Mysterious Purpura & Neutropenia in a Cocaine User

Introduction:
The following clinical presentation is increasingly being seen in urban hospitals throughout the country. The main findings are:

• Purpuric skin lesions
  
• Severe neutropenia
  
• High fever
• Swollen glands
• Painful sores on the mouth or anus
• Lingering infections, including sore throat, mouth sores, skin infections, abscesses, thrush, or pneumonia

It's diagnosis is challenging given that this condition has many shared exam and laboratory findings as many vasculitides. In particular, these patients often have lupus anticoagulant and c- or p-ANCA positivity.

Case:
Ms. F is a 38-year-old woman with hepatitis C, remote miscarriage, and active polysubstance abuse who was admitted for MRSA endocarditis. A toxicology screen performed on admission was positive for cocaine, opiates, and benzodiazepines. She developed a deep venous thrombosis on her right lower extremity that was initially treated with heparin and an appropriate bridge to warfarin. On day 27 (day 12 of warfarin), she developed multiple discrete, stellate, purpuric macules, papules, and plaques with a bright erythematous border on her pinna, earlobes, cheeks, right breast, and bilateral proximal upper and lower extremities. Concomitant unexplained episodic tachycardia and neutropenia were noted. Skin biopsy specimens revealed leukocytoclastic vasculitis with mural fibrin deposition, neutrophilic infiltrate, nuclear dust, and extravasated erythrocytes involving superficial small vessels with pauciinflammatory luminal thrombosis in a few vessels. Synchronous tests revealed positive platelet factor IV antibody (but negative serotonin release assay), mixing studies (noncorrecting), lupus anticoagulant, and Russell viper venom time. Antineutrophil cytoplasmic antibodies (ANCAs) directed against proteinase-3 (PR-3) were 39.2 EU/mL (normal, <4.0>9/L). A subsequent urine toxicology screen on hospital day 33 was positive for cocaine, confirming in-hospital cocaine use.

Diagnosis:
This is a case of levamisole toxicity. Levamisole is an antihelmithic drug that has increasingly been used as a cutting agent for cocaine. In July 2009, the Substance Abuse and Mental Health Services Administration (SAMHSA) found the drug in over 70% of cocaine analyzed. Another recent analysis in Seattle, WA found individuals who tested positive for cocaine also tested positive for levamisole nearly 80% of the time.The Drug Enforcement Agency says it has seen a steady increase in the amount of the medication found in cocaine since 2002. Complete clinical resolution of skin lesions occurs 2 to 3 weeks after stopping levamisole and serologies normalize within 2 to 14 months. Detection of levamisole is challenging, because specific testing is necessary but not routinely available; levamisole's half-life is so short (5.6 hours) that only 2% to 5% of the parent drug is detected in urine; and the sensitivity of available testing is low. However, the clinical constellation of retiform purpura, neutropenia, lupus anticoagulant and ANCA positivity, and temporal association with cocaine use is nearly pathognomonic for levamisole toxicity. So, next time you're in the ED and you see a patient with purpura and neutopenia in the context of a positive utox for cocaine, add this to your differential!

Reference: Trimarchi et al. Cocaine-induced midline destructive lesions: clinical, radiographic, histopathologic, and serologic features and their differentiation from Wegner granulomatosis. Medicine (Baltimore) 2001; 80:391-404.

Drinking Blood

During the medicine sub-i, Dr. Ashar handed us a really interesting paper from 1940s.  I will review this paper here for both the strangeness and the results derived. 

Paper: Schiff, L., Stevens, R.J., Shapiro, N., and Goodman, S.  Observations on the oral administration of citrated blood in man. March, 1941 - Presented at American Society of Clinical Investigation

Objective: In the context of an upper GI bleed, tarry stool or frank blood in stool is assumed to signify sever hemorrhage.  The paper goes about trying to quantify how much blood actually needs to be lost in an upper GI bleed before stool quality changes.

