Review of the case: 54 year-old male with Granulomatosis with Polyangitis on prednisone, ESRD, and myelodysplastic syndrome present with fever, chills and one week of hemoptysis. Has had decreased exercise tolerance and recent hospital admission showing EF 50% and small pericardial effusion. Today, he is febrile, hypertensive, tachycardic, and borderline hypoxic with some wheezes and crackles on lung exam. Labs are notable for elevated creatinine, BNP, leukocytosis, and stable anemia. Bedside ultrasound shows diffuse B lines. CT chest is notable for infiltrates.
What is a differential diagnosis?The differential for this patient is a bit complicated. Given his vague symptoms and non-specific findings on exam, labs, and imaging, we find ourselves in a difficult situation. Let’s begin our discussion with the hemoptysis. This blood can be categorized by its location or origin (airway, parenchyma, vasculature, etc.) Airway diseases are the most common cause in developed countries. They include bronchitis, neoplasm, and foreign body (Suk, 2013). Vascular disorders include pulmonary embolism, arteriovenous malformations, and elevated pulmonary capillary pressure from mitral stenosis or LV failure. Intraparenchymal causes include infections and connective tissue disorders. (Connective tissue disorders are slightly more difficult to classify and can be both vasculitis and intraparenchymal etiology.)
Our patient has a relatively normal echocardiogram, making “pump” failure less likely. With any dialysis patient, hypervolemia is a concern when there are B-lines. Fluid retention doesn’t typically present with fevers and myalgia making this improbable. So, is this just a simple pneumonia? After all, the patient is febrile, tachycardic, and has myalgias. It would be reasonable to start empiric antibiotics until more information is gathered. However, there is no sputum and the timing of the symptoms seems to be a little prolonged for pneumonia. This appears to be a subacute or possibly even acute worsening of a chronic condition. This is a case of Diffuse Alveolar Hemorrhage (DAH), which is often seen in Granulomatosis with Polyangitis (GPA).
Diffuse Alveolar Hemorrhage (DAH) is a syndrome describing the accumulation of intra-alveolar red blood cells originating from the alveolar capillaries. DAH can initially present with fever, cough, and hemoptysis. Despite what the name might imply, DAH might present as anything but buckets of blood pouring out of the lungs. In approximately one third of cases, patients might not have any hemoptysis at all (Lara and Schwarz, 2010). Imaging and exam are often not helpful as they are non-specific. The confirmation of DAH will likely not take place in the ED because it requires bronchoalveolar lavage. However, read below as the treatment is important.
Does the BNP level help with the diagnosis? How?
Our patient has a BNP of 1316 and diffuse B lines that are suggestive of an intraalveolar fluid. In CHF, this would be all but diagnostic. However, in a patient on dialysis, this is not the case. Plasma BNP concentrations are elevated in patients on dialysis regardless of acute heart failure and trend in the opposite direction of GFR (Mueller et al., 2005). This may be due to volume overload and stretch on the ventricular wall or decreased BNP clearance from the kidney. There is a conflicting philosophy as to whether or not BNP can be trended. According to literature on this topic, most of the data seems to exclude patients on dialysis, making it less applicable. Because the etiology of the B lines is not clear, renal should be consulted for possible dialysis. If pulmonary edema is the cause, resolution or reduction in B lines should occur. That brings us to our question. Does BNP help in the diagnosis? Probably not as helpful as bedside lung ultrasonography. However, had BNP been low, we could essentially rule out CHF and volume overload as a cause and save time and resources.
How would you treat this patient?
References:
Lara, Abigail R., and Marvin I. Schwarz. “Diffuse Alveolar Hemorrhage.”CHEST 137.5 (2010): 1164-171.
