Abstract
Background: There are no studies on the clinical significance of asymptomatic hypoglycemia detected incidentally during routine testing.
Methods: Baseline fasting serum glucose was determined in 1333 individuals without diabetes (43.3% males, median age 50 years, range 18 to 91 years) to investigate the prevalence of hypoglycemia (fasting serum glucose <70 mg/dL) and the associated demographic, lifestyle, and metabolic factors. Individuals with baseline hypoglycemia were followed (median follow-up, 8.7 years) and assessed for hypoglycemia symptoms. Seven-day continuous glucose monitoring was performed in a subsample of 489 individuals.
Results: Baseline hypoglycemia was observed in 20 individuals (weighted prevalence, 1.58%, 95% confidence interval 0.87%–2.28%). Hypoglycemia was mild and asymptomatic in all cases (median, 67 mg/dL, range 63 to 69 mg/dL). The characteristics of those with hypoglycemia were similar to those with fasting serum glucose 70 to 80 mg/dL. Hypoglycemia was associated with female sex, younger age, and a more favorable metabolic profile (lower body mass index, glycohemoglobin and insulin resistance) than individuals with fasting serum glucose >80 mg/dL. Individuals with baseline hypoglycemia showed no distinct hypoglycemia features in continuous glucose monitoring (n = 9). During follow-up (n = 19), hypoglycemia in routine determinations, always mild, recurred in 42.1% of individuals, although the mean of successive glucose concentrations was higher than baseline in all cases. None of the individuals had symptoms that could constitute Whipple's triad (low serum glucose, symptoms of hypoglycemia, and symptomatic improvement after correction of hypoglycemia) during the follow-up period.
Conclusions: Detection of asymptomatic, mild hypoglycemia in routine blood tests is not indicative of disease and does not require further investigation.
Introduction
In contrast to patients with diabetes treated with insulin or secretagogues, hypoglycemia is rare in people without diabetes.1 In a retrospective study of 37,898 hospital admissions of patients without diabetes who were hospitalized outside an intensive care unit, glucose concentrations <60 mg/dL (as measured by both laboratory and point-of-care blood glucose values throughout hospitalization) occurred in only 0.19% (n = 71) of admissions.2 The prevalence of hypoglycemia in outpatients without diabetes has been less studied and might be even lower than in hospitalized patients, given that acute illness is a major cause of hypoglycemia in patients without diabetes, mainly due to hepatic or renal dysfunction.3 Hypoglycemia in people without diabetes can occur in the fasting or postprandial state.4 Causes of fasting hypoglycemia in people without diabetes include consumption of alcohol or certain drugs,4–6 hormone deficiencies,7,8 consumption of glucose by blood cells in cases of leukocytosis or thrombocytosis,1 and endogenous hyperinsulinism.4,9 The most common cause of endogenous hyperinsulinism is insulinoma, a rare neuroendocrine tumor (estimated incidence, 1.3 cases/million/year),10 which can have malignant potential.11 Endogenous hyperinsulinism is therefore the most feared cause of fasting hypoglycemia in individuals without diabetes. On the other hand, postprandial hypoglycemia is typically reactive to carbohydrate intake and is less concerning for the diagnosis of endogenous hyperinsulinism.4,12
The glucose threshold for defining hypoglycemia is not clearly defined in people without diabetes, in whom serum glucose concentrations >70 mg/dL are considered normal.1 For this reason, and because hypoglycemia in people with diabetes is defined as serum glucose concentrations <70 mg/dL (further classified as level 1 [54 to 69 mg/dL] and level 2 [<54 mg/dL] hypoglycemia),13 a great number of laboratories will consider 70 mg/dL to be the lower limit of glucose normality. Thus, the incidental detection of fasting serum glucose concentrations <70 mg/dL in tests performed on outpatients could raise questions about its meaning and the need to perform further testing or consider specialty referral. In people without diabetes, current Endocrine Society guidelines4 recommend an etiologic study of fasting hypoglycemia to exclude endogenous hyperinsulinism only in those who meet Whipple’s triad: low serum glucose concentrations, presence of symptoms compatible with hypoglycemia, and improvement of these symptoms after raising serum glucose levels.14 Therefore, it is not recommended to study asymptomatic hypoglycemia because it is assumed to be benign (ie, not indicative of an insulinoma).1,4 However, this recommendation is based on expert opinion,1,4 given that, to the best of our knowledge, no studies have investigated the significance of incidental fasting hypoglycemia in outpatients without diabetes. New technologies such as continuous glucose monitoring (CGM), which are increasingly being used in people without diabetes,15⇓–21 could help to assess the frequency and timing of hypoglycemia in individuals with a history of asymptomatic fasting hypoglycemia in blood samples.
