Diabetic Ketoacidosis Treatment Outcome and Associated Factors Among Adult Patients Admitted to the Emergency Department and Medical Wards at King Abdulaziz Medical City, Riyadh, Saudi Arabia

Background Diabetic ketoacidosis (DKA) is a life-threatening condition with high morbidity and mortality rates. It should be diagnosed immediately and managed intensively to prevent its significant complications. Objectives The aim of this study to assess DKA treatment outcome and associated factors among adult patients at King Abdulaziz Medical City Emergency Department and Medical Wards, Riyadh, Saudi Arabia. Materials and Methods A retrospective cross-sectional study was conducted using a chart review to assess DKA treatment outcome and associated factors. All patients who were admitted as DKA cases from September 2017 to August 2019 were selected by simple random sampling except those with incomplete charts or younger than 14 years. Data were entered and analyzed using SAS Version 9.4 (SAS Institute, Cary, NC, USA). Results A total of 223 reviewed charts were collected. The frequency of DKA recurrence in most of the patients was once per year (126 [56.5%]). The most common precipitating factor was inappropriate insulin therapy (104 [46.64%]). More than half of the patients (120 [53.81%]) got out of DKA management protocol within 24-72 hours with a hospital stay of less than or equal to five days. The mortality rate was 1.83%. Patients with two or more DKA episodes per year tended to be admitted to ICU more frequently than those with one episode (p=0.001). It was found that patients who had a duration of one to five years of diabetes mellitus were almost five times more likely to get out of DKA in more than 72 hours when compared with those who had a duration of more than five years (adjusted OR: 4.7; 95% CI: 1.34-16.60; p=0.01). Conclusions The findings of this study highlight that majority of DKA patients showed improvement and discharged with a very low mortality rate. Inappropriate insulin therapy was the most common precipitating factor; thus, educating diabetic patients about the complications of treatment non-compliance is an important part of management.


Results
A total of 223 reviewed charts were collected. The frequency of DKA recurrence in most of the patients was once per year (126 [56.5%]). The most common precipitating factor was inappropriate insulin therapy (104 [46.64%]). More than half of the patients (120 [53.81%]) got out of DKA management protocol within 24-72 hours with a hospital stay of less than or equal to five days. The mortality rate was 1.83%. Patients with two or more DKA episodes per year tended to be admitted to ICU more frequently than those with one episode (p=0.001). It was found that patients who had a duration of one to five years of diabetes mellitus were almost five times more likely to get out of DKA in more than 72 hours when compared with those who had a duration of more than five years (adjusted OR: 4.7; 95% CI: 1.34-16.60; p=0.01).

Introduction
Diabetic ketoacidosis (DKA) is one of the life-threatening acute hyperglycemic complications of diabetes mellitus (DM) carrying high morbidity and mortality among type 1 diabetics and less commonly but recognized in type 2 diabetics. It is characterized by massive hyperglycemia, ketonemia, and acidosis [1]. It is most commonly precipitated by infections, non-compliance to insulin therapy, or first presentation of diabetes. Other factors include stressors such as cerebrovascular accidents, cardiac ischemia, trauma, and pancreatitis [1][2][3].
DKA can result in significant complications as volume depletion and severe acidosis may lead to cardiac arrest and acute kidney injury (AKI); thus, prompt diagnosis and initiation of intensive treatment protocol by experienced staff are important for successful management [3,4] followed by careful monitoring to prevent the occurrence of iatrogenic complications with insulin and fluid administration, such as hypoglycemia, hypokalemia, and cerebral edema. Appropriate early fluid therapy and insulin administration show a significantly better outcome and fewer recurrences. DKA can be successfully managed within 12-36 hours with appropriate treatment, with which the complications and mortality can be prevented and reduced effectively [5].
To address this, a few studies were conducted to draw the actual picture regarding the DKA outcome and their risk factors in Saudi Arabia [6,7]. Therefore, this study helps health care workers to build a uniform treatment protocol. The gained information will help also both patients and health care providers to know the most common reasons for DKA recurrence to prevent future episodes. Furthermore, the results of this study can be used as a database for future researches. Keeping the aforesaid facts in view, this study aims to assess DKA treatment outcome and associated factors among adult patients at King Abdulaziz Medical City (KAMC) Emergency Department and Medical Wards, Riyadh, Saudi Arabia.

