Background and Objectives:
Among otherwise healthy adults, there is a subgroup of individuals who develop symptoms of hypoglycemia during episodes of food restriction. The aim of the present study was to investigate whether such individuals develop hypoglycemia or react abnormally in other metabolic aspects during a 24-hour fast.
Subjects and Methods:
Ninety medical students were asked if they wanted to participate. Sixteen were selected; none dropped out. A 24-hour fast was performed at a hospital ward. Blood samples and questionnaires were taken at eight specific times.
During the fast, the sensitive group reported significantly higher scores on ‘irritation’ and ‘shakiness’. However, no hypoglycemia occurred and the lowest detected blood glucose concentration was 3.7 mmol/l. There were no differences between the groups in plasma glucose, cortisol, growth hormone (GH), insulin, β-hydroxybutyrate (β-OH) and lactate levels. The blood pressures and heart rates were also similar.
Adults, despite subjective signs of hypoglycemia, can fast without any metabolic or endocrine derangement.
There is a subgroup of healthy young adults who develop symptoms associated with hypoglycemia if they fail to get a regular intake of energy throughout the day. They claim that a prolonged time between meals will make them shaky, irritated, unable to concentrate and sweaty. According to a general belief among the affected subjects, the reason for these experiences is stress induced by low blood glucose concentrations. This makes them unable to carry out their normal daily life. This was noted in some individuals as early as the late 1970s (Yager and Young, 1974). A regular eating pattern is of utmost importance and they often express worries that their blood sugar will fall. The extent to which such symptoms are associated with verifiable hypoglycemia is not known, but blood glucose measurements below 3 mmol/l among otherwise healthy individuals have been reported (Marks, 1987).
On the other hand, there are individuals who claim that they can ignore their energy intake for long periods without any apparent trouble. They do not hesitate if they are asked to remain fasting and their only consideration is the feeling of hunger. Thus, there is a virtual difference between these two groups of apparently healthy subjects. In general, healthy individuals are considered to be capable of maintaining normoglycemia under prolonged fasting (Wiesli et al., 2004, 2005).
The underlying mechanisms that cause hypoglycemic symptoms and possible hypoglycemia in some individuals are unknown. A benign form of hypoglycemia in children is the so-called ketotic hypoglycemia (Grunt et al., 1970; Haymond et al., 1983). This is characterized by symptomatic hypoglycemia in combination with high concentrations of plasma β-hydroxybutyrate (β-OH) after insufficient energy intake in otherwise healthy children (Dahlquist et al., 1979). We hypothesized that adults with symptoms of hypoglycemia during fasting might have a similar condition.
The aim of this study was to investigate whether subjects who develop symptoms associated with hypoglycemia during periods of reduced energy intake will develop hypoglycemia during a 24-h fast and, secondly, to investigate whether they have an altered hormonal and ketone body response to the fast, compared with a control group consisting of subjects who can manage without a regular food intake without any complaints.
Subjects and methods
Two groups of healthy volunteers were recruited to perform a 24-h fast. A structured personal interview was used for the selection procedure, the aim being to evaluate their ability to manage without a regular energy intake, and according to their self-reported experience, to stand a prolonged fast. To be included in the ‘resistant’ group (n=8), the following criteria had to be fulfilled; no subjective symptoms of hypoglycemia that is headache, cold sweat, weakness, shakiness, dizziness and/or irritation when leaving out food. They were also questioned whether they had experienced longer periods of fasting before, and whether they had any subjective changes in mood during such episodes. Those eight individuals assured that they had no such problems. All participants freely volunteered to take part in the study. They were reinsured that the fast should be terminated if they developed hypoglycemia and/or severe symptoms of hypoglycemia. Five women and three men were enrolled in the sensitive group. The resistant group were composed of three women and five men.
To be included in the other group (n = 8), the subjects had to have experienced subjective emotions associated with hypoglycemia if they did not have meals regularly throughout the day. They were asked if they were used to skipping meals and what happened if they did. The eight persons included claimed that they could not miss a meal without experiencing specific symptoms connected with hypoglycemia, as mentioned above. They were always watchful not to miss a serving of food and felt like they were ‘falling in blood glucose’ if they did. This ‘sensitive’ group had never tried to fast for a longer period of time since they truly believed that they could never sustain such a test without a subsequent fall in glucose. All participants were healthy otherwise and did not use any medication that could affect the investigation. Three of the women used contraceptives (all in the sensitive group). The Ethics Committee of the Karolinska Institutet approved the study and the informed consent of all participants was obtained.