How much blood is necessary to produce tarry stool?
The authors used 18 subjects (3 controls and 15 other patients in hospital for non-GI related diseases), who drank varying amounts of venous blood mixed with 100 to 200 mL of Vichy water (mineral water) to "help disguise the taste."  The result of the study was that grossly tarry stools were detected starting from subjects who drank at least 100 mL of blood (4/7 patients who drank 100 mL of blood had tarry stools).  Everyone who drank 200 mL of blood had tarry stools.

After a bleed in upper GI tract, how long does it take for blood to show up in stool?
On five subjects (unknown identities), 1000 to 2000 mL of blood was poured down nasogastric tubes. To prevent reflex diarrhea from irritation by stored blood products, subjects were pre-treated with codeine and atropine.  After blood was introduced through NG tube, the first bloody stool appeared starting from 4 hours to 12 hours post procedure.  Bloody or tarry stools lasted for three to five days.

After an upper GI bleed, how long does occult blood last in stool?
When two subjects drank anywhere from 75 to 250 mL of blood, the average number of days with guaiac positive stool ranged from three (for 75 mL of blood) to ten days (for 250 mL of blood).  When more than 1000 mL of blood was introduced via the NG tube, guaiac positive stool lasted for five (for 1000 mL of blood) to 12 days (for 2000 mL of blood).

Conclusion: Tarry stools can appear after as little as 100 mL of blood from upper GI bleed - not a massive hemorrhage.  It takes roughly about four to twelve hours for the blood to show up in stool after a big bleed of one to two liters (although patients were pre-treated with anti-motility agents).  Subsequent to a bleed, the occult blood can last for one to two weeks.

Hepatorenal Syndrome

Hey Shamik,
I have a patient with alcohol-induced cirrhosis and ascites. In addition, he was found to be in acute renal failure at admission with a Cr 1.5 from baseline 0.8. My concern was his kidneys, and my question was regarding hepatorenal syndrome, the most dreaded complication of cirrhosis.

I did some research, which I thought was interesting. The mechanism of hepatorenal syndrome is multifaceted and still not fully elucidated. However, following the development of cirrhosis and portal hypertension, there is increased pressure within the splanchnic circulation, leading to increased production of nitric oxide. This causes local vasodilation of the splanchnic vasculature and pooling of blood. As such, there is reduced effective arterial blood volume, which kicks in the RAAS system. Consequently, there is renal vasoconstriction and decreased GFR.

There are two types of hepatorenal syndrome: type I is very bad with a survival of 2 weeks. With type I, there is a doubling of Cr usually above 2.5 and Cr clearance is generally less than 20. Type II, however, is more insidious; Cr is usually around 1.5 and clearance around 40. It is classically associated with ascites refractory to diuretics.

So, with my patient, I gave him a fluid challenge. This was somewhat counterintuitive because he had ascites and 2+ LE edema; he appeared outwardly volume up. However, in order to rule-out hepatorenal syndrome, the key is to rule-out renal failure of other causes, namely pre-renal. Notably, I also ordered a FEurea, which is more accurate than a FENa in the context of recent diuretic use; less than 35% is pre-renal. My patient was at 34%.

His albumin was 1.0, and he had a stable iron deficiency anemia. Instead of giving 1L NS, I gave him 1 unit of blood to both increase his circulating volume as well as his oncotic pressure. Then, I gave him a 500cc fluid bolus with IV NS. I figured that it would be useful to increase his oncotic pressure first and then give fluid as the NS would stay intravascular better--no evidence but it sounded logical.

His kidneys responded, and his Cr came down to baseline. He was pre-renal. What did I learn from this case:

1. Cirrhotics can be outwardly volume up (eg., ascites and edema) but intravascularly volume down (often from aggressive diuresis).
2. To rule-out hepatorenal syndrome, first consider the history, ie, has this patient been aggressively diuresed?
3. In the absence of history, check a FEurea and/or give a fluid bolus. I liked the idea of giving blood instead of NS in the context of anemia; kill two birds with one stone (ie, anemia and pre-renal failure). Obviously though, there are risks to giving blood products...
4. If the patient doesn't respond, further assess for hepatorenal syndrome. Apparently, dopplers of the renal arteries can assist with the diagnosis. I never got to that point, thankfully.
5. Lastly, a word about diuresis. Never forget to add a spironolactone to your lasix. Remember the RAAS pathway and the importance of inhibiting aldosterone!

