{
  "abstract": "Background Despite the high prevalence of diabetes in Qatar, Ramadan-related glycemic events remain inadequately characterized. This study examined sociodemographic, clinical and behavioral factors independently associated with hypoglycemic or hyperglycemic events among patients with diabetes during Ramadan in Qatar.Methods A retrospective analysis was conducted using routinely collected data from the Qatar Diabetes Association (QDA) Ramadan support program covering three Ramadan seasons (2022–2024). Patients with type 1 or type 2 diabetes who attended QDA diabetes education or dietetic clinics were included, regardless of fasting intention. Those with pre-diabetes or gestational diabetes were excluded. Assessments began 6–8 weeks before Ramadan, with structured follow-up throughout the month. Primary outcomes—based on patient-recorded blood glucose readings—were defined to capture the overall risk of glycemic instability during Ramadan and included (1) any hypoglycemic (<70 mg/dL) or hyperglycemic (>250 mg/dL) event and (2) any such events during fasting hours. Generalized estimating equation logistic regression models identified factors associated with hypoglycemic or hyperglycemic outcomes. Missing data were <5% and handled by complete-case analysis. Statistical significance was defined as 95% CIs excluding 1.Results The study included 374 patients; 95.5% of observations were from patients with type 2 diabetes and 43.3% reflected longer diabetes duration. Adjusted ORs indicated higher odds of hypoglycemia or hyperglycemia at any time among patients with a longer diabetes duration, a high International Diabetes Federation–Diabetes and Ramadan risk score, those who discussed a fasting plan with a physician, those who used medications known to increase hypoglycemia risk, and those who used continuous glucose-monitoring devices. During fasting hours, higher odds were observed among patients with type 1 diabetes and those with a higher daily number of recorded blood glucose readings.Conclusions The findings underscore the importance of structured pre-Ramadan risk assessment, patient-centered education, close and rigorous follow-up and individualized treatment adjustment—particularly for high-risk groups—to optimize safety and enhance the overall Ramadan experience among persons with diabetes.",
  "authors": [
    {
      "affiliations": [
        "Infectious Disease Epidemiology Group, Weill Cornell Medicine–Qatar, Cornell University, Doha, Qatar"
      ],
      "name": "Zahir M Tag"
    },
    {
      "affiliations": [
        "Infectious Disease Epidemiology Group, Weill Cornell Medicine–Qatar, Cornell University, Doha, Qatar",
        "Department of Population Health Sciences, Weill Cornell Medicine, Cornell University, New York, NY, USA",
        "Department of Public Health, College of Health Sciences, QU Health, Qatar University, Doha, Qatar",
        "College of Health and Life Sciences, Hamad bin Khalifa University, Doha, Qatar"
      ],
      "name": "Laith J Abu-Raddad"
    },
    {
      "affiliations": [
        "Qatar Diabetes Association, Doha, Qatar"
      ],
      "name": "Amel A Mustafa"
    },
    {
      "affiliations": [
        "Infectious Disease Epidemiology Group, Weill Cornell Medicine–Qatar, Cornell University, Doha, Qatar",
        "Department of Population Health Sciences, Weill Cornell Medicine, Cornell University, New York, NY, USA"
      ],
      "name": "Hiam Chemaitelly"
    },
    {
      "affiliations": [
        "Qatar Diabetes Association, Doha, Qatar"
      ],
      "name": "Abdulla O Alhamaq"
    },
    {
      "affiliations": [
        "Qatar Diabetes Association, Doha, Qatar"
      ],
      "name": "Katie G El-Nahas"
    }
  ],
  "full_text": "WHAT IS ALREADY KNOWN ON THIS TOPIC During the Islamic holy month of Ramadan, many persons with diabetes choose to fast despite an elevated risk of glycemic instability. Although glycemic outcomes during Ramadan have been studied and informed the International Diabetes Federation–Diabetes and Ramadan (IDF–DAR) guidelines, outcome heterogeneity persists and real-world adherence to recommended self-management remains suboptimal. This study addresses this gap by evaluating sociodemographic, diabetes-related clinical and glucose-monitoring and self-management factors associated with hypoglycemic or hyperglycemic events among persons with diabetes during Ramadan in Qatar.WHAT THIS STUDY ADDS The study identifies key factors associated with higher odds of hypoglycemic or hyperglycemic events, including type 1 diabetes, a longer diabetes duration, a high IDF–DAR risk score, having discussed a fasting plan with a physician, use of medications known to increase hypoglycemia risk, use of continuous glucose-monitoring devices and a higher daily number of recorded blood glucose readings.HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY The findings emphasize the importance of jointly considering clinical, behavioral and self-management factors when assessing glycemic outcomes and fasting safety during Ramadan, and highlight the value of routine-care longitudinal data. They support strengthening structured pre-Ramadan risk assessment, patient-centered education, close and rigorous follow-up and individualized treatment adjustment—particularly for high-risk groups—and scaling up standardized, guideline-aligned Ramadan diabetes care programs within routine health services to improve the overall Ramadan experience.Introduction Diabetes mellitus represents a major global public health challenge, with an estimated 589 million adults affected worldwide in 2024, according to the International Diabetes Federation (IDF). 