There were a handful of big name critical care studies published the last few weeks (LOGICAL, ARISE FLUIDS, SODa-BIC to name a few) but many of you probably already saw those so we're gonna stick with something a little more off the beaten path.

A retrospective observation trial was published recently by Gastanadui et al in Journal of Intensive Care looking at the mortality impact of corticosteroids in cardiogenic shock.  They looked at 167,721 patients from the Vizient database admitted with cardiogenic shock (excluded patients with other indications for steroids).  The unadjusted mortality was a whopping 48.8% in the steroid group and 29.6% in the non-steroid group.  They did Inverse Probability Treatment Weighting (IPTW), which attempts (imperfectly, but best as you can retrospectively) to control for treatment selection bias, and still found a roughly 3% mortality increase in the steroid group.

Bottom Line: Corticosteroids were already controversial (at best) in cardiogenic shock, but this further adds to the reasons to avoid them in this condition.  Of note, this trial excluded patients with classic hard-indications (e.g. adrenal insufficiency, COPD exacerbation, etc) and steroids should still be given to those groups, but in my opinion if you feel cardiogenic is the primary driver of the patient's shock and they do not have another indication, I would consider avoiding steroids solely for the shock.

Show Additional Information

Show References

PaO2 to FiO2 (P:F) ratios, are often considered the gold standard in critical care for assessing the degree of oxygen-refractory hypoxia in various pathologies, particularly ARDS.  P:F does have some limitations, including not accounting for the PEEP, but probably the most limiting is that it requires collecting an ABG, which is invasive and not always feasible or a top priority when resuscitating a critically ill hypoxic patient.  On the other hand, SpO2 (pulse ox saturation) is routinely available, and of course the FiO2 should be known, so many have suggested perhaps using an SpO2 to FiO2 (S:F) ratio instead.  But how S:F maps to P:F and how well they correlate is not fully known.  Chaudhuri et al recently conducted a meta-analysis, published in Critical Care Medicine this month, which reviewed the literature on this. 

Bottom Line: Yes, S:F ratios correlate well with P:F ratios, especially when the SpO2 is less than 97%, but you can't just substitute the S:F for P:F, you have to use one of the accepted formulas.  See additional info on the website for the actual formula to apply and how a given S:F translates to P:F.

Show Additional Information

Show References

It is a common scenario in the ICU, and occasionally in the ED, to be asked which pressor you would like to wean first, norepinephrine or vasopressin.  This is mostly an “art not science” question, but is there a right answer?  Does picking one vs the other to wean first lead to less hypotension?

Bottom Line: This meta-analysis doesn't suggest that either the norepi-first or vasopressin-first strategies for vasopressor wean are associated with an increased incidence of hypotension, although the literature is mixed.  Whatever your current practice is, it's probably reasonable to stick with that.  See the additional information for my personal approach.

Show Additional Information

Show References

Perintubation hypotension is a major problem, and can precipitate hemodynamic collapse and cardiac arrest for a multitude of reasons.  To prevent this, many different strategies have been explored (some of which work and some of which don't), including empiric IV fluid boluses, additional resuscitation before intubation, switching or dose-reducing induction agents and much more.  But we know pressors  like norepinephrine raise blood pressure effectively, so should we just put everybody on a norepinephrine drip before we intubate them?

Probably not.  The EPITUBE trial included 210 patients at a single-institution undergoing cardiac surgery, and randomized them to empirically starting a norepinephrine infusion before induction vs just rescue ephedrine when needed (fairly standard anesthesia practice).  For the empiric norepinephrine group, they started at 0.06 ug/kg/min, and once the drip was up and running, they titrated for a MAP of 65-80 (which could include stopping the norepi if that the patient remained above 80 despite downtitration)

The incidence of severe hypotension (MAP < 55) did not differ between the groups, although fewer empiric norepinephrine patients had a MAP < 65 at any point (which was a secondary outcome).  Naturally, the differences between this practice setting (the cardiac surgery OR) and the emergency department should be noted and are not addressed by this study.

