Comparing Insulin Sliding Scale Starting Points 150 mg/dL vs. 180 mg/dL in the Neuro Critical Care Unit​

ABSTRACT

Background

To provide better insulin sliding scale management, the University of Utah Hospital (UU) updated the neuroscience insulin sliding scale protocol to incorporate a higher hyperglycemia threshold of 180 mg/dL compared to previous blood glucose (BG) of 150 mg/dL. The previous and updated protocol are both within the American Diabetes Association (ADA) recommendation guidelines, yet it is unknown if this update has an increase or decrease effect regarding safety.

Objective

The primary objective of our study is to determine if initiating insulin sliding scale therapy among the UU Hospital neurocritical care patients at a BG level of ≥ 180 mg/dL compared to a BG of ≥ 150 mg/dL reduced hyperglycemia or hypoglycemia events and decreased the length of stay.

Methods

This is a descriptive study conducted as a retrospective chart review performed at a single-center, UU Hospital.

Results

Hypoglycemia episodes have no statistical difference (p=0.749271) between both groups. Whereas more hyperglycemia events in the ≥180 mg/dL group compared to the ≥ 150 mg/dL group with a statistical difference (p=0.000672). Long-acting insulin usage resulted in no difference (p=0.092123). The length of stay was 0.23 less days in the ≥ 180 mg/dL group, but was not of statistical significance (p = 0.406057).

Conclusion

Based on the results, the ≥ 150 mg/dL group had a decreased incidence of hyperglycemia compared to the ≥ 180 mg/dL group, which prompts for consideration of changing the current protocol. Further studies may need to be conducted for correlation of clinical outcomes such as increased infection rate or length of stay due to hyperglycemic episodes. Regular insulin was the formulation studied; therefore, future studies would be necessary to compare other rapid-acting insulin formulations.

Published in College of Pharmacy, Virtual Poster Session Spring 2020

Responses

  1. Vannary, A poster! Well done! I’m curious why the protocol was changed in 2015 to treat once glucose was 180 or more? It would seem like there was no benefit and perhaps some harm from the change? Also, can you clarify the relation between the numbers you have in your Table 2 and the numbers in the flow chart for hypo- and hyperglycemia?

    1. Dr. Keefe! Thank you for expressing your interest.

      1. )Primarily the protocol changed after the landmark trial, NICE-SUGAR, where there was debate on intensive versus conventional insulin therapy. The bottom line from that study concluded in medical ICU patients, intensive glycemic control (target 81-108 mg/dL) led to more deaths compared to conventional control (target≤180). This particular trial led to further discussion of how glycemic control among other ICU patients. There are several available studies that suggest that a higher glycemic index may be plausible for NCCU patients due to the brain’s demand for glucose increases with brain injury. (References: 1, 4, 8, 9)

      2.) How the numbers are calculated differ based on the units. For example, the flowchart is expressed in units of patients, whereas, table 2 is represented by the percentage. The 10% is from 5 patients had hypoglycemia out of 50 patients reviewed in the 150 mg/dL protocol. Alternatively, the 12% is calculated from the 6 patients who had hypoglycemia out of the 50 patients reviewed in the 180 mg/dL protocol. I hope this clarifies your question. Again, thank you for participating in the discussion.

  2. Vannary – a nice poster. I have a question on Length of stay. Do you expect a a shorter of Length of stay in the new protocol (150) compared to the old (180) protocol? Has this been demonstrated in previous literature (both in RCT or real world evidence study)? Can you provide any reasons why your findings are different or similar to what you expected?

    1. Nui! Thank you for asking your questions.

      1.) I would expect no significant difference in the length of stay. Based on the observed results, I would anticipate that the >150 mg/dL group would have a slightly longer length of stay (approximately 0.23 days) compared to the >180 mg/dL group.

      2.) I did find some evidence in the literature that has measured the outcome. There are two examples I found that studied the length of stay as an outcome. Both studies concluded there was no significant difference seen in either group.
      1. Investigators, N.-S.S., et al., Intensive versus conventional glucose control in critically ill patients. N Engl J Med, 2009. 360(13): p. 1283-97.
      2. Green, D.M., et al., Intensive versus conventional insulin therapy in critically ill neurologic patients. Neurocrit Care, 2010. 13(3): p. 299-306.

      3.) Green et al. roughly had about 45 patients who were trending similar to the results I found. Additionally, the NICE-SUGAR study had over 3000 patients in their respective groups. I think if I were to increase my sample size, I would think the findings will maintain to be similar, but only including more patients would sufficiently answer the question.

      Thank you for participating in this discussion.