1 This number is projected to increase substantially, reaching 853 million by 2050.1 The burden of diabetes is disproportionately high in the Middle East and North Africa region, which has the highest age-adjusted prevalence rates worldwide among adults aged 20–79 years.1 Qatar exemplifies this regional pattern, with diabetes prevalence approaching 20% among adults, a burden largely driven by the high prevalence of obesity.2–4Ramadan, the ninth month of the Islamic lunar calendar, is observed annually by more than two billion Muslims worldwide and involves daily fasting from dawn to sunset for 29–30 consecutive days.5 6 For individuals with diabetes, fasting during Ramadan poses physiological and clinical challenges. Prolonged fasting, along with changes in medication timing and lifestyle behaviors, may disrupt glycemic control and increase the risk of both hypoglycemia and hyperglycemia.7 Despite religious exemptions for individuals with diabetes, many still choose to fast during Ramadan.8Evidence from a large epidemiological study has demonstrated an increased risk of glycemic events during Ramadan compared with non-fasting months.8 In response, international organizations have developed evidence-based guidelines to support safer fasting practices.7 The IDF–Diabetes and Ramadan (IDF–DAR) Practical Guidelines emphasize individualized risk stratification, shared decision-making, pre-Ramadan education, medication adjustment and frequent glucose monitoring as core components of optimal diabetes management during Ramadan.7 Nevertheless, real-world adherence to these recommendations remains heterogeneous, with persistent gaps between guideline recommendations and lived experience as well as routine clinical practice, particularly in everyday care settings.9–11Despite the high prevalence of diabetes and widespread observance of Ramadan in Qatar, local real-world evidence on glycemic outcomes and their determinants remains poorly characterized. Existing Qatar-based studies have primarily focused on metabolic changes, the incidence of hypoglycemia or the prediction of glucose variability during Ramadan,12–15 whereas clinical, behavioral and self-management factors influencing hypoglycemic or hyperglycemic events remain underexplored. To address this gap, the present study examines sociodemographic, diabetes-related clinical, and glucose-monitoring and self-management factors associated with hypoglycemic or hyperglycemic events among patients with diabetes during Ramadan in Qatar.Methods Study design and population This study was reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines ( online supplemental Table S1). Data were drawn from the Qatar Diabetes Association (QDA) Ramadan support program, implemented as part of QDA’s broader mission to provide multidisciplinary diabetes care and support for individuals living with or at risk of diabetes.16 The program offers medical and nutritional guidance to help individuals with diabetes fast safely and has been active since 2022. The analysis incorporates clinical and behavioral data collected during program implementation between 2022 and 2024 to examine sociodemographic characteristics, diabetes-related clinical factors and glucose-monitoring and self-management practices associated with hypoglycemic or hyperglycemic events among patients with diabetes during Ramadan in Qatar.SP110.1136/bmjdrc-2026-006076.supp1Supplementary dataPatients were eligible if they attended either a QDA diabetes education or dietetic clinic and had a diagnosis of type 1 or type 2 diabetes. Patients were excluded if they had a diagnosis of pre-diabetes or gestational diabetes.As the analysis was based on routinely collected data from clinical encounters within the QDA Ramadan support program, all eligible patients during the study period were included. Therefore, no formal sampling frame, active recruitment process or a priori sample size calculation was undertaken.Data sources and collection procedures Clinical, behavioral and self-management information was documented using a structured data sheet for all patients, regardless of whether they planned to fast. Assessments began 6–8 weeks before Ramadan, when patients presenting for diabetes management guidance were proactively informed about the upcoming fasting period and its implications for their care, and continued through scheduled follow-up visits throughout Ramadan.All data were collected by a team of nine trained diabetes educators using a standardized workflow. During the pre-Ramadan visit for each Ramadan period, educators administered the IDF–DAR Fasting Risk Assessment,7 a validated tool that stratifies patients with diabetes into low-risk, moderate-risk or high-risk categories based on clinical characteristics, including diabetes type and duration; hypoglycemia frequency and severity; glycemic control; treatment regimen; self-monitoring of blood glucose practices; complications (acute, macrovascular and renal); pregnancy status; frailty and cognitive function; physical labor intensity; prior Ramadan fasting experience and fasting duration. Educators then provided individualized education and recommendations based on each patient’s resulting risk score. Patients classified as high risk are generally advised not to fast, whereas those at lower risk may fast with appropriate medical guidance.Study outcomes In this study, two primary and three secondary outcomes were prespecified to assess hypoglycemic or hyperglycemic events during Ramadan. Hypoglycemia and hyperglycemia were combined into composite primary outcomes to maximize statistical power and, importantly, to capture the overall burden of glycemic instability during Ramadan, as both