The Danger of Hospital Food

In the book we discuss how high blood sugar is a powerful predictor of poor outcomes in hospital patients.

Well, Abby’s grandmom is in the hospital with pneumonia. She is being fed with intravenous glucose, and has developed high blood sugar – 160 mg/dl – despite provision of insulin.

Abby says the doctors aren’t concerned. But they should be.

High Blood Sugar Causes Pneumonia

High blood glucose and high insulin both suppress immune function. As a result, pneumonia rates are much higher in hospital patients with high blood sugar. If hyperglycemia helps cause pneumonia, it’s surely a danger in someone who already has pneumonia.

A recent paper investigating the health effects of high blood sugar during total parenteral (meaning intravenous) nutrition gives us a quantitative assessment of the risks of elevated blood sugar. [1]

They report that pneumonia was the number one consequence of elevated blood sugar from intravenous feeding:

In multivariate analysis adjusting for age, sex, and history of diabetes, the blood glucose within 24 h of TPN >180 mg/dl was associated with increased risk of pneumonia (OR 3.6, 95% CI 1.6–8.4) and acute renal failure (2.2, 1.02–4.8 1) compared with patients with blood glucose <120 mg/dl. [1]

In other words: If 24 hours after intravenous feeding is started, blood glucose rises over 180 mg/dl, the likelihood of subsequently contracting pneumonia is increased 3.6 fold.

Abby’s grandma had blood glucose of 160 mg/dl, which is still at elevated risk [1]:

Here the measure is mortality, not pneumonia, but we can see that blood glucose in the range 150-180 mg/dl 24 hours after initiation of TPN is quite a bit more dangerous than blood glucose levels below 150 mg/dl.  Mortality rates are almost 50% higher.

This is not a new finding:

TPN therapy has been associated with increased risk for infections and mortality (2,1013). The increased risk of complications appears to be related, among other factors, to the development of hyperglycemia (4,14). Observational studies have reported a 33% mortality rate in TPN patients who developed hyperglycemia (15), as well as an increased risk of cardiac complications, infections, systemic sepsis, and acute renal failure (3,4,6). In agreement with these reports, we found a strong correlation between TPN-induced hyperglycemia and poor clinical outcome. [1]

Some of the reasons hyperglycemia is dangerous involve immune suppression:

The mechanisms underlying the detrimental effects of hyperglycemia relate to alterations in immune functions and inflammatory response (16,17). Hyperglycemia impairs leukocyte function, phagocytosis, and chemotaxis (18). Hyperglycemia also increases counterregulatory hormones, inflammatory cytokines, and oxidative stress (16,17), which can lead to endothelial dysfunction and cardiovascular complications (17). [1]

Significance for Treatment

Patients who need intravenous feeding, in pneumonia cases presumably due to high choking risk or need for oxygen, have great difficulty avoiding negative health consequences. But the risks of intravenous feeding are increased by suboptimal formulations that have too much sugar and too much omega-6 polyunsaturated fat. Such formulations can strongly suppress immune function, especially if they produce hyperglycemia.

I hesitate to second-guess doctors who are on the scene and privy to case knowledge, but I think the evidence is pretty strong for the dangers of hyperglycemia. If Abby’s grandma has enough muscle and fat, fasting and receiving parenteral water, vitamins, and minerals might be a better strategy. At a minimum, glucose provision should be reduced or insulin increased to reduce blood glucose levels.

Doctors are Enslaved to Faulty Practice Standards

The paper I cited is from 2010, but the dangers of hyperglycemia have been known for decades. Yet hospitals still commonly induce it.

Our current medical institutions seem to have left doctors terrified of deviating from standard practice, even if standard practice is known to be harmful. And new, better practices can’t be adopted until proven in costly clinical trials, even though the existing practices were adopted without such trials.

I wonder what Dr. House would do in Abby’s grandma’s case? And if Dr. House would do the right thing for the patient in defiance of standard practice, how long would he keep his medical license or his malpractice insurance?


[1] Pasquel FJ et al. Hyperglycemia during total parenteral nutrition: an important marker of poor outcome and mortality in hospitalized patients. Diabetes Care. 2010 Apr;33(4):739-41.