  3. Vannary–congrats on your project and poster. It is one I’m interested in. What were the specific sliding scales used? Did patients get a lot more insulin in the >150 group? There was no difference seen in the incidence of hypoglycemia in terms of % of patients with the lower threshold SSI, but were there differences in how low patients got when on the > 150 SSI? Basically, if you used a lower threshold for hypoglycemia, like, say < 60 mg/dL or even < 40 mg/dL, do the percentages between the two groups differ?

    1. Dr. Nyman, thank you for your questions!

      1.) The sliding scales that were used were specific to the NCCU at the hospital using regular insulin. This was a nurse driven protocol. An excerpt of the 180 mg/dL protocol range is denoted here: BG 181-200 give 4 units, 201-230 give 6 units, 231-250 give 8 units, 251-280 give 10 units, 281-300 give 12 units, and >300 give 15 units. The 150 mg/dL protocol range was: BG 150-179 give 4 units, 180-199 give 7 units, 200-229 give 10 units, 230-269 give 13 units, and >270 give 15 units.

      2.) Patients in the >150 mg/dL group did not necessarily received more insulin. Based on the patients that were reviewed the >150 group, There were 225 values that received insulin compared to the 753 recorded glucose values. Alternatively, the >180 group had 226 values that received insulin compared to the 740 recorded glucose values. Because of the small sample size collected this may have been disproportionate.

      3.) There was also no difference in how low the patients got when on the >150 SSI. The range of hypoglycemia recorded values was BG 62-80 with a mean of 72.75. Whereas for the other group, >180 the range of hypoglycemia was BG 66-80 with a mean of 76.14. Therefore, the data that I have collected would not be able to answer your question if there was a lower threshold. Potentially if I reviewed more patients this may occur.

      Thanks for your questions and generating this discussion.

  4. Vannary,
    You did it! Good job! Based on the baseline characteristics that you reported, the two groups seem pretty comparable. What are your thoughts on additional characteristics that might have resulted in bias in your observed results?

    1. Dr. Witt! Thank you and I appreciate your enthusiasm.

      1.) Correct. The groups seem pretty comparable, but there may have been additional characteristics that might have bias my observed results. For example, insulin is eliminated in the urine and I was unable to review patients’ kidney function. Additionally other alternatives, would have been underlining poor control of their T2DM or how sick at baseline the patients are. Hopefully this answers your question. Thank you for chiming in.

  5. Thank you for presenting this poster! Were there 2 separate ISS protocols used? If so what were the glucose cut-offs and amount of insulin administered? When assessing glucose and ISS, I was curious if you identified other glucose containing products such as dextrose containing medications or the use of continuous tube feeding? Thank you.

    1. Dr. Terry,
      Thank you for taking the time to view my poster.

      1.) Yes, there were two separate ISS protocols used. Please see my reply to Dr. Nyman above for a further breakdown of the differences.

      2.) I did look into if patients received dextrose containing medications, but did not adequately review the results. Although, when I was data collecting, the majority of the patients’ charts did not receive dextrose containing products. I apologize I do not have a definitive answer. Unfortunately, I did not assess for continuous tube feeds usage in this population. I do think your question does raise awareness of how this particular project could have improved.

  6. Interesting poster. What results would you expect with rapid acting insulin?

    1. Dr. Lim!
      Thank you for stopping by and engaging in the conversation.

      1.) I would expect that the rapid-acting insulin potentially would require more usage. Based on the pharmacokinetics of rapid-acting insulin, the peak effect is less than regular insulin, approximately 3 hours compared to 6 hours respectively. One of the reasons why I think it would be fascinating is less than a year ago the NCCU switched their protocol from regular insulin to insulin lispro.

  7. Beautifully done! It will be interesting to see in future studies how initiation of SSI at the different cutoff points will turn out in different patient populations.

    1. Angie!
      Thank you for your kindness and for stopping by. Sometimes I feel the projects lead to more questions than answers.

  8. Hi Vannary,
    Great poster and all your research presented.
    Sure excited you’ll be finishing soon! Congratulations

    1. Hello Judy,
      Thank you for taking the time to view my poster. I appreciate your kindness throughout the years.

  9. Great job on this project. The results are really practical in caring for these patients.

    1. Dr. Tyler,
      Thank you for taking the time to compliment the project. I don’t think I would have been able to continue the project without the help of my mentor, Chad.

  10. Hi Vannary – this is a very important issue and one that has been controversial and debated for some time. Thank you for addressing this topic.

    1. Hello Dr. Shane-McWhorter,
      Thank for viewing my poster. I believe the controversy and the debate are what sparked this project.

  11. Vannary, this study is a really nice example of clinical care and investigation science. great work.

Comments are closed.