BEEM Cases 3 on EM Cases – Acute Respiratory Failure. BEEM Cases is a collaboration between Andrew Worster of Best Evidence in Emergency Medicine (BEEM) and Emergency Medicine Cases’ Anton Helman, Rory Spiegel and Justin Morgenstern.
Written by Justin Morgenstern (@First10EM), edited by Anton Helman (@EMCases), September 2016
Dypnea & Acute Respiratory Failure: Sometimes the Cause is Not So Obvious
The Case…
A 73-year-old woman presents to the emergency department via EMS with increasing shortness of breath and cough over the past day. She has a history of COPD, CHF, hypertension, and hyperlipidemia. On arrival, she is breathing rapidly at 34 breaths a minute and is using all her accessory muscles. Her heart rate is 115, BP 155/95, Temp 37.5 and oxygen saturation 89% on 4L via nasal cannula. You perform a rapid physical exam, but you still aren’t sure exactly what is causing her dyspnea. Your RT turns to you and asks what you’d like to do.
Shortness of breath is a very common chief complaint in the emergency department, but despite our familiarity with this symptom, management is not always straightforward. The differential diagnosis is extensive, including the common cardiorespiratory conditions, but extending to toxicologic, hematologic, neuromuscular, metabolic, and psychiatric causes. Over the past decade, we have seen the widespread adoption of new technologies to help us manage these patients. This post will look at some new evidence on two of those technologies: noninvasive positive pressure ventilation (NIPPV) and ultrasound (POCUS). We will answer 3 questions based on 3 systematic reviews using the BEEM critical appraisal framework:
Question #1
Does noninvasive positive pressure ventilation (NIPPV) reduce mortality in acute respiratory failure?
Question #2
Does prehospital CPAP or BiPAP improve clinical outcomes for patients in acute respiratory failure?
Question #3
What is the sensitivity and specificity of POCUS using B-lines in diagnosing acute cardiogenic pulmonary edema in patients presenting to the ED with acute dyspnea?
Question #1 Does noninvasive positive pressure ventilation (NIPPV) reduce mortality in acute respiratory failure?
The Paper
Cabrini L, Landoni G, Oriani A. Noninvasive ventilation and survival in acute care settings: a comprehensive systematic review and metaanalysis of randomized controlled trials. Critical care medicine. 43(4):880-8. 2015.
Study details (PICO)
Systematic review and meta-analysis
| Population | Adult patients with acute respiratory failure including COPD and CHF |
|---|---|
| Intervention | NIPPV |
| Control | Any other mode of ventilation |
| Outcomes | Mortality at the longest available follow-up |
Key results
This meta-analysis found 78 trials that fit the inclusion criteria, with a total of 7365 patients. For the primary outcome of mortality, they found that noninvasive positive pressure ventilation decreased overall mortality (RR=0.73 [95% CI: 0.66, 0.81]) with a NNT=19.
BEEM critique
This is the largest review of NIPPV to date and its primary outcome is mortality, the ultimate clinical outcome. Although 60% of the data is from the ICU setting, the results are probably still applicable to the ED and provide convincing evidence that patients with acute respiratory distress (except asthma) should be considered for NIPPV as a first line therapy. The suggestion that early NIPPV is better than late requires further study.
Key EBM point: Heterogeneity. The trials included in this meta-analysis displayed high heterogeneity. This simply means that the trials were different from each other in some way. There are two key types of heterogeneity. Clinical heterogeneity occurs when there is variability in key clinical aspects of trials. For example, two trials may look at different populations of patients or measure different outcomes. Statistical heterogeneity refers to the likelihood that the variability among the different results (one trial might report a 2% benefit whereas another reports a 18% benefit) is due to chance alone. Heterogeneity matters because if trials are too dissimilar it may not be appropriate to combine them into a single statistical analysis.
Case continued…
You start the patient on BiPAP and within 10 minutes her numbers have improved and she looks a lot better. One of the paramedics who brought her in is surprised by the rapid improvement and asks you if they should be starting some kind of non-invasive ventilation in the ambulance before arriving at the emergency department.
Question #2 Does NIPPV improve clinical outcomes in acute respiratory failure?
The Paper
Goodacre S, Stevens JW, Pandor A. Prehospital noninvasive ventilation for acute respiratory failure: systematic review, network meta-analysis, and individual patient data meta-analysis. Academic emergency medicine : official journal of the Society for Academic Emergency Medicine. 21(9):960-70. 2014.