The fast started at 1800 hours and continued for 24 h. During this time, the participants were in a hospital ward and were not allowed to smoke or use any other form of tobacco. One to two hours before the fast, the subjects had a meal of their own choice. Furthermore, they had been told to avoid any strenuous training the day before they were admitted to the hospital. During the fasting period, they were only allowed to drink water.
Before the fast, two intravenous cannulas were inserted in the forearms. At 1230 hours on day 2, the volunteers took a 30 min supervised walk in an attempt to make the stay in the ward more like a regular day.
Venous samples were collected at 1800, 2200, 0200, 0600, 0900, 1200 and 1500 hours and again at 1800 hour. The following hormones and metabolites were measured: glucagon (Euro-Diagnostica AB, Malmö, Sweden), cortisol (Beckman Coulter Inc., Fullerton, CA, USA), GH (growth hormone), insulin (PerkinElmer Inc., Wellesley, MA, USA), FFAs (free fatty acids; Wako Chemicals, Richmond, VA, USA), lactate, glucose (F Hoffmann-La Roche Ltd, Basel, Switzerland) and β-OH. All analyses except β-OH were performed at the Department of Clinical Chemistry, Karolinska University Hospital Huddinge. β-OH was analyzed at The Center for Inherited Metabolic Diseases, Karolinska University Hospital Huddinge by using the method of Wildenhoff (1970). Blood pressure and pulse were measured at 1800, 2000, 2200, 1200 and 1500 hours and again at 1800 hours in all the participants, who were not informed of their blood glucose concentrations during the fast.
Subcutaneous microdialysis was used to follow short-time variation in glucose (Kamel et al., 1999). A CMA 60 catheter (length 30 mm, cutoff 60 kDa, CMA Microdialysis, Solna, Sweden) was inserted in the umbilical subcutaneous fat in every subject. The perfusate, that is Ringer's solution (Perfusion fluid, CMA Microdialysis, Solna, Sweden) entered the catheter at a rate of 0.3 μl/min administered with a CMA 106 pump. Samples were collected every 30 min during the fast and analyses were performed bedside at the CMA 600 (CMA Microdialysis, Solna, Sweden).
Throughout the fast, the participants were asked repeatedly to fill out a form to get a subjective account of how they experienced the fast. The questionnaires comprised questions pertaining to sensations and emotions that are often connected with hypoglycemia. The form was filled out at the time of blood samplings, that is eight times. The VAS (visual analogue scale) was used with the following variables: shaky, tiredness, dizziness, weakness, irritation, cold sweat, headache and hunger. The questionnaires were collected immediately after they were filled out. In addition, the subjects were not permitted to weigh their experience against that of the other participants. The checks were translated into numbers on a 0–100 scale.
Statistical analyses for changes in hormones were calculated using two-way analysis of variance (ANOVA; Statistica version 6, Assessment Systems Corporation, St Paul, MN, USA). The Mann–Whitney U-test was used for the experience when the groups were compared. The Wilcoxon-signed rank test was used for differences within the group, both for blood values and experience (SPSS version 14.0, SPSS Inc., Chicago, IL, USA). The digits from the VAS were regarded as interscale comparisons, not continuous values.
Short-time variations in subcutaneous glucose level were calculated by using the formula described by Horal et al. (1995). The absolute percentage difference between the previous and the subsequent value was calculated throughout the fast for every individual. The mean value over the 24 h was then used when the groups were compared with t-test. The values are expressed as mean±s.d. (SPSS version 14.0, SPSS Inc., Chicago, IL, USA).
All participants had a normal body composition and the groups were very similar. The clinical data on the two groups of subjects are summarized in Table 1. All subjects completed the 24-h fast and in none of them, plasma glucose level fell below 3.5 mmol/l at any time during the fasting period. Thus, none of the participants satisfied the criteria required for the diagnosis of hypoglycemia as characterized by Whipple's triad: (1) symptoms of hypoglycemia, (2) plasma glucose level of 2.5 mmol/l or less, and (3) reversibility of symptoms when glucose is administered (Karam, 2001). Glucose concentrations after the 24-h fast and changes in glucose level are presented in Figure 1. The mean decrease in glucose level was 0.9 and 1.05 mmol/l, respectively, for the sensitive and resistant group. At 1800 hours, when the fasting period started, the postprandial plasma glucose level was in the range of 4.3–6.8 mmol/l. At the end of the 24-h fast, plasma glucose level had decreased to 4.1–5.1 mmol/l. The greatest fall was found in one of the sensitive individuals (2.5 mmol/l, from 6.6 to 4.1 mmol/l). A significant difference was seen between the groups (ANOVA, P<0.05). This was due to the great divergence among the individuals at the beginning of the fast and could be explained by postprandial changes.
