortality.

Fusion Beat Bottom Line Impression: This observational study demonstrates an association between IV lasix within 60 minutes and improved in-hospital morality in AHF patients, however, given the differences in group presentations and gut feeling/face validity, it seems more of this survival benefit was potentially due to early recognition of disease. No causation of early lasix with improved mortality can be demonstrated by this study.