Study details (PICO)
Systematic review and meta-analysis
| Population | Prehospital patients in respiratory distress |
|---|---|
| Intervention | Prehospital continuous positive airway pressure (CPAP) or BiPAP |
| Control | Standard care |
| Outcomes | Primary: MortalitySecondary: Endotracheal intubation |
Primary outcome (mortality):Key results
- CPAP reduced morality (OR=0.41; 95% credible interval [Crl] 0.20 to 0.77)
- The effect of BiPAP on mortality was unclear (OR=1.94; 95% Crl = 0.65 to 6.14)
Secondary outcome (intubation):
- CPAP reduced intubations (OR=0.32; 95% Crl 0.17 to 0.62)
- The effect of BiPAP on intubation was unclear (OR=0.40; 95% Crl = 0.14 to 1.16)
BEEM critique
The benefits of NIPPV for patients in acute respiratory failure are well documented. Also, NIPPV is likely to be most effective when introduced early. The evidence supporting at least CPAP from this study is encouraging but differences in outcomes between CPAP and BiPAP reflects more upon the lack of large RCTs rather than the actual clinical difference between them. Regardless, the cost of equipping ambulances with NIPPV gear has to be taken into consideration when assessing its effectiveness in the prehospital setting.
Case continued…
The patient is improving, but you still aren’t sure about the diagnosis. There might have been an elevated JVP, but her neck isn’t easy to examine. The lungs sound a little wheezy, but there were probably some fine crackles there as well. You are resigned on waiting for the chest x-ray, when your resident asks if lung ultrasound might help diagnose pulmonary edema.
Question #3 Accuracy of POCUS for Diagnosing Acute Heart Failure
What is the sensitivity and specificity of point of care ultrasound (POCUS) using B-lines in diagnosing acute cardiogenic pulmonary edema in patients presenting to the ED with acute dyspnea?
The Paper
Al Deeb M, Barbic S, Featherstone R, Dankoff J, Barbic D. Point-of-care ultrasonography for the diagnosis of acute cardiogenic pulmonary edema in patients presenting with acute dyspnea: a systematic review and meta-analysis. Academic emergency medicine : official journal of the Society for Academic Emergency Medicine. 21(8):843-52. 2014.
Study details (PICO)
Systematic review and meta-analysis
| Population | Prospective cohort and prospective case-control studies that recruited patients presenting to hospital with symptomatic acute dyspnea, or where there was a clinical suspicion of congestive heart failure, and reported the sensitivity and specificity of B-lines in diagnosing acute cardiac pulmonary edema (ACPE)Excluded: Studies of asymptomatic individuals or in patients where there was no suspicion of ACPE |
|---|---|
| Intervention | Bedside ultrasonography for ACPE (final diagnosis of ACPE) and evaluating B-lines |
| Control | None |
| Outcomes | Sensitivity and specificity, diagnostic odds ratio |
Key results
They identified 7 studies that included a total of 1075 patients. Two of the studies were ED studies. The other 5 took place in the ICU, hospital wards, or prehospital environment.
Diagnostic characteristics:
- Sensitivity of B lines of POCUS to diagnose acute pulmonary edema: 94% [95% CI: 81.3%, 98.3%]
- Specificity of B lines of POCUS to diagnose acute pulmonary edema: 92% [95% CI: 84.2%, 96.4%]
- Positive likelihood ratio 12.4 [95% CI: 5.7, 26.8]
- Negative likelihood ratio 0.06 [95% CI: 0.02, 0.22]
BEEM critique
The question asked in this review is relevant but as the authors admit, there is no standardized threshold for the diagnosis of acute cardiac pulmonary edema (ACPE) and no definitive gold standard. Like the first study reviewed in this BEEM Cases, this one was too heterogeneous. While this study was exhaustive in searching for ultrasound diagnostics performed at the bedside it was not restrictive in settings, patient demographics, or ultrasound training of provider and this would lead to heterogeneity. Another issue that contributes to the heterogeneity and challenges the validity of the results is the lack of standardization of the ultrasound exam: The identification of ACPE using B-lines via the Volpicelli method is dependent upon patient position as well as position duration.
The conclusion that B-line on ultrasound can confirm the diagnosis when the pretest probability of disease is high or low has little utility. Diagnostic tests are valuable when they can confirm or refute a diagnosis when the pretest probability is indeterminate.
Case Resolution…
The patient rapidly improves after being placed on BiPAP. Your bedside ultrasound was consistent with CHF, but understanding the limitations of the test you also ordered your traditional work-up including blood work, ECG, and chest x-ray. Within a few hours in the department, after treatment with nitroglycerin and furosemide, you are able to titrate down and then discontinue the positive pressure ventilation. On a repeat bedside ultrasound, the b-lines have disappeared. Combining the ultrasound findings with the remainder of your tests, and most importantly your clinical judgement and frequent reassessments of the patient, you diagnose her with an exacerbation of CHF and admit her to the medical team for monitoring and adjustment of her medications.