events may occur within the same patient.The primary outcomes assessed were (1) hypoglycemia or hyperglycemia at any time and (2) hypoglycemia or hyperglycemia during fasting hours. The secondary outcomes further characterized these glycemic events by examining the individual components of the composite endpoints within distinct time periods, including (1) hypoglycemia during fasting hours; (2) hyperglycemia during fasting hours and (3) hyperglycemia after fasting hours.In addition to the primary and secondary outcomes, exploratory analyses were conducted to evaluate Ramadan-related outcomes that provide additional context for interpreting glycemic patterns; these analyses should be considered hypothesis-generating. These included patients’ reporting of positive Ramadan experience, interruption of fasting for 1 day or more, refusal to break the fast despite experiencing hypoglycemia or hyperglycemia and treatment adjustments made during Ramadan to manage hypoglycemia or hyperglycemia.Outcome measurement and validation During each Ramadan period, patients were instructed to use their own glucose-monitoring devices to measure and record their glucose levels as part of routine self-monitoring. Educators maintained regular contact with patients to collect information on hypoglycemic and hyperglycemic events. For patients using self-monitoring blood glucose meters, data were based on self-reported readings. For those using continuous glucose monitoring (CGM) devices, data-sharing features enabled educators to review glucose values in real time or retrospectively.All outcomes were treated as binary variables coded as 1 for ‘yes’ and 0 for ‘no’. Hypoglycemia was defined as a blood glucose reading <70 mg/dL, consistent with the American Diabetes Association (ADA) definition of level 1 hypoglycemia and adopted in the literature,17–19 capturing hypoglycemia irrespective of symptom status. Although lower thresholds (eg, <54–55 mg/dL) may better identify clinically significant events, the <70 mg/dL cut-off was selected to maximize sensitivity, particularly given the reliance on patient self-monitoring data. Notably, none of the hypoglycemic events required external intervention (eg, glucagon administration or emergency department visits). All patients remained conscious and oriented, and self-managed their episodes, either by breaking their fast or, in some cases, by continuing to fast with close glucose monitoring until the fast-breaking meal at sunset. Hyperglycemia was defined as a reading >250 mg/dL during or after fasting hours.Fasting hours were defined based on local astronomical times, beginning at dawn (Fajr) and ending at sunset (Maghrib) for each Ramadan season.Independent variables Independent variables potentially associated with hypoglycemia or hyperglycemia were determined a priori based on theoretical frameworks and existing literature. 20–25 Definitions and categorizations of all variables are provided in online supplemental Table S2. Variables were selected to ensure that each predictor preceded the outcome of interest.For the primary and secondary outcomes, independent variables included sociodemographic characteristics, specifically age in years and sex.Diabetes-related clinical factors encompassed hemoglobin A1c (HbA1c) level, reflecting average glycemic control over the preceding 2–3 months before Ramadan; diabetes type; diabetes duration; the IDF–DAR risk score7; whether a fasting plan was discussed with a physician; and the use of medications known to increase hypoglycemia risk.Medications were categorized as ‘no risk’, defined as diet alone or oral agents with no risk of hypoglycemia (metformin, dipeptidyl peptidase-4 inhibitors and sodium–glucose cotransporter-2 inhibitors), or ‘at risk’, defined as oral agents associated with hypoglycemia (sulfonylureas and meglitinides), and all insulin regimens (short-acting, long-acting and premixed insulins).Glucose-monitoring and self-management factors included the use of CGM devices and the daily number of recorded blood glucose readings.For exploratory analyses, the above factors were included, and selected primary, secondary or other exploratory outcomes could additionally serve as independent variables for another exploratory outcome, depending on their temporal sequence and the specific analytic model.Statistical analysis Sociodemographic, diabetes-related clinical (including use of insulin) and glucose-monitoring and self-management factors were described using frequency distributions. The prevalence of each hypoglycemic or hyperglycemic outcome was estimated first at the observation level to describe its distribution across Ramadan periods and then at the patient level, defining patients as having an event if they experienced at least one episode of the corresponding glycemic outcome across the three Ramadan periods.Associations with each outcome were first examined using univariable logistic regression models fitted with generalized estimating equations (GEE) and an exchangeable correlation structure to account for the non-independence of repeated observations within patients across Ramadan periods.Variables with a p value ≤0.20 in univariable analyses were subsequently entered into multivariable GEE logistic regression models. Statistical significance in the multivariable model was determined at a p value <0.05. Effect estimates were reported as ORs and adjusted ORs (aORs) with corresponding 95% CIs, in line with the logistic regression framework and the study objective of examining associations with hypoglycemic or hyperglycemic events.Absolute risks and predicted probabilities were not estimated, as the analysis was not intended for risk prediction or individualized risk assessment. Multicollinearity was assessed using the variance inflation factor (VIF), with values ≥5 indicating multicollinearity.26 Interactions were not investigated, and no adjustment for multiple testing was applied.Descriptive and regression analyses were performed using R V.4.4.127 (R Foundation for Statistical Computing, Vienna, Austria) with the ‘geepack’28 package.Missing data handling Missing values for all selected variables were below 5%, which is generally considered an acceptable threshold for complete case analysis. 