Gruber’s Follow-Up Notes (featuring Dr. Jason West)

IMG_20131228_084611

Case 1

Patient with RUQ pain 7 weeks after gastric bypass

The first case is a 41 year-old obese male presenting for one day of RUQ pain 7 weeks s/p Roux-en-Y gastric bypass surgery. He denies nausea, vomiting, post-prandial pain exacerbations, fever, and cough. He reports weight loss after the surgery, now weighing in at 350lbs from 500lbs prior to the procedure. Despite this weight loss, he mentions that he has “not been doing much other than playing video games” in the last few days.

Temp 99.5, Pulse 103, RR 20/min, BP 121/74, and 96% on room air.

On exam, he has clear breath sounds and tenderness to his RUQ. His labs are unremarkable, and his ECG is shown below. You have no prior ECG for comparison.

Later, his vitals were taken again; and his pulse rate improved to 96.

You go to reassess the patient, and he develops shortness of breath when you help direct him to the location of the bathroom.

CASE 1 - West Followup Rounds

What abdominal problems are patients likely to have 7 weeks after gastric bypass?

Bleeding from anastomotic sites, which present with tachycardia and a drop in hematocrit, typically occur within the first few weeks after surgery. The bleeding is most commonly intraluminal and may present with melena. It occurs more frequently in laparoscopic bypass surgeries than with open ones.

Wound infection is more common with open gastric bypass surgery, and it generally occurs 1-2 weeks post-op.

Anastamotic leaks of gastric contents and GI obstruction most commonly occur 1-4 weeks post-op. These patients may present with peritonitis and eventually develop fistulas if the leak is not recognized early.

Problems occurring later include internal hernias, obstruction, distention of a blind pouch, cholelithiasis, and stomal ulcers. Internal hernia formation results from herniation of the small bowel through potential spaces created during the surgery. Most patients with obstruction will present with a persistent inability to tolerate PO, yet the intermittent nature of the obstruction may make its detection difficult with imaging. The gastric pouch created during the gastric bypass can become distended from impaired emptying or from obstruction and result in rupture or spillage of bowel contents causing severe peritonitis. A post-surgical gastric pouch may be seen as large gastric bubble on imaging, and it may present with shortness of breath.  There is a high mortality with this blind pouch syndrome unless it is recognized early and explored operatively.

Stomal obstruction and cholecystitis can present months to years after the initial surgery. Stomal ulcers can present 2-4 months post-op, and stomal stenosis is usually seen 6 months post-op. Rapid weight loss predisposes to gallstones, and up to 50% of gastric bypass patients develop gallstones. 40% of the 50% who will get gallstones will become symptomatic.

Dumping syndrome occurs after eating simple carbohydrate meals and presents with nausea, vomiting, and generalized weakness (1).

He is unlikely to have an infection, bleed, or leak and only have developed his problem 7 weeks after his surgery. Obstruction is also unlikely without vomiting. But there are other complications of the surgery that can present within this time range.

Do normal labs dissuade you from going after the diagnosis of acute cholecystitis?

The short answer is no.

Liver function tests (LFTs) and alkaline phosphatase are generally not elevated, as the common bile duct is not usually blocked during acute cholecystitis. In fact, if there is even slight elevation in LFTs or alkaline phosphatase, one should consider that the common bile duct is blocked. Furthermore, elevated LFTs do not have adequate sensitivity to dissuade you from a diagnosis of acute cholecystitis. Both leukocytosis and elevated LFTs have positive LRs lower than 2, and neither are useful in determining which patient with RUQ pain will have acute cholecystitis diagnosed by imaging (2,3).  Additionally, as we know from Gruber’s publication, both fever and elevated WBC are frequently (>25%) not seen in patients with confirmed acute cholecystitis (4).

What should we do about his abnormal ECG?

Get a troponin? Sure. But think about other causes of diffuse flipped T waves.

What is the most common cause of flipped T waves?