The examined counterregulatory hormones were in the normal range at every measuring time point (Table 2). Three women were using contraceptives and had high cortisol concentrations due to their medication. However, no statistical difference between the groups could be seen. Mean cortisol level was approximately 30% higher after 24-h fast in both groups, but only in the resistant group was the rise significant. The same pattern was observed for GH excretion: the resistant group had a more pronounced rise during the fast (P<0.05). Plasma insulin level fell in a similar manner in all subjects during the fast, and reached a stable level after 7 h. Owing to the stable pattern in glucose concentration, there was no increase in glucagon level during the fast (Karam, 2001). Lactate level remained fairly constant in all subjects during the fast, but among the sensitive individuals a significant fall (P<0.05) could be observed. All 16 subjects had initially undetectable concentrations of β-OH, that is <0.10 mmol/l. Maximum concentrations were observed at the end of the 24-h fast: medium values of 0.56 and 0.77 mmol/l for the sensitive and the resistant group, respectively. These findings are comparable to those of earlier studies in adults (Haymond et al., 1983). Only one subject exceeded 1.0 mmol/l (1.02 mmol/l, stable). No significance could be found with regard to the hormones and metabolites when the groups were compared. The results are presented in Table 2.
Lipolysis was evaluated by means of plasma FFA level. Nocturnal rises were seen in both groups, in line with earlier findings in healthy subjects (Hagstrom-Toft et al., 1997; Meyer et al., 2003). As long as the fast continued, FFA level inclined to reach a maximum of 1.0 and 1.2 mmol/l for the sensitive and the resistant group, respectively. Thus, there was no statistical difference between the groups.
Variability in subcutaneous glucose level measured with microdialysis showed no difference between the groups. No short-term hypoglycemic episodes were found during the 4-hour periods between the blood glucose measurements. Percentage variabilities during the fast were 7.9 (1.2) (mean, s.d.) and 7.0 (2.1), P=0.57, for the sensitive and the resistant group, respectively. In three of the subjects (two in the resistant and one in the sensitive group), it was possible to detect a substantial short-term drop in subcutaneous glucose level (0.5, 1.4 and 0.5 mmol/l, respectively). No clear connection between changes in glucose concentration and mood, in general, could be found among the subjects.
The sensitive group expressed more sensations of shakiness and irritation (P=0.045 and P=0.040, respectively, Table 3) at the end of the fast. For the remaining variables – fatigue, dizziness, weakness, cold sweat, headache and feeling of hunger – no differences between the groups were found, but an increase in most of the variables was observed in the sensitive group (Table 3). Blood pressure and pulse were normal and similar at all times. At the end of the fast, most of the subjects in the sensitive group verified that they had experienced the fast to be easier to perform than they had expected.
In the present study, no subjects developed hypoglycemia during the 24-h fast. All blood glucose measurements remained within the normal range. Furthermore, no major differences were found in the metabolic or hormonal response between the sensitive and the resistant groups. Thus, the present study indicates that the presumed symptoms of hypoglycemia that frequently occur among healthy subjects are not necessarily associated with a reduced fasting capacity. On the other hand, it is reported that cognitive deterioration may occur after a prolonged fast in spite of normoglycemia (Wiesli et al., 2005).
Healthy individuals generally have the ability to manage without food for 24–48 h without any remarkable changes in blood glucose level (Wiesli et al., 2004, 2005). In case of depressed food intake during longer periods, the human body has several systems to provide additional energy such as ketone bodies as well as substrates for gluconeogenesis, which will keep the blood glucose within the normal range (Champe and Harvey, 2005). There is no doubt that some individuals may experience a reduced energy intake to be more stressful than others, but as long as no other metabolic disease is present, the ability to maintain blood glucose in the normal range does not seem to be affected. However, otherwise healthy individuals do not infrequently claim that they experience hypoglycemic symptoms during their ordinary daily life.