29 30 Therefore, no imputation was performed.Ethics This study was reviewed and approved by the Institutional Review Board (IRB) of Weill Cornell Medicine–Qatar (IRB Number 25–00047), with a waiver of informed consent, as it involved the use of existing retrospective coded data not collected specifically for this study. The funder of the study had no role in study design, data collection, data analysis, data interpretation or writing of the article. The authors had full access to all the data in the study and had the final responsibility for the decision to submit for publication.Results Study population characteristics Of the 379 patients initially identified, 374 were included after excluding three with pre-diabetes and two with gestational diabetes. These patients contributed a total of 397 observations, as 37 individuals were observed across more than one Ramadan season ( table 1).Table 1Characteristics of study participantsAt any timeDuring fastingAfter fastingTotal N (%*)Hypoglycemia or hyperglycemiaHypoglycemia or hyperglycemiaHypoglycemiaHyperglycemiaHyperglycemiaTotal n (%*)165†(41.6%)107† (27.0%)65† (16.4%)59† (14.9%)101† (25.4%)397‡ (100.0%)Sociodemographic factorsAge (years)      <5075 (40.8)44 (23.9)26 (14.1)24 (13.0)47 (25.5)184 (46.3)≥5090 (42.3)63 (29.6)39 (18.3)35 (16.4)54 (25.4)213 (53.7)Sex       Male139 (41.4)88 (26.2)53 (15.8)49 (14.6)91 (27.1)336 (84.6) Female26 (42.6)19 (31.1)12 (19.7)10 (16.4)10 (16.4)61 (15.4)Diabetes-related clinical factorsHbA1c level (%)       <738 (29.9)23 (18.1)18 (14.2)8 (6.3)23 (18.1)127 (32.0) ≥7127 (47.0)84 (31.1)47 (17.4)51 (18.9)78 (28.9)270 (68.0)Diabetes type       Type 2148 (39.1)92 (24.3)54 (14.2)52 (13.7)95 (25.1)379 (95.5) Type 117 (94.4)15 (83.3)11 (61.1)7 (38.9)6 (33.3)18 (4.5)Diabetes duration (years) <1070 (31.1)43 (19.1)29 (12.9)20 (8.9)39 (17.3)225 (56.7) ≥1095 (55.2)64 (37.2)36 (20.9)39 (22.7)62 (36.0)172 (43.3)Risk score       Low58 (27.0)31 (14.4)19 (8.8)14 (6.5)39 (18.1)215 (54.2) Moderate53 (46.9)33 (29.2)20 (17.7)17 (15.0)30 (26.5)113 (28.5) High54 (78.3)43 (62.3)26 (37.7)28 (40.6)32 (46.4)69 (17.4)Fasting plan discussed with physician No92 (35.8)60 (23.3)34 (13.2)33 (12.8)62 (24.1)257 (64.7) Yes73 (52.1)47 (33.6)31 (22.1)26 (18.6)39 (27.9)140 (35.3)Medications with hypoglycemia risk No risk41 (22.9)25 (14.0)13 (7.3)13 (7.3)26 (14.5)179 (45.1) Possible risk124 (56.9)82 (37.6)52 (23.9)46 (21.1)75 (34.4)218 (54.9)Insulin use       No79 (29.0)44 (16.2)25 (9.2)21 (7.7)51 (18.8)272 (68.5) Yes86 (68.8)63 (50.4)40 (32.0)38 (30.4)50 (40.0)125 (31.5)Glucose-monitoring and self-management factorsUse of continuous glucose monitors No95 (32.5)52 (17.8)33 (11.3)26 (8.9)64 (21.9)292 (73.6) Yes70 (66.7)55 (52.4)32 (30.5)33 (31.4)37 (35.2)105 (26.4)Number of blood glucose readings§ Lower71 (30.9)18 (10.8)23 (10.0)8 (4.8)47 (21.3)230 (57.9) Higher94 (56.3)89 (38.7)42 (25.1)51 (22.2)54 (30.7)167 (42.1)The definitions and categorizations of variables used in the analysis are detailed in online supplemental Table S2.*Missing values were excluded from the analysis. Some patients contributed repeated measurements; therefore, observations were not fully independent, and generalized estimating equations were applied in the regression analyses to account for this non-independence.†Indicates the total number of observations with the assessed hypo- or hyperglycemic outcome during the study period.‡Indicates the total number of observations included during the study period.§Patient frequency of daily blood glucose monitoring. For hypoglycemic or hyperglycemia at any time, the total number of daily blood glucose readings was used, with monitoring frequency classified as lower (<4 readings/day) or higher (≥4 readings/day). For analyses restricted to fasting hours or after fasting hours, only blood glucose readings recorded during the corresponding period were considered, and monitoring frequency was classified as lower (<2 readings/day) or higher (≥2 readings/day). Although not shown in the table, 167 patients (42.1%) had lower-frequency monitoring and 230 (57.9%) had higher frequency-monitoring during fasting hours. After fasting hours, 221 patients (55.7%) had lower-frequency monitoring and 176 (44.3%) had higher frequency monitoring.HbA1c, hemoglobin A1c.Observations were distributed across the three Ramadan seasons as follows: 38.5% in 2022, 34.6% in 2023 and 26.9% in 2024. The vast majority fasted for the full duration of Ramadan (86.4%), while 6.3% missed only 1 day and fewer than 2% missed 5 days or more. Slightly more than half of the observations were from patients aged 50 years or older, and the vast majority were men (84.6%) (table 1).Regarding diabetes-related clinical factors, most observations reflected an HbA1c level ≥7%, type 2 diabetes and a shorter diabetes duration (table 1). Nearly half were classified as moderate or high risk according to the IDF–DAR risk score, of whom 90.6% chose to fast. A fasting plan had been discussed with a physician in just over one-third of observations, including 37.9% of those with a moderate or high IDF–DAR risk score, and 54.9% used medications known to increase