This study aimed to investigate the frequency of asymptomatic fasting hypoglycemia on routine outpatient testing, the factors associated with asymptomatic fasting hypoglycemia, and the long-term clinical significance.
Methods
Study Population and Setting
This observational study was conducted in the municipality of A-Estrada (Spain), as reported elsewhere.22,23 An outline of the study (A-Estrada Glycation and Inflammation Study [AEGIS]) is also available at www.clinicaltrials.gov (code NCT01796184). A flowchart of the study is represented in Appendix 2. Briefly, the study included an age-stratified sample (n = 1,516) of the adult (18 years and older) population from the municipality. The sample was drawn from Spain’s National Health System Registry, which covers more than 95% of the population. From November 2012 until March 2015, participants were successively convened at the A-Estrada Primary Care Centre (A-Estrada, Galicia, Spain) where they completed an interviewer-administered questionnaire (which collected demographic and anthropometric data; and provided information about their lifestyle, including alcohol consumption, smoking, and physical activity) and blood collection for analytic studies. For the present study, individuals with diabetes (n = 183)24 were excluded to rule out hypoglycemia secondary to hypoglycemic treatment. Of the remaining 1333 individuals, 577 (43.3%) were males; the median age was 50 years, with a range of 18 to 91 years. A subsample of 489 individuals underwent 7-day interstitial CGM (see below).
Ethical Issues
The study was approved by the Regional Ethics Committee (codes 2010-315 and 2012-025). We obtained written informed consent from all the participants.
Assessment of Lifestyle Variables
Alcohol Consumption
We evaluated the number of glasses of wine, bottles of beer, and units of spirits consumed per week. According to Spanish validation studies, each glass of wine and each bottle of beer contains approximately 10 g of alcohol, and each unit of spirit contains 20 g of alcohol.25 Participants with an alcohol consumption of 10 to 139 g/week, 140 to 279 g/week, and ≥280 g/week were considered light, moderate, and heavy drinkers, respectively. The remainder, comprising alcohol abstainers and occasional alcohol drinkers (ie, less than one drink a week), were grouped together.
Smoking
We considered consumers of at least one cigarette per day to be smokers. Individuals who had quit smoking during the preceding year were still considered smokers, whereas those who had quit more than one year before the study were considered ex-smokers.
Physical Activity
All the study participants completed the freely available International Physical Activity Questionnaire (short version), which has been validated in Spain.26 The questionnaire allows for the calculation of metabolic equivalents of task27 and for the stratification of habitual physical activity as low, moderate, or high.22,28
Laboratory Determinations and Definition of Metabolic Abnormalities
All laboratory determinations were performed for all participants in fresh samples obtained after a minimum of 8 hours under fasting conditions. For baseline determinations, blood was drawn at the primary center on the day of the study and sent to the reference laboratory for analysis on the morning of collection, following the same protocol in all cases (see above, Study population and setting).