Materials And Methods
This retrospective cross-sectional study used a chart review to assess DKA treatment outcome and associated factors among adult patients at KAMC Emergency Department and Medical Wards. KAMC is a tertiary hospital located in Riyadh, the capital city of Saudi Arabia. It has a bed capacity of 690 beds and provides all types of care to all National Guard soldiers and their families, starting from primary health care up to tertiary specialized care.
To determine the sample size, the following assumptions were considered in this study. According to a study was conducted in Ethiopia to assess the DKA outcome, 84.90% of the patients were discharged with improvement with 95% confidence level and 5% margin of error [8]. The sample size was calculated by using Raosoft software (http://www.raosoft.com/samplesize.html). Based on this, the final expected sample size was approximately 196 patients. All patients who were admitted as DKA cases from September 1, Data were presented as mean ± standard deviation. Frequencies and percentages were used to describe categorical variables. We used Fisher's exact test or chi-square test for association between categorical variables, and the Wilcoxon two-sample test, Kruskal-Wallis test (for not normally distributed data), t-test, or one-way analysis of variance (ANOVA) test (for normally distributed data) for continuous variables. We examined outcome "time within which patients get out of DKA" predictors using multivariate logistic regression. The regression model included age, type of DM, duration of DM, treatment regimen for DM, and comorbidities such as hypertension, dyslipidemia, and chronic kidney disease (CKD). All statistical tests were considered significant at p<0.05. Data were analyzed using the statistical program SAS Version 9.4 (SAS Institute, Cary, NC, USA).
Ethical approval was obtained from the Institutional Review Board at King Abdullah International Medical Research Center.

Comparison among admitted ICU patients and associated factors
Patients with a frequency of recurrent DKA two times or more tended to be admitted to ICU more frequently than those who had only one episode per year (p=0.001). In addition, patients who were admitted with Kussmaul sign had a propensity to be admitted to ICU than those who were not (p=0.005). In addition, those who showed altered mental status at the time of presentation were likely to be admitted to ICU (p=0.002). Regarding the vital signs, patients with an initial pulse rate of 113.42±18.68 beats/min tended to be admitted to the ICU when compared with those with an initial pulse rate of 102.65±18.95 beats/min (p≤0.0001). Regarding the initial lab results, patients who presented with an investigation lab result of anion gap calculation of 28.24±7.01 mmol/L (p-0.03), serum bicarbonate of 9.61±7.49 mEq/L (p≤0.0001), and WBC count 1of 6.72±6.02 x 10 9 /L (p≤0.0001) were more likely to be admitted to ICU than the other group (26.11±6.08 mmol/L, 14.6±5.4 mEq/L, and 11.67±5.98 x 10 9 /L, respectively). The time within which the patients got out of DKA was found to be longer (>24 hours) in ICU patients when compared with those who got a DKA free in less than 24 hours (p≤0.0001). Also, those patients who were admitted to the ICU appeared to have a prolonged hospital stay (>five days) when compared with those who only stayed for less than two days (p≤0.0001). Table 6 shows the details for the comparison of response among patients who were admitted to ICU.