hypoglycemia risk, with insulin accounting for 31.5%.For glucose-monitoring and self-management factors, just over one-quarter of observations involved the use of CGM devices, and more than 40% had a higher daily number of recorded blood glucose readings (table 1).Prevalence of hypoglycemic or hyperglycemic outcomes The prevalence of hypoglycemia or hyperglycemia at the observation level varied by the timing of assessment ( table 1). Estimates were 41.6% (95% CI 36.8 to 46.5%) for hypoglycemia or hyperglycemia at any time; 27.0% (95% CI 22.8 to 31.5%) for hypoglycemia or hyperglycemia during fasting hours; 16.4% (95% CI 13.1% to 20.3%) for hypoglycemia during fasting hours; 14.9% (95% CI 11.7% to 18.7%) for hyperglycemia during fasting hours; and 25.4% (95% CI 21.4% to 29.9%) for hyperglycemia after fasting hours. Most observations (83.6%) had no hypoglycemia during fasting hours, whereas 11.1% and 5.3% had 1–2 and ≥3 episodes, respectively.Prevalence of all hypoglycemic or hyperglycemic outcomes was higher among observations with HbA1c ≥7%, type 1 diabetes, a longer diabetes duration, higher IDF–DAR risk scores, having discussed a fasting plan with a physician, use of medications known to increase hypoglycemia risk and use of insulin (table 1). In contrast, prevalence was lower among observations that did not use CGM devices and had a lower daily number of recorded blood glucose readings.At the patient level (online supplemental figure S1), prevalence estimates were 40.3% (95% CI 35.5% to 45.4%) for hypoglycemia or hyperglycemia at any time; 25.9% (95% CI 21.8% to 30.6%) for hypoglycemia or hyperglycemia during fasting hours; 16.0% (95% CI 12.7% to 20.1%) for hypoglycemia during fasting hours; 15.0% (95% CI 11.7% to 18.9%) for hyperglycemia during fasting hours; and 25.7% (95% CI 21.5% to 30.3%) for hyperglycemia after fasting hours.Factors associated with hypoglycemia or hyperglycemia Results of the multivariable logistic regression analyses are presented in tables 2 and 3 and online supplemental Tables S3 and S4. VIF was <5 across all models, indicating no multicollinearity among independent variables.Table 2Factors associated with hypoglycemia or hyperglycemia at any timeCharacteristicsAt any timeHypoglycemia or hyperglycemiaUnivariable regression analysisMultivariable regression analysisOR (95% CI)P valueWald test P value*aOR (95% CI)P value†Sociodemographic factorsAge (years)      <501.00 0.936---- ≥501.02 (0.67 to 1.54)0.936 ----Sex      Male1.00 0.821---- Female1.07 (0.62 to 1.85)0.821 ----Diabetes-related clinical factorsHbA1c level (%)      <71.00 0.0031.00  ≥71.99 (1.26 to 3.16)0.003 0.81 (0.46 to 1.44)0.474Diabetes type      Type 21.00 0.0031.00  Type 127.98 (3.04 to 257.27)0.003 5.21 (0.76 to 35.72)0.093Diabetes duration (years) <101.00 <0.0011.00  ≥102.63 (1.73 to 4.01)<0.001 2.21 (1.37 to 3.55)0.001Risk score      Low1.00 <0.0011.00  Moderate2.34 (1.44 to 3.80)0.001 1.74 (0.96 to 3.13)0.066 High9.31 (4.76 to 18.22)<0.001 4.70 (1.98 to 11.14)<0.001Fasting plan discussed with physician No1.00 0.0031.00  Yes1.82 (1.22 to 2.71)0.003 1.85 (1.14 to 2.99)0.012Medications with hypoglycemia risk No risk1.00 <0.0011.00  Possible risk4.25 (2.72 to 6.64)<0.001 2.44 (1.38 to 4.33)0.002Glucose-monitoring and self-management factorsUse of continuous glucose monitors No1.00 <0.0011.00  Yes3.87 (2.35 to 6.38)<0.001 3.68 (1.67 to 8.12)0.001Number of blood glucose readings‡ Lower1.00 <0.0011.00  Higher2.73 (1.78 to 4.19)<0.001 1.07 (0.56 to 2.02)0.842The definitions and categorizations of variables used in the analysis are detailed in online supplemental Table S2.*Covariates with p value≤0.2 in the univariable analysis were included in the multivariable analysis.†Covariates with p value<0.05 in the multivariable analysis were considered as showing statistically significant evidence for an association with hypoglycemia or hyperglycemia.‡Patient frequency of daily blood glucose monitoring. For hypoglycemia or hyperglycemia at any time, the total number of daily blood glucose readings was used, with monitoring frequency classified as lower (<4 readings/day) or higher (≥4 readings/day). For analyses restricted to fasting hours, only blood glucose readings recorded during the corresponding period were considered, and monitoring frequency was classified as lower (<2 readings/day) or higher (≥2 readings/day).aOR, adjusted OR; HbA1c, hemoglobin A1c.Table 3Factors associated with hypoglycemia or hyperglycemia during fasting hoursCharacteristicsDuring fastingHypoglycemia or hyperglycemiaUnivariable regression analysisMultivariable regression analysisOR (95% CI)P valueWald test P value*aOR (95% CI)P value†Sociodemographic factorsAge (years)      <501.00 0.339---- ≥501.26 (0.79 to 2.00)0.339 ----Sex      Male1.00 0.405---- Female1.28 (0.71 to 2.31)0.405 ----Diabetes-related clinical factorsHbA1c level (%)      <71.00 0.0311.00  ≥71.90 (1.06 to 3.42)0.031 0.78 (0.38 to 1.57)0.483Diabetes type      Type 21.00 <0.0011.00  Type 113.00 (3.65 to 46.35)<0.001 3.81 (1.06 to 13.71)0.041Diabetes duration (years)      <101.00 <0.0011.00  ≥102.30 (1.44 to 3.67)<0.001 1.89 (1.09 to 3.26)0.023Risk score      Low1.00 <0.0011.00  Moderate2.39 (1.35 to 4.24)0.003 2.16 (1.04 to 4.51)0.040 High8.28 (4.23 to 16.19)<0.001 5.28 (2.19 to 12.74)<0.001Fasting plan discussed with physician No1.00 0.0571.00  Yes1.51 (0.99 to 2.32)0.057 1.35 (0.82 to 2.24)0.241Medications with hypoglycemia risk No risk1.00 <0.0011.00  Possible risk3.53 (2.05 to 6.09)<0.001 1.48 (0.70 to 3.13)0.306Glucose-monitoring and self-management factorsUse of continuous glucose monitors No1.00 <0.0011.00  Yes4.74 (2.78 to 8.08)<0.001 2.86 (1.55 to 5.29)0.001Number of blood glucose readings‡ Lower1.00 <0.0011.00  Higher4.69 (2.68 to 8.21)<0.001 2.62 (1.34 to 5.13)0.005The definitions and categorizations of variables used in the analysis are detailed in online supplemental Table S2.