It may be speculated that even a minor decrease in plasma glucose may cause a distinct neurohormonal response, which will be unpleasant for some individuals. The decrease in plasma glucose may not reach hypoglycemic levels, but still give an uncomfortable feeling. It has been shown that minute variations in glucose concentrations can influence feelings of hunger and satiety (Chapman et al., 1998; Gielkens et al., 1998). However, we were not able to detect any increased variability in the peripheral glucose levels measured in subcutaneous adipose tissue in the sensitive subjects. Thus, the difference between the groups is probably explained by an increased sensitivity rather than differences in glucose regulation. An alternative explanation could be that differences in neuro- and gastrointestinal hormonal response to fasting can affect the experience of hunger and satiety (Williams et al., 2004; Levin et al., 2006).
Hypoglycemia, defined as a blood glucose concentration below 3 mmol/l, has indeed been observed in previous studies (Sasaki et al., 1996; Brun et al., 2000; Inoue et al., 2004). The number of subjects in the previous studies widely exceed the number in the present study, and the present study does not exclude the possibility of the occurrence of hypoglycemia among healthy subjects. However, there are many problems associated with blood glucose analyses that may cause false low blood glucose measurements (Elimam et al., 1997; Marcus, 2001) and to the best of our knowledge, no studies have shown that clinically relevant hypoglycemia, as defined by Whipple, occurs among otherwise healthy individuals. The present study, although based on a small number of patients, confirms that symptoms attributed to hypoglycemia are unlikely to represent true hypoglycemia in healthy adults.
There are several known causes of hypoglycemia. The most common is insulin-treated diabetes mellitus, but diseases such as insulinoma, glycogen storage diseases, adrenal disorders and mitochondrial diseases can also cause hypoglycemia. Many of the metabolic diseases may go undiagnosed as long as the subject's energy is balanced, but once the subject becomes catabolic due to, for example, an infection, the system becomes unbalanced and symptoms can be observed.
Ketotic hypoglycemia is the most common variant of hypoglycemia in childhood. The cause is unknown and it seems that the hypoglycemic tendency spontaneously remits before puberty (Dahlquist et al., 1979). Whether this is due to a change in metabolic needs as the child grows up or is due to the fact that more mature individual adapts better to the situation is unknown. As far as we know, there have been no reports on adult ketotic hypoglycemia. There is a marked resemblance between ketotic hypoglycemia during childhood and the symptoms described by adult individuals such as those selected in the sensitive group. However, this group had no signs of a more pronounced increase in ketone bodies during the fast than the resistant group.
Before the fast, the sensitive group had advertised doubts as to whether they could manage the 24-h fast, but nevertheless all participants (n=8 in both the sensitive and the resistant group) had by their own request asked to participate in the present study. Consequently, the subjects were not matched into the two groups. In general, the sensitive group managed surprisingly well during the fasting period. The sensitive group, in contrast to the resistant group, had significantly increased scores on most of the charted emotions during the fast, and they scored significantly higher than the resistant group on both ‘irritation’ and ‘shakiness’.
In conclusion, we observed no signs of reduced fasting capacity in a small group of subjects who reported that they develop symptoms associated with hypoglycemia during energy restriction. Thus, it remains to be established whether hypoglycemia really occurs among otherwise healthy individuals after a prolonged fast and, if so, what causes this condition.
Brun JF, Fedou C, Mercier J (2000). Postprandial reactive hypoglycemia. Diabetes Metab 26, 337–351.
Champe PC, Harvey RA (2005). Lippincott's Illustrated Reviews: Biochemistry. Lippincott, Williams & Wilkins: Philadelphia.
Chapman IM, Goble EA, Wittert GA, Morley JE, Horowitz M (1998). Effect of intravenous glucose and euglycemic insulin infusions on short-term appetite and food intake. Am J Physiol 274, R596–R603.
Dahlquist G, Gentz J, Hagenfeldt L, Larsson A, Low H, Persson B et al. (1979). Ketotic hypoglycemia of childhood – a clinical trial of several unifying etiological hypotheses. Acta Paediatr Scand 68, 649–656.
Elimam A, Horal M, Bergstrom M, Marcus C (1997). Diagnosis of hypoglycaemia: effects of blood sample handling and evaluation of a glucose photometer in the low glucose range. Acta Paediatr 86, 474–478.