Predictors of time within which the patient gets out of DKA
Different factors have been studied as determinants of time within which the patient gets out of DKA. It was found that patients who had a duration of one to five years of DM were almost five times more likely to get out of DKA in more than 72 hours when compared with those who had a duration of more than five years (adjusted OR: 4.7; 95% CI: 1.34-16.60; p=0.01). Patients who had a treatment regimen of both oral medication and insulin were likely four times to get free of DKA in more than 72 hours as compared with those only were on insulin (adjusted OR: 4.5; 95% CI: 0.97-21.15; p=0.05). Patients with CKD had a likelihood to get out of DKA in more than 72 hours by more than four times as compared with those who were not (adjusted OR: 4.0; 95% CI: 0.86-18.68; p=0.04). Table 7 shows the predictors of time within which the patients got out of DKA.  showed improvement and were discharged [8]. In this study, the mortality rate is 1.83%, which is relatively close to that reported in a study conducted at Chang Gung Memorial Hospital, where only 0.67% of the admitted patients with DKA showed no improvement and died [9]. The mortality rate in this study is considered relatively low as compared to other studies that were conducted in Zambia and Malaysia with a mortality rate of 16.66% and 17.6%, respectively [10,11]. Concerning the hospital stay, majority of the patients had a hospital stay of less than or equal to five days, which is considered less than a study that was conducted locally at King Fahad Medical City with an average hospital stay of seven days [12]. This difference in improvement, mortality, and length of hospital stay can be explained mainly by the advanced age, associated comorbidities of the patients, and the different treatment protocols that were provided by different medical institutions.
In this study, the frequency of DKA recurrence in majority of patients was found to be only one episode per year. This shows lower recurrence as compared with a study was conducted at Adama University Hospital, in which the recurrence was found to be two or more episodes per year in most (65%) of the patients [8]. It is well known worldwide that the most common precipitating factor for DKA is infection followed by inappropriate insulin therapy [2,8]. In contrast, the most common precipitating factor in our study is inappropriate insulin therapy (46.64%) due to poor compliance followed by infections (31.39%). Also, this was reported in other various studies that were conducted in Saudi Arabia, which showed a similar pattern in poor insulin compliance, which was the most common among patients with recurrent DKA admissions [12][13][14][15]. This variation in precipitating factors was attributed to the difference in population across the world as explained in an article review that was published by Kitabchi et al. [1]. Further researches are needed to find out the reason behind fewer recurrence of DKA that is precipitated by inappropriate insulin therapy as found in our study.
Concerning the clinical features of DKA patients, abdominal pain (69%) was reported among the patients in a study published in 2015 at KAMC [15]. Vomiting (61.6%) was reported in another local study that was conducted among DKA adults in a tertiary hospital in Riyadh, Saudi Arabia [14]. Polyurea (26.3%) and polydipsia (28.2%) were found as initial clinical presentations among patients. In addition, dehydration signs including hypotension and tachycardia were reported in 61.8% of the patients in a study conducted locally at King Fahad Medical City [12]. These clinical features are consistent with our results.
Nearly three-quarters of the patients who were admitted with DKA did not develop complications associated with the disease. Majority (13.45%) of our patients were complicated with AKI, which is considered low as compared to a study conducted in Bangladesh showing that 29.5% of their patients developed AKI [16]. This variation could be explained by the severity of dehydration and how early fluid replacement is initiated. We found that 3.59% of our patients were complicated with hypokalemia, which is consistent with various studies reporting the incidence of hypokalemia in DKA patients to range from 0% to 4% during various stages of DKA management [17,18]. This is explained by insulin administration and correction of acidosis and hyperosmolarity that drive potassium intracellularly, resulting in hypokalemia; thus, careful monitoring of potassium is an important aspect of management [19].
Most of the DKA patients in our study were treated in the general medical wards (68.16%), which is consistent with a study conducted at Auckland Hospital in which almost 70% did not need to be admitted to ICU [20]. Similarly, other studies showed that more than half of their DKA patients were treated in the general medical wards [21,22]. Similar to a Libyan study, we found that patients who were admitted to ICU either due to severity of the episode, development of acute complications, or having other comorbidities stayed significantly longer in hospital (>five days) when compared to those who were treated in the general medical wards (<two days) [21].
With appropriate treatment of DKA, patients are expected to be out of DKA, which is defined as glucose less than 200 mg/dL and at least two of the following: venous pH over 7.30, serum bicarbonate more than or equal to 15 mEq/L, or anion gap less than or equal to 12 mEq/L within 48 hours [23,5]. In KAMC, the treatment protocol is generally according to the recent American Diabetic Association guidelines for DKA management [24]. The majority of DKA patients in our study got out of DKA within 24-72 hours (53.81%) or less than 24 hours (34.08%). We found that patients who were already diagnosed with DM and being treated with both oral and insulin were more likely to be out of DKA in more than 72 hours when compared to those who were on insulin alone. This can be explained by the concept that patients who are on both oral and insulin therapy are likely to be type 2 diabetics with advanced or uncontrolled disease and thus higher rates of complications and comorbidities. We could not find studies that support this result; thus, further studies are needed to identify factors that play a role in a prolonged time to be out of DKA.
Each study has its limitations. A limitation of our study is that we did not have much data regarding the treatment protocol in detail due to a poor chart review in this part. Therefore, we could not assess the management and its associated factors. Another limitation is that our study was only limited to one hospital (KAMC) and not generalized to other hospitals in the Riyadh region.

Conclusions
The findings of this study highlight that majority of DKA patients showed improvement and were discharged with a very low mortality rate. Most of the patients received their medical treatment in medical wards, with a hospital stay of less than or equal to five days. Our data suggest that the time within which the patient got out of DKA is 24-72 hours. The frequency of DKA recurrence was found to be only one episode per year and precipitated commonly by poor insulin therapy and infections. Further studies are needed to better understand the management and its associated factors.