Incidence of diabetic ketoacidosis does not differ in Ramadan compared to other months and seasons: results from a 6-year multicenter study

Abstract

Background

Diabetic ketoacidosis (DKA) is a life-threatening adverse complication of patients with diabetes mellitus (DM). It is postulated that fasting during Ramadan can increase the risk of DKA; however, there are contradicting data in this regard. Furthermore, studies from Western countries have suggested a seasonal variation in the incidence of DKA. This study examines the differences in the number of DKA episodes during Ramadan compared to the rest of the year in patients with type 1 DM (T1D) and type 2 DM (T2D). Besides, we aim to examine the seasonal difference in the incidence of DKA.

Methods

We included consecutive index-DKA admissions from 2015 to 2021 and used descriptive statistics to compare the episodes of DKA in Ramadan vs other months and seasons.

Results

Of 922 patients, 480 (52%) had T1D, whereas 442 (48%) had T2D. The median age (IQR) was 35 (25–45) years, with the majority being Arab (N = 502, 54.4%). There were 94 DKA admissions in six collective Ramadan months, whereas the DKA admissions ranged from 61 to 88 episodes in other months (p = .3). The highest DKA admissions were observed in Autumn (N = 236) and the lowest in Spring (N = 226) with no statistically significant difference (p = .4). There were no differences in DKA severity or new-onset diabetes rates when analyzed based on Hiji months, Roman months, or seasons.

Conclusions

DKA occurrence is not increased during Ramadan. We found no evidence of seasonal variations in the rates of DKA in the State of Qatar.

Keywords:

• Diabetic ketoacidosis
• DKA
• type 1 diabetes
• type 2 diabetes
• Ramadan
• months
• seasons

Introduction

The pathophysiology of diabetic ketoacidosis (DKA) is complex and involves multiple metabolic and hormonal changes¹. The primary driver for DKA is insulin deficiency, either absolute (such as in T1D or advanced T2D with beta-cell dysfunction) or relative (such as in T2D or under-treated T1D)². Insulin deficiency, coupled with the body’s inability to utilize glucose, results in multifactorial and unrestricted hyperglycemia². Besides, the lack of insulin results in the loss of the storage of free fatty acids in the adipose tissue, resulting in excess free fatty acids, which are then converted into ketones by the liver². The lack of insulin also triggers an uprise in the counter-regulatory hormones, such as cortisol, glucagon, epinephrine, norepinephrine, and growth hormone which further propagate hyperglycemia and ketonemia¹.

Ramadan is the ninth month of the Islamic lunar calendar (Hijri). Adult Muslims, fast from dawn to sunset during Ramadan. Muslims with chronic medical conditions – such as diabetes – are exempt from fasting³. However, many Muslims with DM still fast during Ramadan. The EPIDIAR study reported that 42.8% of patients with T1D and 78.7% of patients with T2D fasted for at least 15 days. Hence, it is essential to provide intensive education and medication adjustment during fasting for better and safe management of their DM^4,5.

In Ramadan, the long periods of fasting (especially in countries with increased fasting hours) and the relevant shift of the body’s metabolism to break down fatty acid into ketones may lead to ketoacidosis in patients with diabetes treated with insulin⁶. The usual insulin doses in patients with T1D and T2D are reduced by 25–40% to counter the increased risk of hypoglycemia during fasting⁷. This iatrogenic reduction in basal insulin, prolonged fasting hours, and concomitant dehydration increase the risk of ketoacidosis. However, whether these factors translate into an increased incidence of DKA during Ramadan remains unclear. Historically, expert opinions have persistently stated an increased risk of DKA during Ramadan^8–17. However, these reviews and opinions were mainly based on anecdotal data. Multiple studies have been conducted to ascertain the incidence of DKA in Ramadan compared to other months, from 1984 to 2022. However, most of the studies included a small number of patients and yielding inconsistent results^3,18–26. A recent review by Beshyah et al. has critically appraised the available literature (including reviews and original studies) and concluded that the risk of DKA in Ramadan is likely not increased²⁷. Furthermore, some studies from Western countries reported an inverse relationship between ambient temperature and the incidence of DKA, with the highest incidence in Winters compared to other seasons^28,29. However, the effect of ambient temperature and seasonal variations on the incidence of DKA has not been studied in the Middle East, where the seasons are different in intensity compared to the West.

