Faculty of Science
QUANTIFICATION OF KETONE BODIES
CHROMATOGRAPHY-MASS SPECTROMETRY
2019 | AUSTIN NOAH PIETRAMALA

B.Sc. Honours thesis – Chemical Biology

IN

SALIVA

USING

GAS

QUANTIFICATION OF KETONE BODIES IN SALIVA USING GAS
CHROMATOGRAPHY-MASS SPECTROMETRY
by
AUSTIN NOAH PIETRAMALA

A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR
THE DEGREE OF
BACHELOR OF SCIENCE (HONS.)
in the
DEPARTMENTS OF BIOLOGICAL AND PHYSICAL SCIENCES
(Chemical Biology)

This thesis has been accepted as conforming to the required standards by:
Kingsley Donkor (Ph.D.), Thesis Supervisor, Dept. Physical Sciences
Mark Rakobowchuk (Ph.D.), Co-Supervisor, Dept. Biological Sciences
Dipesh Prema (Ph.D.), Committee Member, Dept. Physical Sciences

Dated this 17th day of May, 2019, in Kamloops, British Columbia, Canada
Ó Austin Noah Pietramala, 2019

ABSTRACT
Type 1 Diabetes Mellitus is an autoimmune disease where pancreatic b-cells within the
Islets of Langerhans are destroyed. A complication associated with type 1 diabetes mellitus is
diabetic ketoacidosis, and if left untreated can result in coma or death. There are commercial
methods for individuals to analyze ketone body concentrations within urine and blood, but these
are invasive, expensive, and aren’t always accurate. Therefore, saliva should be examined as a
potential alternative to the current commercial methods on the market. This experiment aimed to
quantify the amount of beta-hydroxybutyrate and acetoacetate, in the saliva, blood, and urine
samples using GC-MS by inducing ketosis in consenting participants. The participants followed a
ketogenic diet for four days, and their biological samples were obtained before and after. The blood
samples were centrifuged to isolate the plasma and deproteinated. All sample matrices were
evaporated, the ketone bodies were derivatized using BSTFA + 1% TMCS at 80°C for 1.5 h, and
the headspace was analyzed. The participants followed an Atkins diet to induce ketosis and the
amount of beta-hydroxybutyrate was able to be quantified for each participant. The urine samples
had high concentrations of beta-hydroxybutyrate, which could be a result of the acidic urine
increasing the derivatization efficiency. There was very little beta-hydroxybutyrate detected within
saliva. Since there was detection, this could potentially be used as a method to quantify ketone
bodies upon further method development. No acetoacetate was detected in any of the samples.
Future work should further optimize the methodology to detect acetoacetate, and the addition of
recovery and internal standards to the samples to ensure all the analyte is derivatized and to
determine if there is any variation between runs.

Thesis supervisor: Dr. Kingsley Donkor

ii

ACKNOWLEDGMENTS
I would like to thank Dr. Kingsley Donkor and Dr. Mark Rakobowchuk for their assistance
in helping me plan and execute this experiment. I would also like to thank Trent Hammer for his
guidance and expertise on the GC-MS. Thank you to NSERC for their generous funding of this
experiment. As well, thank you to the Thompson Rivers University Chemistry Department and
Honours coordinators for helping make this research possible. Thank you to Dr. Dipesh Prema for
agreeing to be on my honours committee. Finally, thank you to all of the participants who
consented to participate in my study.

iii

TABLE OF CONTENTS
ABSTRACT……………………………………………………………………….
ii
ACKNOWLEDGMENTS…………………………………………………………
iii
TABLE OF CONTENTS ….………………………………………………………
iv
LIST OF FIGURES..………………………………………………………………..
v
LIST OF TABLES……………………………………………………………….....
vi
1.0 INTRODUCTION……………………………………………………………….. 1
1.1 TYPE 1 DIABETES MELLITUS …………………………………………. 1
1.2 KETOGENIC DIETS AND KETONE BODIES…………………………… 3
2.0 METHODS ……………………………………………………………………... 9
2.1 STANDARDS AND MATERIALS…………………………………………. 9
2.2 KETOGENIC DIET………………………………………………………….. 9
2.3 SAMPLING……………………………….………………………………….. 10
2.4 SAMPLE ANALYSIS……………………………………………………….. 11
2.5 INSTRUMENTAL CONDITIONS……………………………….………….. 12
2.6 PREPARATION AND ANALYSIS OF KETONE BODY STANDARDS…. 13
3.0 RESULTS AND DISCUSSION………………………………………………….. 13
3.1 DETECTIONS AND CALIBRATION OF THE KETONE BODIES……….. 13
3.2 THE QUALITATIVE RESULTS OF THE DIET…………..……………….. 20
3.3 THE QUANTIFICATION OF KETONE BODIES …………………..…….. 27
3.3.1 QUANTIFICATION IN BLOOD…………………………………. 27
3.3.2 QUANTIFICATION IN URINE…………………………………... 29
3.3.3 QUANTIFICATION IN SALIVA…………………………………. 31
4.0 CONCLUSIONS…………………………………………………………………….. 33
5.0 FUTURE WORKS…………………………………………………………………… 34
6.0 LITERATURE CITED………………………………………………………………. 35
7.0 APPENDICES……………………………………………………………………….. 39
THE KETOGENIC DIET OUTLINE……………………………..……………. 39
PARTICIPANT 2 BLOOD AFTER-THE-DIET (TRIAL 2)……..……………. 40
PARTICIPANT 2 BLOOD BEFORE-THE-DIET (TRIAL 2)……..……………. 40
PARTICIPANT 2 SALIVA BEFORE-THE-DIET (TRIAL 1)……..……………. 41
PARTICIPANT 2 URINE ABEFORE-THE-DIET (TRIAL 1)……..……………. 41
PARTICIPANT 2 URINE AFTER-THE-DIET (TRIAL 1)……..…………….
42
COMPARISON BETWEEN ACIDIFIED BETA-HYDROXYBUTYRATE (a) AND
NON-ACIDIFIED BETA-HYDROXYBUTYRATE AT 5000 ppm EACH…… 42
COMPARISON BETWEEN ACIDIFIED ACETOACETATE (a) AND NONACIDIFIED ACETOACETATE AT 1600 ppm EACH………………………… 43
ETHICS APPROVAL……………………………..…………………………..
44
BIOSAFTY APPROVAL……………………………..……………………….
45

iv

List of Figures
Figure 1. The ketone bodies that is produced within the human body…………………………...4
Figure 2. The chromatogram (a) of 5000 ppm derivatized BHB with its corresponding mass
spectrum (b). The BHB signal is found at 6.521 min……………………………………………15
Figure 3. The chromatogram (a) of 5000 ppm derivatized AcAc with its corresponding mass
spectrum (b). The BHB signal is found at 7.197 min……………………………………………17
Figure 4. The derivatized acetoacetate (a) and beta-hydroxybutyrate (b) that were to be analyzed
by the GC-MS. …………………………………………………………………………………..18
Figure 5. The beta-hydroxybutyrate standard curve using the GC-MS…………………………19

v

List of Tables
Table 1. The breakfast consumed by each participant during the course of the diet………...21
Table 2. The lunch consumed by each participant during the course of the diet…………….22
Table 3. The dinner consumed by each participant during the course of the diet……………24
Table 4. The snacks consumed by each participant during the course of the diet……………26
Table 5. The experimentally determined amounts of beta-hydroxybutyrate in the blood
samples………………………………………………………………………………………28
Table 6. The experimentally determined amounts of beta-hydroxybutyrate in the urine
samples………………………………………………………………………………………30
Table 7. The experimentally determined amounts of beta-hydroxybutyrate in the saliva
samples………………………………………………………………………………………32

vi

1.0 INTRODUCTION
1.1 Type 1 Diabetes Mellitus
Diabetes mellitus is an epidemic in today’s world with 600 million people worldwide being
diagnosed as diabetic with 10% of those patients being diagnosed with type 1 diabetes mellitus
(T1D) (Dholakia et al. 2016). T1D is an autoimmune disease where pancreatic b-cells within the
Islets of Langerhans are destroyed, and the individual can no longer produce insulin.
Unfortunately, the mechanism that triggers the autoimmune destruction of b-cells is unknown, but
T1D pathogenesis is being hypothesized to be in three stages (Butalia et al. 2016; Pociot and
Lernmark 2016). In the first stage, b-cell autoantibodies are present, but inactive (Pociot and
Lernmark 2016). The second stage involves the autoantibodies attacking the b-cells once some
sort of environmental trigger occurs, and the third stage involves the symptoms of T1D to be
present (Pociot and Lernmark 2016). The pathogenesis of T1D seems to be a combination of many
factors. For instance, there are many genetic loci that are attributed to increasing the chances of
being diagnosed with T1D like the human leukocyte antigen class II molecules, DQ and DR, within
the major histocompatibility complex locus, and more are being discovered still (Butalia et al.
2016; Pociot and Lernmark 2016). As well, certain environmental factors contribute to the
increased risk of T1D like the mumps, rubella, microorganisms, and other factors which are still
being studied (Butalia et al. 2016). Individuals with T1D have to monitor their blood glucose level
every 4-8 h with the goal of being in a range of 4.0 to 7.0 mmol/L (Giugliano et al. 2008; Morales
and Schneider 2014). If they are less than 3.9 mmol/L, they would have hypoglycemia and
typically are treated with juice (Morales and Schneider 2014). Blood sugar levels consistently
greater than 7.0 mmol/L while fasting refers to hyperglycemia and is treated with exogenously
administered insulin (Giugliano et al. 2008).