Ventricular hypertrophy is the most common cause of TWI, and LVH presents with TWI in the lateral leads.  This patient’s ECG has TWIs in the anterior, lateral, and inferior leads; and it does not meet criteria for either LVH or biventricular hypertrophy.

What besides myocardial ischemia presents with anterior and inferior TWIs?

Marchick (5) performed the best and largest study on the ECG findings in patients tested for pulmonary embolism (PE), and it included over 6,000 patients with a 6% PE prevalence rate. Anterior TWIs were one of only two ECG findings the authors found to have relatively useful LRs, and the positive LR of anterior TWIs is 3.7.

Witting and Mattu (6) found that simultaneous TWI in the inferior and anterior leads is highly predictive of PE.  In their retrospective review, they selected 97 patients with PE, 89 with ACS, and 105 with non-cardiac chest pain as controls.  2mm TWIs in III, aVF, V1, and V2 were poorly sensitive (4%), yet highly specific enough to drive up the positive LR to 16.  More subtle and less extensive inferior, right, and anterior precordial TWIs had LRs ranging from 2-5.  The prevalence of PE is high in their study, and the positive predictive value may be lower than expected despite the very high positive LR of 16 in our patient population, which is likely to have a lower rate of positive PE studies (closer to the 6% prevalence rate from Marchick’s study).

On further exam of our patient’s ECG, it is noted he has subtle S1Q3T3

How useful is the S1Q3T3 finding for PE?

Like TWIs in the anterior leads, Marchick’s paper found S1Q3T3 to be poorly sensitive but highly specific for PE. The positive LR of S1Q3T3 is 3.7.

However, much earlier literature has shown that any cause of acute cor pulmonale, such as pneumothorax or acute bronchospasm, can cause this ECG finding (7).

Does our patient have risk factors for PE?

He is post surgery, and PE is responsible for >50% post gastric bypass deaths (8); but post-surgical risk is classically considered up to 4 weeks post surgery. He is very sedentary, and this was considered to be relative immobilization. He has no history of DVT, PE, or family history of either. He denied hemoptysis and known malignancy.

Our patient’s tachycardia improved upon re-examination, and there is some literature to suggest that the normalization of abnormal vital signs does not reduce your pre-test probability of PE (9).

We also have a better diagnosis for RUQ pain available, but you are concerned that the clinical decision rule is not as good as your judgment.

Our patient has a Wells Score of 3, placing him at either moderate or “unlikely” pre-test probability of PE depending on how you rank the score. A score of 3 gives you a pre-test probability of 16% in our American population; and if you use the positive LR ratio (3.7) of either S1Q3T3 or anterior TWI and apply it to a Fagan nomogram, you get a new pre-test (of CTA for PE) probability of almost 40%, which is the highest pre-test probability that the Wells Score provides.

How useful is wells criteria if your gestalt or suspicion is high?

Gestalt has been shown to outperform wells criteria in a recent Annals of EM paper by Penezola (10), who compared unstructured clinical gestalt to the Wells and Revised Geneva scores. They found that clinical gestalt outperformed both clinical decision rules due to higher accuracy in patients with either low or high pre-test probabilities. However, this is a European study with a high PE prevalence rate (31%); and it is unclear (if not unlikely) that these results will translate into the American setting, where the prevalence in our literature is much lower. A change in prevalence will alter your pre-test probability and the positive predictive value (PPV) of your gestalt.

However, gestalt overrules the Wells Score in this obese, sedentary patient who easily develops SOB and has ECG findings highly specific for PE.

What is the best initial approach to this patient?

Ideally, you should (have time to) perform bedside ultrasounds of the RUQ to look for acute cholecystitis, of the legs to look for DVT, and of the heart to identify right ventricle dilatation.  The sensitivity of your sonogram will be dependent on your experience, and you may not have time to sonogram multiple regions.

If you found a DVT on ultrasound, would you stop and admit?  Or would you CT scan the abdomen believing that the DVT is a red herring distracting you from finding an abdominal pathology with worse consequences?