Bottom line: There isn't good evidence to support empirically starting all patients on a norepinephrine infusion prior to intubation as a method to prevent perintubation hypotension.  You should always have rapid access to vasopressors when intubating, and should continue to tailor your therapy to the individual patient, but probably don't start just putting everyone on norepinephrine before you intubate them.

Show References

Another month, another study of hemodynamic targets in sepsis…  The age-old questions: is a MAP > 65 a good target for everybody, or should we individualize?  Should we just give a bolus of fluids to everyone and then move to pressors, or should this strategy change patient to patient?  

Huet et al have a preprint that'll appear in Intensive Care Medicine looking at this question in 517 patients.  I can't reprint it here due to copyright (follow link below, go to full PDF and scroll to figures at bottom if curious), but basically their algorithm was 1) check if patient is fluid responsive via either echo or swan, 2) give fluid if yes, 3) do something else (pressors) if no.  

Importantly the differences were not statistically significant, but they found a strong, nearly significant, trend towards benefit on SOFA score, ICU and hospital LOS in the “personalized therapy” group (also of note, these are dubious as patient oriented outcomes).  The sickest patients (by SOFA) showed the most benefit.

Bottom Line: The “personalized hemodynamic therapy” literature continues to show a modest benefit of using tools like echo (e.g. LVOT VTI) to determine if the patient is fluid responsive (or fluid tolerant) and NOT give fluid (instead using pressors) if that is not the case, but for now there's relatively limited support for hyper-personalized approaches like varying MAP goals or otherwise mixing up your strategy.  Some day we'll likely find a more nuanced approach, but for now I think a reasonable strategy in critically ill septic patients is to use ultrasound to determine if the patient needs fluid, if yes give fluid and reassess, and if not move to pressors, to maintain a MAP > 65.

Show References

We've got supine positioning and prone positioning... what about something in-between?  Ye et al studied 2,159 patients coming out of anesthesia in a PACU after extubation.  As sedation wore off, they placed one group in lateral decubitus, and left the other group supine.  The lateral decubitus group had less hypoxia, a higher lowest SpO2, and required fewer airway rescue maneuvers.  

Of note, the investigators didn't compare lateral or supine to prone positioning, which is often felt to be the best position for oxygenation (depending on patient characteristics and pathophysiology).  And of course, this study represents a very specific scenario quite different from the ED (PACU patients post-extubation), so it's not clear how broadly extrapolatable this is.  But this does add to the argument that supine is a poor position for oxygenating patients.  

Bottom Line: If your supine patient is oxygenating marginally and you want a small bump without going all the way to prone positioning, consider lateral positioning.  May make the most sense for procedural sedation and post-extubation patients in terms of similarity to this particular study.

Show References

Following up Dr. Flint's pearl from the other day, the largest study to date looking at a lower Mean Arterial Pressure (MAP) target in elderly ICU patients is the “65” Trial, published in JAMA in 2020.  This trial compared a MAP target of 60-65 to the usual goal of >65, in critically ill patients age 65 and older.  It included 2,455 patients in 65 ICUs in the UK, and found no difference between the groups.  

Bottom Line: Although most intensivists still target a MAP > 65 regardless of patient age, you do have some evidence to support you if you want to target 60-65 in patients over age sixty-five.  However, there are some important limitations (well outlined in the PulmCrit article linked below), and therapy should always be optimized to the patient and markers of end organ perfusion.

Show References

Duan et al recently published in Intensive Care Medicine the results of a trial looking at a PEEP of 5 cm H2O vs 10 cm H2O and impact on failure rate (progression to intubation) when using non-invasive ventilation (NIV).  In their trial, the high PEEP group had a lower rate of intubation (32% vs 43%), and this was statistically significant.  It is important to note that they excluded patients whose indication for NIV was heart failure, asthma, or COPD exacerbation.

Ultimately, how to choose the right PEEP is a very complex question and requires tailoring to your patient's physiology and clinical circumstances.  For example, hypercarbic patients may benefit more from a maximization of their driving pressure (Pplat - PEEP), which can involve lowering their PEEPs, especially when trying to avoid gastric insufflation (remember, pressures of 30 cm of H2O or higher are very likely to open the LES).