1

There are many complications associated with T1D. For instance, those with T1D are prone
to end-stage renal disease, blindness, foot and leg amputations, and cardiovascular diseases (Pociot
and Lernmark 2016). Another complication of T1D is diabetic ketoacidosis. Essentially, it is a
metabolic state in which the diabetic individual does not have enough insulin to stimulate glucose
uptake by target cells, thus excreting glucose via urine, and increasing concomitantly the blood
glucagon concentration which stimulates ketogenesis within the liver, and the excess of ketone
bodies ends up acidifying the individuals blood (Laffel 1999; Mullins et al. 2011). By clinical
definition, one is diagnosed with diabetic ketoacidosis when they have a blood glucose level
greater than 11 mmol/L, a pH less than 7.3, ketonaemia, and ketonuria (Usher-Smith et al. 2011).
One is considered in ketonaemia when blood ketone concentration is greater than 3 mM (Laffel
1999). The amount of ketone bodies can be as high as 25 mM in those suffering from T1D (Laffel
1999; Kanikarla-Marie and Jain 2016). In non-diabetic populations, ketoacidosis generally not a
concern since these individuals have two feedback loops to maintain plasma ketone levels below
3 mM (Mullins et al. 2011).
Ketone bodies are ionized at physiological pH, so the hydrogen ions bind to the bicarbonate
molecules, which tends to overwhelm the blood-buffering capacity and subsequently acidifies the
blood (Laffel 1999; Mullins et al. 2011; Kanikarla-Marie and Jain 2016). This also leads to the
loss of sodium ions, loss of ammonium, severe dehydration, and the loss of blood volume (Mullins
et al. 2011). Those with T1D also struggle to clear ketone bodies through urine, whether or not
they are suffering from ketoacidosis, and have decreased succinyl CoA-oxoacid transferase
(SCOT) and 3-hydroxybutyrate dehydrogenase activities thus increasing ketone concentration
within the blood (Kanikarla-Marie and Jain 2016; Sherwin et al.). If left untreated, this can result
in coma or death and is the leading cause of death among children with T1D (Mullins et al. 2011;

2

Usher-Smith et al. 2011). This is concerning because between 10% and 70% of children recently
diagnosed with T1D are in diabetic ketoacidosis, so it is crucial that the symptoms of diabetic
ketoacidosis is acted upon and subsequently treated (Usher-Smith et al. 2011). Ketoacidosis is
treated using insulin, potassium bicarbonate, and saline, which treat hyperglycemia, the lack of
electrolytes within the individual and balance pH, and treat dehydration, respectively (Laffel 1999;
Kanikarla-Marie and Jain 2016).

1.2 Ketogenic Diets and Ketone Bodies
Ketogenic diets (KD) have become very popular for those wishing to lose weight. KDs are
a high-fat, moderate protein, and low carbohydrate intake diet that reduces glucose supply to
decrease glycolysis and increase fatty acid breakdown through beta-oxidation (Branco et al. 2016;
Roehl and Sewak 2017). Initially, acetyl-CoA is generated from fatty acid oxidation and can be
used by the citric acid cycle to produce ATP. Using acetyl-CoA, the ketone bodies acetoacetate
(AcAc), acetone, and beta-hydroxybutyrate (BHB) can be produced which serves as an energy
source for the brain, the heart, the kidney cortex, and skeletal muscle (Laffel 1999; Roehl and
Sewak 2017). The production of ketone bodies are necessary since they are water soluble, allowing
energy sources to travel systemically, and can subsequently travel through the brain-blood barrier
to enter neurons and generate ATP through their mitochondria (Mullins et al. 2011; Wibisono et
al. 2015). There are four different types of KD which are the long-chain triglyceride KD (classic
KD), modified Atkins diet, medium-chain triglyceride KD, and low glycemic index treatment
(Branco et al. 2016; Roehl and Sewak 2017). The diets differ in the ratio of grams of fat consumed
to grams of protein and carbohydrates consumed, percentage of calories consumed, and in the
foods that are allowed to be consumed (Branco et al. 2016; Roehl and Sewak 2017). Originally,

3

KD was created in 1921 as a way to treat epilepsy, as it is believed to induce anticonvulsant effects
in humans which affects neuronal firing and the spread of seizures (Branco et al. 2016). It has also
been used as a therapeutic diet for those suffering from type II diabetes mellitus, obesity, and
cancer (Wibisono et al. 2015; Branco et al. 2016; Roehl and Sewak 2017). The diet affects
cancerous cells by reversing redox signaling pathways within tumors (Branco et al. 2016). Though
it has been used therapeutically, there are potential short-term and long-term harms of the diet. In
the short-term, participants may experience constipation, hypoglycemia, dehydration, lethargy,
and acidosis (Branco et al. 2016). Long-term effects include hypercholesterolaemia,
nephrolithiasis, and cardiomyopathy (Branco et al. 2016).

Figure 1. The ketone bodies that is produced within the human body.

Individuals on KDs usually go into ketosis, which refers to a metabolic state where ketone
bodies are metabolized to produce ATP (Branco et al. 2016). Under normal physiological
conditions, the ketone concentration within plasma is less than 0.2 mM (Garber et al. 1974;
Courchesne-Loyer et al. 2013). Under ketosis, the concentration of BHB is around 1-2 mM and
around 0.5 mM for AcAc after 3 days of fasting, and can reach as high as 8 mM, and between 1-2
mM for AcAc for obese individuals after 42 days of fasting (Owen et al. 1967; Owen et al. 1969;
Garber et al. 1974). Further, it was found that the ketone bodies also can provide up to two-thirds

4

of the brain’s energy requirement during ketosis (Owen et al. 1967; Courchesne-Loyer et al. 2013).
Ketone bodies are generated in response to a change in the molar concentration of glucagon to
insulin, as a result of a lower concentration of glucose circulating throughout the body (Mullins et
al. 2011; Kanikarla-Marie and Jain 2016). Because the insulin content is so low, lipolysis is no
longer inhibited, nor is hormone-sensitive lipase deactivated, allowing for fatty acids from adipose
tissue to be metabolized (Laffel 1999; Kanikarla-Marie and Jain 2016). Therefore, the increase of
cAMP, due to a high glucagon content, activates protein kinase A which signals for the enzyme
adipose triglyceride lipase to begin hydrolysis of the triglycerides within adipose tissue (O’Neill
et al. 2013; Steensels and Ersoy 2019). The newly formed diglycerides are then hydrolyzed by
hormone-sensitive lipase to make monoglycerides, and glycerol (O’Neill et al. 2013; Steensels and
Ersoy 2019). From here, the fatty acids are then released into circulation, and travels to various
tissues through albumin (Laffel 1999; Steensels and Ersoy 2019). The fatty acid translocase/CD36,
the most effective fatty acid transporter for increased fatty acid content, uptakes the fatty acids,
and carnitine acyltransferase 1 then transports the fatty acids across the inner mitochondrial
membrane, where beta-oxidation of fatty acids takes place to form acetyl CoA (Laffel 1999;
Nickerson et al. 2009; Kanikarla-Marie and Jain 2016). Normally, under non-ketotic conditions,
the acetyl CoA would enter the citric acid cycle, and condense with oxaloacetate (Laffel 1999).
Since there is a lower content of glucose within the bloodstream, hepatic cells therefore utilize the
oxaloacetate within gluconeogenesis, effectively slowing down the citric acid cycle (Laffel 1999).
Further, under non-ketotic conditions, acetyl CoA could leave the inner mitochondria and be
converted to malonyl CoA via acetyl CoA carboxylase, which is used for fatty acid biosynthesis
(Laffel 1999). Though during ketosis, acetyl CoA carboxylase is inhibited by the increase in
glucagon and lectin and decrease in insulin and adiponectin within the body, thus shunting these

5

pathways and using the acetyl CoA for the generation of ketone bodies (Laffel 1999; O’Neill et al.
2013). Acetyl CoA is converted to acetoacetyl CoA by 3-ketothiolase, followed by the formation
of β-hydroxy β-methylglutaryl CoA through mitochondrial β-hydroxy β-methylglutaryl CoA
synthase, and cleaved to form AcAc by β-hydroxy β-methylglutaryl CoA lyase (Laffel 1999;
Kanikarla-Marie and Jain 2016). AcAc is reduced to BHB by 3-hydroxybutyrate dehydrogenase,
which oxidizes an NADH as a result (Laffel 1999; Kanikarla-Marie and Jain 2016). Acetone is
also produced from AcAc by spontaneous decarboxylation (Laffel 1999). BHB and AcAc are the
two dominant ketone bodies that are used as energy sources within the body (Laffel 1999; Fujii et
al. 2014; Kanikarla-Marie and Jain 2016). The ratio between the two ketone bodies is usually three
BHB molecules to one AcAc molecule, which is a result from the NADH/NAD+ ratio and activity
of the 3-hydroxybutyrate dehydrogenase (Laffel 1999; Kanikarla-Marie and Jain 2016).
The ketone bodies, being water soluble, travel throughout the blood stream to serve as an
energy source (Laffel 1999; Kanikarla-Marie and Jain 2016). The ketones enter organs and are
transported into the cell by monocarboxylate transporter 1 and 2 (Puchalska and Crawford 2017).
Subsequently, the ketones travel through to the inner mitochondria for ketolysis, but the
mechanism is unclear (Puchalska and Crawford 2017).The abundant BHB molecules revert back
to AcAc by 3-hydroxybutyrate dehydrogenase (Branco et al. 2016). From here, the AcAc
molecules are used to regenerate acetoacetyl CoA through the enzyme SCOT and an acetyl group
is cleaved from methylacetoacetyl CoA thiolase to regenerate acetyl CoA which enters the citric
acid cycle (Laffel 1999; Branco et al. 2016; Kanikarla-Marie and Jain 2016). It should be noted
that SCOT is the rate limiting step of ketolysis, and the highest expression of this enzyme is in the
heart and the brain (Laffel 1999; Kanikarla-Marie and Jain 2016). Further, SCOT is downregulated