The consensus from follow-up rounds was that the patient should get a CTA of the chest for PE and then look at the abdomen and pelvis to address abdominal pathology related to the gastric bypass given the high complication rate. We have been told conflicting information from the radiology department on whether a CTA protocoled for detecting PE can simply be continued down to the abdomen and pelvis. Depending on your radiologist’s confidence that the IV contrast will be timed appropriately for the abdomen and pelvis as well, you may have to perform two scans so the timing is correct.

Outcome:

CTA showed a considerable thrombus within the artery supplying the right lower and middle lobes.

Teaching points

-Always reassess your patients. This patient only became SOB after a repeat, albeit unintentional, re-examination while walking.

-Don’t limit your differential to abdominal pathology in a patient with RUQ pain.

-Normal labs shouldn’t dissuade you from pursuing a diagnosis of acute cholecystitis with imaging.

-Correction of tachycardia with fluids is not reassuring that your patient doesn’t have PE.

-S1Q3T3 and anterior TWIs are uncommon (poorly sensitive) in patients being evaluated for PE but are specific with likelihood ratios of 3.7 each, and these LRs can be used to increase your pre-test probability of PE.

-Simultaneous anterior and inferior TWIs are likely less common and may be even more specific if chest pain is the main chief complaint, but their positive predictive value (PPV) is unknown when applied to any patient in which PE is considered regardless of their presenting complaint.

-Your clinical gestalt of PE may be better than the Wells score in certain subsets of patients depending on the prevalence of PE in your population.

Case 2
Patient with diarrhea and supra-sternal pain radiating to the back after a bad meatball sandwich

The second case is a 55 year old man with obesity and HTN presenting with an initial ED triage complaint of diarrhea after eating a “bad meatball sandwich” earlier that night. He was quickly assigned to a zone and evaluated, because his secondary complaints included nausea, vomiting, diaphoresis, and suprasternal pain radiating bilaterally to his shoulders and back. On exam, he appeared diaphoretic with clear breath sounds.

His ECG is shown below and has ST elevations in the precordial leads with what may appear to be a left bundle branch block (LBBB).

His Chest XRay is normal, and the labs are not yet back.

CASE 2 - West Followup Rounds

Is this pt having an ST elevation MI (STEMI) or is this LBBB?…or is it both?

LBBB has monophasic, monolithic R waves in the lateral leads I, aVL, V5, and V6. In this case, there is a smaller r wave and a big S wave in V6 along with large anterior Q waves. Q waves normally seen in the lateral leads are not seen in LBBB due to the reversal of repolarization; and a LBBB pattern with a Q wave in I, V5, or V6 is likely to either be pathologic or make you consider it to be an alternative ECG pattern. In the ECG for this case, it would be considered to be an incomplete block if the QRS were < 120ms; but the wide QRS makes it much more likely be be a complete LBBB versus some other intraventricular conduction delay. Traditional emergency medicine resources such as Tintinalli’s textbook say that no Q waves are allowed in leads I, V5, and V6; and the QS pattern in V5 and the deep S wave of V6 in our ECG could make you think this a variant LBBB. Despite the variant LBBB, we can assume that the ST elevation (STE) is due to ischemia. Had the clinicians on this case recognized that the block did not meet the traditionally used LBBB criteria, the Sgarbossa criteria could have been applied to distinguish a STEMI from a LBBB pattern. Sgarbossa’s data came from a subset analysis from the large GUSTO-1 trial intended to study different thrombolytic strategies for acute MIs (11). Of this large registry, only 131 had acute MI with LBBB. LBBB in this study was largely defined as a QRS > 125ms, S in V1, and r without Q waves in the lateral leads. The authors found 3 ECG findings with low to moderate sensitivity (31-73%) and high specificity (92-96%), and they created a scoring system with a low diagnostic sensitivity (52%) and high specificity (98%). Of the three, excessive discordant STE > 5mm had the worst performance.