Leave a comment ?


  1. Paul,

    Interesting topic, and one far more complicated than that paper or your analysis would lead one to believe.

    yes for decades the association of hyperglycemia with higher mortality has been known. unfortunately as with many things association and/or correlation does not mean causation.

    Greet van der Berghe published a series of excellent trials in the new england journal of Medicine in 2000, 2005 and a bunch of papers elsewhere that seemed to indicate that treating patients to normoglycemia in the ICU reduced mortality and improved a number of other markers. however, as many people pointed out, the control mortality in her trials was about 50% higher than in the rest of the ICUs in the world and perhaps there was a control group issue. the issue that most of us felt was the problem was TPN administration. Most doctors i know would never give TPN. last year the German SepNet group published 10 years of data showing that patients who are NEVER fed have superior outcomes to those given TPN. terrible stuff that TPN.

    last year, the NICE-SUGAR trial was published in NEJM showing that targeting blood sugars of 140-180 mg/dL for ICU patients was superior to 70-110 mg/dL, with 1/30 patients in the normoglycemia arm dying due to the treatment.

    so, after a decade of treating patients based on a single center trial we find out that we are likely doing more harm than good.

    for non-icu patients, there are no meaningful data, though most hospitals are or have adopted a standard of keeping the blood glucose under 150mg/dL (also in the absence of good data).

    the take home point being, that paper is interesting. many of us have interpreted the van den Berghe trials from Belgium as demonstrating that anyone on TPN should receive insulin to bring their glucose into the normal range though that is purely speculative. your anecdote is interesting and sad, is likely to indicate that her doctors should be more focused on feeding her gut rather than using TPN, but that the control of blood sugar to normal should not be assumed to be helpful because a paper shows a correlation with high mortality. you might just find out that hyperglycemia is an epiphenomenon related to something we aren’t smart enough to understand yet and that treating it hastens death.

  2. Paul-I’m reading a great book by Jenny Ruhle called Blood Sugar 101. She explains in the book that there is a significant increase in neuropathy in people who’s 2 hour BG is over 140 mg/dl after an OGTT. When under 200 mg/dl, the small nerve fibers tend to be most significantly damaged.

    As a chiropractor and acupuncturist, the majority of my patients are overweight and de-conditioned. Jenny Ruhle claims that people are overweight due to blood sugar regulation problems, most notably insulin resistance (as opposed to their blood sugar problems being due to their being overweight). If this is true, I wonder how many of my overweight patient’s pain issues are caused by various levels of neuropathy of small nerve fibers? Could it be that much of people’s insidious onset, soft tissue complaints are due to their diet (and gene) induced peripheral neuropathy? Considering that most people (even the diabetics!)stop to take a piece (or several) of candy from an open candy bowl we having available to all patients (not my idea), I think this might be a possibility. The SAD is alive and well.

  3. Very good topic Paul, I had an aunt who got paralyzed due to vascular dementia or Alzheimer or a mix of both. Guess what they gave her? Yes, it was intravenous glucose; a high concentration of the stuff that brought her this stroke.

    The SAD-meals in hospitals are not really helping the patients to get better. Good point that you brought it under the attention. I’ll put a link on our Facebook page. VBR Hans

  4. My father was in the hospital some time ago for pneumonia. His blood sugars were elevated because he was given steroids so that he could breathe better. The steroids likely saved his life or kept him of a ventilator – the blood sugars weren’t in excess of 150 and he seemed to have no long term sequelae, though he was septic at one point and thus was at high risk of dying.

  5. Dear Dr. Dan,

    Thanks for the expert commentary. It’s always risky to comment without a thorough literature review and I’m grateful that you’ve taken the time to provide an overview.

    Our book discusses the dangers of the lipids in TPN, where the need for liquid oils and the bypassing of the digestive tract create great problems. So there’s no doubt that TPN is a problem apart from the hyperglycemia. I would agree that an extended fast is better than TPN.

    I’m not arguing that normoglycemia is optimal; but there seems in the data to be a breakpoint around 150 mg/dl, maybe a bit lower, above which risk increases rapidly. We know from other experiments that significant cytotoxicity begins around 140-160 mg/dl – some of those experiments are cited by Jenny Ruhl in her Blood Sugar 101 book which Thomas mentions. It seems likely that 160 mg/dl is too high.