Antibiotics and COPD

Situations like the following occur often: A 58 year old woman with history of smoking and COPD presents with increased dyspnea and cough.  She appears exhausted.  She is treated with bronchodilators, steroids, and antibiotics.  But, why antibiotics?  Are most COPD exacerbations associated with bacterial infections?

A randomized placebo controlled trial examining the utility of antibiotics in addition to corticosteroids for acute exacerbation of COPD in a patient population without asthma or pneumonia was published by a Dutch group in 2010. 

Paper: Daniels, J.M.A. et. al., Antibiotics in Addition to Systemic Corticosteroids for Acute Exacerbations of Chronic Obstructive Pulmonary Disease.  Am. Journal of Resp. and Crit. Care Med. 181 (2010): 150-157.

Objective: Randomized, placebo controlled trial to test whether addition of doxycycline to corticosteroids improved clinical outcomes.

Setting: Two hospitals in Netherlands

Intervention: Patients in the intervention group received 7 day course of doxycycline while all patients received corticosteroid treatment starting with intravenous steroids for six days followed by slow oral taper in addition to bronchodilators.

Patient Population: Patients were all at least 45 years of age with diagnosed COPD presenting with an acute exacerbation of COPD - described as increased dyspnea, sputum volume, or sputum purulence. Patients were excluded if they were febrile or had radiographic evidence of pneumonia.  265 patients were randomized to get doxycycline (128 patients) or placebo (137 patients).  87% of the doxycycline group completed the trial while 80% of the placebo group completed the trial.  The most common reason for dropout from the trial was perceived lack of efficacy, which was more common in the placebo group.

Outcome Measured: Primary measure was treatment response (either cure or improvement in symptoms) at day 30.  Secondary measures were treatment response at days 10 and cure rates on days 10 and 30.  Other measures tracked were FEV1, CRP, and microbiological response.  Analysis was by intention-to-treat.

Results: At day 30, treatment response was observed in 61% in doxycycline group and 53% in the placebo group, which was not a significant difference (p = 0.32).  The group tracked many secondary outcomes.  At day 10, treatment response was significantly better for the doxycycline group (80%) compared to the placebo group (69%) with p = 0.03.  Regarding clinical cures, there was no difference between the groups at day 30 (51% in doxy group vs. 41% in placebo) while at day 10, more patients in the doxycycline group were cured.   On lung function tests, there was no difference  in FEV1 increases between the two groups in the short term (day 10) or long term (day 30).  Symptomatically, patients with cough or sputum purulence benefited with doxycycline compared to placebo at day 10 but not at day 30.  Microbiologically, the most commonly isolated bacteria were H. influenza (41%), S. pneumoniae (24%), and M. catarrhalis (22%).  In the intervention group, bacterial persistence rates were 31% for H. influenza, 17% for S. pneumoniae, and 9% for M. catarrhalis after completion of doxycycline course.

In subgroup analysis, those patients who did not have all three symptoms of dyspnea, increased sputum, and sputum purulence did not benefit from doxycycline treatment at day 10 or at day 30.

Conclusion: During an acute exacerbation of COPD, treatment with course of antibiotics probably does not make a whole lot of difference in the long term but results in better response rates short term.  At day 10, even with steroids alone, there was an impressive response rate of about 69% compared to 80% response rate with steroids plus doxycycline.  So, the benefit of adding antibiotics is mainly incremental, symptomatic, and experienced mostly by patients with cough and sputum purulence.  In the study, despite doxycycline treatment, there was considerable persistent bacterial colonization not wholly explained by antibiotic resistance (in vitro resistance rate for H. influenza was only 1%).  

So, in our patient presenting with COPD flare with cough and dyspnea without increased sputum or sputum purulence, there is little added benefit in prescribing antibiotics with corticosteroids.  If her cough is particularly bad, she may benefit from the addition of antibiotics.