6

by high concentrations of AcAc, which is the reason for an increase in circulating ketone bodies
during the initial phases of ketosis (Laffel 1999).
There are commercial methods for individuals to analyze ketone body concentrations.
Commonly, urinary ketones are measured to screen individuals for diabetic ketoacidosis (Laffel
1999). The urine kit relies on the Legal reaction, in which AcAc reacts in presence of an alkaline
buffer and nitroferricyanide to produce a purple color on a test strip (Laffel 1999; Brooke et al.
2016). This test does not react with BHB, nor does it measure the amount of ketones present (Laffel
1999; Brooke et al. 2016). Measuring AcAc through urine is a cheap method to determine if
someone is in a state of ketoacidosis; although it can be perceived as intrusive and awkward for
people and is less accurate than measuring blood ketone levels, which primarily focuses on BHB
(Brooke et al. 2016). Using a blood ketone monitor, the amount of BHB is quantitatively
determined (Brooke et al. 2016). It is believed that BHB levels within blood provide fewer false
positives than urine kits that test for AcAc. Also there is a higher correlation between blood BHB
concentrations and the clinical markers of diabetic ketoacidosis (Brooke et al. 2016). Though,
measuring blood has its challenges since the meters have difficulties measuring concentrations
greater than 5 mM, can be seen as invasive, and the blood strips and meter is expensive to purchase
(Brooke et al. 2016). Thus, it would be in the best interest of the public to develop a less invasive
and affordable way to analyze ketone bodies, especially BHB, within the body.
Saliva plays an important role within the body as it clears substances from the mouth,
contains salivary amylase to break down carbohydrates, buffers pH, and protects, hydrates, and
lubricates oral mucosal surfaces (Proctor 2016). Saliva consists mostly of water, but also contains
hormones, microbes, enzymes, and other metabolites (Elmongy and Abdel-Rehim 2016;
Viswanath et al. 2017). Most notably, saliva is isotonic to plasma (Aps and Martens 2005;

7

Elmongy and Abdel-Rehim 2016; Proctor 2016). The concept of using saliva as a diagnostic model
is slowly gaining popularity within the literature, but there is very little research on the
quantification of ketone bodies in the literature, not to mention quantifying the amount of BHB or
AcAc within individuals undergoing ketosis (Elmongy and Abdel-Rehim 2016). Previously, it was
noted that both saliva and blood contain BHB, and there is a positive correlation between them
(Liu et al. 2015). The study done by Liu et al. also observed BHB within urine, but their subjects
used were not in ketosis (2015). In another study, saliva was analyzed in diabetic patients and
healthy volunteers for BHB, which was found to be around 25 ng/mL and 16.5 ng/mL,
respectively, though the participants were also not in ketosis either (Tsutsui et al. 2012). AcAc was
not explored in both experiments (Tsutsui et al. 2012; Liu et al. 2015).
The instrument technique of gas chromatography-mass spectrometry (GC-MS) involves
producing and separating gas phase analytes, and detecting these analytes by first ionizing them
using a mass spectrometer (Niwa 1995). This technique has been used in previous experiments to
quantify ketone bodies within biological fluids (Paul et al. 2006; Hassan and Cooper 2009; Holm
et al. 2010; Føreid and Gadeholt 2017), and it is a very powerful technique to detect analyze the
analytes of interest. The gas chromatograph separates gaseous analytes which are carried through
a stationary column by an inert gas (Niwa 1995). In the column, the analytes are separated based
on their affinity with the column and vapor pressure (Niwa 1995). The stationary phase interacts
with the analytes and separates them based on their affinity to the non-polar column (Niwa 1995);
the more non-polar the analyte is, the longer it takes for it to elute through the column (Niwa 1995).
If the analyte has a high vapor pressure, it will travel through the column quickly as well (Niwa
1995). Once it is separated, the analytes will be bombarded by electrons and ionized by an electron
impact mass spectrometer, which then separates the charged ions through a mass analyzer by its

8

mass-to-charge ratio to obtain a mass spectrum (Niwa 1995). This allows for the investigators to
determine what eluted from the gas chromatograph (Niwa 1995).
Therefore, this experiment looks to quantify the amount of BHB and AcAc, the two most
common ketone bodies, found in the saliva, blood, and urine via GC-MS before and after inducing
ketosis in consenting participants. Through quantification, the development of methods that are
non-invasive and cheap to purchase to observe the ketone body content within individuals can be
explored. Within this study, method development was conducted to try and optimize the detection
of the ketone bodies. From there, nine participants followed a ketogenic diet for four days, where
their blood, urine, and saliva were sampled before-the-diet and after-the-diet. The ketone bodies
were then extracted, analyzed, and quantified for each individual participant.

2.0 METHODS
2.1 Standards and Materials
Lithium acetoacetate, (±)-sodium 3-hydroxybutyrate, and BSTFA + 1% TMCS were
purchased from Sigma-Aldrich Canada Ltd. (Oakville, Ontario, Canada). The acetonitrile was of
HPLC grade. The trichloroacetic acid and Tris were purchased from Sigma-Aldrich Canada Ltd.
(Oakville, Ontario, Canada). The headspace vials were 20 mL and clear, while the caps were 18
mm magnetic screw caps. Both were purchased from Canadian Life Science (Peterborough,
Ontario, Canada). They were previously never used before the experiment.

2.2 Ketogenic Diet
Participants were recruited to participate in a KD through posters and word of mouth
around the Thompson Rivers University campus. The participants contacted the investigators and

9

arranged a meeting time to further discuss the experiment. The participants were given a numbered
consent form upon arrival for them to read over and was signed if they consented to participating
in the study. They were then given a numbered screening questionnaire to complete so the
investigators could determine if they were healthy enough to participate in the study. Both the
consent forms and questionnaires were numbered in effort to keep the participant’s anonymous,
and their name and corresponding number were recorded in a secure notebook. There were 10
participants originally, but Participant 1 removed themselves from the study prior to starting the
diet.
The diet that the participants were following was a classic ketogenic diet that followed a
3:1 ratio of fats to protein and carbohydrates. The participants were encouraged to follow the meal
plan that was created by the investigators (see Appendix), but they were able to eat other food as
long as they received permission from the investigators and recorded it in a meal log. The diet was
followed for four days. Because the participants have different eating habits, they were encouraged
to eat until they were full as long as they were maintaining a 3:1 ratio of fats to proteins and
carbohydrates.

2.3 Sampling
The participants fasted for a minimum of 8 h prior to sampling. The participants were
instructed to urinate a minimum of 10 mL into a centrifuge tube. The tubes were sealed, wiped,
and given to the investigators. The participants were to passively drool a minimum of 500 µL into
a 1.5 mL Eppendorf tube. This was done by sitting down, and arching their back to a 45° angle,
and further directing their head downward to allow for the saliva to pool in their mouth. They then
had to manipulate their mouth to allow for the saliva to enter the Eppendorf tube. They had to

10

ensure that the saliva was not bubbly as it would not give an accurate representation of the amount
currently in the tube itself. The participants then had 5 mL of their blood drawn into lithium heparin
tubes. The blood was placed on ice after the initial collection. This was conducted before the
participants started the diet and 4 days after following the ketogenic diet.
The urine was split into three 1.5 mL Eppendorf tubes as storage, and 500 µL of the sample
was pipetted into a sterile Eppendorf tube for subsequent use within the experiment. The saliva
was stored in its original container. The blood, after being drawn, was then centrifuged at 3000 rcf
for 10 min at 4°C to separate the plasma. The plasma was then pipetted into 1.5 mL Eppendorf
tubes as storage, and a 500 µL sample was taken to be used in the experiment. All fluids were then
stored at -80°C until the samples were to be analyzed.

2.4 Sample Analysis
All of the sample tubes that were stored were dethawed on ice. The plasma was then
deproteinated to remove potential protein interferences within the plasma. There was 500 µL of
2 M trichloroacetic acid that was added to the Eppendorf tube. It was then centrifuged at 13,000
rpm for 3 min at 4°C. The supernatant was moved, and 250 µL of 4 M Tris was then added. The
solution was then centrifuged at 13,000 rpm for 15 min at 4°C. It was also brought back to pH 7
using KOH. Once the saliva had dethawed, the sample was centrifuged at 13,000 rpm for 10 min
to pellet the mucins within the fluid. Afterwards, 200 µL of each fluid was pipetted into a new
headspace vial, along with 10 µL of 70% ethanol. The solution was capped, mixed, and it sat for
1 h in order to sterilize the fluids. The vials had their caps removed, and the solutions inside were
evaporated in an oven at 80°C. After the evaporation, 75 µL of BSTFA + 1% TMCS and 150 µL
of acetonitrile were added to each headspace tube, were capped, and derivatized at 80°C for 1.5 h.