To investigate the differences in the occurrence of DKA during Ramadan compared to other Hijri months and the seasonal variations in the incidence of DKA, we aimed to conduct a retrospective analysis on patients with T1D and T2DM admitted with DKA We studied the differences in the occurrence of DKA in Hijri months, Roman months, and seasons over 6 years. The differences in the occurrence of DKA were studied based on total DKA episodes, DKA episodes categorized by the type of DM, the duration of DM, and lastly, based on the severity of DKA.

Research design and methods

Study design

This retrospective cross-sectional study included consecutive index DKA admissions at four hospitals (three secondary care and one tertiary care hospital) of Hamad Medical Corporation (HMC) from January 2015 to March 2021.

Ethical approval

The institutional review board at the medical research center (MRC), Qatar, approved this study (MRC-01-21-476). Due to the retrospective nature of the data review and anonymous data sharing, the IRB waived the need for informed consent.

Inclusion criteria and data collection

The methodology has been detailed in previous publication³⁰. Inclusion criteria were adult patients (>14 years of age) with index DKA admissions, identified through electronic medical records (EMR) Cerner Millennium (Oracle). The following criteria were applied to validate the diagnosis of ketoacidosis: pH 7.3 or less, anion gap > 10 mmol/L, and ketonemia (beta hydroxybutyrate more than 3 mmol/L) or positive ketones in the urine. Included patients had a diagnosis of DM based on a glycated hemoglobin (HbA1c) level > 6.5% or a fasting glucose level > 7.0 mmol/L at the time of DKA diagnosis or before the relevant hospitalization. In order to include patients with euglycemic DKA in the study group, the glucose level was not used to classify DKA cases. Exclusion criteria were patients who had ketoacidosis due to other conditions, such as starvation or alcohol-induced ketoacidosis, and pregnant patients with DKA.

International Classification of Diseases (ICD) codes were used to identify T1D and T2D in electronic medical records. The coding accuracy was validated in the subgroup of patients by looking at C-peptides and Anti-GAD antibodies where available. The study divided the patients into different categories based on the status of DM. The category “Newly diagnosed versus pre-existent T1D and T2D” differentiated between patients diagnosed with diabetes on the studied admission (T1D or T2D) and those with a known diagnosis of DM. The Newly diagnosed DM group included patients who presented with DKA without a previous diabetes diagnosis and were subsequently diagnosed with either T1D or T2D by clinical impression, the presence or absence of anti-GAD antibodies and levels of c-peptide. DKA was categorized using the American diabetes association (ADA) DKA diagnostic criteria³¹. Mild DKA was defined as PH 7.25–7.3, moderate DKA was defined as PH 7–7.24, whereas severe DKA was defined as PH < 7.

Clinicodemographic data, including age, sex, ethnicity, body mass index (BMI), comorbid conditions, and relevant laboratory investigations at admission, were collected manually and cross-checked for data quality. There was no missing data among the variables that diagnosed DKA. Data pertaining to the time of admission and discharge were also collected, which were then used to categorize admissions in Islamic months, Roman months, and seasons. During the study period (6 years), the average duration of fasting was between 14 and 15 h, with an average temperature of 40 degrees Celsius.

Study outcomes

Study outcomes were the frequencies of index admissions with DKA in Islamic months, Roman months, and seasons based on the DM type, duration, and severity of DKA.

Statistical analyses

The study cohort’s demographic data were presented using descriptive statistics. We categorized DM as type 1 or type 2 and differentiated whether it was newly diagnosed or pre-existing based on the time of diagnosis. The cohort was divided into four ethnic groups: Arab, Asian, African, and others. BMI was classified as normal, overweight, or obese using ethnic-specific cut-off points. Patients were classified into high metabolic risk and low metabolic risk based on the presence of either obesity, hypertension, or dyslipidemia. Continuous variables were summarized using mean (SD) and median (IQR), while categorical variables were expressed as percentages. The Chi-square test was used to compare categorical variables. We used STATA 15 for all analyses.