Serum Glucose
Glucose concentrations were determined in fresh serum samples via the glucose oxidase peroxidase method. Participants were considered as having hypoglycemia when baseline fasting glucose concentrations were <70 mg/dL. For the present study, individuals with baseline fasting hypoglycemia were compared with those having baseline fasting glucose concentrations between 70 and 80 mg/dL and those having baseline fasting glucose concentrations >80 mg/dL. In addition to the baseline determination, the electronic medical records of participants with fasting hypoglycemia (<70 mg/dL) at baseline were reviewed to record routine outpatient fasting glucose determinations in subsequent years until April 2024. Those performed during intercurrent acute processes (including emergency treatment and hospitalization), during periods of pregnancy, or in postprandial conditions were excluded. These follow-up glycemic determinations were requested by either primary care clinicians or hospital specialists, and were analyzed in the same reference laboratory as baseline samples in all cases. The clinical records of participants with baseline hypoglycemia were also reviewed for the presence of Whipple’s triad.14
Glycated Hemoglobin
Glycohemoglobin (HbA1c) was determined by high-performance liquid chromatography using a Menarini Diagnostics HA-8160 analyzer; all HbA1c values were converted to Diabetes Control and Complications Trial-aligned values (ie, HbA1c was expressed as a percentage of total hemoglobin).29 Individuals were deemed to have a diagnosis of diabetes if they had been previously diagnosed as such, had an Hb1Ac level ≥6.5%, and/or had a fasting serum glucose concentration ≥126 mg/dL, based on the American Diabetes Association criteria.24
Serum Insulin
Insulin concentrations were determined using the ADVIA Centaur XP immunoassay system (Siemens, Erlangen, Germany).
Insulin Resistance
Insulin resistance was estimated using the homeostasis model assessment method for insulin resistance (HOMA-IR) as the fasting concentration of serum insulin (µ units/mL) × serum glucose (mg/dL)/405.30
Body Mass Index
Body mass index (BMI) was calculated as weight (in kg) divided by the square of height (in meters). Individuals were divided into 3 groups according to the BMI: BMI ≤25 kg/m2, BMI >25 to 30 kg/m2, and >30 kg/m2.
Continuous Glucose Monitoring
Interstitial CGM was performed in a subsample of 489 individuals (Appendix 1), as reported elsewhere.15–20 At the start of the 7-day monitoring period, a research nurse inserted an Enlite™ sensor (Medtronic Inc., Northridge, CA, USA) subcutaneously in the participants’ abdomen and instructed the participants in the care of the connected iPro™ model CGM device (Medtronic Inc). The iPro™ is a professional blind CGM system that provides retrospective information on glucose profiles. Given that the system does not provide data on glucose levels during its use, the participants’ routines should not have been affected by the CGM use. Moreover, the participants were instructed not to change their usual diet and physical exercise routines during the CGM period. The sensor continuously measures interstitial space glucose concentrations in the subcutaneous tissue, recording values (range 40 to 400 mg/dL) every 5 minutes and storing them in the device. To assure the accuracy of the CGM data, participants were instructed to perform ≥3 capillary blood glucose measurements per day to calibrate the CGM system. Across all participants, CGM data were calibrated a total of 9980 times during the CGM period (median, 18 calibrations per participant). Sensor accuracy was measured as the mean absolute relative difference (MARD).31 MARD was ≤7.6%32 on all days of the CGM period except for the first day (day 1 MARD, 12.1%). On day 7, the sensor was removed and data were downloaded for analysis, disregarding results from the first day (due to the suboptimal sensor accuracy noted above), results from the last day (because the data for day 7 did not reach 24 hours, since the sensor was removed during that day), results from days in which data acquisition failure totaled more than 2 hours, and results from days in which CGM data were not calibrated with ≥3 capillary glucose controls. The following parameters were used to describe the CGM data:
Mean Glucose
Mean of all the glucose values that were recorded in the system during the CGM period.
Time in Range
Percentage of the day during which the participant had interstitial glucose concentrations between 70 mg/dL and 140 mg/dL.