    Dear Emily,

    I’m very glad he’s well! Must have been scary.

    If the blood sugars never got above 150 then I wouldn’t have been too concerned either. 140 is a fairly normal postprandial number. 160 is an alarm bells level for me because of cytotoxicity and immune suppression.

  6. Paul,

    several years ago my colleagues and i tried for 2 years to get funding to do a trial comparing 70-110 to 111-150 for hospitalized patients. the hypothesis being that outcomes would be equivalent but it would be easier for nurses/patients with fewer episodes of hypoglycemia. no one would fund us. too bad, really.

    TPN is a true disaster, from the composition of the fats (in the US it is all vegetable oil omega 6) to the use of 25% dextrose IV to meet 50% of daily calories. the stuff is just poison. Unfortunately, there are a small number of conditions where there is no other option and my hope is that we will be able to get the fat solutions used in southeast asia or australia for the US.

    if you have a family member in the hospital i would fight tooth and nail to keep them from getting TPN, you just might save their life.

    best, Dan

    • Dear Dan,

      Thanks much for your comments, it’s great to have expert commenters.

      In our book, in the section on PUFA safety, we talk about the relative toxicity of the various formulations used in parenteral nutrition. I’m sure you’re familiar with the Children’s Hospital Boston experience. The following cites are from our discussion in the book:

      “Old-fashioned lifeline,” Boston Globe, Jan. 9, 2009,
      “Fishing for the right solution,” Children’s Hospital Boston,
      Puder M et al. Parenteral fish oil improves outcomes in patients with parenteral nutrition-associated liver injury. Ann Surg. 2009 Sep;250(3):395-402.
      Gura KM et al. Reversal of parenteral nutrition-associated liver disease in two infants with short bowel syndrome using parenteral fish oil: implications for future management. Pediatrics. 2006 Jul;118(1):e197-201.

      Definitely the catastrophic liver and other damage from TPN is something to be avoided! Unless you’re eager for a liver transplant.

      We closed the discussion with this paragraph:

      Other studies in parenteral nutrition have noted that soybean oil kills immune cells in the blood. To reduce this destruction of lymphocytes, researchers are considering switching from a 100% soybean oil formula to one with 80% olive oil, 20% soybean oil. The new formula still kills immune cells, but less prodigiously than the pure soybean oil formula.

      (Citing: Cury-Boaventura MF et al. Effect of olive oil-based emulsion on human lymphocyte and neutrophil death. JPEN J Parenter Enteral Nutr. 2008 Jan-Feb;32(1):81-7.

      It’s frustrating the glacial pace of progress in improving parenteral formulas. And then when new formulas are studied, they use half-measures. Is there any reason to include soybean oil in parenteral nutrition? A coconut oil, olive oil, fish oil blend would seem both much safer and a better match for human fatty acid ratios.

  7. You’ve just solved the mystery for me of why, when my mother went into the hospital a few years ago with pneumonia, she suddenly started having high blood sugar when she had not been diabetic before entering the hospital. Her condition deteriorated steadily, and then they decided she “needed” a heart bypass. She couldn’t rebound from that, and then had a stroke that left her aphasic and paralyzed on one side of her body. The nurses were always incredulous when I insisted she had not been diabetic before going into the hospital, since her blood sugar was so high while on the feeding tube. She was on a ventilator and finally underwent a terminal vent weaning and died. It took three different hospitals 3 1/2 months to kill her but they eventually succeeded, but not before getting as much medicare and supplemental insurance out of her as they could. If only I had known then what I’ve since learned, I could have made some different decisions on her behalf. Instead I assumed the medical advice I was being given was trustworthy. That won’t happen again with anyone else for whom I may have responsibility in the future. I only pray that, if ever I need to be in a hospital, I am mentally lucid enough to question every move they make.

  8. ethyl,

    What a horror story. My heart goes out to you.

  9. My mother had a heart attack at 54, and 5 months after her hospital stay, she had type 2 diabetes. 10 years later, she had a stroke.

Leave a Comment

NOTE - You can use these HTML tags and attributes:
<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>