Results

Clinical characteristics of the study population

The baseline characteristics of the study cohort are summarized in Table 1. The median age of the patients was 35 years (IQR 25–45), out of which 575 (62%) were males. The mean body mass index (BMI) was 24.8 ± 6.2 kg/m². Of the 922 patients, 480 (52%) had T1D, whereas 442 (48%) had T2D. Of the 332 individuals with pre-existent T1D, basal insulin, bolus insulin, and mixed type of insulins were prescribed to 155 (46.7%), 161 (48.5%), and 43 (12.9%) patients, respectively. Of the 266 patients with pre-existent T2D, the prescription patterns revealed that 77 (30%), 61 (23%), and 33 (12.4%) individuals were prescribed basal insulin, bolus insulin, and mixed insulins, respectively. Additionally, prescriptions for non-insulin medications included sodium-glucose co-transporter-2 inhibitors (SGLT2i) in 33 (12.4%) patients, sulfonylureas in 29 (11%) patients, metformin in 126 (47.4%) patients, thiazolidinediones in 8 (3%) patients, Glucagon-like peptide-1 agonists (GLP-1A) in 8 (3%) patients, and dipeptidyl peptidase-4 inhibitors (DPP-i) in 57 (21.4%) patients. Some patients received insulin (and non-insulin prescriptions) from private healthcare facilities, and consequently, the corresponding data could not be extracted from the electronic medical records (EMR) system. The mean HBA1C in patients with T1D was 11.7% ± 2.6, while HBA1C in patients with T2D was 12.3% ± 2.8. Of the 922 DKA episodes, 78 (8.5%) were euglycemic DKAs. The median PH at admission was 7.19 (7.08–7.27), median serum bicarbonate level was 11 (7.1–14.7), and the median anion Gap at admission, 22 (18–27). The median beta hydroxybutyrate (BHB) at admission was5.8 (4.5–7.3) mmol/L.

Table 1. Characteristics of the patients admitted with index DKA episode.

Baseline characteristics	Results (N = 922)
Median age in years	35 (25–45)
Males (%)	575 (62.4%)
Ethnicities, N (%)	Arabs	502 (54.4%)
	Asians	300 (32.5%)
	Africans	90 (9.7%)
	Others	30 (3.2%)
Mean BMI (kg/m^2)	24.8 ± 6.2
T1D	Pre-existing	332 (36%)
	New	148 (16%)
	Total	480 (52%)
T2D	Pre-existing	266 (28.8%)
	New	176 (19%)
	Total	442 (47.9)
Duration of DM in years	6 (3–12)
DM Complications	Microvascular complications	95 (10.3%)
	Macrovascular complications	56 (6%)
	High Metabolic risk	394 (42.7%)
HBA1c on Admission	12.1 ± 2.7 %

Abbreviations. BMI, Body mass index, DM, Diabetes mellitus, T1DM, Type 1 diabetes, T2DM, Type 2 diabetes, COPD, Chronic obstructive pulmonary disease, HIV, Human immunodeficiency virus, Metabolic risk = dyslipidemia, ethnic-specific obesity, hypertension).

Frequency of admissions with DKA based on Islamic months

Figure 1(A) shows the frequencies of admissions with DKA in each month during the study period (6 years). The admissions ranged from 61 (6.6%) in the collective months of Dhu Elhajj to 94 (10.2%) in the collective months of Ramadan. Of the 598 patients with pre-existing DM, admissions with DKA ranged from 35 (5.9%) Dhu Elhajj to 61 (10.2%) Ramadan. On the other hand, among the 324 patients with newly diagnosed DM, admissions with DKA ranged from 15 (4.6%) Muharram to 33 (10.2%) Ramadan and the collective months of Shabaan (Figure 1(B)). These differences based on the onset of DM were statistically insignificant (p = .24). Of the 480 patients with T1D, admissions ranged from 32 (6.7%) Muharram to 54 (11.2%) Ramadan. Among the 442 patients with T2D, admissions ranged from 27 (6.1%) Dhu Elhajj to 42 (9.5%) Ramadan. The differences in the frequencies of admissions with DKA in different Islamic months were insignificant (p = .3) (Figure 1(C)). The cohort consisted of 202 (22%) admissions with mild DKA, 332 (36%) with moderate DKA, and 388 (42%) with severe DKA. Admissions with mild DKA ranged from 11 (5.4%) in Jamada Althani to 24 (11.8%) in Safar. Admissions with moderate DKA ranged from 22 each in Jamada Althani, Dhu Elhajj, and Shawal to 38 (11.4%) in Safar. The admissions with severe DKA ranged from 23 (5.9%) in Jamada Al Awwal to 44 (11.3%) in Rabi Althani. The admissions based on the severity of DKA were not statistically different (p: .4) (Figure 1(D)).

Figure 1. (A) Frequency of Overall DKA admissions in Islamic months. (B) Frequency of DKA admissions in Islamic months based on new versus pre-existent DM. (C) Frequency of DKA admissions in Islamic months based on the type of DM. (D) Frequency of DKA admissions in Islamic months based on the severity of DKA. X-axis represent number of DKA admissions and Y-axis represent Islamic calendar months.