Time Below Range
Percentage of the day during which the participant had interstitial glucose concentrations <70 mg/dL.
Time Above Range
Percentage of the day during which the participant had interstitial glucose concentrations >140 mg/dL.
Coefficient of Variation
The size of the glucose oscillations occurring throughout the day.33 Therefore, it measures intraday glucose variability, calculated as the standard deviation of the glucose values divided by the mean glucose.34
Mean of Daily Differences
Mean of all valid absolute value differences between glucose concentrations measured at the same time of day on consecutive days.35 Therefore, it measures interday glucose variability.
Statistical Analyses
To account for the stratified sampling, a design-based analysis including compensatory weights was performed for the estimation of hypoglycemia prevalence in the overall population. Given the non-normal distribution of some variables, nonparametric statistical tests were used. We employed the Chi-Square test to compare proportions (with trend analysis, when appropriate; ie, to evaluate a change in a dependent variable with respect to an independent ordinal category). We used the Mann-Whitney test and the Kruskal-Wallis test to compare numeric variables between 2 and among more than 2 independent categories, respectively. The Wilcoxon signed-rank test was employed to compare numeric variables in paired samples. Logistic regression was used for the multivariate analysis of factors associated with fasting hypoglycemia. Covariates were forced to enter the equation in all models. We used Spearman’s rank test to assess the correlation between 2 numeric variables. A generalized additive mixed model36 was employed to analyze the evolution of fasting serum glucose levels over time in the 19 individuals with baseline fasting hypoglycemia and available follow-up, capturing both the overall time trend and individual patient differences. The model uses penalized splines to estimate the nonlinear trend of glucose over time, while adjusting for each individual’s unique baseline fasting glucose level with a random intercept. All tests were 2-tailed. P-values lower than 0.05 were considered statistically significant.
Results
Prevalence of Fasting Hypoglycemia
Of the total 1333 participants, 20 showed serum glucose concentrations lower than 70 mg/dL in baseline fasting determinations (absolute prevalence 1.50%; weighted prevalence 1.58%, 95% confidence interval 0.87%-2.28%). Serum glucose in these cases ranged from 63 to 69 mg/dL (median, 67 mg/dL). Thus, all the low glucose values detected were compatible with level 1 hypoglycemia (54 to 69 mg/dL), and there were no cases of level 2 hypoglycemia (<54 mg/dL).13
Factors Associated with Fasting Hypoglycemia
Univariate Analyses
The frequency of fasting hypoglycemia in baseline blood samples decreased significantly with age: 2.6% (11/420) among individuals aged 18 to 40 years, 1.2% (6/483) among those aged 41 to 60 years, and 0.7% (3/430) among those aged more than 60 years (P = .022; Figure 1). The majority (14/20, 70.0%) of fasting hypoglycemia cases were female.
Prevalence of hypoglycemia (fasting serum glucose lower than 70 mg/dL) in relation to age and body mass index (BMI) categories.
Table 1 represents a comparison of demographic, lifestyle, and metabolic factors among individuals without diabetes with different categories of fasting serum glucose (<70 mg/dL, 70 to 80 mg/dL, and >80 mg/dL) at baseline. Significant differences among groups were observed for age, sex, smoking, alcohol consumption, BMI, HbA1c, and HOMA-IR, although differences were primarily restricted to the >80 mg/dL group (Table 1). The group of individuals with fasting hypoglycemia had the highest frequency of current smokers (Table 1).
