When to stop or restart antiplatelets and anticoagulation during GI bleed?

Let’s say one of your patients is an 80 year old male with a history of CAD, HTN, HLD, and CKD stage III presenting with nausea and abdominal pain. He also complains of a “choking” sensation for the past week. You notice that his hemoglobin is around 9 g/dL; a couple of months ago it was 13 g/dL. Given his symptoms, you consult GI and he undergoes EGD that shows LA grade C esophagitis and some gastritis that gets biopsied. He is started on a PPI BID, and as you are going through his discharge med list, you wonder if his aspirin, which was held on admission, is safe to be restarted–and if so, when?

Luckily, to this particular scenario, the answer is pretty clear. He doesn’t have a high bleeding risk otherwise, and in patients with CAD, as shown in two meta-analyses cited by this Hospitalist piece, it’s safer to restart aspirin immediately rather than hold it indefinitely. (Aspirin probably didn’t need to be held in the first place.)

Other scenarios are not so clear. What if this wasn’t just a baby aspirin, but warfarin or another agent like dabigatran or apixaban? What if the patient had an NSTEMI and got a stent 3 months ago? What if the bleeding were so severe he was getting daily blood transfusions? In the graphic below, I have tried to show that stopping and resuming antiplatelets and anticoagulation is about weighing benefits of stopping anticoagulation vs. risks of thrombosis for each individual patient–there is no blanket statement that can be made.

Generally speaking, for resuming antiplatelets and anticoagulation, you should do it as soon as hemostasis has been achieved, or when endoscopy shows no active bleeding. The ACC recommends restarting warfarin within 24 hours after hemostasis (on a bridge if needed). There’s no data to support restarting a DOAC within a certain amount of time, but, extrapolating, it’s probably fine to restart within a day.

For information about specific medications and specific endoscopic procedures, the ASGE has an excellent practice guideline published in 2016. Just to entice you, here’s an example of one of the tables:

Checking residuals (gastric residuals, that is)

Ah, the fun tradition of “residual 250 cc…hold TF?” This national survey shows that there is very wide variance in how residuals are checked, and how much is “concerning.” 200 cc? 250 cc? 500 cc? Going once, going twice…

Disclaimer: the information below is about adults. Who knows if the answers are the same for a pediatric population? Not me.

This ACG clinical guideline on nutrition in the hospital setting opened my eyes (mostly to how much we don’t know about managing nutrition). The authors write that checking residuals “has been shown to be a poor marker of true gastric volume, gastric emptying, risk of aspiration, pneumonia, and poor outcomes.” The sensitivity of residuals for predicting aspiration is 2-8%. This JAMA study shows that in the ICU, checking residuals didn’t affect development of ventilation-associated pneumonia. In other papers, one of the authors has advocated for the abolishment of checking residuals in nursing practice (see this link for good, practical information).

When is checking residuals useful?

Checking residuals is thought to be most useful in the first 48 hours of initiating tube feeds, as patients are getting increased to goal rate. If the residuals are building up (see below), it might be an indication to pause the tube feeds and make sure there’s nothing wrong with the tube, there are no tube feeds leaking into the peritoneum (this is actually something terrifying that can happen), etc. If your patient is intubated/sedated/obtunded, residuals are the only measure you have–but if the patient is awake and alert, you could talk to them and ask if they’re having worsening nausea, bloating, etc.

How high is too high? When do tube feeds need to be held for residuals?

I don’t know where the adage, “hold TF for residuals >200 cc” comes from. Abbott Nutrition states that when residuals > 500 cc, consider holding tube feeds. (For perspective, the human stomach can accomodate +/- 1 liter.) Check your institutional hospital policy, but…the general message of this post is that you shouldn’t be too worried.

Does checking residuals have harmful effects?

It is known that protein clots when the pH <5.0 (stomach acid pH is usually between 1.5-3.5). Could checking residuals lead to refluxing clotted protein, increasing the likelihood of clogging the tube? This small study (n=30) says, yes. In the group undergoing regular residual checks, there were 10 clogging events, compared to 1 in the no-residual checks group.

Checking residuals can also lead to tube feeds being held unnecessarily–if higher residuals aren’t a good sign of impending aspiration, we’re just holding tube feeds for hours at a time, depriving patients of the nutrition they’re supposed to be getting.

When is colonoscopy indicated for colonic ischemia (ischemic colitis)?