Frequency of admissions with DKA based on Roman months

Figure 2 shows the frequencies of admissions with DKA based on Roman months. Based on the severity of DKA, mild DKA admissions were most prevalent in July (24, 11.4%), moderate DKA admissions were most prevalent in June (35, 10.5%), and severe DKA admissions were most common in January and May (41, 10.6%) (P: 0.9) (Figure 2(A)). Comparison based on the onset of DM showed that DKA admissions with pre-existing DM were most common in October (61, 10.2%). In contrast, DKA admissions with newly diagnosed DM were commonest in January (35, 10.8%) (p: .6) (Figure 2(B)). Of the 480 patients with T1D, the highest admissions with DKA were observed in June (52, 10.8%), whereas most admissions among the 442 patients with T2D were seen in July (47, 10.6%) (p: .3) (Figure 2(C)).

Figure 2. (A) Frequency of DKA admissions in Roman months based on the severity of DKA. (B) Frequency of DKA admissions in Roman months based on new versus pre-existent DM. (C) Frequency of DKA admissions in Roman months based on the type of DM. X-axis represent the Roman calendar months and Y-axis represent number of DKA admissions.

Frequency of admissions with DKA based on seasons

Figure 3 shows the frequencies of admissions with DKA based on seasons. Mild DKA cases were admitted the most in Summer (55, 27.2%), moderate DKA in Autumn (91, 27.4%), whereas severe DKA cases were admitted the most in Spring (103, 26.5%) (p: .7) (Figure 3(A)). When compared based on the type of DM, most patients with T1D and DKA were admitted in winter (132, 27.5%). On the other hand, most of the patients with T2D and DKA were admitted in Spring (114, 25.8%) (p: .4) (Figure 3(B)). Lastly, Of the 598 patients with pre-existing DM, the highest admissions were seen in August (168, 28%). The highest admissions among the 324 patients with newly diagnosed DM were seen in winter (88, 27.2%) (P: 01) (Figure 1(C)).

Figure 3. (A) Frequency of DKA admissions in different seasons based on the severity of DKA. (B) Frequency of DKA admissions in different seasons based on the type of DM. (C) Frequency of DKA admissions in different seasons based on new versus pre-existent DM. X-axis represent the four seasons and Y-axis represent number of DKA admissions.

There was no difference in MICU admission based on Hijri, Roman, or seasonal analysis. These results were similar across all ethnic groups.

Discussion

This study represents the largest cohort from the region that studied the risk of DKA during Ramadan, based on Roman months; and based on seasons. The frequency of admission with DKA and DKA severity were similar in all Hijri months, irrespective of type and duration of DM. Similarly, the frequencies of admissions with DKA were not significantly different when analyzed based on Roman months or during the four seasons.

The combination of prolonged fasting, changes in diet, altered medication regimens, and dehydration during Ramadan was believed to increase the risk of DKA in people with diabetes⁶. Our study supports the findings of other large cohort studies. Beshyah et al. reviewed data from 283 patients with 432 DKA episodes over 10 years and showed that DKA occurrence in Ramadan was not different than in other months²⁵. Baeno et al. prospectively followed 301 DM patients and reported on the safety of anti-diabetic drug regimen adjustments during Ramadan fasting²⁴. No excess DKA was seen in Ramadan, and hypoglycaemias were less in Ramadan. On the other hand, Alshahrani et al. published a retrospective study on 51 patients with DM. They showed an increased rate of DKA admissions during Ramadan (37.3%) compared to the previous month Shabaan (15.7%). However, the admissions with DKA in Shawaal were even higher (47%)²⁶. The study had a small sample size and only looked at three months in one calendar year. Abdelgadir et al. reported increased DKA admission during Ramadan compared to previous Hijri months³. Although the study was multicounty, the small sample size was small (170 patients from four countries) and included a few months before and after Ramadan during a single calendar year. When studied in larger cohorts, the similarities in the frequency of DKA in Ramadan compared to other months is more in line with its pathophysiology, which entails dehydration as a consequence of DKA instead of a trigger¹. Taken together, fasting during Ramadan does not seem to increase the risk of DKA.