11

They were subsequently left at room temperature for three days afterwards. The samples were
agitated at 250 rpm for 6 min at 80°C prior to analysis. The PAL headspace syringe, at 85°C, then
penetrated through the vial cap and sampled 100 µL of headspace and injected it into the GC-MS.

2.5 Instrumental Conditions
Analysis was performed using an Agilent 7890B-GC coupled 5977A-MS (Agilent
Technologies, Santa Clara, CA). The instrument operated using a PAL RSI 85 autosampler system
equipped with a headspace function and an agitator with an oven. The column was an Agilent HP5MS column (30 m x 250 µm x 0.250 µm) (Agilent Technologies, Santa Clara, CA). An agitator
on the autosampler was heated to 80°C, and the syringe was heated to 85°C. The split was 2:1, and
the split flow was 2 mL/min. The injector temperature was 250°C. The temperature program used
was developed from a previous method (Hassan and Cooper 2009). The initial oven temperature
was 60°C and was held at this temperature for 2 min. The temperature increased to 180°C at
20°C/min, and then increased to 250°C at 50°C/min which was then held for 1 min. The total run
time was 10.4 min. The amount injected was 100 µL.
A scan on the mass spectrometer was used to detect the ions from 40-350 m/z. The ion
chromatograms were extracted using selected ion monitoring to determine where the ketone bodies
were eluting at. The ions 147 and 233 were used to identify the BHB, and 147 and 231 were used
to identify AcAc; the 147 ion was used to determine the concentration of the ketone bodies. As
well, the NIST database was used in elucidating the mass spectrum to help estimate the compound
that eluted from the gas chromatograph (Linstrom and Mallard 2001). This software estimates the
compounds that eluted from the instrument, giving a probability of accuracy, based on the mass
spectrum of the peak at that specific point (Linstrom and Mallard 2001).

12

2.6 Preparation and Analysis of the Ketone Body Standards
A stock solution of 5000 ppm was made for BHB and AcAc, individually. The BHB stock
solution was made up in a 50 mL volumetric flask with 18 MW water, while the AcAc solution
was made up to 25 mL with 18 MW water. These stocks solutions were used to develop standards
with concentrations of 50 ppm, 100 ppm, 250 ppm, 500 ppm, 1000 ppm, and 2000 ppm, made up
to 10 mL with 18 MW water in a volumetric flask. The standards were mixed, and 225 µL of each
solution was added to a new headspace vial. The solution was evaporated in an oven at 80°C. The
solution then had 75 µL of BSTFA + 1% TMCS and 150 µL of acetonitrile added to each vial.
They were derivatized in the oven for 1.5 h at 80°C and were run three days afterwards. Before
analysis, the samples were agitated at 250 rpm for 6 min at 80°C. The PAL headspace syringe then
was used to inject 100 µL of the standards headspace into the GC-MS.

3.0 RESULTS AND DISCUSSION
3.1 Detection and Calibration of the Ketone Bodies
The development of a method to detect the ketone bodies was investigated using the BHB
standard. This standard was used for method development because of its prominence within the
blood of individuals producing ketone bodies (Laffel 1999; Fujii et al. 2014). A mass of BHB was
added into hexane, acetonitrile, water, and methanol. It was not soluble in the hexane since nothing
seemed to be dissolved, but it was soluble in methanol, water, and acetonitrile; it was most soluble
in methanol and water. Therefore, the BHB standard was dissolved in 18 MW water to create the
stock solutions. The solution was then pipetted into a glass headspace vial, where the solution was
evaporated to dryness using nitrogen gas. This was inefficient and extremely difficult to do, so the

13

headspace vials were transported to an oven which was set to 80°C to evaporate the solvent. The
derivatizing agent was then pipetted into the vial, it was capped, and the vial was heated to promote
the derivatization reaction. Upon analysis, there was no signal that was ascertained. Therefore, the
stock solution would be acidified to ensure that all of the hydroxyl groups were protonated to help
with the derivatization reaction. The same procedure was then conducted, but there were still no
results that were ascertained. Acetonitrile was then added as a catalyst for the derivatization to take
place, especially since BHB is soluble in it. There was a BHB signal detected with the GC-MS.
The signal obtained had a mass-to-charge ratio at 147 and 231, and the NIST database predicted
this molecule to be a trimethylsilyl beta-hydroxybutyrate molecule as result of the derivatization.
The chromatogram in Figure 2 displays BHB found within the chromatogram and subsequent mass
spectra.

14

(a

(b
Figure 2. The chromatogram (a) of 5000 ppm derivatized BHB with its corresponding mass
spectrum (b). The BHB signal is found at 6.521 min.
Next, we determined how much acetonitrile to be added to help the derivatization occur.
In each reaction, 225 µL of 5000 ppm BHB stock solution was added to a headspace vial and 75
µL of BSTFA + 1% TMCS was added; the amount of acetonitrile added was 75, 150, and 225 µL
to each reaction afterwards. All were derivatized for 25 min at 80°C, and it was found that the 150
µL solution produced the highest signal for BHB. The optimal time of the derivatization was
determined afterwards. A sample that was produced during the method development for the
optimal amount of acetonitrile added was analyzed one month following to its initial derivatization
was found to have an extremely large peak area. Because of this, it was assumed that the amount

15

of time the initial derivatization in the oven was for was irrelevant as the derivatization seemed to
occur for longer than anticipated. A solution was then derivatized in an oven for 1.5 h at 80°C to
aid in as much derivatization as possible and was then kept at room temperature and analyzed
every day. The solution seemed to have the same peak area on the third and fourth day of analysis.
Therefore, the solutions were to be derivatized for 1.5 h at 80°C and then placed in a room
temperature environment for the next three days before they were analyzed by the GC-MS.
The next standard that was used was AcAc. This standard was difficult to use since its
chromatograms had many impurities, and the peak area was not as high as it was for BHB. A 2000
ppm stock solution was made for method development nonetheless, and the peak obtained was
very small, especially in comparison to the BHB peak area. This standard had two derivatizations
occur instead of one. AcAc has one hydroxyl group, but based on the mass spectra and the NIST
database, it seemed like there were two derivatizations occurring. The carboxylic acid remained
the same, but the lone carbonyl tautomerized and was subsequently converted to a hydroxyl group
which was also derivatized.
The AcAc standard solution was not acidified to see if it had any effect on the end result.
This produced a chromatogram that had a greater peak area than the one previously seen with the
acidified AcAc. Therefore, the experimental procedure for BHB was reexamined since both ketone
bodies were predicted to be within the biological fluids. The BHB sample solution was not
acidified, derivatized, and subject to analysis; the results showed that it still had a distinct signal,
but the peak area was not as high as it was for the acidified BHB. In an effort to detect quality
ketone body signals, the solutions were not acidified to aid in detecting AcAc among the biological
samples. Figure 3 includes a chromatogram and mass spectrum for AcAc.

16

(a

(b
Figure 3. The chromatogram (a) of 5000 ppm derivatized AcAc with its corresponding mass
spectrum (b). The AcAc signal is found at 7.197 min.
A calibration curve was produced for both analytes separately. This was done by preparing
standards of increasing concentrations and obtaining the peak areas of each signal and plotting it
on a graph. The concentrations used were 50 ppm, 100 ppm, 250 ppm, 500 ppm, 1000 ppm, 2000
ppm, and 5000 ppm. The range of concentrations was very wide in an effort to account for the
variation in ketone bodies produced by each individual. These solutions were analyzed in triplicate,
save for the 100 ppm standard of BHB which was analyzed in duplicate; this is because the run

17

had an error which stopped the run before the BHB signal was analyzed. The derivatized ketone
bodies that the mass spectrometer detected are found in Figure 2.

Figure 4. The derivatized acetoacetate (a) and beta-hydroxybutyrate (b) that were to be analyzed
by the GC-MS.
For the BHB standard curve, there was a trend of linearity based on the coefficient of
variation (R2=0.9742). The limit of detection (LOD) for the BHB standard curve was found using
the standard deviation of the lowest concentration standard used divided by the slope of the curve
and multiplied by 3. Likewise, the limit of quantification (LOQ) was found in the same manner
but was multiplied by 10 instead of 3. The LOD and the LOQ was found to be 26.48 ppm and
88.28 ppm, respectively. Other papers have determined the LOD to be 0.833 ppm (Holm et al.
2010), and 10 ppm (Paul et al. 2006). The closest to the LOD obtained in this experiment is by
Paul et al., but that result is much smaller than the results experimentally determined. Because of
this, the method should be further optimized since this method isn’t able to detect low levels of
ketone bodies as effectively as the other methods do. One way to do this could be to run many
trials, and then average the spectra together to lower the signal-to-noise ratio.

18

Figure 5. The beta-hydroxybutyrate standard curve using the GC-MS.

For the AcAc standard curve, the GC-MS was not able to detect the ketone body standards
below 1000 ppm. It also produced unreliable signals for standards that were greater than 500 ppm
as the 2000 ppm standard seemed to have a peak area consistently larger than 5000 ppm. Because
the chromatogram had many foreign peaks as well, it was suspected that the standard was
contaminated. Upon obtaining a second batch of AcAc, it also seemed to have foreign peaks and
difficulties detecting the AcAc ketone body. Thus, there was no standard curve generated for
AcAc.

