Predictors of Diabetic Ketoacidosis among Patients with Type 1 Diabetes Mellitus Seen in the Emergency Unit

Aims: To identify the factors that can predict diagnosis of T1DM, presence of acute recent illness, missing insulin doses and frequency of the dose missing, number of previous DKA attacks, using of syringes or pens as a tool of insulin delivery, however, the most important two predictors are the source of insulin supply to patients, presence of glucometer at homes and frequency of its uses. Conclusion: This result provided evidence that multiple factors interact together to play a vital role in the development of DKA among patients with T1DM in Basrah.


INTRODUCTION
Diabetic ketoacidosis (DKA) is a serious acute complication of diabetes mellitus. According to American Diabetes Association, the mortality rate in patients with DKA is reported to be less than 5% [1]. However, it is reported that there has been a 4.2% decrease in the death rate per year in the United States during 1985-2002 [2]. Mortality in patients developing DKA is predominantly due to the underlying morbidities such as sepsis or myocardial infarction [3].
The most common precipitating factor in the development of DKA is missed insulin doses [4]. Other factors that may precipitate the incidence of DKA are infections, cerebrovascular accident, alcohol abuse, pancreatitis, myocardial infarction, trauma and drugs such as corticosteroids, thiazides, and sympathomimetic agents.
In young patients with type 1 diabetes mellitus (T1DM), psychological problems with eating disorders may be a contributing factor for 20% of recurrent DKA [5].
The majority of DKA cases occurs in patients with previously diagnosed diabetes [5], and only a small proportion of patient accounts for the majority of hospitalizations [6]. Thus, it is estimated that 50% of hospital admissions could be prevented by improved outpatient treatment and better adherence to self-care [7].
Recent studies found higher glycosylated hemoglobin (HbA 1 C), female sex, older age, family and school problems, higher insulin dose, underinsurance, and psychiatric disorders to be associated with an increase in DKA rate [5,8,9].
The early identification and treatment of patients at risk is essential to be targeted in intervention programs aimed at preventing DKA [10].
Despite major advances in the care of diabetes, DKA remains a leading cause of hospitalization and the leading cause of morbidity and death in children and adolescents with T1DM.
It is important to try to prevent DKA in order to reduce morbidity and mortality associated with severe metabolic decompensation [11]. This prevention can be accomplished through appropriate education, improved self-care and adherence, and consistent self-monitoring of blood glucose and ketones [12]. DKA as an initial manifestation of T1DM is less amenable to prevention [13], other than through surveillance in youth with a positive family history of diabetes, and increased public awareness of the symptoms of diabetes.
The study aimed to identify the factors that can predict the development of DKA among patients with T1DM in Basrah.

Design and Data Collection
A cross-sectional case-control study was carried out to analyze the predictors of DKA among patients with type 1 diabetes mellitus seen in the emergency unit that involved four hundred patients with type1 diabetes mellitus seen in the emergency units of Al-Faiha and Al-Basrah General Hospitals in Basrah, Southern Iraq during a one-year period (June 2013-June 2014).
The ethical committee of the Basrah College of Medicine approved the study.

Measurements
All patients are sometimes their relatives (parents, brothers, sisters) were subjected to specific questionnaires already prepared for this study. Written informed consent was taken from the patients or their relative if they age under 19 years.
Patients were selected according to inclusion and exclusion criteria; the inclusion criteria for the patients were: patients who were 15 years (visiting adult emergency unit) and above and diagnosed with type 1 diabetes mellitus.
Patients who were not interested in questionnaires and those who's their first diagnosis of type 1 diabetes mellitus were as DKA during the carrying of the study in the emergency units were excluded.
T1DM was defined as insulin-dependent less than six months from diagnosis [14].
DKA was defined as a tetrad of blood glucose>200 mg/dl (11 mmol/L), ketonemia and ketonuria, venous PH<7.3 and or bicarbonate<15 mmol/L [15], in addition to the clinical features of DKA in the patients. In this study arterial, blood gas analyzer {IRMA (TRU POINT)} was used to check the PH and bicarbonate level.