Bottom Line: PEEP and other vent settings should be tailored to the patient's pathophysiology, but this trial suggests that in hypoxemic patients not getting NIV for heart failure, asthma, or COPD exacerbation, a higher PEEP (10 vs 5) may reduce the risk of intubation.

Show References

Intubation and mechanical ventilation of brain injured patients, which is extremely common in the Emergency Department, can be very challenging and subject to significant practice variation.  It is often said that brain injured patients “can't take a joke”, meaning that they are less tolerant to hemodynamic and metabolic perturbations, and these perturbations tend to be associated with very large swings in their clinical outcomes.  For example, hypo/hyperglycemia, hypo/hypernatremia, hypo/hypertension, hypo/hyperoxia, hypo/hypercapnea, etc are all extremely important to avoid.  This is probably the one patient population where “euboxia” (the notion that we obsess too much about making all the numbers pretty in the EMR) is probably not as applicable.  As such, there is at least good physiologic rationale, and now increasing empirical evidence, that ventilating these patients very thoughtfully is extremely important and likely to have meaningful impact on patient-oriented outcomes (mortality, neurologic outcome, etc).

The VENTIBRAIN study was a prospective observation trial of 2,095 intubated patients in 26 countries who had TBI, ICH (including SAH), or acute ischemic stroke.  Interestingly, they found that patients with lower tidal volume (TV) per predicted body weight had higher mortality (although the majority of their TVs were well controlled and in a fairly tight range), which is contrary to conventional thinking in pulmonary pathologies like ARDS.  They also found that higher driving pressure (DP) was associated with higher mortality, which agrees with data from other conditions.  PEEP and FiO2 had U-shaped curves, but FiO2 in particular tended to favor lower FIO2, also similar to current thinking for ICU patients in general.  

Take Home Points:

1. Although most brain injury patients have relatively normal pulmonary function, lung compliance, ventilator waveforms, etc, their ventilatory parameters (TV, PEEP, DP, pCO2/pH, oxygenation, etc) should be carefully monitored and a deliberate strategy to manage these parameters is essential.  Haphazard ventilatory strategies in these patients are clearly associated with poorer patient-oriented outcomes.
2. It's possible (although not definitively proven) that aggressively low TVs in these patients may lead to hypercapnea - which we know is poorly tolerated in brain injured patients - and worse outcomes.  The role of classic “permissive hypercapnea” (ala ARDS management, goal pH > 7.2) in these patients is unclear, and one should probably be more judicious in letting these patients get overly acidotic or hypercapneic, as opposed to other pathologies like ARDS where this is probably more allowable.  
3. Despite the paradoxical finding with low TVs, high driving pressure remains an important predictor of mortality in essentially all critical patient populations.   Care should be taken to minimize DP (guidelines say < 15 cm H2O, but goal should be minimum achievable value while meeting pCO2/pH targets).  DP/PEEP titrations should be carried out regularly when feasible (not all providers are comfortable with this practice, but it is safe and easy to learn, see references below).
4. Hypoxia and hyperoxia are both extremely dangerous for this population.  The minimum FiO2 needed to achieve a pulse oximetry reading of around 90-96% (exact numbers vary slightly by guideline and any underlying pulmonary pathology) should be used.  Be very wary of the pulse ox sitting constantly at 100% in these patients.

Show References

Post-Intensive Care Syndrome (PICS) is an increasingly recognized phenomenon of impairment of physical, cognitive, and/or mental health after intensive care admission.  Even more recently, similar deficits in caregivers of patients admitted to the ICU, often called Post-Intensive Care Syndrome Family (PICS-F) is increasingly recognized.  A study recently published by Watland et al in Critical Care Medicine looking at reducing PICS-F through a “caregiver pathway” got me wondering if there's any literature out there about reducing PICS-F via interventions in the emergency department.  Patients' treatment course in the ED is a highly stressful and uncertain time for both the patient and family members, so it stands to reason this is an impactful period where intervention may help, and even in patients where their condition is too advanced for us to make a medical difference, our actions could have a positive impact on long term outcomes for the family members.