19

3.2 The Qualitative Results of the Diet
The participants were to follow a ketogenic diet in an effort to produce ketone bodies within
their body. The participants were to eat 3 g of fats for every 1 g of protein and carbohydrates. Each
participant documented what they ate, and the time they ate the food at. They were encouraged to
eat food until they were full, and not eat a specific amount of calories.
The first meal that was recorded was breakfast. It was defined as a substantial meal that
was ingested before 10:30 am. Because each participant had different eating habits, this meal was
not forced upon the participants. The food that was consumed by the participants were recorded
and displayed in Table 1. It was the least consumed meal as there were two participants (2 and 4)
who exclusively did not eat breakfast; Participant 5 and 6 each had breakfast only on one day, and
Participants 7 and 8 seemed to have only tea in the morning save for the Day 2 breakfast for
Participant 7. For those who did not eat breakfast, their body would be in a fasting state which
would increase the amount of triglycerides that were eventually oxidized to produce ketone bodies
(Mullins et al. 2011; Kanikarla-Marie and Jain 2016). As for those who ate a breakfast, they mainly
ate some form of eggs with or without bacon which are both higher in protein and fat content. This
is fine since biochemical pathways will prefer to use triglycerides as an energy source before
proteins (Kanikarla-Marie and Jain 2016). Though the protein content is high, the amount of
carbohydrates ingested were very low. This is beneficial since the amount of glucose circulating
within the blood is minimized. Therefore, it seemed like the participants were following the
ketogenic diet during breakfast.

20

Table 1. The breakfast consumed by each participant during the course of the diet.
Participant 2

Participant
3

Participant
4

Participant
5

Participant
6

Participant 7

Participant 8

Participant
9

Participant
10

Breakfast
Day 1

N/A

Bacon and
Eggs

N/A

Omelette

N/A

English tea
and 2 cheese
strings

N/A

N/A

Breakfast
Day 2

N/A

Bacon and
Eggs

N/A

N/A

Bacon and
peanut
butter

Mint
tea

green

Breakfast
Day 3

N/A

N/A

N/A

N/A

Mint
tea

green

Greek
yogurt and
strawberries

Bacon and
eggs

Breakfast
Day 4

N/A

2
egg
omelette
with
onions,
green
peppers,
bacon,
cheese,
and
chicken
Bacon and
Eggs,
berries,
and whip
cream

English tea
with some
almond
milk,
omelette
cooked in
butter with
pepper,
onions,
hams, and
mozzarella
cheese
English tea
with some
almond milk

Scrambled
eggs with
spinach,
avocado,
and
pepperoni
stick
Scrambled
eggs

N/A

N/A

N/A

N/A

Mint
tea

green

Greek
yogurt and
strawberries

Bacon and
eggs and
black
coffee.

Red pepper
slices,
black
coffee

The second meal that was recorded was lunch. It was defined as a substantial meal that was
ingested after 10:30 am and before 4:00 pm since each participant would have different eating
habits and schedules. The food that was consumed by the participants at lunch were recorded and
displayed in Table 2. Lunch was eaten by all of the participants, but Participant 3 only consumed
lunch 2 out of the 4 days. It’s worth considering that this participant was also not consuming
breakfast; there is a chance that this participant was producing many ketone bodies as a result of
this fast. The meals that were eaten seem to consist of a lot of vegetables, mainly in the form of a
salad. There was also a noticeable amount of bacon and pork chops consumed. The vegetables
contain many vitamins that the participants would be ingesting during the course of the diet; they
themselves have little protein, fat, and carbohydrates. Pork chops have a high protein content, but

21

there still is a substantial amount of fat which is still relevant for the purpose of the diet. Bacon
has more fat then protein, which is extremely beneficial for the purpose of this diet. There does
not seem to be any particular meal with a high content of carbohydrates which would therefore
hinder the results of this experiment. Thus, it seemed like the participants were following the
ketogenic diet during lunch.
Table 2. The lunch consumed by each participant during the course of the diet.
Participant
2

Participant
3

Participant 4

Participant 5

Participant
6

Participant 7

Participant 8

Participant 9

Participant
10

Lunch
Day 1

Black
Coffee,
Beef Jerky,
Cucumber

Chicken
breasts with
hot sauce

Scrambled
eggs cooked
in
butted
with cheese
and ketchup

Lettuce
wrap burger

2 eggs with
siracha,
bacon with
mustard

4 eggs with
bacon

N/A

Red pepper
and
cucumber

Lunch
Day 2

Pork
Chops
(cooked in
butter),
broccoli,
bell
peppers,
zucchini,
beans
(called
vegetable
medley)
Pork chops
with
vegetable
medley

N/A

Chicken,
bacon,
lettuce,
cheese salad

N/A

N/A

Omelette
with ham

Brie cheese,
pepperoni
stick, some
red pepper,
and
strawberries

Keto
Avocado,
Bacon, and
goat-cheese
salad with
ranch
dressing

Raspberries
with
whipping
cream

Scrambled
eggs, butter,
shredded
cheese, and
strawberries

Bacon with
mustard and
an apple

Bacon, apple,
and
dark
chocolate

N/A

Pepperoni
sticks and
pork chop

Spinach
and goat
cheese
salad with
olive oil

Tea
with
Steak
and
parmesan
cheese chips

Steak
and
dark
chocolate

Cheddar
cheese, chia
tea
with
heavy
cream,
½
green pepper
Salad with
cheddar
cheese, feta,
eggs, bacon,
and Caesar
dressing

Keto
pancakes
with
raspberries

Cucumber
and cheese

Green salad
with
Chinesestyled
chicken and
beef
Chicken
Caesar salad

Celery
with
peanut
butter, red
bell
pepper,
pepperoni
sticks,
bacon
Pepperoni
sticks,
chicken
sausage,
Brie
cheese,
celery
sticks with
peanut
butter, and
1 red bell
pepper
Zucchini
boats

Lunch
Day 3

Lunch
Day 4

Keto
pancakes
with
raspberries
and
strawberries

The third meal that was recorded was dinner. It was defined as a substantial meal that was
ingested after 4:00 pm and before 8:30 pm since each participant would have different eating habits
and schedules. The food that was consumed by the participants at dinner were recorded and
displayed in Table 3. Every participant ate dinner for the duration of the ketogenic diet, except for
Participant 9 who did not eat dinner on Day 4. There seemed to be a large amount of protein eaten

22

during this time; the protein includes burgers and chicken mainly, but other options like pork chops
and steak. The burgers all seem to have been eaten with a lettuce bun, save for Participant 8 on
Day 2. This is beneficial since the bread is a source of carbohydrates which would have potentially
tamper the results of the experiment. As for Participant 8, the participant seemed to still consume
a high enough content of fats in comparison to the amount of carbohydrates eaten, so the effect of
glucose within the bloodstream may be minimized. Chicken has a higher content of protein in it,
but there still is a fat content making it suitable for the diet. Other than the previously mentioned
dinner, there does not seem to be any particular meal with a high content of carbohydrates which
would therefore hinder the results of this experiment. The tacos consumed by Participants 7 and 8
slightly raise a flag because the taco shell contains carbohydrates, but there should be enough fat
within the meal that minimizes the effects of the carbohydrates. Thus, it seemed like the
participants were following the ketogenic diet during dinner.

23

Table 3. The dinner consumed by each participant during the course of the diet.
Participant
2

Participant
3

Participant 4

Participant 5

Participant
6

Participant 7

Participant 8

Participant 9

Participant
10

Dinner
Day 1

Pork
Chops,
green
Beans, bell
peppers,
broccoli,
zucchini

2
beef
patties,
lettuce
wrap,
onion, hot
sauce

Green salad
and chicken
wings

Hardboiled egg.
A
salad
with
bacon,
avocado,
and goat
cheese.

3 tacos with
lettuce as a
bun, ground
beef, onion,
red and green
peppers,
cheese, and
hot sauce

3 tacos with
lettuce as a
bun, ground
beef, onion,
red and green
peppers,
cheese, and
hot sauce

Spinach
salad with
bacon, goat
cheese, and
avocado

Bacon and
eggs

Dinner
Day 2

Pork chops,
bacon,
vegetable
medley,
cheese

Rotisserie
chicken

2
Burgers
with cheese,
pickles,
bacon,
onions,
mayo,
mustard, and
lettuce as a
bun
2
Burgers
with cheese,
pickles,
bacon,
onions,
mayo,
mustard, and
lettuce as a
bun

Chicken
burger (with
lettuce as the
bun), Green
salad

Bacon and
avocado
salad with
spinach
and goat
cheese,
and
2
pepperoni
sticks

Cheeseburger
and Caesar
salad without
croutons

Zucchini
boats with
pepperoni
sticks,
parmesan,
and
goat
cheese

Pork chops
with green
beans and
pickles

Dinner
Day 3

Chicken
and
avocado
salad (with
cucumbers,
bell
peppers,
kale,
lettuce)

Chicken
breasts and
bacon

Caesar salad
(bacon
cheese,
lettuce,
grilled
chicken, and
Caesar
dressing)

Burgers with
lettuce bun

Jerk chicken
legs,
pepperoni,
and
cheese
string.
Later
was
steak

Zucchini
boats and a
pepperoni
stick

Pork chops
with green
beans

Dinner
Day 4

Avocado
Salad (with
cheese,
cucumber,
lettuce,
bacon,
olive oil)

