We’re all pretty familiar with the banana bag: intravenous (IV) fluids with the addition of thiamine, folic acid, multivitamins, and sometimes magnesium. Banana bags are commonly utilized in patients at risk for alcohol withdrawal symptoms or those who present to the emergency department (ED) acutely intoxicated.
The rationale behind ordering banana bags for these patients is relatively simple–alcoholics are likely to have nutritional deficiencies related to their dietary preferences for alcohol over nutrient-dense foods, putting them at risk for complications. Furthermore, the administration of fluids is conventionally believed to help speed up sobriety. But it seems the combination of these components may be unnecessary, let’s break it down piece by piece:
1. IV fluids
Dr. Salim Rezaie did some recent mythbusting on REBEL EM regarding the utility of IV fluids for intoxicated patients1. He discussed four papers that evaluated ethanol clearance with IV fluid administration. The majority of his review focused on a recent trial that found no difference in ED length of stay, treatment times, or breath alcohol levels at 2 hours with the administration of IV fluids (compared to no IV fluids).2 The bottom line from his literature review: there’s no evidence that IV fluids expedite sobriety in patients with acute alcohol intoxication. Obviously if patients show signs of dehydration, IV fluids are a reasonable intervention.
2. Folic acid
Ethanol inhibits the absorption of folate and impairs its enterohepatic cycle.3 Ethanol abuse causes rapid declines in folate stores (within 2-3 days) and eventually leads to the development of megaloblastic anemia. Non-alcoholic individuals typically take four to five months to develop megaloblastic anemia with low dietary folate intake; this timeline can be shortened to five to ten weeks in alcoholics.3
In a recent evaluation by Li and colleagues, investigators evaluated serum folate, thiamine, and vitamin B12 levels in acutely intoxicated patients with a detectable blood alcohol level and clinical evidence of intoxication.4 A total of 77 patients, with a mean age of 46 years and mean blood alcohol level of 280 mg/dL, were included in their analysis. The authors found that no patient evaluated had a low serum folate level. However, serum folate levels are not a perfect marker of total body folate stores, as serum concentrations can be acutely elevated or depressed by a single meal or ethanol intake, respectively.5,6 A more accurate measure of folate stores is the red blood cell folate level, although this is a more expensive test.
Ethanol also inhibits the absorption of thiamine, decreases its hepatic storage, and impairs systemic utilization. This places alcoholic patients at risk for developing Wernicke’s Encephalopathy (WE) or later, Korsakoff’s psychosis. Many alcohol-related ED visits have a component of altered mental status, so WE may be in the differential diagnosis–especially if gait ataxia or oculomotor dysfunction is present. In my experience, I’ve seen providers be tempted to rule out WE because the patient has a normal ammonia level. Let’s be clear here, WE is different from hepatic encephalopathy; the term “encephalopathy” simply means a syndrome of brain dysfunction.
Difficulty in diagnosis of WE is related to the lack of quick and easy objective testing. The Caine Criteria have been proposed to aid in diagnosis, with suggestion of WE if two of the following are present7:
- Dietary deficiency
- Oculomotor abnormalities
- Cerebellar dysfunction
- Altered mental status/memory impairment
Early case series estimated the incidence of missed diagnosis of WE in a series of post-mortem necropsies at 67-80%.8 In the evaluation by Li and colleagues, a small portion of patients (6 of 39) with measurements performed had low thiamine levels (mean 130 nmol/L, SD 69 nmol/L, reference range 87-280 nmol/L).4
When left untreated, WE carries a mortality rate of up to 20% and Korsakoff’s psychosis develops in as many as 85% of survivors.8,9 As the diagnosis relies more on clinical judgment and the cornerstone of treatment, thiamine, is very low risk, empiric therapy with thiamine is likely warranted if WE is suspected.
Here’s where a pitfall comes in: using banana bags for treatment of suspected WE. The standard banana bag comes with 100 mg IV thiamine, and if this is diluted in a liter of IV fluid, it may take several hours for the whole dose to be administered. Evidence supporting “high dose” thiamine treatment is largely conjecture, and derived from several case series and a correspondence in which patients were treated with doses of thiamine greater than those found in banana bags, yet still developed WE.8,10 As a result, Chataway and Hardman recommend 500 mg of IV thiamine every 8 hours for two days, then 500 mg IV thiamine daily until oral therapy can be tolerated.10 Oral therapy with 100 mg twice daily is recommended to continue at that point until the patient is able to abstain from alcohol. A recent Cochrane Review addressing this topic found insufficient evidence to guide clinicians on optimal dosing for prevention or treatment of WE.11
4. Parenteral multivitamins
Parenteral multivitamins are what give banana bags their bright yellow color and may also be used in parenteral nutrition formulations. In addition to folic acid deficiency, low levels of vitamin B12 can lead to megaloblastic anemia. Alcoholics are commonly thought to be vitamin B12 deficient due to their dietary choices; however, this might not be the case. In the study by Li and colleagues mentioned previously, the authors found that no patient analyzed had a low vitamin B12 level.4
In addition to dietary deficiencies, ethanol inhibits tubular reabsorption of magnesium and may increase renal excretion due to secondary hypoaldosteronism.12,13 Some guidelines recommend against routine magnesium supplementation in these patients unless hypomagnesemia is confirmed.14 However, strong evidence is lacking to offer guidance on this point.