The short answer is no, to this author's knowledge and based on my review of the literature, there is no good evidence for reducing PICS-F by ED interventions (hint, hint: if anyone's looking for a good area to study…)  Based on evidence from the critical care realm, the following are probably reasonable approaches that would translate well to the ED:

1. Recognize, especially when you have a patient who likely has a very poor prognosis, that for our critical patients it is important to treat the family, as well as the patient.  
2. Update the family early and often.  Uncertainty is a key contributor to PICS-F.  
3. Consider developing a brochure for family of critically ill patients at your facility.  Basic information such as where to park, how to get into the hospital, where their loved one may go after the ED, where they can get food, what visiting hours are allowed, whom to contact with questions, etc seem exceptionally simple to us but are often early points of stress for family.  
4. Consider screening family members for PICS-F (probably better left to the ICU, but could be considered for longer ED stays or if patient prognosis is extremely poor).  There are multiple validated screening tools available.
5. Consider encouraging patient (if they are able) or family to keep a diary.  ICU diaries have been shown to decrease incidence of both PICS and PICS-F.  See also icu-diary.org
6. If feasible, consider follow up with family members at high risk of PICS-F.  Could be done as a joint venture between the ED and inpatient services or as a hospital-wide initiative.  
7. Engage ancillary services such as pastoral care, palliative care, integrative medicine, and others early and often to foster a multi-disciplinary approach.  Also, make sure to communicate well with your nursing team, who are at the bedside and often more in tune with family signs of future PICS-F.

Show References

In out of hospital cardiac arrest (OHCA), does it matter if you choose an intraosseous (IO) vs intravenous (IV) approach to getting access and giving meds?

No, according to a recent study by Couper et al, just published in NEJM.  No significant difference in any clinically meaningful outcome including survival, neurologically intact discharge, etc.  Technically the IV group had slightly higher rates of ROSC, which just met statistical significance, and to be fair that group did trend very slightly towards better outcomes in some categories, but really well within the range expected by statistical noise.  

Interestingly, the median time from EMS arrival to access being established was the same in both groups (12 minutes), which I think raises some face validity questions.  Furthermore, of course, previous trials have raised questions as to whether ACLS meds even work or impact outcomes anyways, so naturally if they don't, the method by which they are given isn't likely to matter either.

Bottom Line: This large, well conducted trial continues to support the notion that either an IV-focused, or IO-focused approach to access and medication delivery in OHCA is reasonable.  You and your prehospital colleagues can likely continue to make this decision based on personal comfort, local protocols, and patient/case circumstances.  At the very least, this continues to support the notion that if an IV is proving challenging, pursuing an IO instead is a very appropriate thing to do.

Show References

Airway management in the pre-hospital setting is a matter of much controversy, and overall I will defer to my EMS colleagues, but several previous studies have failed to show a benefit to endotracheal intubation in the field as opposed to alternate approaches like a supraglottic airway.  Another nod in this direction has recently come out, with Battaglini et al performing a post-hoc analysis of one of the larger studies in the history of cardiac arrest, TTM-2, looking specifically at outcomes stratified by pre-hospital airway management strategy.  

Do patients who undergo endotracheal intubation in the field do better than those who get a supraglottic airway?

No, they don't.  TTM-2 included 1900 patients, of whom 1702 had enough data to be included in this re-analysis.  28% got supraglottic airways, and 72% got endotracheal intubation.  The groups were reasonably well matched on most characteristics, and if anything most well-known prognostic factors favored the endotracheal intubation group (very slightly).  It should be noted that several outcome metrics, including modified Rankin scale, did show slight signs of benefit for the endotracheal intubation group, even sometimes in a statistically significant fashion, but fell out when a multi-regression analysis, which was the primary endpoint, was done.  

Bottom Line: In pre-hospital cardiac arrest, there remains limited data to support the notion that endotracheal intubation results in better outcomes than supraglottic airway placement.  You should defer to your local protocols and continue to work with your paramedics and EMS directors as evidence continues to evolve.  For now, I don't think there's sufficient data to suggest that a given patient should be intubated vs undergoing supraglottic airway placement, and it is probably best to defer to the judgement, training, and protocols of your folks on scene.