*Covariates with p value≤0.2 in the univariable analysis were included in the multivariable analysis.†Covariates with p value<0.05 in the multivariable analysis were considered as showing statistically significant evidence for an association with hypoglycemia or hyperglycemia.‡Patient frequency of daily blood glucose monitoring. For hypoglycemia or hyperglycemia at any time, the total number of daily blood glucose readings was used, with monitoring frequency classified as lower (<4 readings/day) or higher (≥4 readings/day). For analyses restricted to fasting hours, only blood glucose readings recorded during the corresponding period were considered, and monitoring frequency was classified as lower (<2 readings/day) or higher (≥2 readings/day).aOR, adjusted OR; HbA1c, hemoglobin A1c.Primary outcomes Hypoglycemia or hyperglycemia at any time. Patients with a longer diabetes duration had more than twice the odds of experiencing hypoglycemia or hyperglycemia at any time (aOR: 2.21, 95% CI 1.37 to 3.55) (table 2). Those with a high IDF–DAR risk score had higher odds of experiencing hypoglycemia or hyperglycemia at any time compared with those at low risk (aOR: 4.70, 95% CI 1.98 to 11.14). Higher odds were also observed among patients who had discussed a fasting plan with a physician (aOR: 1.85, 95% CI 1.14 to 2.99), who used medications known to increase hypoglycemia risk (aOR: 2.44, 95% CI 1.38 to 4.33) and who used CGM devices (aOR: 3.68, 95% CI 1.67 to 8.12).Hypoglycemia or hyperglycemia during fasting hours. Patterns observed during fasting hours were largely consistent with those for events occurring at any time, although some of the associations did not reach statistical significance (table 3). Moreover, during fasting hours, statistically significant higher odds were observed among patients with type 1 diabetes (aOR: 3.81, 95% CI 1.06 to 13.71) and among those with a higher daily number of recorded blood glucose readings (aOR: 2.62, 95% CI 1.34 to 5.13).Secondary outcomes Patterns observed for the secondary outcomes were generally consistent with those for the primary outcomes, although some associations did not reach statistical significance, likely because of the smaller number of events ( online supplemental Table S3).For hypoglycemia during fasting hours (online supplemental Table S3), higher odds were observed among patients with type 1 diabetes (aOR: 3.67, 95% CI 1.27 to 10.58) and those who had a higher daily number of recorded blood glucose readings (aOR: 2.92, 95% CI 1.36 to 6.26).For hyperglycemia during fasting hours (online supplemental Table S3), higher odds were observed among patients with a longer diabetes duration (aOR: 2.15, 95% CI 1.13 to 4.11), a high IDF–DAR risk score (aOR: 6.44, 95% CI 2.55 to 16.25), who used CGM devices (aOR: 2.61, 95% CI 1.30 to 5.24) and who had a higher daily number of recorded blood glucose readings (aOR: 2.89, 95% CI 1.22 to 6.84).For hyperglycemia after fasting hours (online supplemental Table S4), higher odds were observed among patients with a longer diabetes duration (aOR: 2.07, 95% CI 1.27 to 3.38), a high IDF–DAR risk score (aOR: 2.28, 95% CI 1.09 to 4.75) and who used medications known to increase hypoglycemia risk (aOR: 1.97, 95% CI 1.06 to 3.67).Notably, no associations were found between age or sex and any of the hypoglycemic or hyperglycemic outcomes.Exploratory analyses Exploratory analyses of Ramadan-related outcomes are presented in table 4.Table 4Exploratory analyses of factors associated with Ramadan-related outcomes: (a) patients’ reporting of positive Ramadan experience, (b) interruption of fasting for 1 day or more, (c) refusal to break the fast despite experiencing hypoglycemia or hyperglycemia, and (d) treatment adjustments made during Ramadan to manage hypoglycemia or hyperglycemiaRamadan-related outcomesUnivariable regression analysisMultivariable regression analysisOR (95% CI)P valueWald test P value*aOR (95% CI)P value†Patients’ reporting of positive Ramadan experienceInterruption of fasting for 1 day or more No1.00 <0.0011.00  Yes0.07 (0.04 to 0.14)<0.001 0.08 (0.03 to 0.22)<0.001Refusal to break the fast despite experiencing hypoglycemia or hyperglycemia No1.00 0.0061.00  Yes0.27 (0.11 to 0.69)0.006 0.13 (0.04 to 0.42)0.001Interruption of fasting for 1 day or moreSex      Male1.00 0.0061.00  Female2.58 (1.31 to 5.07)0.006 3.81 (1.65 to 8.79)0.002Fasting plan discussed with physician      No1.00 0.0031.00  Yes2.21 (1.31 to 3.75)0.003 1.92 (1.02 to 3.62)0.044Hypoglycemia or hyperglycemia at any time No1.00 <0.0011.00  Yes14.18 (6.09 to 33.00)<0.001 7.39 (2.45 to 22.27)<0.001Refusal to break the fast despite experiencing hypoglycemia or hyperglycemiaRisk score      Low1.00 <0.0011.00  Moderate1.72 (0.65 to 4.56)0.278 1.48 (0.49 to 4.47)0.492 High10.12 (4.33 to 23.66)<0.001 7.88 (2.39 to 26.03)0.001Treatment adjustments made during Ramadan to manage hypoglycemia or hyperglycemia‡Risk score      Low1.00 <0.0011.00  Moderate3.29 (2.05 to 5.28)<0.001 3.04 (1.77 to 5.23)<0.001 High9.75 (5.17 to 18.39)<0.001 6.92 (3.43 to 13.96)<0.001Interruption of fasting for 1 day or more No1.00 <0.0011.00  Yes6.13 (3.07 to 12.22)<0.001 3.86 (1.79 to 8.35)0.001Only variables that remained significant in the multivariable analyses are included in the table.The definitions and categorizations of variables used in the analysis are detailed in online supplemental Table S2.*Covariates with p value≤0.2 in the univariable analysis were included in the multivariable analysis.