Chicken
breasts

Caesar salad
(bacon
cheese,
lettuce,
grilled
chicken, and
Caesar
dressing)

Southern
Chicken and
Beef salad

Cucumber,
green bell
pepper
with cream
cheese,
peanut
butter,
bacon,
cheese,
and baconcheese
balls.
2
beef
burgers in
iceberg
lettuce
with Brie,
garlic aioli,
and onion

Burger with
lettuce bun,
cheddar
cheese,
tomato,
pickles,
bacon,
special sauce,
served with a
side Caesar
salad
with
bacon and no
croutons
Steak
and
left-overs
from
the
tacos from
Dinner Day 1

Jerk chicken
leg, one hot
Italian
sausage,
veggie
stir
fry

chicken leg,
one
hot
Italian
sausage,
veggie stir fry

N/A

Keto
Avocado,
Bacon, and
goatcheese
salad with
ranch
dressing

The last meal that was recorded was snacks that would occur throughout the day. To be a
snack, it had to be eaten at a separate time from when the three major meals were consumed and
in a lesser amount. Snacks also included beverages that were had throughout the day that weren’t
designated as a breakfast drink or water. The food that was consumed by the participants as snacks
were recorded and displayed in Table 4. Majority of the participants had snacks throughout the
day. The snacks seem to be potentially worrisome in regard to the carbohydrates being consumed.
For instance, Participants 7 and 8 both consumed beer and carrots which are extremely high in
carbohydrates. Their dinners also had a higher amount of carbohydrates in comparison to the other

24

participants. Because of this, they may have ingested more carbohydrates than fats, and
compromised their results for the experiment. Of course, this would have to be determined based
on the analysis of their biological fluids. Pepperoni sticks were the most common snack that was
ingested. This food contains an extremely high concentration of fat in it, a moderate amount of
protein, and essentially no carbohydrates. This was an ideal food to snack on during the duration
of the diet. Participants 7 and 8 did eat these, but it is unclear if they ate enough for it to make the
carbohydrates they consumed to follow the 3:1 ketogenic diet ratio. In general, it seemed like the
snack foods had very little carbohydrates in it.

25

Table 4. The snacks consumed by each participant during the course of the diet.
Participant
2

Participant
3

Participant 4

Participant 5

Participant
6

Participant 7

Participant 8

Participant 9

Participant
10

Snacks
Day 1

Pepperoni
Sticks

Beef jerky

N/A

N/A

Coffee,
pepperoni
sticks, and
bacon

Carrots, tea,
cheese
strings,
pepperoni
sticks, apple,
dark
chocolate

Red pepper,
pepperoni
sticks, and a
cup
of
halotop ice
cream

Pepperoni
sticks

Snacks
Day 2

Cucumber,
pepperoni,
cheese,
pork rinds

Almonds
and beef
jerky

3
eggs
cooked in
butter with
shredded
cheese and
pepper

Raspberries,
strawberries
,
and
blueberries

Coffee

Dark
chocolate,
apple,
pepperoni
sticks, and
cheese
strings

Haltotop ice
cream, Brie
cheese

Pepperoni
sticks and
cucumber

Snacks
Day 3

Pepperoni
sticks,
cheese

Almonds
and beef
jerky

N/A

N/A

4 beers

Peanut
butter and
halotop ice
cream.

Halotop ice
cream,
pepperoni
sticks

Snacks
Day 4

Pepperoni
sticks,
cheese

N/A

Pepperoni
sticks,
strawberries
,
peanut
butter,
ketogenic
pancakes

N/A

Coffee,
baconcheese
balls,
celery
sticks with
peanut
butter, red
bell
pepper
with
cream
cheese
Coffee,
Baconcheese
balls,
celery
stick with
cream
cheese,
and
3
pepperoni
sticks

Parmesan
cheese chips,
strawberries,
salami,
cheese
string, some
dark
chocolate,
and tea.
Strawberries
,
carrots,
cheese
strings, hot
pepperoni
sticks,
pickles,
double gin
and tonic
3
beers,
strawberries,
dark
chocolate

Tea
with
almond milk,
hot salami,
cheese
strings,
strawberries

Berries

Brie cheese,
cheddar
cheese,
strawberries
,
and
pepperoni
sticks,
peanut
butter

Salad with
avocado,
cheese,
beans,
tomatoes,
and corn

In general, the participants overall ate minimal carbohydrates, but they did consume a large
amount of protein and fat. The targeted ratio of fat to protein and carbohydrates was supposed to
be 3:1. Based on the food that was ingested, it seems like the amount of fats consumed was
essentially the same compared to the amount of fats, and that the carbohydrates consumed were
much lower. It seemed like the participants ended up following the Atkins diet eating a ratio closer
to 60% fats, 30 % protein, and 10% carbohydrates, as opposed to the classical diet (Kang et al.
2007). Even though the diet has changed, blood ketosis is still probable as previous studies have
recorded patients having BHB levels greater than 3 mM within 3 days of following the diet (Kang

26

et al. 2007). But it is worth noting that the amount of food and the way it was cooked was not
recorded, so there is a chance that there are more carbohydrates within the meals than was written
by the participants. There is also a chance that the participants did not write down all of the food
they ate, and they cheated on the diet. Though difficult to prove, it is worth considering when
observing the results from the analysis.

3.3 The Quantification of Ketone Bodies
3.3.1 Quantification in Blood
Blood is one of the most accurate ways to determine the concentration of ketone bodies
within the individual as ketone bodies circulate through the body via blood (Laffel 1999;
Kanikarla-Marie and Jain 2016; Brooke et al. 2016). The sample was then brought back to
physiological pH because acetoacetate was better able to be derivatized and detected at pH 7
compared to below pH 2. The samples were analyzed in duplicate on the GC-MS, and the results
are displayed in Table 5. Based on the results, it seems like Participant 2, 4, 5, and 8 seem to have
increased the amount of BHB within their blood. Each trial of Participant 2’s blood was able to
detect BHB, which strengthens the argument that there was an increase in the levels of BHB
circulating their body. Both Participants 4 and 8 did not have a BHB signal detected before the
diet occurred, which could mean that the amount of BHB within their system was below the LOD
and LOQ of the instrument, or that there was variation within the instrument during the different
runs. Participants 6, 7, 9, and 10 seem to have a lower amount or similar amount of BHB in the
after-the-diet compared to the before-the-diet samples. This could result from the participants not
following the diet that was prescribed to them. By eating carbohydrates, their body would not have
shunted glycolysis and the citric acid cycle, which would have resulted in little amounts of ketone

27

bodies produced. Another reason why BHB may have been detected could be a result of the
participants fasting longer than 12 h as the concentration of BHB increases over time (Marinou et
al. 2011). Statistics were not conducted since not every participant was able to generate average
values of BHB before and after the diet.
Table 5. The experimentally determined amounts of beta-hydroxybutyrate in the blood samples.
Participant Trial
BHB
Retention Average BHB
Retention Average
Number
Number Before
Time
BHB
After
Time
BHB
Diet
(min)
Before
Diet
(min)
After
(ppm)
Diet
(ppm)
Diet
(ppm)
(ppm)
2
1
245.46
6.520
274.96
390.89
6.521
398.09
2
2
304.48
6.519
405.29
6.520
3
1
0
N/A
N/A
0
N/A
N/A
3
2
248.90
6.519
268.83
6.520
4
1
0
N/A
N/A
240.48
6.522
245.45
4
2
0
N/A
250.43
6.520
5
1
0
N/A
N/A
0
N/A
N/A
5
2
246.69
6.519
339.99
6.520
6
1
0
N/A
N/A
0
N/A
N/A
6
2
247.04
6.520
258.12
6.520
7
1
250.64
6.519
261.75
247.34
6.518
247.05
7
2
272.86
6.520
246.75
6.5117
8
1
0
N/A
N/A
240.24
6.52
242.22
8
2
0
N/A
244.21
6.518
9
1
0
N/A
N/A
0
N/A
N/A
9
2
266.06
6.521
283.55
6.646
10
1
0
N/A
N/A
239.60
6.519
254.09
10
2
257.57
6.519
268.58
6.518
Interestingly enough, it seems like the results agreed with the results that were expected
within the blood. By converting the concentrations in ppm to mM, the amount of BHB within the
blood samples seem to agree with the results of previously found by Owen et al. (1967), Owen et
al. (1969), Galvin et al. (1968) and Garber et al. (1974) being in between around 1-2 mM after 4
days. There is variation between individuals as everyone has a slightly different metabolism and
ate slightly different foods. It is also worth acknowledging that Participant 2 had a concentration
28

of 3.82 mM of BHB which refers to potential ketoacidosis arising based on the symptom
ketonaemia (Laffel 1999; Usher-Smith et al. 2011). It is unknown if that was the situation since
the blood glucose and blood pH was not measured, but it is possible.
Within the blood, there was no AcAc signals detected in any of the samples. It is possible
that there was AcAc within the blood samples, but the concentration was below the LOD. In the
future, certified reference material should be utilized during method development to help quantify
the AcAc from within the blood.
There are many runs where there is no signal detected. It does not necessarily mean that
there were no ketone bodies present within the fluid, but that it may have been below the LOD. As
well, there was no internal standard used which would have been used to determine if there was
any variation between runs. It does seem like there was variation between runs since some samples
have BHB detected in one trial, but not the other trial. It is unknown what could have caused that
variation between runs to occur.