MUELLER, CHRISTIAN, KIRSTEN LAULE-KILIAN, ANDR´E SCHOLER, CHARLY NUSBAUMER, THOMAS ZELLER, DANIEL STAUB, and ANDR´E P. PERRUCHOUD. “B-type Natriuretic Peptide for Acute Dyspnea in Patients with Kidney Disease: Insights from a Randomized Comparison.”Kidney International 67 (2005): 278-84.
Park, Moo Suk. “Diffuse Alveolar Hemorrhage.” The Korean Academy of Tuberculosis and Respiratory Diseases 74 (2013): 151-62.
Steven Greenstein
Latest posts by Steven Greenstein (see all)
- Case of the Month 19: Inhaled Steroids for Asthma - March 16, 2017
- Case of the Month 19 - March 3, 2017
- Case of the Month 18: Refractory Epistaxis - February 13, 2017
8 Comments
ablumenberg · October 14, 2016 at 6:49 pm
Would you start steroids on this patient in the ED?
I would not. I would treat for HCAP, start broad antibiotics, avoid immunosuppressives, and have blood on hold.
Though I can’t quote numbers I don’t think absence of sputum production significantly decreased likelihood of pneumonia.
Given the relative dangers of sepsis secondary to HCAP in a hemodialysis patient with an autoimmune disorder and on immunosuppressive therapy versus the possibility of a relatively rare complication of Wegener’s I think it would be dangerous to further immunosuppressive this patient in the ED.
In short this may be sepsis or DAH. He’s not drowning in blood. Sepsis can’t wait. Mild DAH can.
Ian deSouza · October 16, 2016 at 9:35 pm
Although I didn’t really think the patient had pneumonia, I considered all the factors that you have mentioned and administered broad-spectrum antibiotics. Since the bedside US showed normal contractility and collapsible IVC, I considered cardiogenic pulmonary edema unlikely. After much thought (and despite a not-so-surprising, unhelpful phone conversation with rheumatology), DAH appeared to really be the most likely diagnosis, so I gave 1g methylprednisolone. 2-3 hours later, the patient reported decreased cough and dyspnea and had improved oxygenation. Vital signs (aside from mild elevated BP) had normalized by 6 hours after steroids. Greenstein, any additional information from the inpatient course?
ablumenberg · October 17, 2016 at 11:22 am
Very interesting points.
I have some counter arguments however, and they all come down to probabilities and risk stratification.
This discussion reminds me of something we talked about in my college physics class. There is a thought experiment called “Schrodinger’s cat.” The basic idea is there is a closed box with a cat, a bottle of poison gas, and a device which may or may not release the poison gas. The device has a 50% chance of releasing the gas and killing the cat. You cannot know whether the cat is alive or dead until you have opened the box. Of course, the cat IS either alive or dead before we open the box, we just do not know yet.
Schrodinger’s cat is a good metaphor for the art of diagnosis and empiric treatment. The patient we have been discussing either has diffuse alveolar hemorrhage (DAH) or sepsis secondary to healthcare associated pneumonia (HCAP). Other less likely processes are possible too, but we will leave them out of the discussion for now. Once we open the box so to speak, i.e., broncheoalveolar lavage or lung biopsy, we will know whether this was DAH or HCAP. I have heard situations like this referred to as the “retrospectoscope,” when empiric treatment was started for what turned out to be a different disease process.
Keeping our patient’s best interests in mind, we need to make a risk assessment regarding two competing highly lethal diagnoses. To do so I think about the relative chance of each diagnosis, the risks of delayed treatment, and the harms of empiric treatment.
On the surface, diffuse alveolar hemorrhage seems like the most likely diagnosis. We have a Wegener’s patient with hemoptysis, multifocal pulmonary infiltrates, and evidence of systemic inflammation. If this were a boards question it would be the obvious answer. Keep in mind that exam questions and case reports are heavily biased. Writers tend to select rare and interesting cases.