Show References

The CLOVERS trial (NEJM 2023) examined one of the eternal questions of critical care, liberal vs restrictive fluid management in sepsis… and found no difference.  But there are criticism of CLOVERS, and while some other trials agreed with this result, there are also signals in the literature that restrictive fluid strategies are beneficial.  Furthermore, we know that these trials suffer from issues of  heterogeneity, and often lump together very different patients.

Jorda et al recently published in Critical Care a posthoc re-analysis of CLOVERS looking specifically at patients with advanced CKD (eGFR < 30).  This is a challenging group of patients to manage.  On the one hand their renal function is already marginal, so the last thing we want to do is potentially deprive starved kidneys of necessary intravascular volume, but on the flip side their septic shock puts them at high risk of full blown renal failure (transient or permanent) and they're thus at very high risk of fluid overload with aggressive resuscitative fluids and potentially limited ability to clear those fluids renally in the next few days.  So how did these patients do in CLOVERS?

They did significantly better with the restrictive fluid strategy (mortality 22% vs 39%, HR CI 0.29-0.85).  They also had more pressor free days and vent free days.  

Bottom Line (my opinion): While a restrictive vs liberal fluid strategy in septic shock remains a bit up for debate, the evidence continues to slowly tip towards restrictive fluids (i.e. earlier pressors) as the preferred approach.  In patients with advanced CKD (eGFR < 30), there is probably now sufficient evidence to favor vasopressors over IV fluid administration when resuscitating septic shock.

Show References

Magnesium is known to relax smooth muscles.  Interestingly, there is also some literature using it as part of Rapid Sequence Intubation (RSI) pre-treatment in general, in hopes that this or other mechanisms might allow it to improve intubating conditions.  Zouche et al recently published an RCT looking at giving IV magnesium as part of RSI pretreatment in cases where neuromuscular blockade (NMB) is not going to be given (e.g. scenarios where it is contraindicated).  IV Magnesium Sulfate, 50 mg/kg in 100 mL of saline given 15 minutes before induction, significantly improved intubating conditions in those getting sedation but not NMB (95% vs 39%).  

In 2013, Park et al did an RCT giving magnesium to all RSIs, even with the use of rocuronium in those patients, arguing that magnesium is also known to potentiate the effects of non-depolarizing NMB agents.  They also found better intubating conditions in the magnesium patients.

In both trials, magnesium was associated with lower heart rates and less hypertension in the peri-intubation and immediate post-intubation periods (of note: high dose magnesium is known to be associated with lower blood pressures, and can induce overt hypotension).  Neither study was really powered for more important measures like first pass success, mortality, or important side effects like peri-intubation hypotension.

Bottom Line: These are two small trials, and while more abundant literature should probably be obtained before we change our practice, one could consider giving magnesium sulfate, 50 mg/kg in 100 mL saline, prior to intubation in an attempt to improve intubating conditions.  In my opinion, this is probably worth considering in the rare circumstance that your patient has a true contraindication to neuromuscular blockade, but I probably wouldn't start doing this in standard RSI where you're going to be giving NMB until more literature confirms the safety of this approach.  Also, I would avoid this in situations where the patient is already hypotensive or at high risk of peri-intubation hypotension.  This may be worth considering in the very rare patient you're not necessarily going to give NMB to right away (maybe awake fiberoptic intubations?) who are also very low risk for hypotension.

Show References

Historically, guideline recommendations have been to use a transfusion threshold of hemoglobin < 7 g/dL for patients unless they are a) undergoing orthopedic surgery or b) have cardiovascular disease (CVD).  

Applefeld et al conducted a meta-analysis in 2018 which suggested that restrictive (i.e. lower hemoglobin trigger, typically 7-8) transfusion targets lead to worse outcomes in CVD patients than liberal (i.e. higher hemoglobin trigger, typically 9-10) targets, and those authors have updated this analysis to include data from newer trials.  Interestingly, the conclusion remains similar: that when you look at the larger studies on restrictive vs liberal transfusion targets, CVD plays an important role, as patients with CVD tend to do better with liberal targets, and patients without CVD tend to do better with restrictive targets.  Of note, CVD is variably defined in these studies, and sometimes limited only to active Acute Coronary Syndromes, and other times refers to all patients with acute or chronic CVD.  However, according to their analysis, the aggregated data suggests that we should continue having higher transfusion targets in patients with CVD, and perhaps even more in the 9-10 range, as opposed to the goals of 7 or 8 which are common.