Comparison of Demographics, Lifestyle and Metabolic Factors in Relation to Fasting Serum Glucose in Individuals Without Diabetes
Demographic, lifestyle, and metabolic variables were similar in individuals with fasting hypoglycemia and individuals with fasting serum glucose concentrations between 70 and 80 mg/dL, with the exceptions of HbA1c, and HOMA-IR, which were slightly lower among individuals with fasting hypoglycemia (P = .024 and P = .018, respectively) (Table 1). In comparison with individuals with a fasting serum glucose >80 mg/dL, individuals with fasting hypoglycemia showed younger age and a more favorable metabolic profile, including lower BMI, HbA1c values, insulin concentrations, and HOMA-IR (P ≤ .001 in every case, Table 1). The prevalence of fasting hypoglycemia decreased significantly with increasing BMI: 2.5% (10/406) among individuals with a BMI ≤25 kg/m2, 1.6% (8/510) among individuals with a BMI >25 to 30 kg/m2, and 0.5% (2/417) among individuals with a BMI >30 kg/m2 (P = .019; Figure 1).
Multivariate Analysis
BMI was associated with the variables associated with fasting hypoglycemia in the univariate analyses: current smokers had a lower BMI (median, 25.8 kg/m2, interquartile range [IQR] 22.6 to 29.2 kg/m2) than ex-smokers and never-smokers (median, 27.7 kg/m2, IQR 24.7 to 31.1 kg/m2; P < .001); heavy drinkers had a higher BMI (median, 28.7 kg/m2, IQR 25.9 to 31.9 kg/m2) than nonheavy drinkers (median, 27.2 kg/m2, IQR 24.1 to 30.7 kg/m2; P = .002); women had a lower BMI (median, 26.9 kg/m2, IQR 23.7 to 31.0 kg/m2) than men (median, 27.9 kg/m2, IQR 24.9 to 30.7 kg/m2; P = .013); and there was a positive correlation between age and BMI (Rho = 0.372; P < .001). In a multivariate analysis (logistic regression) examining the presence of fasting hypoglycemia in relation to age, sex, BMI, alcohol consumption, and smoking status, only BMI maintained a negative association with hypoglycemia (Table 2).
Multivariate Analysis of Factors Associated with Fasting Hypoglycemia (<70 mg/dL)
Continuous Glucose Monitoring in Relation to Baseline Fasting Serum Glucose
Seven-day CGM was performed in 489 individuals, including 9 individuals with baseline fasting serum hypoglycemia. After disregarding data according to the previously mentioned criteria, 424 (86.7%) participants had 5 days of valid CGM data, 36 (7.4%) participants had 4 days of valid CGM data, 17 (3.5%) participants had 3 days of valid CGM data, 9 (1.8%) participants had 2 days of valid CGM data, and 3 (0.6%) participants had 1 day of valid CGM data. Table 3 represents a comparison of CGM parameters among individuals with different categories of fasting serum glucose (<70 mg/dL, 70 to 80 mg/dL, and >80 mg/dL). There were no significant differences in the time below range among the 3 groups (Table 3). Although the coefficient of variation was similar among the 3 groups, individuals with fasting serum hypoglycemia tended to have a higher mean of daily differences. Possibly related to this greater interday glycemic variability, the group of individuals with fasting serum hypoglycemia had the highest time above range and, consequently, the lowest time in range (Table 3).
Comparison of Continuous Glucose Monitoring Parameters in Relation to Fasting Serum Glucose in Individuals Without Diabetes
When CGM parameters were specifically compared between individuals with fasting serum glucose <70 mg and individuals with fasting serum glucose 70 to 80 mg/dL, statistically significant differences in CGM parameters were found only for the time above range (P = .034; Table 3).
Outcomes of Individuals with Baseline Fasting Hypoglycemia
Follow-up fasting serum glucose determinations were available for 19 of 20 individuals with baseline fasting hypoglycemia in the follow-up years (median follow-up 8.7 years; IQR 8.1 to 9.1 year; range 3.7 to 10.7 years). The median number of subsequent glucose determinations per participant was 6 (IQR 2 to 10, absolute range 1 to 21). The individual results and the mean trend of fasting glucose over time are represented in Figure 2.