We could not find any effect of the holidays impacting DKA admissions, as the rates were not different throughout the year. In addition to winter holidays (which are fewer in number compared to Western countries where extremes in weather make commute difficult), Qatar also witnesses holidays during the two Eid occasions. The first Eid (Eid-ul-Fitr) ranged from May to July in the 6 studied years, whereas the second Eid (Eid-ul-Adha) ranged from July to September. It is known that on occasions such as Eid, patients usually sway from dietary restrictions and are less compliant with diet and medicine than usual³².

Our study showed no evidence of seasonal variations in DKA admissions or DKA severity rates. Lu et al. reported more DKA episodes in winter than in other seasons in Taiwan²⁸. Bogar et al. studied the effects of ambient temperatures on DKA occurrence in the USA and showed more DKA episodes in cold temperatures compared to hot climates²⁹. A Canadian study of around 21 thousand hospitalizations examined the occurrence of DKA and hypoglycemia in patients with T1D and showed increased DKA admissions during winter seasons³³. Another study from Pakistan on 104 patients with DM also reported an increased frequency of DKA admissions in winter mostly due to noncompliance with treatment and infections³⁴. The possible rationale for an increased DKA incidence in colder temperatures is the effect of temperature on glucose regulation in the body, inflammation, thermoregulatory changes, differences in renal clearance of the drugs based on temperatures, variations in air pollution, and the holiday season in winter associated with dietary noncompliance^29,33. Our study had numerically higher rates of DKA episodes in the six consecutive winter seasons, especially in patients with T1D and newly diagnosed with DM. Whether viral infections are the trigger of seasonal variations in DKA remains unclear. While Lui et al. showed that the seasonal variation in DKA in Taiwan was independent of viral infections, Boagar et al. suggested a role for viral infections in USA^28,29. However, the estimated prevalence of the most common viral infection during the winter seasons-influenza is similar between Qatar (22.6%) and the USA (20.0%)^35,36. Furthermore, the extreme winter conditions in some countries might affect access to healthcare^37,38. On the contrary, healthcare in Qatar is accessible throughout the seasons.

The main strength of this study is its extensive examination of various aspects of DM concerning DKA in different Islamic and Roman months and seasons. Additionally, a relatively large sample size with an almost equal proportion of T1D and T2D increased the validity and generalizability of our results. The study’s generalizability is also enhanced by including a diverse range of ethnicities, making the findings more applicable to the Middle East region and beyond. The study used the ADA DKA diagnostic criteria to exclude patients with incorrect coding with DKA, resulting in a more accurate patient population. However, the study is limited by its retrospective nature, which might influence some of the results, such as anti-GAD Abs and C-peptide levels. This also limited our ability to validate the type of DM for each patient. Some of the factors that may influence the frequencies of DKA admissions, such as anti DM medication adjustments, compliance rates, among other factors could not be studied due to the retrospective nature of the study. It was not possible to identify individuals who were fasting; hence it was assumed that the patients admitted during Ramadan months were fasting, whereas others were non-fasting. This was assumed based on the cultural and religious obligations of the predominantly Muslim population of Qatar and also based on the results of studies showing that most patients with DM fast in Ramadan in this region. Also, there may be patients whose DKA diagnosis was not coded in the EMR, and thus, they were not included in the study. Although this could potentially impact the results, the number of such cases is expected to be small compared to the study’s size, thereby minimizing the effect on the results.

Future studies

Authors plan to publish studies from the 6-year index DKA cohort explaining the differential evolution of DKA based on the type and duration of diabetes, which will also include clinical and demographic data of the patients with T1D and T2D in more detail.

Conclusion

This is the largest regional cohort from Qatar that explored the occurrence of DKA in Ramadan compared to other months and seasons based on the type of DM, duration of DM, and severity of DKA. Evolving evidence suggests no difference in the incidence of DKA during Ramadan as well as in various seasons. Prospective studies on larger cohorts are needed to validate these results and to strengthen the guidelines for safe fasting in Ramadan by patients with T1D and T2D.

Transparency

Author contributions

Conceptualization and Methodology: FA, MB; Literature review: All authors; Manuscript writing: All authors; Critical Review and Revisions: FA, MB; Data Collection: FA, AK; Data analysis and Data interpretation: MB; Supervision: MB; Review and approval of the final manuscript: All authors.

Ethics statement

The study was approved by the Medical Research Centre (MRC) Qatar (MRC-01-21-476).

Acknowledgements

Open Access funding provided by the Qatar National Library. None.

Declaration of financial/other relationships

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Data availability statement

Available from the corresponding author upon reasonable request.

Additional information

Funding

The publication of this article was funded by the Qatar National Library.