Bottom Line: We will likely continue to see different transfusion targets recommended for patients with cardiovascular disease (CVD), and may even see guideline and blood bank recommendations raise the target for these patients more into the 9-10 range, or expand this group to include chronic CVD.  This would mean a substantial increase in recommended RBC transfusions, and as emergency physicians it is important for us to monitor these recommendations, especially since transfusions are not harmless and raising hemoglobin thresholds could lead to complications that are difficult to measure in the literature.

Show References

Multiple studies have suggested differences in patient outcomes with balanced solutions (e.g. plasmalyte) vs unbalanced solutions (e.g. normal saline) when large volumes are administered.  But what about when giving smaller volumes of fluid?  Does it matter which one you choose?

A recent study by Raes et al in the Journal of Nephrology looked at urine and serum effects of administering 1L of normal saline, vs 1L of plasmalyte, to ICU patients needing a fluid bolus.  Chloride levels, strong ion difference (SID), and base excess were all significantly different between the two groups.  There was no difference in blood pressure or need for vasopressors.  As best I can tell, other clinically significant differences such as kidney injury were unfortunately not reported.

Bottom Line: When giving small (e.g. 1L) volumes of IVF, there ARE real physiologic differences seen between balanced and unbalanced solutions.  Whether these differences translate to patient-oriented or clinically significant outcomes remains unclear.

Show References

As is well known, fluid resuscitation strategy ("liberal" vs “restrictive”) in sepsis is a controversial topic.  An RCT in NEJM called CLOVERS that looked at this and found no difference was recently re-analyzed to answer the following question… should my choice of strategy change if the patient presents with an Acute Kidney Injury (AKI)?  

For the most part, the answer is no.  In the group with AKI, the restrictive group did slightly, but non-statistically-significantly, better.  Interestingly, in the group without AKI, the relationship reversed, and in fact of the 4 groups (AKI vs no AKI, Restrictive vs Liberal), the no AKI but liberal strategy group did best (liberal vs restrictive in the no AKI group almost reached statistical significance in favor of the liberal strategy, but not quite).

Bottom Line: In septic patients presenting with an AKI, we don't know whether liberal or restrictive strategy is better, but either is probably reasonable.  In patients presenting without an AKI, it may be more ok to lean more towards liberal fluid resuscitation than in non-AKI patients*.  

*There are several important caveats here: 1) they didn't closely evaluate for potential side effects of over-resuscitation such as hypoxia or pulmonary edema (the primary outcome was need for renal replacement therapy), 2) as mentioned above, this trended towards but did not reach statistical significance, 3) this is one small study which did a subgroup secondary-analysis of a larger trial.

Show References

For the folks who have been in practice for a while, you may be aware of the roller-coaster evidence base looking at steroids for pneumonia.  Once thought to be beneficial and clearly indicated, of late steroids for pneumonia have fallen out of favor.  Hamad et al have published an excellent (and brief) review in Clinical Infectious Diseases which suggests the pendulum might be swinging back in favor of giving steroids to patients with pneumonia.  It's a ~5 minute read, so I recommend glancing through it yourself, but below are my two cents (solely my opinion) on where we are with steroids for pneumonia.