3.3.2 Quantification in Urine
The participants also had their urine collected before and after conducting the diet. The
samples were analyzed in duplicate on the GC-MS, and the results are displayed in Table 6. There
was no AcAc detected in the urine. This is somewhat surprising since it is measured in commercial
methods currently (Brooke et al. 2016). A possible reason could be that the urine had an acidic pH
and was not adjusted to pH 7. Because of this, there needs to be further method development
conducted to ensure that AcAc is observed within urine. As well, there was an extremely high
amount of BHB recorded within urine. The amount of BHB was negligible for all of the
participants before the diet, but they almost all seemed to have higher BHB levels within their

29

urine than in their blood. Statistics were not conducted since not every participant was able to
generate average values of BHB before and after the diet.

Table 6. The experimentally determined amounts of beta-hydroxybutyrate in the urine samples.
Participant Trial
BHB
Retention Average BHB
Retention Average
Number
Number Before Time
BHB
After
Time
BHB After
Diet
(min)
Before
Diet
(min)
Diet (ppm)
(ppm)
Diet
(ppm)
(ppm)
2
1
0
N/A
N/A
7461.37 6.521
9572.11
2
2
0
N/A
11682.8 6.520
5
3
1
0
N/A
N/A
871.75
6.520
663.56
3
2
0
N/A
455.37
6.521
4
1
0
N/A
N/A
294.94
6.523
369.18
4
2
0
N/A
443.423 6.521
5
1
0
N/A
N/A
276.17
6.522
278.74
5
2
0
N/A
281.31
6.523
6
1
0
N/A
N/A
1133.22 6.520
1042.82
6
2
0
N/A
952.41
6.519
7
1
0
N/A
N/A
324.80
6.520
349.49
7
2
0
N/A
374.19
6.520
8
1
0
N/A
N/A
269.51
6.523
275.54
8
2
0
N/A
281.56
6.524
9
1
0
N/A
N/A
350.65
6.520
366.65
9
2
0
N/A
382.65
6.520
10
1
0
N/A
N/A
1103.14 6.519
1553.35
10
2
0
N/A
2003.56 6.520
In a study conducted by Liu et al., they observed participants biological fluids after fasting
a minimum of 9 h and found 0.12 µM of BHB in urine compared to 68.8 µM in plasma (2015).
Based on this piece of data, it seems like BHB is at a higher concentration within the bloodstream
than it is within urine. But in the past, as ketosis occurs for a longer period of time, the amount of
AcAc in urine decreases and the amount of BHB excreted increases (Galvin et al. 1968). Though,
it is difficult to compare how much AcAc and BHB were analyzed since it was in units of µmol/min

30

(Galvin et al. 1968). Nonetheless, it is still believed that the amount of BHB detected is positively
skewed since the urine is acidic. Having acidic urine would likely derivatize more BHB within the
vial, thus a higher concentration was seen by experiment. To know for sure, the experiment would
have to be conducted again, except there would have to be a control for the pH of urine.

3.3.3 Quantification in Saliva
Saliva was also collected and analyzed to see how much ketone bodies were present. Saliva
is becoming a more popular diagnostic model to use as it is less invasive and isotonic to plasma
(Aps and Martens 2005; Elmongy and Abdel-Rehim 2016; Proctor 2016). The concept of
quantifying ketone bodies within saliva in ketotic participants is rather novel. The samples were
analyzed in duplicate on the GC-MS, and the results are displayed in Table 7. There was no AcAc
detected in the saliva. It was expected that it would have very little amounts of AcAc since saliva
is isotonic to blood which contains a small amount of AcAc in it (Liu et al. 2015). As for BHB,
there was very little signal that was detected overall. There were 4 signals that were obtained from
the before-the-diet samples. The concentration of BHB in these samples are very similar to the
results obtained in these participants bloods samples. There were only two signals obtained in the
after-the-diet samples by Participant 3 and 4. These signals were smaller than the signal detected
in the before-the-diet sample for these participants. Theoretically, it should be higher since the
ketogenic diet would have occurred, increasing the amount of ketone bodies within the individuals.
Statistics were not conducted since not every participant was able to generate average values of
BHB before and after the diet.

31

Table 7. The experimentally determined amounts of beta-hydroxybutyrate in the saliva samples.
Participant Trial
BHB
Retention Average BHB
Retention Average
Number

Number Before
Diet

Time

BHB

After

Time

BHB After

(min)

Before

Diet

(min)

Diet (ppm)

Diet

(ppm)

0

N/A

N/A

0

N/A

0

N/A

243.84

6.519

240.48

6.522

0

N/A

(ppm)

(ppm)
2

1

253.70

6.520

2

2

245.48

6.520

3

1

253.22

6.521

3

2

0

N/A

4

1

244.44

6.518

4

2

0

N/A

249.59

N/A

N/A

N/A

N/A

In a study conducted by Liu et al., they previously observed participants biological fluids
after fasting a minimum of 9 h and found 1.74 µM of BHB in saliva compared to 68.8 µM in
plasma (2015). Based on this result, it would be expected that there is some BHB detected, but not
as much as in the blood. There are two potential reasons why the amount of signal detect was very
low. The first could be that the amount of BHB within the sample was below the LOD, which is
logical considering the amount of BHB in the plasma is almost 40x larger than that observed in
the saliva (Liu et al. 2015). The other reason is that there was not a lot of BHB that was derivatized
within the sample, so the signal was very low as a result. Likewise, the samples that did produce a
signal could have had a very effective derivatization by chance. A way to improve upon this would
be to add a recovery standard to determine how effective the derivatization was to ensure that all
of the potential analyte was derivatized and detected.

32

4.0 CONCLUSIONS
In conclusion, participants underwent a ketogenic diet to induce ketosis; samples of their
blood, urine, and saliva were collected before and after the diet to determine if BHB and AcAc
could be quantified through the GC-MS. The conditions were moderately optimized to obtain the
best possible signals for the BHB and AcAc ketone bodies for the time being. A calibration curve
for BHB was produced (R2 = 0.9742) and was used to quantify the amount of ketone bodies present
in the biological samples. For the participants whose blood samples were able to be quantified, it
seems like they were in ketosis and the amount of BHB seems within range of what was expected.
The urine samples had a high concentration of BHB, which could be a result of the acidic urine
which would help in the derivatization step, so it could be analyzed. There was very little signal
detected within saliva, but the fact that signal was detected means that this could potentially act as
a method to quantify ketone bodies upon more method development.

33

5.0 FUTURE WORK
Future work for this research would look to optimize the detection of AcAc from
biological fluids. Possible ideas include trying to separate this ketone body from the BHB and
using a different derivatizing agent to make it volatile. Future work should also consider using
two recovery standards to know how effective the derivatization is, and if any analyte is lost
during the matrix evaporation step. Adding an internal standard would be useful to determine if
there is any variation between runs. Another idea for future work would be to control the pH
across all biological fluids since it seems like urine produced a higher signal because the acidity
of the matrix favored the derivatization reaction to occur. Lastly, optimizing a better method to
extract, derivatize, and analyze the BHB from saliva. Possible ideas include acidifying the pH of
a sample to protonate the analyte in an effort to do liquid-liquid microextraction with hexane.
Then, concentrating the solution and derivatizing it. This may be effective for detecting BHB, so
the future work on AcAc will dictate if the method would change to detect the amount of AcAc
in the biological fluid. Potentially, this method could be used on all biological fluids as well.

34

6.0 LITERATURE CITED
Aps JKM, Martens LC. 2005. Review: The physiology of saliva and transfer of drugs into
saliva. Forensic Sci. Int. 150:119–131. doi:10.1016/j.forsciint.2004.10.026.
Branco AF, Ferreira A, Simões RF, Magalhães-Novais S, Zehowski C, Cope E, Silva AM,
Pereira D, Sardão VA, Cunha-Oliveira T. 2016. Ketogenic diets: From cancer to mitochondrial
diseases and beyond. Eur. J. Clin. Invest. doi:10.1111/eci.12591.
Brooke J, Stiell M, Ojo O. 2016. Evaluation of the Accuracy of Capillary Hydroxybutyrate
Measurement Compared with Other Measurements in the Diagnosis of Diabetic Ketoacidosis : A
Systematic Review. :1–9. doi:10.3390/ijerph13090837.
Butalia S, Kaplan GG, Khokhar B, Rabi DM. 2016. Environmental Risk Factors and Type
1

Diabetes:

Past,

Present,

and

Future.

Can.

J.

Diabetes

40:586–593.

doi:10.1016/j.jcjd.2016.05.002.
Courchesne-Loyer A, Fortier M, Tremblay-Mercier J, Chouinard-Watkins R, Roy M,
Nugent S, Castellano CA, Cunnane SC. 2013. Stimulation of mild, sustained ketonemia by
medium-chain triacylglycerols in healthy humans: Estimated potential contribution to brain energy
metabolism. Nutrition 29:635–640. doi:10.1016/j.nut.2012.09.009.
Dholakia S, Mittal S, Quiroga I, Gilbert J, Sharples EJ, Ploeg RJ, Friend PJ. 2016. Pancreas
Transplantation:

Past,

Present,

Future.

Am.

J.

Med.