The Bottom Line
- If you’re seeing an acutely intoxicated patient in the ED, IV fluids aren’t going to help them sober up faster.
- If your patient has symptoms suggestive of Wernicke’s Encephalopathy, don’t be afraid to be aggressive with thiamine dosing (500 mg IV infused over at least 25 minutes). For prevention, a single dose of thiamine 100 mg IV over 5 minutes is a reasonable intervention.
- It’s likely sufficient to leave multivitamin and folate supplementation for outpatient, discharge prescriptions, as the consequences of these deficiencies don’t develop overnight (and aren’t likely to be fixed quickly with IV supplementation).
- Wait until confirmed hypomagnesemia before jumping to IV magnesium supplementation.
This is an excellent post. Mythbusts a lot of issues around the banana bag. Haven\'t seen the banana bag at my institution for over 8 years now, so glad that we are consistent with the literature. Notes:
- The title using \"(Myth)Busting\" is challenging for search engines to find your term. Would just list as \"Mythbusting\".
- I sometimes give MVI, folate, thiamine PO to chronically intoxicated patients who I know have poor dietary intake. They have very poor access to primary care and are poor followup candidates. Is this overkill?
- The 4 Caine Criteria, to follow along the F-style writing that Scott alludes to, can be listed in bullet points.
Expert Peer Review
Meghan, this is a very good post (I would actually recommend submitting it as a review article). I have several comments to be addressed:
- When referring to Dr. Rezaie\'s post, please add the link.
- Please add a reference for the statement in the folic acid section beginning \"Non-alcoholic individuals...\"
- When describing the study by Lit, please describe whether the patients were acute intoxication or chronic alcoholics.
- Please add a reference for the statement in the folic acid section beginning \"However, serum folate levels...\"
- Please add a reference for the first sentence in the thiamine section.
- For the sentence in the thiamine section beginning \"The common pitfall...\", Please add a reference or state \"in my experience...\" I don\'t see this pitfall. Also, hepatic encephalopathy can present with normal ammonia levels, so this really isn\'t specific to the WE diagnosis.
- Please add the normal thiamine reference range in the sentence \"In the evaluation by Li...\"
- Regarding the sentence starting \"Evidence supporting high dose...\", I don\'t think this should be stated as a common pitfall (ie, mistake). While it is true that higher doses of thiamine are not particularly harmful, the recommendation to use such high doses is largely conjecture and not based on evidence. The case series of autopsy patients really can\'t be used since doses were unknown. Please reword this section and include this reference: PMID 23818100.
- I think the word \"deficiency\" needs to be added after Vitamin B12 in the first sentence of section 4.
- In the bottom line section, 2nd paragraph: If making this recommendation, please provide administration instructions. At my institution, thiamine 100 mg IV has to go in over at least 5 minutes. 500 mg would have to be put in some fluid to go in anyway, so whether fluid helps them sober up faster may not be relevant in a sicker patient with possible WE.
- In the last paragraph. 1st sentence: While this statement may be true, there wasn\'t any evidence provided to support that IV administration doesn\'t acutely raise levels.
Thank you for the opportunity to review this submission. My overall recommendation is to accept with the suggested changes and just make sure not to overstate the evidence with too much opinion.
Bryan, I appreciate all your feedback on the post! I think with the revisions you suggested it will be much stronger. I\'ve addressed your suggestions below with the number it responds to (above) in your comments:
- This link is available in the reference section. If we’d like to also link to it in this section that’s fine, I’m just not sure how to do that
- Reference added
- The description of patients was added
- Reference added
- Reference added, no PMID available as this journal was not indexed by Pubmed until one year after this article was published
- Reworded as suggested
- The normal reference range used in the paper was added
- Reworded as suggested, with Cochrane Review included
- Not sure this reads correctly? I reworded the sentence to hopefully aid in clarification.
- Recommendations for how long to infuse 500 mg of IV thiamine were added
- Not entirely sure which section you\'re referring to here, perhaps the bullet point about thiamine in the bottom line section? I reworded this to hopefully be more in line with what I think you\'re suggesting.