If I present the same patient with a different emphasis (all facts taken from the presentation): a dialysis patient with fever, tachycardia, multifocal pulmonary infiltrates, myelodysplastic syndrome, on chronic prednisone, recent hospitalization, and cough, then the most likely diagnosis is HCAP. If we add the additional history of Wegener’s and hemoptysis we have introduced DAH as a potential alternative diagnosis, but not substantially decreased the likelihood of HCAP. A history of Wegener’s puts DAH on the table as a possible diagnosis, but also increases the chance of pneumonia due to further immune and pulmonary dysfunction. That is to say the increased chance of DAH has not significantly decreased the already high chance of HCAP.
Alright, it is either HCAP or DAH. The cat is either alive or dead. While Occam’s razor suggests that today’s presentation is an exacerbation of what has been going on for months, Hickam’s dictum states it could be a second entity. HCAP is common, and by definition is superimposed on other diseases. We cannot really know in the ER which disease entity we are fighting. The results of a tissue diagnosis several days later would be falsely validating because those results could easily be different for a patient presenting the same way in the hospital down the street. So what do we actually do?
If our patient has sepsis secondary to HCAP he is implicitly unstable. Early administration of antibiotics and fluid resuscitation (used judiciously in this dialysis patient) has a number needed to treat of only 6 (Source: Rivers, Promise, Process, Arise). That is to say without prompt antibiotics this patient has a very high mortality, and our early intervention would have a profound impact on his prognosis.
If our patient has DAH he is stable. Stable?? Well, stable enough to wait. Here is why. The chronicity of his symptoms are subacute. He was hospitalized a week ago and his symptoms have been progressive over months. His oxygen saturation is perfectly adequate even without supplemental oxygen, his hemoglobin/hematocrit are at baseline, and his hemoptysis has been trending towards improvement. Furthermore, I am aware of no evidence for early treatment of DAH as robust as that for sepsis.
In the ER we have to hedge our bets and this patient is a really interesting example of that thought process. Initially I chose to treat HCAP and not provide steroids based on my knowledge of pathophysiology from medical school. I felt it would be unwise to treat what could be an infectious process with immunosuppressive agents. I looked into this more and found Cochrane reviews that specifically addressed the role of steroids in sepsis and pneumonia, respectively.
http://www.cochrane.org/CD002243/ANAESTH_corticosteroids-treating-sepsis
http://www.cochrane.org/CD007720/ARI_corticosteroids-for-pneumonia
These reviews found no significant harm from giving steroids to patients with sepsis and pneumonia. I guess the old saying “one experiment is worth a thousand expert opinions” is true. If the diagnosis is HCAP steroids probably will not hurt, but they will help if it is DAH. I now agree that treating with steroids in addition to antibiotics is probably the right thing to do.
Great case!
Ian deSouza · October 18, 2016 at 2:00 am
Great comment, Blumenberg. You can add one more “logical diagnosis” to the the narrowed differential that was investigated. Tachycardia, tachypnea, hypoxemia, blood-tinged sputum…. pulmonary embolus. A CTPA evaluated for this possibility and “further characterized” the infiltrates seen chest radiography.
As for your Schrodinger’s Cat metaphor. It’s nice, but it fails in the sense that I could kind of see the f-in cat. How many immunosuppressed, dialysis patients with abnormal vital signs and multiple infiltrates IF in fact due to pulmonary sepsis would be smiling at you, eating in bed, and able to offer a full history? This is an area where physics and medicine do not overlap.
As for waiting for the inpatient service to treat DAH the next day, we all know when that would have happened. There would need to be an official rheumatology consult first (likely only after clinic has finished), 2-3 sets of rounds with an extensive discussion (not unlike yours I would surmise), and a noon conference. If I have a high enough suspicion for even a subacute process that may compromise airway, breathing, or circulation within a time course that I cannot predict, I tend to treat that process – especially if the most effective therapy (steroids) has demonstrated no real downside. In this situation, I dont like to wait/rely on others to possibly do a confirmatory test at some later time.
ablumenberg · October 18, 2016 at 7:34 am
I think that’s fine and I would probably also start steroids. All I’m saying is we are acting on our best guess and it’s important to acknowledge that. You just can’t say this was absolutely DAH, so you can’t really see the f-ing cat.
