Saturday, January 26, 2013

Your urine tox screen was positive for cocaine...

Motivation: Awkwardness arises when a patient with bacterial endocarditis denies use of IV drugs, but his or her urine tox screen comes back positive for cocaine. Urine tox screens come with the disclaimer from the lab that these tests are not definitive, and further confirmatory testing is required. When a urine tox screen comes back positive for cocaine, how accurate is this result? Surprisingly, few studies have investigated this common issue, and it is difficult to compare study results to the exact test used in the hospital, as there are many different urine tox screen assays...but I did manage to find one article that dealt with this question:

Study: Linder MW, Bosse GM, Henderson MT, Midkiff G, Valdes R. Detection of cocaine metabolite in serum and urine: frequency and correlation with medical diagnosis. Clin Chim Acta. 2000 May;295(1-2):179-85.

Study design and results: A retrospective chart review was performed at a level 1 trauma center. 500 sequential subjects with urine tox screens in the ED were included. Urine specimens positive for the cocaine metabolite (benzoylecgonine-BE) were confirmed with GC/MS technique. 54 patients were positive for BE in the urine; of these patients, 10 had a medical diagnosis of acute cocaine intoxication. 25% of these 54 patients reported using cocaine in the past, and 1 reported use of cocaine on day of testing. Based on their chart review, the authors estimated sensitivity of 100%, specificity of 90.6%, NPV of 100%, but a PPV of only 18.5% for identifying acute cocaine intoxication in the urine test. (Linder et al., 2000)

Thoughts: This study addresses the relationship between cocaine positivity in the urine test and acute cocaine intoxication. I would have been interested in the relationship between cocaine positivity in the urine test and recent cocaine intake in general-but this is difficult because it would rely on actually knowing whether or not a patient took cocaine (which could be practically impossible to know with 100% certainty) or comparing to a gold standard (which is possible, but once again, no test is perfect). From this study, I take home the point that a negative result is useful, but a positive result may not be accurate and requires further testing. This is a key point to remember in keeping the trust in a therapeutic patient-physician relationship, when a patient lets a physician know in good faith that he or she has not taken cocaine, but a positive cocaine test results. One should then verify with further tests if a true positive result would change medical management.

Monday, January 21, 2013

Lyme cardiomyopathy

Question: Can Lyme disease be associated with cardiomyopathy?
Motivation: A patient presents with new onset cardiomyopathy and heart failure. She lives in a Lyme endemic area. EKG does not show any conduction abnormalities - which is the more common manifestation of Lyme disease in cardiac disease - but ECHO shows dramatically decreased LVEF. This study examines the evidence for Lyme cardiomyopathy from a pathological perspective.
Design and results: This study was a pathological/molecular analysis of endomyocardial biopsy (EMB) specimens for the Borrelia (Borrelia burgdorferi sensu lato) genome in cases of dilated cardiomyopathy (DCM). The authors compared EMB of patients with new-onset DCM and of patients with CAD (as controls). Specimens were analyzed by PCR and EM. There was a higher frequency of Borrelia burgdorferi sensu lato in the DCM (24%) versus the control group (0%) (p=0.035), while CMV and parvovirus B19 were similar in both groups. (Kubanek et al., 2012)
Thought: The presence of the Borrelia genome in EMB specimens of DCM does not PROVE causation - it does not necessarily prove that Lyme disease causes DCM. Lyme disease may be incidentally found in DCM, and/or related to another etiological factor for DCM. It is also important to ensure in these cases of DCM that there does not exist a more obvious etiology for cardiomyopathy. Nevertheless, I find these data supportive of the hypothesis that Lyme disease can be an underlying pathological etiology for cardiomyopathy. I will be sending off Lyme serologies for patients from endemic areas who present with new cardiomyopathy, especially if it is DCM.

Sunday, January 20, 2013

Restricting the transfusion threshold in acute UGIB

Question: When to transfuse in acute upper GI bleed (UGIB)?

Motivation: A common presentation in medicine, and unclear guidelines for when to transfuse. 

Study: Villanueva C, Colomo A, Bosch A, Concepción M, Hernandez-Gea V, Aracil C, Graupera I, Poca M, Alvarez-Urturi C, Gordillo J, Guarner-Argente C, Santaló M, Muñiz E, Guarner C. Transfusion strategies for acute upper gastrointestinal bleeding. N Engl J Med. 2013 Jan 3;368(1):11-21. doi: 10.1056/NEJMoa1211801.