Definition of Variables
Variables were defined as: • The address of the patients: all the centered areas in Basrah are considered as urban while all the peripheral areas are considered as rural. • In regard to the insulin starting after diagnosis of T1DM: Patients who started prescribing insulin therapy on the same day of T1DM diagnosis by the doctor are considered as immediate insulin users otherwise they are considered nonimmediate insulin users. • In regard to the regimen of insulin: in the study any patient used prandial regular human insulin three times daily with additional doses of basal insulin (including NPH insulin) at bed time is considered as basal-bolus insulin user otherwise they are considered non basal bolus insulin users. • In the study, patients who use syringes for insulin delivery are considered as insulin syringe users otherwise they are considered non-insulin syringes users including those using insulin pens or using both syringes and insulin pens for insulin delivery. • What is considered as "recent acute illness" in this study is that any patient who had an illness of belongs to or occurring at a time immediately before the present, and his disease is of rapid onset and or a short course. • Regarding missed insulin dose: any doses of daily insulin missed intentionally or unintentionally were considered as missing an insulin dose. • Previous DKA in this study, is that any patient who had DKA diagnosis documents by a doctor with hospital admission and discharging card.

Statistical Analysis
Data obtained from the questionnaires was analyzed using the Statistical Package for Social Sciences (SPSS) software version 15.0. Data are presented as a mean± standard deviation in case of continuous variables and as absolute numbers (percentage) in case of dichotomous variables with statistical analysis was carried out using Student's t test for continuous variables and Chisquare test for dichotomous variables.
The (p-value) less than 0.05 was considered significant. Univariate analysis was performed to determine each variable's association with diabetic ketoacidosis. In 400 patients, univariate analysis was performed for each of the factors in the presence or absence of diabetic ketoacidosis after the presence of diabetic ketoacidosis was used as the dependent variable. Variables found to be associated with diabetic ketoacidosis in univariate analysis were then entered into a multivariable model using logistic regression to determine the power of each variable for associated with diabetic ketoacidosis.

RESULTS
The baseline characteristics of all patients are given in (Table 1)  Regarding the address of the patients included in this study, patients were divided into those living in rural areas 183 (45.8%) and those who live in urban areas 217 (54.2%).
According to starting on insulin therapy after diagnosis of T1DM, patients subdivided into 2 groups: group1 those who start insulin immediately after diagnosis and group2 included patients who not start insulin immediately and were found that their number and percentage 213 (53.3%), 187 (46.7%) respectively.
The regimen of the used insulin subdivided the studied patients into 2 groups: Group 1 those who used the regimen of basal-bolus insulin186 (46.5%), group 2 those who used non basalbolus insulin regimen (Premix insulin 70:30 twice daily or modified basal-bolus) were 214 patients (53.5 %).
There were 214 (53.5%) of the studied patients were having a recent acute illness, whereas 186 (46.5%) did not have.
According to the tool used for insulin delivery, the studied patients were subdivided into: Those who used syringes only to deliver insulin 186 (46.5%), those who not used syringes only (using pens or both pens and syringes) 214 (60.3%).
According to the presence of insulin missing doses; the studied patients were subdivided into two groups: Group 1 those who had missed insulin dose 166 (41.5%), and group 2 were those who not missed insulin dose 234 (58.5%), and according to the frequency of missed insulin doses; patients were subdivided into 2 groups also: Group 1 those who had one missed insulin dose 101 (60.8%) and group 2 were those who had multiple missed insulin doses 65 (39.2%).
According to the availability of glucometer at home and frequency of its use for monitoring blood sugar, we subdivided the participated patients into 2 groups: Group   So variables associated with DKA with statistical significance were all the studied variables. Predictors of DKA identified by multivariate analysis are shown in Table 3.
Variables that remain statistically significantly associated with diabetic ketoacidosis were: age, gender, duration of type1 diabetes mellitus, the address of the patients, insulin regimen, tool of insulin administration, presence of acute recent illness and previous diabetic ketoacidosis attacks and its number, missing of insulin and its frequency, frequency of glucometer usage at home to check blood sugar and the source of insulin supply to the patients with type1 diabetes mellitus.