The underlying pathophysiology of colonic ischemia (frequently called “ischemic colitis”) is much like that of ischemia anywhere else in the body: ischemia of the heart (MI), ischemia of the kidney, ischemia of the brain (stroke), etc. The reason it should be called “colonic ischemia” is because not every case of ischemia results in colitis, but the type of injury is the same.

The incidence of colonic ischemia cannot be accurately stated, because so many cases are relatively “meh” or benign. Someone gets some cramping, maybe they pass a little blood in their stool, things get better after a day or two on their own. That makes strongly evidence-based guidelines hard to come by. This is my caveat for the answers below. My source for this post is the 2015 ACG clinical guideline.

But you’re probably wondering, “Do I need to consult GI for every single case of suspected colonic ischemia? Do I need to ask if colonoscopy is indicated for every single patient?”

Appropriate consultation is an art, of course. GI consultation should not be automatic, but for ischemia, which can be vague and multifactorial, never shy away from consulting GI just because “it’ll get better on its own.” A consultant can help identify contributing factors, like meds that should be discontinued, workup for rheumatologic or hematologic disease, suggest other input from cardiology, etc. Think about the following factors:

  • How confident are you about the diagnosis? Some cases are slam dunks: cramping abdominal pain, some red stools, double halo sign on CT in a patient with a history of vascular disease. But a lot of these findings are non-specific. The differential includes diverticulitis, IBD, infectious colitis, and even malignancy. If you want to rule out another cause and prove it’s ischemic injury, ask GI about more imaging vs. colonoscopy.
  • How severe is the case? Mild cases will likely resolve on their own after a couple of days. Moderately severe cases (see algorithm below) seem most likely to benefit from the additional diagnostic of colonoscopy. Colonoscopy can show the extent of ischemic damage; in particular colonoscopy can visualize right-sided ischemia and pancolitis, which are associated with worse outcomes. As discussed below, anyone with severe findings/peritonitis should NOT have colonoscopy.
  • Is this the first time it happened, or is there a pattern of recurrence? If it’s a mild, first-time case, it’s probably not a big deal. Multiple episodes deserve a more detailed workup with endoscopic evaluation and histopathology to prove ischemic injury.
  • Are there implications for the future? Because of the differential diagnosis above, consider your patient’s other symptoms, medical conditions, and overall state of health. Would it matter for this particular case whether other types of colitis or masses were excluded?
From Brant, Feuerstadt, ACG guideline

NB: no prep is indicated for colonoscopy being done for colonic ischemia.

When should colonoscopy NOT be performed? The yield decreases substantially >48 hours after symptoms begin (to about 30%) so colonoscopy won’t be as helpful then. Colonoscopy is not recommended in severe cases, peritoneal signs, or diverticulitis because of theoretical risks of insufflation worsening pressure-related injury or causing perforation.

As a corollary–the ACG guidelines from 2015 also state that there is no evidence to support specific rules on when antibiotics are indicated and how long to give them for. Antibiotics not only prevent bacteremia/sepsis, they probably also help reduce the body’s inflammatory response to bowel injury…but there’s not a whole lot of human studies to show how helpful or unhelpful they are. Therefore, the guideline authors recommend antibiotics that cover bowel flora (like a third-gen cephalosporin + metronidazole) in “moderate” and definitely “severe” cases of colonic ischemia. They suggest 72 hours, but I think if someone is rapidly clinically improving, it’s reasonable to stop antibiotics after 1-2 days.

Do you have to check an ammonia level if you’re concerned about encephalopathy?

Sometimes, you might be told to order a serum ammonia level on a patient who is encephalopathic, whether they have a history of liver disease or not. Why? Does the ammonia level actually matter?

I would argue for most cases, no. There are specific situations in which serum ammonia is a good prognostic or diagnostic test:

  • Acute liver failure (or acute fatty liver of pregnancy)–associated with risk of cerebral herniation and poorer outcomes
  • Patients who have inborn errors of metabolism–can suggest a diagnosis of urea cycle disorders
  • Reyes syndrome–can be suggestive of this diagnosis
  • Monitoring of ammonia-lowering therapy–in the research phase as far as I know

If you have a patient with cirrhosis or chronic liver disease who comes in with hepatic encephalopathy, though, the serum ammonia level is almost certainly not going to change your management. Ask yourself: If the ammonia level is 20 in a patient with major hand flapping, will you stop their lactulose? If it’s 140 in an alert cirrhotic will you get a head CT to look for cerebral herniation? Probably no, and no. It’s generally agreed that a serum ammonia level >100 is probably bad. And then there’s the question of arterial, venous, or partial pressure–better to just not get it in the first place. It also costs anywhere from $30-50.