Take Home Points (OPINION ALERT):

1) When you have a condition present that you consider an indication for steroids (e.g. severe COVID-19 for sure; septic shock, s. pneumo infection, and ARDS depending on how you feel about the existing literature) --> strongly consider giving steroids unless there's a contraindication

2) When you have an undifferentiated patient who MAY have one of these conditions (e.g. pneumonia with COVID pending, patient potentially in ARDS or high risk of going into ARDS, etc) who is very sick --> it is reasonable to give steroids (if no contraindication) or not give steroids.  My tendency is to lean towards giving steroids in these cases, but do be aware that society guidelines recommend against steroids here (although debatable if they just haven't caught up to more recent literature)

3) When you have an undifferentiated patient who may have one of these conditions, but is NOT very sick --> I do not think there is significant enough evidence to support empiric steroids

4) Factors that might push you one way or another:

• Severity of disease (more severe favors giving steroids),
• Pathogen (COVID-19 and s. pneumo favor steroids),
• What formulation of steroids you have availabile.  Some of these studies used continuous hydrocortisone infusions, for example, which most hospitals don't routinely do.
• Comorbidities (uncontrolled diabetes, wound healing issues, risk for opportunistic infections might argue against giving steroids)

Show References

When patients fail simple respiratory support therapies like nasal cannula or non-rebreather, it is often a point of debate whether to move next to High Flow Nasal Cannula (HFNC) or Noninvasive Positive Pressure Ventilation (NIPPV).  This study randomized patients in acute respiratory failure (ARF) to CPAP, a form of NIPPV, vs HFNC.  They looked at all comers in ARF, and primary outcome was need for intubation.  Importantly, they excluded asthma/COPD exacerbation, for which BiPAP is typically considered the first line therapy due to improved CO2 clearance.

They found a significantly lower number of patients required intubation in the CPAP (28.9%) group than the HFNC (42.6%) group (p=0.006).  They hypothesized that the enhanced PEEP improved oxygenation (hypoxia being a common trigger for moving to intubation), but as opposed to BiPAP,  the lack of additional driving pressure limited tidal volumes and Patient Self-Inflicted Lung Injury (P-SILI), which is a known mechanism of ARDS and mortality.  They use this argument to explain why trials like FLORALI, pitting HFNC vs BiPAP, tend to not find an advantage for the NIPPV arm.  While this rationale makes sense, it should be noted that the study does not directly investigate if this was the reason for the difference, and for what its worth the inverse argument that using driving pressure to reduce respiratory rate, hypercarbia, and work of breathing (other very common indications for intubation) would also theoretically reduce intubations.  Furthermore, it's not clear why reducing P-SILI, which tends to cause mortality on a much longer duration, would improve the short-term outcome of need for intubation.

Bottom Line: This study demonstrated a benefit to CPAP over HFNC in terms of decreasing need for intubation amongst non-asthma/non-COPD patients with acute respiratory failure, and offered a physiologic rationale but one that requires further verification and discussion.  While it may be reasonable to choose CPAP instead of HFNC in marginal patients at risk of intubation (but stable enough to trial noninvasive support first), in my opinion more studies are likely needed before a wholesale change in practice.  The study also does not take into consideration the enhanced comfort and compliance we tend to see with HFNC over NIPPV, which should be considered as well.  

Show References

Central Venous Catheter (CVC; aka central line) placement is a common procedure in both the ED and ICU, and while overall quite safe, does carry some risk.  In particular, many of us regularly are confronted with the challenge of placing a line in a patient with profound thrombocytopenia, which can result in significant bleeding.  In these cases, should we give platelets before we place the line?

Van Baarle et al published a randomized study in NEJM comparing an empiric 1u platelet transfusion vs no transfusion in patients with a platelet count of 10,000-50,000, prior to line placement.  The study included both HD and non-HD (e.g. TLC) lines, from all three major access sites, in patients in their ICU or hematology ward.  They found statistically fewer serious bleeding events in the transfusion group (4.8%) vs no transfusion group (11.9%).  The study wasn't powered to look at more patient oriented outcomes like mortality, but I'm sure we can all agree less bleeding is probably a good thing.  Also importantly, this study did not evaluate the risks/benefits of delaying line placement to obtain platelets when the line is urgently needed, so I would not recommend extending this to conclude platelets must be given before line placement if the line is needed for something highly time-sensitive (e.g. only available access to infuse pressors in a hypotensive patient).  

Bottom Line: It is probably beneficial and appropriate to provide prophylactic platelet transfusion prior to CVC placement in patients with a platelet count less than 50,000, assuming circumstances allow.  

Show References