†Covariates with p value<0.05 in the multivariable analysis were considered as showing statistically significant evidence for an association with the corresponding Ramadan-related outcome.‡Treatment adjustments made during Ramadan to manage hypoglycemia or hyperglycemia refers to any treatment modification for the patient initiated by the Qatar Diabetes Association team or the patient’s physician during Ramadan, including changes to medication or dietary regimen aimed at controlling blood glucose levels.aOR, adjusted OR.Patients who interrupted their fast for 1 day or more had lower odds of reporting a positive Ramadan experience compared with those who did not (table 4). Similarly, patients who refused to break their fast despite experiencing hypoglycemia or hyperglycemia had lower odds of reporting a positive Ramadan experience than those who did not.Female patients, those who discussed a fasting plan with a physician, and those who experienced hypoglycemia or hyperglycemia at any time had higher odds of interrupting their fast for 1 day or more (table 4).Patients with a high IDF–DAR risk score had higher odds of refusing to break their fast despite experiencing hypoglycemia or hyperglycemia (table 4).Patients with a high IDF–DAR risk score and those who interrupted their fast for 1 day or more had higher odds of having treatment adjustments made during Ramadan to manage hypoglycemia or hyperglycemia (table 4).Discussion This study assessed sociodemographic, diabetes-related clinical and glucose-monitoring and self-management factors associated with hypoglycemic or hyperglycemic events among patients with diabetes during Ramadan in Qatar, providing real-world evidence in a high-prevalence setting.Overall, type 1 diabetes, a longer diabetes duration, a high IDF–DAR risk score, having discussed a fasting plan with a physician, use of medications known to increase hypoglycemia risk, use of CGM devices and a higher daily number of recorded blood glucose readings were associated with higher odds of hypoglycemic or hyperglycemic events. The direction and magnitude of these associations were generally consistent across primary and secondary outcomes.Type 1 diabetes and longer diabetes duration were associated with higher odds of hypoglycemic or hyperglycemic events, likely reflecting greater dependence on exogenous insulin, impaired counterregulatory mechanisms and the cumulative pathophysiological effects of long-standing disease, all of which predispose to glycemic instability during prolonged fasting.7 31 32In secondary analyses, type 1 diabetes was associated with hypoglycemia during fasting hours, consistent with this group’s increased susceptibility to fasting-induced glucose decline,8 33 whereas longer diabetes duration was associated with hyperglycemia during and after fasting hours, possibly reflecting progressive beta-cell dysfunction and reduced endogenous insulin reserve34 35 as well as suboptimal medication adjustment during Ramadan.Patients with a high IDF–DAR risk score had higher odds of hypoglycemic or hyperglycemic events than those at low risk, consistent with evidence from the Middle East and North Africa and South Asia supporting the tool’s predictive validity.22 36–38 This pattern extended also to hyperglycemia during and after fasting hours, which may reflect increased hyperglycemia in high-risk patients with longer diabetes duration.11Exploratory analyses further indicated that these patients were more likely to refuse to break their fast despite experiencing hypoglycemia or hyperglycemia and to require treatment adjustments during Ramadan to manage such events. Collectively, these findings suggest that the IDF–DAR high-risk profile—characterized by greater comorbidity and diabetes-related complications22 38—is associated with increased physiological vulnerability, reluctance to break the fast and greater clinical management needs during Ramadan.An important and seemingly paradoxical finding was that patients who had discussed a fasting plan with a physician exhibited higher odds of hypoglycemic or hyperglycemic events. This pattern may reflect confounding by indication, as individuals who seek medical advice are typically those at higher baseline risk and may nonetheless choose to fast despite medical caution, remaining vulnerable to glycemic events even with education and close monitoring.7 38 39 The higher odds of interrupting the fast for 1 day or more among these patients in exploratory analyses support this interpretation and suggest that structured counseling may improve symptom recognition, glucose monitoring and timely fast-breaking to prevent complications.7 40Use of medications known to increase hypoglycemia risk was associated with higher odds of hypoglycemic or hyperglycemic events, a biologically plausible finding given the prolonged daytime fasting, delayed meals and altered dosing schedules during Ramadan that may potentiate the glucose-lowering effects of these agents.8 41 The association with hyperglycemia after fasting hours may reflect conservative daytime dosing to avoid hypoglycemia, followed by glucose excursions after breaking the fast related to compensatory eating or high-glycemic index foods.42 Skipping the pre-dawn meal (Suhoor) and underlying comorbidities may also predispose patients—even those using medications with no hypoglycemia risk—to hypoglycemia during fasting.41 43Interestingly, use of CGM devices and a higher daily number of recorded blood glucose readings were associated with higher odds of hypoglycemic or hyperglycemic events. This may reflect surveillance bias, whereby more frequent glucose monitoring increases the likelihood of detecting