129:667–673.

doi:10.1016/j.amjmed.2016.02.011.
Elmongy H, Abdel-Rehim M. 2016. Saliva as an alternative specimen to plasma for drug
bioanalysis. A review. TrAC - Trends Anal. Chem. 83:70–79. doi:10.1016/j.trac.2016.07.010.
Føreid S, Gadeholt G. 2017. Beta-hydroxybutyrate and pyroglutamate can be included in
a rapid GC-MS screening method for differential diagnosis of metabolic acidosis. Scand. J. Clin.
Lab. Invest. 77:149–152. doi:10.1080/00365513.2016.1278261.
Fujii S, Maeda T, Noge I, Kitagawa Y, Todoroki K, Inoue K, Min JZ, Toyo’oka T. 2014.
Determination of acetone in saliva by reversed-phase liquid chromatography with fluorescence
detection and the monitoring of diabetes mellitus patients with ketoacidosis. Clin. Chim. Acta
430:140–144. doi:10.1016/j.cca.2014.01.006.

35

Galvin RD, Harris JA, Johnson RE. 1968. Urinary Excretion of Beta‐Hydroxybutyrate and
Acetoacetate During Experimental Ketosis. Q. J. Exp. Physiol. Cogn. Med. Sci. 53:181–193.
doi:10.1113/expphysiol.1968.sp001958.
Garber AJ, Menzel PH, Boden G, Owen OE. 1974. Hepatic ketogenesis and
gluconeogenesis in humans. J. Clin. Invest. 54:981–989. doi:10.1172/JCI107839.
Giugliano D, Ceriello A, Esposito K. 2008. Glucose metabolism and hyperglycemia. Am.
J. Clin. Nutr. 87:217S–222S. doi:10.1093/ajcn/87.1.217S.
Hassan HM a, Cooper G a a. 2009. Determination of -Hydroxybutyrate in Blood and Urine
Using

Gas

Chromatography--Mass

Spectrometry.

J.

Anal.

Toxicol.

33:502–507.

doi:10.1093/jat/33.8.502.
Holm KMD, Linnet K, Rasmussen BS, Pedersen AJ. 2010. Determination of Ketone
Bodies in Blood by Headspace Gas Chromatography-Mass Spectrometry. J. Anal. Toxicol.
34:549–554. doi:10.1093/jat/34.9.549.
Kang HC, Lee HS, You SJ, Kang DC, Ko TS, Kim HD. 2007. Use of a modified Atkins
diet

in

intractable

childhood

epilepsy.

Epilepsia

48:182–186.

doi:10.1111/j.1528-

1167.2006.00910.x.
Kanikarla-Marie P, Jain SK. 2016. Hyperketonemia and ketosis increase the risk of
complications

in

type

1

diabetes.

Free

Radic.

Biol.

Med.

95:268–277.

doi:10.1016/j.freeradbiomed.2016.03.020.
Laffel L. 1999. Ketone bodies: a review of physiology, pathophysiology and application
of monitoring to diabetes. Diabetes. Metab. Res. Rev. 15:412–26.
Linstrom PJPJ, Mallard WGG. 2001. NIST Chemistry webbook; NIST standard reference
database No. 69 . NIST Chem. Webb.:20899. doi:10.18434/T4D303.
Liu SL, Oyama T, Miyoshi Y, Sheu SY, Mita M, Ide T, Lindner W, Hamase K, Lee JA.
2015. Establishment of a two-dimensional chiral HPLC system for the simultaneous detection of
lactate and 3-hydroxybutyrate enantiomers in human clinical samples. J. Pharm. Biomed. Anal.
116:80–85. doi:10.1016/j.jpba.2015.05.036.
Marinou K, Adiels M, Hodson L, Frayn KN, Karpe F, Fielding BA. 2011. Young women
partition fatty acids towards ketone body production rather than VLDL-TAG synthesis, compared
with young men. Br. J. Nutr. 105:857–865. doi:10.1017/S0007114510004472.

36

Morales J, Schneider D. 2014. Hypoglycemia. Am. J. Med. 127:S17–S24.
doi:10.1016/j.amjmed.2014.07.004.
Mullins G, Hallam CL, Broom I. 2011. Ketosis, ketoacidosis and very-low-calorie diets:
putting the record straight. Nutr. Bull. 36:397–402. doi:10.1111/j.1467-3010.2011.01916.x.
Nickerson JG, Alkhateeb H, Benton CR, Lally J, Nickerson J, Han XX, Wilson MH, Jain
SS, Snook LA, Glatz JFC, et al. 2009. Greater transport efficiencies of the membrane fatty acid
transporters FAT/CD36 and FATP4 compared with FABPpm and FATP1 and differential effects
on fatty acid esterification and oxidation in rat skeletal muscle. J. Biol. Chem. 284:16522–16530.
doi:10.1074/jbc.M109.004788.
Niwa T. 1995. Basic theory of mass spectrometry. Clin. Chim. Acta. 241–242:15–71.
doi:10.1016/S0009-8981(00)89104-4.
O’Neill HM, Holloway GP, Steinberg GR. 2013. AMPK regulation of fatty acid
metabolism and mitochondrial biogenesis: Implications for obesity. Mol. Cell. Endocrinol.
366:135–151. doi:10.1016/j.mce.2012.06.019.
Owen OE, Morgan AP, Kemp HG, Sullivan JM, Herrera MG, Cahill GF. 1967. Brain
metabolism during fasting. J. Clin. Invest. 46:1589–1595. doi:10.1172/JCI105650.
Owen OE, Wahren J, Jr GFC, Owen OE, Felig P, Morgan AP, Wahren J, Cahill GF. 1969.
Liver and kidney metabolism during prolonged starvation Find the latest version : Liver and
Kidney Metabolism during Prolonged Starvation. 48:574–583. doi:10.1172/JCI106016.
Paul R, Tsanaclis L, Kingston R, Berry A, Guwy A. 2006. GC-MS-MS determination of
gamma-hydroxybutyrate

in

blood

and

urine.

J.

Anal.

Toxicol.

30:375–379.

doi:10.1093/jat/30.6.375.
Pociot F, Lernmark Å. 2016. Genetic risk factors for type 1 diabetes. Lancet 387:2331–
2339. doi:10.1016/S0140-6736(16)30582-7.
Proctor GB. 2016. The physiology of salivary secretion. Periodontol. 2000 70:11–25.
doi:10.1111/prd.12116.
Puchalska P, Crawford PA. 2017. Multi-dimensional Roles of Ketone Bodies in Fuel
Metabolism,

Signaling,

and

Therapeutics.

Cell

Metab.

25:262–284.

doi:10.1016/j.cmet.2016.12.022.

37

Roehl K, Sewak SL. 2017. Practice Paper of the Academy of Nutrition and Dietetics:
Classic and Modified Ketogenic Diets for Treatment of Epilepsy. J. Acad. Nutr. Diet.
doi:10.1016/j.jand.2017.06.006.
Sherwin RS, Hendler RG, Felig P, Haven N. Effect of Diabetes Mellitus and Insulin on the
Turnover and Metabolic Response to Ketones in Man. 25.
Steensels S, Ersoy BA. 2019. Fatty acid activation in thermogenic adipose tissue. Biochim.
Biophys. Acta - Mol. Cell Biol. Lipids 1864:79–90. doi:10.1016/j.bbalip.2018.05.008.
Tsutsui H, Mochizuki T, Maeda T, Noge I, Kitagawa Y, Min JZ, Todoroki K, Inoue K,
Toyo T. 2012. Simultaneous determination of DL-lactic acid and DL-3-hydroxybutyric acid
enantiomers in saliva of diabetes mellitus patients by high-throughput LC-ESI-MS/MS. Anal.
Bioanal. Chem. 404:1925–1934. doi:10.1007/s00216-012-6320-0.
Usher-Smith JA, Thompson MJ, Sharp SJ, Walter FM. 2011. Factors associated with the
presence of diabetic ketoacidosis at diagnosis of diabetes in children and young adults: a
systematic review. BMJ 343:d4092–d4092. doi:10.1136/bmj.d4092.
Viswanath B, Choi CS, Lee K, Kim S. 2017. Recent trends in the development of
diagnostic tools for diabetes mellitus using patient saliva. TrAC - Trends Anal. Chem. 89:60–67.
doi:10.1016/j.trac.2017.01.011.
Wibisono C, Rowe N, Beavis E, Kepreotes H, Mackie FE, Lawson JA, Cardamone M.
2015. Ten-year single-center experience of the ketogenic diet: Factors influencing efficacy,
tolerability, and compliance. J. Pediatr. 166:1030–1036.e1. doi:10.1016/j.jpeds.2014.12.018.

38

7.0 APPENDICES
The Ketogenic Diet Outline (3 Figures below)

39

Participant 2 Blood After-The-Diet (Trial 2)

Participant 2 Blood Before-The-Diet (Trial 2)

40

Participant 2 Saliva Before-The-Diet (Trial 1)

Participant 2 Urine Before-The-Diet (Trial 1)

41

Participant 2 Urine After-The-Diet (Trial 1)

Comparison Between Acidified Beta-Hydroxybutyrate (a) and Non-Acidified BetaHydroxybutyrate at 5000 ppm Each

(a

(b

42

Comparison Between Acidified Acetoacetate(a) and Non-Acidified Acetoacetate at 1600 ppm
Each

(a

(b

43

Ethics Approval

44

Biosafety Approval

45