Pathway to Pre-Diabetes: A Model for the Assessment of Stress and Glycemic Dysregulation in Normoglycemic Subjects

Date of Award


Degree Type


Degree Name

Master of Science in Biomedical Sciences

First Advisor

Brian Matayoshi PhD

Second Advisor

Mary P Owen, PhD, JD

Third Advisor

Francis Jenney, PhD


Background: Several factors contribute to the development of type 2 diabetes mellitus. The relationship between emotional stress and plasma hyperglycemia in non-diabetics has been documented in controlled animal studies; however, this relationship has yet to be conclusively established via clinical studies on human populations. The purpose of this research is to establish a model for assessing the relationship between stress and glycemic dysregulation in a population ofnormoglycemic subjects. Our subjects were obtained from a group of graduate students at the Philadelphia College of Osteopathic Medicine-GA campus. We hypothesized that the administration of fmal exams at the end of the academic term would lead to the establishment of a pre-diabetic state (as defmed by the World Health Organization) in study participants. Furthermore, we hypothesized that glycemic dysregulation would manifest itself either in the form of elevated preprandial glucose levels, or in the form of a significant disparity between fasting and 2- hour postprandial glucose levels.

Materials and Methods: A total of 16 graduate students were enrolled in this study, which was administered from May 16th, 2012 to May 31st 2012. The sixteen days of the study were divided into Period I (May 16th-May 23rd) and Period II (May 24th-May 31st). Plasma glucose data from Period I was designated as the control data, providing a comparison point for subsequent data. Plasma glucose data from Period II was designated as stress data, due to the fact that study participants had two fmal exams scheduled consecutively on May 30th and May 31st. This allocation of data was based on the premise that students would be exposed to progressively greater amounts of stress as the final exams approached. Two sets of data were collected for the purposes of this study: daily plasma glucose readings, and oral glucose tolerance test data. On a daily basis, the participants took four glucose readings. Reading A was a fasting reading, taken before the first meal of the day. Reading B was taken before the second meal of the day. Reading C was taken before the third meal of the day, and reading D was taken 2 hours after the last meal of the day. Readings A, B, and C were pre-prandial while reading D measured 2 hour postprandial glucose levels. In addition to daily glucose readings, three oral glucose tolerance tests were taken by study participants. The first oral glucose tolerance test was administered during Period I on May 18th, 2012. The second and third oral glucose tolerance tests were scheduled on May 30th, 2012 and May 31st 2012, respectively; both of these dates fell within the data interval from Period II. All plasma glucose data was obtained from the study participants themselves. Participants were provided with sterile alcoholic wipes, Abbott Freestyle Freedom Lite meters, Abbott Freestyle Lite test strips, and Freestyle 28-gauge sterile lancets for the purposes of this study (Abbott Park, Illinois). In addition, participants were provided logbooks in which to record all glucose readings and personal stressors.

Results: The daily glucose readings were analyzed as follows. Using PRISM software, a cine-way ANOVA was run on readings A, B, and C for each subject. Out of 16 participants, only 1 subject displayed significant changes. Using Vassar Stats software, a Tukey's honestly significant difference test was run on this subject in order to identify the specific days on which this change occurred. The days on which significant changes in pre-prandial glucose levels occurred had no correlation with Period II. Next, a paired ttest was run (using PRISM software) on readings C and D for each student. A total of 4 subjects showed a significant difference between pre-prandial arid 2 hour post-prandial glucose levels. Out of those 4 subjects, 2 appeared to exhibit inhibited ability to return to pre-prandial levels during the week of stress, while 2 exhibited about the same inability to retumto pre-prandial levels during stress and during the control week. Unfortunately, comparisons between Period I and Period II were severely hampered by the fact that large intervals of daily glucose readings were missing. The oral glucose tolerance test data was analyzed primarily in order to determine the presence of a pre-diabetic state (Valdes, 2005). The Period I oral glucose tolerance test showed 4 of the 16 participants to be prediabetic at the beginning of the study. This left 12 normoglycemic subjects. The Period II oral glucose tolerance tests showed that out of the 12 initially normoglycemic students, 3 students developed a pre-diabetic state during the administration offmal exams. Next, the oral glucose tolerance test data for all the subjects were analyzed together as a group. Using VassarStats software, a two-way repeated measures ANOVA was run. Significant differences were found between the fasting reading of the first oral glucose tolerance test and that of the third oral glucose tolerance test, between the 1-hour reading of the first oral glucose tolerance test and the 1-hour reading of the third oral glucose tolerance test, between 1.5 hour reading of the first oral glucose tolerance test and the same of the second oral glucose tolerance test, between the 2-hour post glucose challenge value of the first oral glucose tolerance test and the 2-hour post glucose challenge value of the second oral glucose tolerance test, and 2-hour post glucose challenge value of the first oral glucose tolerance test and 2-hour post glucose challenge value of the third oral glucose tolerance test.

Conclusion: The oral glucose tolerance test results showed that the stress of fmal exams correlated with 25% of the initially normoglycemic participants (3/12) developing a prediabetic state (Valdes, 2005). In addition to the pre-diabetic subjects, five subjects who did not develop pre-diabetes still experienced fluctuations in glycemic regulation: one subject exhibited significant changes in preprandial glucose readings, while four subjects exhibited significant differences in pre-prandial and 2-hour postprandial plasma glucose readings. Although these changes did not occur between Period I and Period II, it must be kept in mind that personal stressors may also have played a role; unfortunately, information provided by the logbook was extremely sparse. It must also be kept in mind that stress is a relative term and that confirmation of stress levels (via psychometric testing) was not available; therefore, we cannot say if Period I and Period II were in fact the control and stress periods for those individuals. All together, 8 of the 16 study participants showed some change in glycemic regulation throughout the course of the study, a number that is consistent with national statistics on the development of diabetes. If the null hypothesis were true, no changes would have occurred at all. Therefore this pilot study did confirm the need for further studies to be carried out. For optimal results, this study would need to be reproduced in a larger population controlled by age, gender, and race. Ideally, dietary intake and hours of .~ sleep would be standardized. Although final exams provided us with a valid method of administering emotional stress, daily psychometric evaluation of stress levels would have been an invaluable addition to this study.

This document is currently not available here.