To fix the Sgarbossa criteria’s low sensitivity, the Modified Sgarbossa criteria uses a calculated ST/S ratio of < -0.25 to increase the rule’s sensitivity performance to 91% without sacrificing much specificity. However, this new modified criteria was a smaller study of 33 patients with LBBB found to have acutely occluded coronary arteries compared against 129 patients with LBBB who did not have a “culprit lesion” in the cath lab. LBBB in the Modified Sgarbossa study was more broadly defined and included a wide QRS, rS in V1, and r in V6 (12). One could argue that the ECG in this case did not meet the inclusion criteria of the original Sgarbossa study but would have been included in the more recent Modified Sgarbossa study. Neither the Sgarbossa rule nor the Modified Sgarbossa rule has been externally validated to be sensitive enough to use in isolation in patients presenting with LBBB and ACS. Our own Dr. Li published that the original Sgarbossa criteria was poorly sensitive to identify a STEMI (13), and the Modified Sgarbossa criteria have yet to undergo external validation. However, had the Modified Sgarbossa been applied to this patient, the rule would have identified this STE pattern as a STEMI. In contrast, the original Sgarbossa criteria would not have identified this ECG as a LBBB with a STEMI. Should this patient get a CT angio (CTA) to rule out dissection before being transferred to the cath lab? Or can the aorta be evaluated on the cath table, too? A cardiac cath angiogram can identify a dissection on the cath table. However, if you are reasonably sure that your patient is having a STEMI, it is very uncommon for an aortic dissection to cause STEs (<5% dissections have STEMI). Nearly all of the STEs from dissection are in Type A, or ascending, dissections (13). A registry of ED STEMI activations was recently published and showed that those who underwent CTA prior to catheterization had a door-to-balloon (D2B) time of 166 minutes versus 76 minutes in those who did not get a CT. This same study showed that CT rarely changed the cardiologist’s decision to go to the cath lab (15). Furthermore, a CTA followed by cardiac catheterization is a double dye load. The literature on contrast-induced nephropathy is controversial, but the current standard is to consider the risk to be real. Should our patient be anticoagulated if dissection is being considered? The consensus is no. Little data exists on inadvertent anticoagulation for aortic dissections, and what evidence we have suggests that it is more common than you would think and has a high rate of serious complication (16). Are there ECG findings that can predict an aortic dissection? The most common abnormal ECG findings are non-specific ST-T wave changes (41%), LVH (26%), and ischemia (15%). 31% of patients with aortic dissections will have a normal ECG (14). But if the dissection involves the coronary arteries, it generally involves the RCA and presents with STE in the inferior leads. In this case, the patient is having anterior STEs, which are very rare in an acute dissection. It would make sense not to delay transfer for a CTA unless you had other reasons to highly suspect the patient was having a dissection. Is there a role for echo in the initial management of a patient suspected of having STEMI? The cardiology fellow we consulted for this article suggested a role for a bedside echo in this patient. If you had a few minutes to perform a bedside echo, wall motion akinesis may occur earlier than troponin elevation in STEMI patients. Without a previous echo, you will not be able to tell if the wall motion abnormality is new or old. Supposedly, a cardiology tech can be available to perform a TTE within 15 minutes during regular working hours. If a bedside echo had been performed on this patient, a lack of akinesis would have gone against a working diagnosis of STEMI. Do Q waves give some prognostic significance in STEMIs? The presence of Q waves gives the cardiologists important information about the 90 day prognosis of STEMI patients. In 2006, a Lancet article reported that Q waves on the baseline STEMI ECG are an independent predictor of higher mortality in STEMIs receiving thrombolytics and of failed reperfusion (17). In 2009, another large study of STEMI patients sent for PCI showed that baseline Q waves surpassed the time from symptom onset (> 3 hours) as a prognostic marker for 90 day death, CHF, or cardiogenic shock (18).

Outcome

Troponin approximately 100

Cath report:
RCA- 50% occlusion of mid and 50% distal
LCX- 75% distal
LAD- 100% proximal and mid first diagonal (D1) 100%

Balloon angio was performed to the LAD and D1 with TIMI grade flow improvement from 0 to 3, which means it went from no perfusion to normal perfusion.