Individual representation of fasting serum glucose trajectories over the years in participants with hypoglycemia (fasting serum glucose lower than 70 mg/dL) at baseline. The solid blue curve represents the mean trend of fasting glucose over time with 95% intervals. M, male; F, female; y, age (years).
In the 19 evaluable individuals, the mean of fasting glucose determinations in the following years ranged from 71.5 to 90.5 mg/dL (median, 77.3 mg/dL; IQR 74.4 to 81.0 mg/dL). The average of the determinations in the following years was higher than the baseline fasting serum glucose in all cases (P < .001). Thus, there was an ascending trend in participants' mean fasting glucose concentrations over the years of follow-up (Figure 2). A fasting serum glucose determination <70 mg/dL was observed during follow-up in 8 (42.1%) cases. The lowest value observed was 60 mg/dL in 1 case. For those 8 participants in whom hypoglycemia recurred, the number of blood tests with fasting hypoglycemia ranged between 8.3% and 33.3% of the determinations (median, 26.3%). None of the individuals with recurrent fasting hypoglycemia had an associated clinical history of Whipple's triad documented in the electronic medical record.
Discussion
Our study examined the prevalence and risk factors for the presence of fasting hypoglycemia in people without diabetes, as well as the clinical impact of asymptomatic hypoglycemia, which to our knowledge has not previously been studied. Fasting hypoglycemia was an infrequent finding in our sample, reaching a prevalence of only 1.5%, confirming that hypoglycemia is a rare condition in individuals who are not receiving hypoglycemic treatment for diabetes.1 We found that the frequency of fasting hypoglycemia was higher in people with low BMI, women, young people, and current smokers. Women, young people, and current smokers all had lower BMIs. When a multivariate analysis was performed including these variables (sex, age, smoking status, and BMI) and alcohol drinking status, only BMI remained negatively associated with the frequency of fasting hypoglycemia. These results suggest that BMI is the main factor negatively associated with the presence of fasting hypoglycemia. This finding is plausible because, as is well known, elevated BMI favors hyperglycemia (and not hypoglycemia).37,38
Insulinemia was lower in individuals with fasting hypoglycemia. Physiologically, insulin concentrations are directly proportional to glucose concentrations, given that hyperglycemia stimulates insulin secretion. Thus, those without endogenous hyperinsulinism would have low insulin concentrations at low blood glucose levels.1,4 In addition, insulin concentrations were probably lower in the hypoglycemic group because individuals in this group also had the lowest BMI, which favors insulin sensitivity.39
The medical records of all participants with baseline fasting hypoglycemia were evaluated for the presence of hypoglycemia symptoms. We did not find any medical visits during which these individuals reported symptoms of hypoglycemia13 during the entire follow-up period. To confirm the benign nature of asymptomatic fasting hypoglycemia, the CGM of these individuals was compared with that of individuals without baseline fasting serum hypoglycemia. When this comparison was made, no differences were found in the time below range. Moreover, mean glucose levels during CGM were not lower in individuals with fasting serum hypoglycemia than in individuals without it. Although HbA1c levels were slightly lower in individuals with fasting hypoglycemia than in those without it, the mean HbA1c levels in the hypoglycemia group (5.0%) were higher than those described in a series of 31 patients with insulinoma (4.7%).40
The evolution of fasting serum glucose levels and potentially associated symptoms were evaluated in the participants with baseline fasting hypoglycemia in the following years. None of them reported symptoms of hypoglycemia13 at any time during follow-up. In routine blood tests during follow-up (median, 8.7 years), fasting hypoglycemia was observed again in some cases; in general, however, the evolution observed was compatible with the phenomenon of regression to the mean, by which extremely high or low variables move closer to the mean on retesting (ie, rare or extreme events are likely to be followed by more typical ones). Moreover, no participant reached severe (level 2) fasting hypoglycemia13 on any of the follow-up tests, and the mean of the fasting glucose values on the follow-up tests was never <70 mg/dL, indicating that most of the follow-up tests did not show fasting hypoglycemia (Figure 2). These results also suggest that mild (level 1) incidental fasting hypoglycemia is a benign finding in asymptomatic individuals.