References
Cabrini L, Landoni G, Oriani A. Noninvasive ventilation and survival in acute care settings: a comprehensive systematic review and metaanalysis of randomized controlled trials. Critical care medicine. 43(4):880-8. 2015.
Goodacre S, Stevens JW, Pandor A. Prehospital noninvasive ventilation for acute respiratory failure: systematic review, network meta-analysis, and individual patient data meta-analysis. Academic emergency medicine : official journal of the Society for Academic Emergency Medicine. 21(9):960-70. 2014.
Al Deeb M, Barbic S, Featherstone R, Dankoff J, Barbic D. Point-of-care ultrasonography for the diagnosis of acute cardiogenic pulmonary edema in patients presenting with acute dyspnea: a systematic review and meta-analysis. Academic emergency medicine : official journal of the Society for Academic Emergency Medicine. 21(8):843-52. 2014.
Hi Anton, great post!
Q # 1
– The majority of trials are on patients with either COPD exacerbation or ACPE. For these 2 categories, NPPV efficacy in terms of ETI reduction and mortality are okay.
Bottom line 1: NPPV is first choice in ACPE or COPDex
– Acute hypoxemic respiratory failure (except ACPE) is still controversial. Few positive RCTs on pneumonia (Confalonieri M goo.gl/K4EHZV, Brambilla AM goo.gl/uKKDpx, Cosentini R goo.gl/rReh7f), one recent positive RCT on ARDS (Patel BK oo.gl/lKNqkv).
Bottom line2: NPPV for pneumonia –> 1. okay in the immunocompromised population, 2. In the immunocompetent population: early application, that is patient selection is the key (and short trial)
– ARDS. Patient selection seems the key, however needs further confirmation
– Asthma. Primum non nocere!
Q # 2
Pre-hospital NPPV modality of choice might be CPAP.
1.The majority of patients have AHF/ACPE, 2. easier to learn and carry, 3. cheaper
Q # 3
LUS is already in every acute dyspnoea algorythm (Lichtenstein blue protocol goo.gl/yEy6ug)
When in doubt (pretest probability indeterminate) more useful if negative (SNOUT) than positive (SPIN), since interstitial syndrome might be due to other causes (pneumonia, ARDS, fibrosis)
Thanks again for your work
Roberto
Thanks for the excellent comment Roberto.
With regards to question one, I think you hit on the key issue. Acute respiratory failure is not a single condition, but actually a collection of many different conditions, and NIPPV might (and probably does) have different effects on different conditions. This is a large part of the BEEM focus on the heterogeneity of the underlying trials. Although not the focus of the paper, NIPPV clearly has an indication in COPD and CHF (as well as ARDS and post-extubation, but those are less relevant in the emergency department). There aren’t great studies in asthma, but I think the evidence favours NIPPV. I definitely use NIPPV early in severe asthma. I would not use NIPPV long term in pneumonia. However, I think the key take home for the emergency department, where we have undifferentiated patients, is that NIPPV seems to lower mortality overall, and should be started early while we work on determining the underlying cause of this patient’s respiratory distress.
With regards to pre-hospital NIPPV, ease of use and cost are definitely important issues. (I would also like to see more compatibility between prehospital equipment and inhospital equipment, both for cost and ease of patient care.) Unless we see large advantages to BiPAP, and I agree that CPAP probably makes the most sense for EMS. However, all of the questions are relatively complex. EMS agencies with longer transport time might benefit from BiPAP – although that assumption currently doesn’t have any evidence to back it up.
In terms of lung ultrasound, I will tell you I use it every shift. However, the widespread use of ultrasound and adoption into protocols does not mean that we are practicing evidenced based medicine. I think the numbers here (which are pretty consistent with all the studies I have seen, including that Lichtenstein paper) show that lung ultrasound is about as accurate for ruling in as it is for ruling out (sensitivity and specificity are both in the low to mid 90s). However, the studies that give us those number have a number of issues that could be inflating the accuracy. The diagnosis in many patients is obvious without ultrasound, and the patients who are less obvious clinically are also less obvious on ultrasound. I love ultrasound an will continue to use it, with the caveat that in the initially undifferentiated patient (pretrest probability of 50%), the numbers reported for ultrasound don’t get be above a 95% post test probability if positive, nor do they get me under a 5% post test probability if negative, so I am am constantly aware that my ultrasound diagnosis might be wrong.
Cheers
Justin