glycemic excursions that might otherwise go unnoticed.7Beyond physiological outcomes, glycemic events also influenced fasting behavior and patient well-being. Patients who interrupted their fast for 1 day or more, as well as those who refused to break their fast despite experiencing hypoglycemia or hyperglycemia, had lower odds of reporting a positive Ramadan experience. Although such events medically and religiously warrant breaking the fast, strong religious and cultural motivations may lead some individuals to prioritize continued fasting over short-term health risks, which may have serious consequences if glycemic events become severe.11 44 45 This tension may contribute to heightened anxiety, reduced quality of life and, ultimately, reduced overall well-being during Ramadan.46 47From a practice and policy perspective, these findings support structured pre-Ramadan assessment, patient-centered counseling and tailored treatment adjustment and education throughout the fasting period—especially for persons with type 1 diabetes, those with long diabetes duration (whether type 1 or type 2), and a high IDF–DAR risk score. Moreover, the high prevalence of hyperglycemia observed in this study underscores the importance of addressing dietary behaviors, including meal portion size and glycemic composition, alongside appropriate medication adjustment, as integral components of targeted Ramadan diabetes interventions.This study has limitations. First, glycemic outcomes relied on patient self-monitoring, which captures real-world glycemic burden but is inherently influenced by testing frequency, monitoring practices, device-related measurement error and variability in fasting duration across days and individuals, potentially introducing event misclassification and limiting endpoint precision. Individuals who monitor more frequently—often those at higher risk or using CGM—may be more likely to detect events and potentially be over-represented in the data, while lower monitoring frequency in others may have led to underestimation of hypoglycemic and hyperglycemic event prevalence.Second, the retrospective design limited the ability to comprehensively capture variables and identifiers that were absent from clinical records or lacked sufficient detail for analysis—such as detailed dietary intake, timing of insulin administration, physical activity, daily fasting duration and detailed event counts for all glycemic outcomes—and precluded adjustment for clustering at the educator or clinic level. It also prevented assessment of temporal clustering of events within Ramadan and, in the absence of pre-Ramadan glycemic excursion data, limited the ability to distinguish fasting-specific effects from habitual glycemic variability or to adjust for underlying glycemic control.Third, although some statistically significant associations were observed in certain analyses, the small sample size in specific subgroups—particularly patients with type 1 diabetes—and a low number of events reduced statistical power and estimate precision, limiting the ability to perform stratified analyses (eg, by ADA hypoglycemia severity). Additionally, no adjustment for multiple testing was applied across the assessed outcomes, and findings should, therefore, be interpreted as nominally significant rather than strictly significant.Fourth, the study population consisted of attendees of the QDA Ramadan support program, who may be more health conscious or more engaged with healthcare services than the broader population of individuals with diabetes in Qatar, thereby introducing potential selection bias.Fifth, only ORs were reported, without estimation of absolute risks or predicted probabilities, which may limit clinical interpretability and applicability for individualized risk assessment.This study has strengths. First, it provides real-world evidence on the intersection between living with diabetes and a religious practice observed by over 20% of the global population and of substantial clinical and public health implications, by capturing the experiences of persons with diabetes during Ramadan in routine care settings. Second, the inclusion of data from three consecutive Ramadan periods enhances the robustness and generalizability of the findings. Third, the study was conducted in Qatar, a country with one of the highest prevalences of diabetes worldwide, rendering the findings particularly relevant to populations most affected by the disease and to regions where Ramadan fasting is practiced.In conclusion, this study provides real-world evidence on sociodemographic, diabetes-related clinical, and glucose-monitoring and self-management factors associated with hypoglycemic or hyperglycemic events among patients with diabetes during Ramadan in Qatar. Type 1 diabetes, a longer diabetes duration (whether type 1 or type 2), a high IDF–DAR risk score, having discussed a fasting plan with a physician, use of medications known to increase hypoglycemia risk, use of CGM devices and a higher daily number of recorded blood glucose readings were associated with higher odds of hypoglycemic or hyperglycemic events. These findings support the integration and scale-up of standardized, guideline-aligned Ramadan diabetes care programs, alongside structured pre-Ramadan assessment, patient-centered education, close and rigorous follow-up and individualized treatment adjustment—particularly for high-risk groups—to improve both safety and the overall Ramadan experience in high-diabetes prevalence settings.",
  "title": "Factors associated with hypoglycemic or hyperglycemic events during Ramadan among persons with diabetes in Qatar",
  "uid": "b27002fe-71af-5c37-bac3-86ae0363e47b"
}