Furthermore, our results cast doubt on whether 70 mg/dL is an appropriate threshold for defining hypoglycemia in people without diabetes. In addition, it is possible that each person without diabetes might have his/her own threshold for symptomatic hypoglycemia, depending on demographic and metabolic characteristics. Therefore, as shown in our study, a blood glucose level slightly below 70 mg/dL might not cause symptoms in certain individuals without diabetes, without necessarily implying that they are insensitive to their internal states. Future studies are needed to determine at what glucose threshold people without diabetes will experience clear symptoms of hypoglycemia.
Our study has limitations that should be recognized. Although the sample size was large, the number of individuals with baseline fasting hypoglycemia was small, given the low prevalence of this finding. CGM was not performed in all the participants in this study; thus, the number of individuals with baseline fasting hypoglycemia who had a CGM sensor placed was limited. CGM systems are known to be less accurate in the hypoglycemic range.41 The assessment of hypoglycemia symptoms was performed retrospectively from the participants’ clinical records. Although serum samples were obtained after a minimum of 8 hours of fasting, a number of participants might have fasted for longer (for instance, not eaten dinner the evening before serum collection), which could favor hypoglycemia. We considered <70 mg/dL as the threshold for defining hypoglycemia; however, similar results were observed when wider definitions were applied. As sensitivity analyses, we redefined hypoglycemia to include those cases with baseline fasting serum glucose equal to 70 mg/dL42 and compare their characteristics with those of individuals with strictly normal baseline fasting serum glucose concentrations (75 to 85 mg/dL), obtaining similar results (Additional file 2). All the hypoglycemias observed were mild (level 1)13; thus, the conclusions cannot be generalized to the potential finding of more severe hypoglycemias (level 2, <54 mg/dL). Given that all the hypoglycemia in our study was recorded in fasting blood samples, the results would not be generalizable to hypoglycemia occurring in the postprandial period. The Spanish population is predominantly White, which could also limit the generalizability of the results to other populations. Strengths of this study include the population-based design and the long follow-up period.
At the beginning of the 20th century, Allen Oldfather Whipple described asymptomatic hypoglycemia as benign and advised against performing etiologic studies in such asymptomatic cases.14 This recommendation, which remains present in current textbooks and guidelines,1,4 has never been validated by epidemiologic studies. To our knowledge, this is the first study to demonstrate the benign nature of incidental asymptomatic hypoglycemia detected by fasting blood tests. According to our results, it is unnecessary to perform an etiologic study in outpatients without diabetes who present with asymptomatic mild (level 1) fasting hypoglycemia, which could be more common in young people with low body weight.
Acknowledgments
The authors would like to thank all the participants in the study and the institutions that funded it.
Notes
This article was externally peer reviewed.
Funding: This study was supported by grants from the Carlos III Institute of Health (Instituto de Salud Carlos III-ISCIII/PI20/01069/Co-funded by European Union), the Network for Research on Chronicity, Primary Care, and Health Promotion (Instituto de Salud Carlos III- ISCIII/RD21/0016/0022/Co-funded by European Union), and the Galician Innovation Agency-Competitive Benchmark Groups (GAIN-GRC/IN607A/2021/02/Xunta de Galicia). This work was also supported by a grant from Medtronic Inc. The funder was not involved in the study design, collection, analysis, interpretation of data, the writing of this article or the decision to submit it for publication. ÓL-B was supported by ISCIII Support Platforms for Clinical Research (ISCIII/PT23/00118/Co-funded by European Union).
Conflict of interest: The authors have no conflicts of interest to declare.
To see this article online, please go to: http://jabfm.org/content/38/3/411.full.
- Received for publication July 20, 2024.
- Revision received November 18, 2024.
- Accepted for publication December 11, 2025.