You say you think this patient would not have been treated for DAH until a lot of deliberating by the medicine team. That’s probably true. If you think he could wait for those discussions it’s probably not critical to treat for DAH immediately. Also just to clarify, he was never hypoxemic in the case presentation and I think this is significant when deciding whether to start treatment in the ED.
As for the expanded differential, CTPA was included in the case presentation and negative for pulmonary embolus. This could have also been endocarditis with septic pulmonary emboli, tuberculosis, PCP, malignancy, etc. They are less likely, but possible and are treated differently than what we’ve discussed.
ablumenberg · October 18, 2016 at 7:39 am
One last thing, there is a spectrum of disease with pneumonia. A lot of pneumonia patients are profoundly ill, but we also discharge patients with pneumonia home. There is a middle group.
Ian deSouza · October 18, 2016 at 10:42 am
I stated that I could “KIND OF” see your f-in cat. This was an attempt to work within your analogy. Although many prefer to work with good EBM, there is also ample evidence that concludes experienced physicians’ gestalt as effective as the use of tools to predict risk (as in PE and chest pain). They can use this to narrow and rank in probability a differential diagnoses (using Bayesian analysis) and also determine test and treatment thresholds. In cases such as this, the ability to examine a patient may be more valuable than in other more straightforward patients.
Your contention that one can wait to treat because there is normoxemia illustrates this point. Hypoxemia alone is not sensitive enough to determine degree of respiratory illness OR risk of disease progression and also should not be used alone as a reason to withhold emergency treatment. This is a classic pitfall in emergency medicine. But aside from this, 96% on room air is not actually normal either.
If you are wary of waiting to treat sepsis until further confirmatory tests (such as BAL), then it would reason that you should be wary of waiting to treat an immunologic condition (albeit a less studied disease) that may lead to serious deterioration in ventilatory function. How long should we wait? I guess one COULD have the discussion with the PGY2 medicine resident and explain to them the importance of airway/respiratory observation, repeated reassessment, and when to pull the trigger on steroids, but good luck with that. I expect that the response you will get would be, “Sounds like this patient needs an ICU consult.”
As far as your claim about early antibiotics and fluids in a patient with the sepsis, this is actually not so clear, even in the patient with septic shock. Im not sure where you calculated your NNT (but also glad you used NNT), but in the recent sepsis trials, ALL patients received some fluid resuscitation and antibiotics prior to randomization. Furthermore, recent evidence suggests that we actually may have up to 5 hours to select APPROPRIATE antibiotics which we often do not do. (Sterling SA, Miller WR, Pryor J, et al. The Impact of Timing of Antibiotics on Outcomes in Severe Sepsis and Septic Shock: A Systematic Review and Meta-Analysis. Crit Care Med. 2015;43(9):1907-1915.)
Yes, there is a spectrum of most illnesses. Again, without seeing the whole f-in cat, its just that patients with multifocal pneumonia typically appear more ill than this patient. But, this is where our debate will lose steam, because I had the opportunity to examine the patient. It is difficult to go further with such a discussion about management unless both parties had been able to examine the patient in order to be more on the same page.
The point about doing the CTPA was not that I didnt think that you had considered PE among the diagnoses. It was meant as an example of doing a test for a disease that was lower in pretest probability (PE) but may have helped to clarify a more likely one (DAH, pneumonia, or septic emboli). (Unfortunately, it didn’t.) There is often additional nuance involved when determining test-treatment thresholds; again in a sense, this is also part of physician gestalt.
Blumenberg, you have more than earned your ASL credit for this discussion.
ablumenberg · October 18, 2016 at 11:57 am
I am a big fan of nuance and gestalt. I think when it comes down to it we’re on the same page here.