Study results vary on whether the serum ammonia level is correlated with encephalopathy. For example, this study of about 120 patients suggests that it is, whereas this study of about 20 patients suggests not,  and this study says yes for ALF but not for patients with chronic liver disease. Even if the majority of evidence tips (pun not intended) towards ammonia levels and encephalopathy being correlated, no one has been able to define specific numerical cut-offs for what levels correlate with mild, moderate, or severe hepatic encephalopathy. So ammonia levels remain clinically not useful for managing most cases of hepatic encephalopathy. I rest my case with this clinical vignette and discussion by Phillip Ge and Bruce Runyon.

NB: I recommend this comprehensive review of the physiology of ammonia: it covers where and how ammonia is made, and how ammonia acts as a neurotoxin in the body. I can honestly say it was the first time I enjoyed reading about glutaminases.

If my patient is on antibiotics and has a history of C. difficile, should I put them on prophylactic oral vancomycin?

Example: you have a lovely 68-year old male with a history of two episodes of C. difficile infections after getting antibiotics in the setting of a hip surgery three months ago. He is admitted to your service with pneumonia and you decide he should get a 5-day course of levofloxacin. He has no abdominal pain or diarrhea (yet). Would PO vancomycin would help prevent recurrence of C. diff? 

In this case, the answer is a weak “yes.” There is no randomized controlled trial data showing that PO vanc prevents recurrence. However, there is a retrospective study of 172 patients in Quebec with a diagnosis of C. diff who were exposed to antibiotics for whatever reason within 90 days of their diagnosis. Those patients who had at least one recurrence of C. diff who got PO vanc had half the risk of getting C. diff again. Because the study looked at patients within 90 days of diagnosis, it’s unclear whether patients who had C. diff from a longer time ago would get the same benefits. A second study from 2016 of over 200 patients showed that patients who got PO vanc had a 4% risk of recurrence, whereas patients who didn’t had a 27% risk. Notably, this study only surveilled patients for 4 weeks after their antibiotic course, and the patients weren’t randomized–I got the sense that the patients who were offered PO vanc were “sicker.”

The regimen that I have seen most often is 125 mg PO vancomycin QID, but the second study, for example, reported 125 mg PO vancomycin BID or daily, so regimens are all over the place.

How long should patients continue PO vanc prophylaxis for? Again, no strong right answer. I have seen patients told to take PO vanc for only as long as they are on other antibiotics, for a week after finishing, or to do a slow taper/pulse down. However, there is a clinical trial  that has set the time course of PO vanc as 5 days after stopping other antibiotics, which sounds pretty good. What if it’s longer than a week, or a month? I did have a patient who required lifelong oral suppressive antibiotics…and we made the decision to keep her on PO vanc lifelong as well.

Do probiotics prevent recurrence of C. diff? The 2018 IDSA guidelines cite unclear evidence on whether probiotics prevent recurrence. While probiotics are fine for the general population (might as well try it if it won’t hurt), there are rare cases of fungemia/invasive infection that have been reported, in patients who are immunocompromised, have PICC lines, or are otherwise severely ill.

How is fecal microbiota transplant (FMT) performed?

Fecal microbiota transplant (FMT) is an effective and durable method of treating recurrent or refractory C. difficile infections. There are several routes of administration:

  • Nasogastric tube: patient takes PPI beforehand, NG tube is inserted, 50-60 cc of fecal slurry is pushed into the tube
  • Endoscopy: 200-250 cc of fecal slurry is delivered by flex sigmoidoscopy or colonsocopy (to the cecum)
  • Capsules: frozen capsules of slurry taken over several weeks (experimental)

Is there any difference between these different routes of delivery? Capsules were first devised at Massachusetts General Hospital; clinical trials are ongoing and they are not widely available. NG tube vs. endoscopy have been compared, and although some have not found a difference, some conclude that endoscopy is superior. For example, this study of 50 patients from the University of Alabama at Birmingham showed that patients who had endoscopy delivery of FMT had a greater rate of “cure” (improvement of symptoms in 2 weeks) and fewer repeat FMT treatments. However, you do have to evaluate each patient individually: some patients may find the idea of an NG tube too repulsive, some may be too sick to undergo endoscopy, etc. Importantly, FMT is felt to be safe for immunocompromised patients, too.

Tangent: according to the 2018 IDSA guidelines for treatment of C. diff, PO vancomycin or fidaxomicin are now considered agents of choice for a first episode of C. diff, no matter the severity. Keep your metronidazole on the shelves! It’s no longer recommended.