Gielkens HA, Verkijk M, Lam WF, Lamers CB, Masclee AA (1998). Effects of hyperglycemia and hyperinsulinemia on satiety in humans. Metabolism 47, 321–324.
Grunt JA, McGarry ME, McCollum AT, Gould JB (1970). Studies of children with ketotic hypoglycemia. Yale J Biol Med 42, 420–438.
Hagstrom-Toft E, Bolinder J, Ungerstedt U, Arner P (1997). A circadian rhythm in lipid mobilization which is altered in IDDM. Diabetologia 40, 1070–1078.
Haymond MW, Howard C, Ben-Galim E, DeVivo DC (1983). Effects of ketosis on glucose flux in children and adults. Am J Physiol 245, E373–E378.
Horal M, Ungerstedt U, Persson B, Westgren M, Marcus C (1995). Metabolic adaptation in IUGR neonates determined with microdialysis – a pilot study. Early Hum Dev 42, 1–14.
Inoue K, Kakehashi Y, Oomori S, Koizumi A (2004). Biochemical hypoglycemia in female nurses during clinical shift work. Res Nurs Health 27, 87–96.
Kamel A, Norgren S, Persson B, Marcus C (1999). Insulin induced hypoglycaemia: comparison of glucose and glycerol concentrations in plasma and microdialysate from subcutaneous adipose tissue. Arch Dis Child 80, 42–45.
Karam JH (2001). Hypoglycemic disorders. In: Greenspan FS and Strewler GJ (eds). Basic & Clinical Endocrinology. Appleton & Lange: Stamford, CT, USA, pp 823.
Levin F, Edholm T, Schmidt PT, Gryback P, Jacobsson H, Degerblad M et al. (2006). Ghrelin stimulates gastric emptying and hunger in normal-weight humans. J Clin Endocrinol Metab 91, 3296–3302.
Marcus C (2001). How to measure and interpret glucose in neonates. Acta Paediatr 90, 963–964.
Marks V (1987). Glycaemic stability in healthy subjects: fluctuations in blood glucose Concentration During the Day. In: Andreani D, Marks V and Lefebvre PJ (eds). Hypoglycemia. Raven Press: NY, USA, pp 19–24.
Meyer EL, Waldenlind E, Marcus C (2003). Diminished nocturnal lipolysis in cluster headache: a sign of central sympathetic dysregulation? Neurology 61, 1250–1254.
Sasaki M, Mogi T, Wada Y, Hirosawa I, Koizumi A (1996). An endemic condition of biochemical hypoglycemia among male volunteers. Ind Health 34, 323–333.
Wiesli P, Brandle M, Zapf J, Seiler H, Zwimpfer C, Spinas GA et al. (2004). Assessment of hyperinsulinaemia at the termination of the prolonged fast. Clin Chim Acta 342, 227–231.
Wiesli P, Schwegler B, Schmid B, Spinas GA, Schmid C (2005). Mini-mental state examination is superior to plasma glucose concentrations in monitoring patients with suspected hypoglycaemic disorders during the 72-hour fast. Eur J Endocrinol 152, 605–610.
Wildenhoff KE (1970). A micro-method for the enzymatic determination of acetoacetate and 3-hydroxybutyrate in blood and urine. Scand J Clin Lab Invest 25, 171–179.
Williams G, Cai XJ, Elliott JC, Harrold JA (2004). Anabolic neuropeptides. Physiol Behav 81, 211–222.
Yager J, Young RT (1974). Non-hypoglycemia is an epidemic condition. N Engl J Med 291, 907–908.
This work was supported by the Freemasons in Stockholm Foundation for Children's Welfare and the Swedish Research Council #K2002-72X-009941-11B.
Contributors: CM has contributed to the initiation of the study, designed the study protocol, participated in the writing of the manuscript and has been the supervisor of the study. EP has contributed to the study design and participated in the data collection. JA has contributed in all parts of the study
About this article
Cite this article
Alkén, J., Petriczko, E. & Marcus, C. Effect of fasting on young adults who have symptoms of hypoglycemia in the absence of frequent meals. Eur J Clin Nutr 62, 721–726 (2008). https://doi.org/10.1038/sj.ejcn.1602785
- blood glucose
- ketone bodies
Growth Hormone & IGF Research (2012)