Study design: This study randomized patients with severe UGIB (and hence high rebleed risk) to transfusion when Hb < 7 g/dL (restrictive) and Hb < 9 g/dL (liberal). The study took place in a hospital in Barcelona (so this was not a multi-center study). Patients were transfused 1 u pRBC until they met their transfusion thresholds, and were also followed up with further treatment (endoscopic therapy, PPI, and/or treatment of portal HTN or esophageal varices when indicated). Deviations for clinically appropriate transfusions (such as symptoms, active bleeding, surgical intervention) were also allowed. Primary outcome was rate of death, while secondary outcomes included any further bleeding or complications. Both groups had similar baseline characteristics (including baseline Hb), and were analyzed with intention-to-treat design. (Villanueva et al, 2003)

Results: Rate of mortality at 45 days was lower in the restrictive group (5%) versus the liberal group (9%) with p = 0.02. Subgroup analysis showed that this difference was more dramatic among patients with cirrhosis and Child-Pugh class A or B. Rate of continued bleeding was lower in the restrictive group (10%) versus the liberal group (16%) (p=0.01). Length of stay in the hospital and rate of complications (especially transfusion reactions and pulmonary edema) were also statistically significantly lower in the restrictive versus the liberal group. Note that deviations from transfusion threshold did occur more in the restrictive group, but in <10% of cases. (Villanueva et al, 2003)

Take home point: For patients who present with acute UGIB, setting a restrictive transfusion threshold of Hb < 7 is appropriate. It will, of course, continue to be important to transfuse (as with some of the patients in this study) when otherwise clinically appropriate (i.e. if necessary for surgical/procedural intervention, symptomatic, or severe active bleeding), but trying to use a lower transfusion threshold is a decision that is justified by the evidence presented in this study.



Saturday, January 12, 2013

Is the 2012 Flu Vaccine Useful?

Motivation: The flu is rampant.  Emergency rooms are full.  And, what about the vaccine?  Did the vaccine miss the important strains or did most patients avoid the vaccine?  Speculation on this topic is common, and the answer is important for public health preparation for the year ahead.

As way of background, this year's vaccine consists of three strains:

  • H1N1 virus: Type A influenza (A/California/7/2009)
  • H3N2 virus: Type A influenza (A/Victoria/361/2011)
  • Type B influenza strain: B/Wisconsin/1/2010 from B/Yamagata strain of viruses
Methods: Weekly Influenza Surveillance Report by CDC available at http://www.cdc.gov/flu/weekly.

Results:
Flu Positive Tests: In the week from December 30 to January 5, 12,876 samples were tested by collaborating laboratories, and 4,222 (32.8%) tested positive for influenza. 

Influenza Strains:  Between December 10 to January 5, the distribution of viral strains as tested in laboratories is as follows:
- 7340 samples (49.77%) of H3 strains of Influenza A
- 4328 samples (29.35%) of Influenza A (subtyping not performed)
- 2959 samples (20.07%) of Influenza B
- 120 samples (0.81%) of H1N1 strain of Influenza A

Detailed Antigenic Subtyping: Since October 2012, 521 influenza viruses have been characterized:
- 17 strains of 2009 H1N1 (included in vaccine)
- Of 327 strains of H3N2, 325 strains (99.4%) were A/Victoria subtype included in vaccine.  2 strains were of different subtype.
- Of 177 Influenza B subtypes tested, 118 strains (66.7%) were of the B/Wisconsin type included in vaccine.  59 additional strains (33.3%) were not included in the vaccine.

Neuraminidase Inhibitor Resistance: All tested strains in the US were susceptible to oseltamivir and zanamivir.  

Discussion: So far this season, the majority of virus strains are of the Influenza type A variety.  At least with the limited strain subtyping, most of the strains that are circulating were included in the year's vaccine.  From the surveillance data, the cause for this year's widespread flu infections is either lack of adequate vaccination or lack of efficacy of the vaccine.  Alternatively, the explanation could be a mix of the two.  In randomized trials, the efficacy of vaccines is about 50-70% with lower efficacy in people with high-risk medical comorbidities.  With a large unvaccinated pool and many elderly people, the disease could be propagated by both the unvaccinated and the elderly with failed efficacy.  In the next year, focus should be on higher levels of vaccination within the population.