Echo:
Significant akinesis of anterior and septal walls with an EF 20-25%

Teaching points

-It is very uncommon to have both STEMI and a dissection

-CTA should only be used in a patient with STEs when you are not sure if the patient is having a STEMI and you are suspicious of dissection.

-If you are reasonably sure your patient is having STEMI and think dissection is a less likely possibility, the patient should be transferred to the cath lab, where the aorta can be viewed by angiogram.

-CTA rarely changes the plan to send the patient to the cath lab and greatly delays the door-to-balloon time.

-The most common STEs in dissection patients are in the inferior leads.

-Specific ECG criteria for LBBB will depend on the literature you are reading, and the information we found in textbooks is not entirely consistent with the primary literature relevant to this case. In this case, the information we learned from Tintinalli’s text may have prevented us from using the Modified Sgarbossa to “rule in” a STEMI immediately after reviewing the ECG.

-The Modified Sgarbossa criteria replaces the excessive discordant STE criteria with an ST/S ratio of < -0.25 to greatly increase the rule’s sensitivity. -Q waves on the initial STEMI ECG are more predictive of failed reperfusion in cath lab than a > 3 hour delay in the time from symptom onset to PCI.

-Q waves on the initial STEMI ECG are a more important independent predictor than time to PCI of which patients will go on to have shock, CHF, or death at 90 days.

References:

1. Ellison SR and Ellison SD. Bariatric Surgery: A Review of the Available Procedures and Complications for the Emergency Physician. J Emerg Med 2008; 34:21-32.

2. Trowbridge RK et al. The rational clinical examination: does this patient have acute cholecystitis? JAMA. 2003;289:80-86.

3. Roe J. Clinical Assessment of Acute Cholecystitis in Adults. Ann Emerg Med. 2006;48:101-103.

4. Gruber PJ et al. Presence of fever and leukocytosis in acute cholecystitis. Ann Emerg Med. 1996 Sep;28(3):273-7.

5. Marchick MR et al. 12-lead ECG findings of pulmonary hypertension occur more frequently in emergency department patients withpulmonary embolism than in patients without pulmonary embolism. Ann Emerg Med. 2010 Apr;55(4):331-5.

6. Witting MD et al. Simultaneous T-wave inversions in anterior and inferior leads: an uncommon sign of pulmonary embolism. J Emerg Med. 2012 Aug;43(2):228-35.

7. Chan TC, Vilke GM, et al. Electrocardiographic manifestations: pulmonary embolism. J Emerg Med. 200; 21(3):263-70.

8. Melinek J et al. Autopsy findings following gastric bypass surgery for morbid obesity. Arch Pathol Lab Med. 2002;126(9):1091.

9. Kline JA et al. Normalization of vital signs does not reduce the probability of acute pulmonary embolism in symptomatic emergency department patients. Acad Emerg Med. 2012 Jan;19(1):11-7.

10. Peneloza A, et al. Comparison of the unstructured clinician gestalt, the wells score, and the revised Geneva score to estimate pretest probability for suspected pulmonary embolism. Ann Emerg Med. 2013 Aug;62(2):117-124.

11. Sgarbossa EB et al. Electrocardiographic diagnosis of evolving acute myocardial infarction in the presence of left bundle-branch block. GUSTO-1 (Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries) Investigators. N Engl J Med. 1996 Feb 22;334(8):481-7.

12. Smith SW et al. Diagnosis of ST-elevation myocardial infarction in the presence of left bundle branch block with the ST-elevation to S-wave ratio in a modified Sgarbossa rule. Ann Emerg Med. 2012 Dec;60(6):766-76.

13. Li SF et al. Electrocardiographic diagnosis of myocardial infarction in patients with left bundle branch block. Ann Emerg Med. 2000 Dec;36(6):561-5.

