September 2015 E-Journal Club

September 1, 2015 by

We had a great group of trainees here in September who hailed from Florida (3) and Texas. They were treated to some really wonderful early Fall weather here in Charlottesville. Journal club this month dealt with the risks of bleeding from feeding tube placement in patients with thrombocytopenia.

June Citation:

Patel RP, Canada TW, Nates JL. Bleeding Associated With Feeding Tube Placement in Critically Ill Oncology Patients With Thrombocytopenia. Nutr Clin Pract. 2015 Aug 21. [Epub ahead of print].

This was a single center, prospective observational study in a 53-bed adult medical oncology ICU that evaluated the incidence of bleeding complications with feeding tube placement in patients with thrombocytopenia (TCP: platelet count <150,000/uL), compared to patients without TCP. Patients who were scheduled to have a feeding tube placed had demographic, laboratory and bleeding and blood product information recorded.

The primary outcome was to compare the incidence of overt and clinically important bleeding complications within 72 hours of attempted feeding tube placement among critically ill oncology patients with and without TCP. Clinically important bleeding was defined as overt bleeding with at least one of the following:
• decreased in hemoglobin by ≥2 g/dL in 24 hour period within 72 hours of feeding tube placement;
• failure of hemoglobin to increase by number of units of packed red blood cells given, minus 2, within 72 hours of feeding tube placement, or
• decrease in systolic blood pressure by ≥20 mm hg within 24 hours of bleeding symptoms.
Secondary outcomes included the total number of attempts for feeding tube placement, number of aborted feeding tube placements, number of displaced feeding tube requiring replacement or readjustment, number of blood product transfusions within 72 hours after feeding tube placement, and number of patients that reach goal nutrition by ICU day 7 or discharge.

Inclusion and Exclusion Criteria were:
Inclusion criteria:
Medical oncology patients aged ≥ 18 years, nasal/oral feeding tube planned, and ICU stay at least 24 hours within the six month study period.

Exclusion criteria:
Pregnancy, ICU admission with GI bleeding, active bleeding or known bleeding diathesis (ex hemophilia), received treatment dose of anticoagulation or antiplatelet therapy.

Major Results:
A total of 64 patients during the 6 month study period had a feeding tube placed; 5 were excluded and 59 completed the study (42 with TCP and 17 without TCP). There were significantly more patients with TCP who had hematologic malignancy and a low hemoglobin at baseline, compared to patients without TCP.

Most of the patients in the TCP group had grade 3 or 4 TCP (plts <50,000/uL). Blood products (packed RBC and platelets) were transfused 24 hours prior to feeding tube placement in 86% of patients with TCP compared to 12% of patients without TCP. Feeding tubes were placed nasally in most patients (95% TCP, 94% non-TCP).

There was no significant difference in total (9.5% TCP vs 5.9% non-TCP, p= 0.57), overt, or clinically important bleeding complications within 72 hours of attempted feeding tube placement between groups. Overt bleeding occurred in 3 patients with TCP and none in the non-TCP group (not significantly different). The TCP patients that experienced bleeding all had hematologic malignancy and 3 had grade 4 TCP the day of feeding tube placement.
There were no differences in the number of total feeding tube placement attempts, patients with repeated and aborted attempts, displaced feeding tubes requiring replacement or adjustment, nor in the number that reached EN goal by ICU day 7 or discharge.

Author’s Conclusions:
“Critically ill oncology patients with TCP do not appear to be at a higher risk for overt or clinically important bleeding complications either GI or point of entry after feeding tube placement as compared with those without TCP, which may be related to blood product transfusion within 24 hours prior to feeding tube placement. A large prospective study may help determine the bleeding risk associated with different grades of TCP and if prophylactic transfusions prior to feeding tube placement decrease bleeding complications.”

This was a single-center, single-ICU observational study with a limited number of patients. Due to the study being conducted within a cancer center, there were far more patients with TCP enrolled and only a small number without TCP. The study enrolled more than the 38 patients that were projected to be needed to show significance between groups, but this power calculation was based on an expected bleeding rate of 50%. The authors point out that due to the low rate of bleeding observed, it would have taken at least 2300 patients per group to be able to detect a 3.6% difference in bleeding complications.

Although this was not a large randomized study, with some unavoidable weaknesses that limit the take-home conclusions, our group felt that is was still a very valuable study.

Our heme-oncology population is one of the last bastions for parenteral nutrition use, in part due to the perceived risks for feeding tube placement in patients with TCP. There is such limited research in this population, which means that we have limited data to counter the dogma that small-bore feeding tube placement constitutes a clear and present danger in patients with TCP. The results of this study are valuable in that the major limitation (very low rates of bleeding observed—and that with multiple attempts in quite a few…) suggest that it may be reasonably safe to place NG access in patients with TCP within 24 hours of a transfusion.

Of course, there is still a need to have randomized studies that compare the use of EN, PN and delaying nutrition support in this population, before we can understand the full effect of these therapies on patient outcomes that matter.

Our group did discuss several other limitations of this study, including the fact that a high number of patients required repeat attempts at feeding tube placement (32 out of 42, or 76% in the TCP group and 11 out of 17, or 65% in the non-TCP group) – and the methods seemed to suggest that only bleeding related to the initial attempt was recorded. Another thing that we discussed was our surprise that in a nutrition support journal, the article came to press without any mention of the type(s) and diameter of the feeding tubes used. Finally, in a 52 bed ICU, 64 feeding tubes seemed like a small number over a 6 month period. It would be interesting to know if patients in this unit ever get salem sump type tubes placed for medications, etc.

Our Take Home Message (s)
1. Nasal or orogastric feeding tube placement in critically ill oncology patients with TCP who have received transfusions within the past 24 hours does not appear to be associated with a high rate of bleeding complications.
2. There is a need for large randomized studies in oncology patients to compare EN vs. PN in patients who require nutrition support.
Other News on the UVAHS GI Nutrition Website: (

Upcoming Webinars 2015:
• Tuesday, September 22: Practical Implementation of a Malnutrition Documentation and Coding System in an Acute Care Hospital–Wendy Phillips, MS, RD, CNSC, CLE, FAND
• Wednesday, October 14: Refeeding Syndrome–Carol Rees Parrish, MS, RD
• Tuesday, November 17: Hyperglycemia in the Acute Care Setting–Joe Krenitsky, MS, RD

Latest Practical Gastroenterology article:
White L. D-Lactic Acidosis: More Prevalent Than We Think? Practical Gastroenterology 2015;XXXIX(9):14.

Joe Krenitsky MS, RD
Carol Rees Parrish MS, RD

PS – Please feel free to forward on to friends and colleagues.