DISCUSSION
The prevalence of T1DM is increasing worldwide [16]. DKA was universally fatal, especially before the days of insulin in the 1920s, after that, the overall mortality decreased to relatively low figures, but there is still the potential of case fatality from DKA either from acidosis or as a complication of therapy [17,18]. There was a predominance of females with DKA in our study that is similar to the universal findings [19][20]. This could be attributed to the fact that insulin purging is frequent to control weight in young women with diabetes [21,22]. Furthermore, adolescent girls with diabetes are at greater risk of developing disordered eating patterns [23]. Adolescent girls and young women are also the group at highest risk to suffer from 'brittle' diabetes, a form of severe unstable diabetes with increased metabolic complications, especially recurrent DKA [24].
In accordance with previous studies [17,20], teenage at diabetes onset was related to elevated incidence of DKA. On one hand, greater personal responsibility diabetes management and less parental monitoring adolescents may lead to a deterioration of metabolic control in this group. On the other hand, endocrine changes associated with puberty lead to greater insulin resistance [24]. Comparable reasons may account for the trend toward a higher incidence of DKA in patients with longer duration of diabetes.
8 The Low educational level of the patients and their parents had significantly associated with the development of DKA in our study, and these results were also demonstrated in other universal similar studies [20,25].
As expected and supported by similar studies in neighboring countries [26][27][28], the patient address significantly associated with DKA occurrence and this is clearly related to the educational level of the patients and their families, their economic status and their access to the medication supply and specialized diabetic centers.
Delayed initiation of prompt insulin therapy following a diagnosis of T1DM has been associated with increased risk of DKA development among patients with T1DM [29]; the same result was found in this study also.
It is found that the initiation of the suitable and the effective insulin regime was associated with reduced DKA rate [29,30], as found in this study that the lowest DKA was among those patients using the prandial regular human insulin before each of the three-day major meals with further dose of basal insulin at bedtime (basal-bolus insulin regimen).
To our knowledge, the effect of how insulin delivered on DKA rate has not been investigated before. In our study, it has been found that using of insulin pens had been associated with increased rate of DKA when compared with the using of the insulin syringes; this is perhaps related to the ignorance of the proper technique for the use of these pens especially among those patients with low educational levels, also in Iraq we use vial and syringes for all, except few used reusable pen ,and with unavailability of insulin cartridge make patients refill them by syringes that definitely end with dosing errors because manufacturer recommendation was against the rifling of insulin cartridge.
The recent acute illness and mainly infections are important DKA precipitators in 109 patients. Most patients presented with nausea, vomiting and abdominal pain, which was interpreted as an indication to reduce or stop their insulin. This is a deep-rooted belief shared by some doctors and diabetic educators. The situation is made worse when such patients consume large amounts of sugar-rich fluid to counteract presumed hypoglycemia [31].
Insulin omission intentionally and unintentionally was found in patients with DKA in the present study. Among patients with poor compliance with insulin therapy, patients stopped insulin therapy either because they are away from insulin supply (lost, broken, etc.), not gave clear reason for stopping insulin or did not know how to handle insulin on sick day. Insulin discontinuation has long been recognized as an important precipitating cause of DKA in retrospective studies [32].
Observational studies in urban African Americans have reported that more than one-half of DKA cases in patients with diabetes were caused by noncompliance with insulin therapy [32,33]. Likewise, a retrospective study in a multiethnic population in Texas listed noncompliance with insulin injections as their most common precipitating cause of DKA [34].
Higher insulin doses were observed in the DKA group, which corresponds to the results of Rewers et al. [35]. They hypothesized that a higher reported insulin dose may represent lower endogenous insulin secretion because of longer duration of diabetes because of puberty, but it could be that in patients who miss insulin injections, a higher prescribed insulin dose may not be consistent with the actual applied insulin dose, and that what is found in this study with high number of DKA among patients with multiple missed insulin doses than those who missed a single dose.
Previous DKA and increase number of previous DKA were associated with increased DKA rate [36] and this, what is found in our study.
The goal of glucose monitoring in diabetes is to obtain useful information about the patient's overall glucose status to normalize glucose and prevent hypoglycemia and minimize hyperglycemia through meaningful and timely interventions. As glucose control is the foundation of diabetic care, self-monitoring of blood glucose (SMBG) is the foundation of glucose monitoring. Studies have shown a direct correlation between SMBG and improve HbA1C levels and reduce DKA rate [37][38][39], the American Diabetes Association recommends that patient with T1DM should self-test at least three times daily. This is identical to what is found in our study of reducing the number of DKA among those who frequently used SMBG (7 patients with DKA out of a total of 72) compared with (16 patient without DKA out of 36) of those who used SMBG infrequently or not used it respectively.
Finally, to our knowledge, the effect of source of insulin supply on DKA rate has not been investigated before. In our study, we found that the source of insulin was significantly associated with the development of DKA, and patients who depend on FDEMC as a source of insulin supply were significantly had lower DKA compared with those who took insulin from the public clinic. This is may be related to the advantage of more experienced staff and multidisciplinary treatment as well as the availability of follow-up investigations and better communication with patients and the dealing with all their aspects.
Study limitations: It was a cross-sectional study; so no data available on mortality. The age of onset of T1DM was not studied, but its play a role in DKA development. The incomes of the family also play a role and was not assessed in this study. The puberty stage was not assessed in the studied patients since puberty plays an important role in the loss of control of diabetes during the transition period [40].

CONCLUSION
The results of this study provided evidence that multiple factors interact together to play a vital role in the development of DKA among patients with T1DM in Basrah.

FUNDING STATEMENT
This research received no specific grant from any funding agency in the public, commercial or notfor-profit sectors.