14. Klompas M. Does this patient have an acute thoracic aortic dissection? JAMA. 2002 may 1:287(17):2262-72.

15. Armstrong EJ, et al. Delaying primary percutaneous coronary intervention for computed tomographic scans in the emergency department. Am J Cardiol. 2012 Aug 1;110(3):345-9. doi: 10.1016/j.amjcard.2012.03.032. Epub 2012 Apr 23.

16. Davis DP et al. The inadvertent administration of anticoagulants to ED patients ultimately diagnosed with thoracic aortic dissection. Am J Emerg Med. 2005 Jul;23(4):439-42.

17. Wong CK et al. Initial Q waves accompanying ST-segment elevation at presentation of acute myocardial infarction and 30-day mortality in patients given streptokinase therapy: an analysis from HERO-2. Lancet 2006 Jun 24;367(9528): 2061-7.

18. Armstrong PW et al. Baseline Q-wave surpasses time from symptom onset as a prognostic marker in ST-segment elevation myocardial infarction patients treated with primary percutaneous coronary intervention. J Am Coll Cardiol. 2009 Apr 28;53(17):1503-9.

Post contributed by Dr. Jason West and Dr. Peter Gruber

2 Comments on "Gruber’s Follow-Up Notes (featuring Dr. Jason West)"

  • these are great cases. good job on the summary!

    curious if there was discussion of diagnostic workup of this same patient if he presented 4 weeks post op when both PE and acute gastric outlet obstruction is more likely, the patient had been vomiting almost everyday, and found to have the same EKG and dyspnea on exertion. had a similar case at our shop, and what we ended up doing was a CTA to r/o PE, but also have the patient swallow contrast just before the CT. This allowed evaluation of stomach to r/o gastric outlet obstruction, yet at the same time, evaluate for PE.

    The alternative is to do a quick gastric swallow study where the patient swallows the oral contrast, gets a single view abdominal x-ray. at our shop, this is time consuming as the radiologist has to be there with the patient.

    doing a runoff to the rest of the abdomen/pelvis might be helpful, although I’m not sure how contrast timing will affect their study since radiology has been more comfortable diagnosing appy’s, diverticulitis, obstruction without IV or PO contrast. unless an intra abdominal abscess is strongly considered, po/iv contrast combo at our shop is not routinely used.

    what about a d-dimer?

    with regards to r/o acute chole, i think if you see stones on your CTA with a negative PE/obstruction study, an official ultrasound would be the next workup. i don’t think the delay in diagnosing acute chole is that significant than say missing the diagnosis of PE simply because you found stones on US.

    Re: 2nd case. The q wave information is helpful. we had a similar case just a couple of days ago, and cardiology was insisting it was not STEMI because pt had q waves. EKG was thought to be an LV aneurysm. Pt still went to cath lab, and had 100% LAD lesion. cardiology for the most part is not aware of the modified sgarbossa criteria.

    Again, great cases and summary! Keep it up!

    LA

  • Thank you for your nice comments LA.

    I think your idea of having the patient drink contrast before the PE study is a smart idea if you had a patient like this presenting a few weeks after some bariatric procedure. Gruber was very interested in discussing the best way to diagnose these post-bariatric surgery patients depending on the time of presentation. My primary interest in EM and in these cases is the cardiopulmonary stuff; but since you both already asked about it, I’ll review what our first reference suggests to do.

    For an anastamotic leak, they recommend a GI series with PO contrast versus a CT with contrast and state they are both sensitive enough.

    For acute GI obstruction, the recommend starting with a plain film to look for a massive gastric pouch with an air:fluid level.

    For stomal stenosis / obstruction, they say start with endoscopy since it can also be therapeutic. An upper GI series is their second choice rec.

    For a post-surgical GI bleed, it is apparently very difficult to find on endoscopy; and our focus should be on resuscitation +/- octreotide, since endoscopy can usually only find bleeds within the gastric pouch.

    I think Uptodate also has a very nice article on this.

    Again, thanks for your comment.

Leave a Reply