Methodology/ Data Collection: After preliminary evaluation and management in the pediatric intensive care unit of this hospital, the detailed assessment of all patients was done and recorded in a pre-designed proforma after obtaining written informed consent from their parents. The proforma contained information on patient's gender, age, area of residence, socioeconomic status of the family according to modified Kuppuswamy scale, Body Mass Index (B.M.I.), level of consciousness, time of admission, duration of symptoms, family history of diabetes, consanguinity, significant presenting signs, symptoms. An attempt to detect the precipitating events were made in all children. The presence of infection/ intercurrent illness as indicated by a positive radiological imaging study or blood culture. This was supported by an elevated white blood cell count and clinical examination by the physician. The measure of compliance regarding insulin was based on the history given by the attendants of the patients. Insulin omission was defined as missing insulin injections on multiple days, especially immediately before or during the period of illness.

Detailed physical examination, including the vitals, anthropometry, and systems examination, was carried out. Essential laboratory parameters done on admission included blood glucose, urine ketone level by dipstick method, arterial/venous blood gas, sodium, potassium, calcium, complete blood counts, blood urea, serum creatinine, chest radiograph and an electrocardiogram. Urine examination was done for routine analysis and for detecting ketone body. C- reactive protein (C.R.P.), blood culture and sensitivity, urine culture, and sensitivity were sent to patients with suspected sepsis. HbA1c was done in all children to look for long-term glycaemic status. Complications including cerebral edema, cardiac arrhythmia, hypoglycemia, hypokalaemia, hypernatremia, infection and renal failure were recorded.

Time duration required for resolution of DKA and insulin infusion duration were recorded. The outcome in the form of survival and death were noted. Resolution of DKA was considered when the consciousness was normal, no vomiting, pH more than 7.3 and serum bicarbonate level more than 15.

DKA is defined as the presence of hyperglycemia (blood glucose >200mg/dL) with a venous pH <7.3 and bicarbonate <15mmol/L with associated

Table 1. Association between severity of DKA with the demographic profile of pediatric patients with DKA.

The most common presenting symptoms were nausea/vomiting in 42 (77.77%), pain abdomen in 36 (66.66%), followed by fever in 35 (64.8%),

by hyponatremia and hypokalaemia in 9 (16.6%), A.K.I. in 8 (14.8%), cerebral edema in 6 (11.1%) and 2 (3.7%) cases had hypernatremia. The mortality rate was 7.4%.

Table 3. Correlation between different clinical, biochemical and socio-demographic parameters in survivor's vs deaths (multivariate logistic regression analysis) 

We observed that only 3 (5.55%) patients had a family history of Type 2 D.M. Similar findings were also noticed by Ababulgu RZ et al. (2020), who found a family history of D.M. in only 7(11.1%) patients [18]. However, Satti AS et al. (2013) reported a family history of diabetes (either type 1 or 2) in 59 (74%) cases which is significantly higher than in our study [19].

In our study, 32 (59.25%) of the DKA cases were already diagnosed with type 1 D.M., while 22 (40.75%) patients were newly diagnosed as type 1 D.M. on admission. Similarly, Bhardwaj P et al. (2017) found that 48.2% were newly diagnosed and 51.8% were previously diagnosed cases of diabetes [20]. But Basavanthapa et al. (2015), in their study, reported that almost 80.7% of patients were newly diagnosed type 1 D.M. [15].

We found that out of 54 cases, 25 (46%) cases presented with severe DKA, 21 (39%) patients with moderate DKA and 8 (15%) cases with mild DKA. Rashid supported this I et al. (2019), where most of the patients had severe DKA [17], whereas Kumar MV et al. (2017) reported 81% moderate DKA cases [21].

Most of the DKA cases, 34 (77.77%) presented to us with nausea/vomiting followed by pain abdomen in 36 (66.66%), fever in 35 (64.8%), weakness in 27 (50%), polyuria in 18 (33.33%), polydipsia in 16 (30%) and headache in 15 (27.8%). This symptomatology was similar to the experience by Satti SA et al., were vomiting and abdominal pain were the common presenting symptoms [19]. However, a study by Kanwal SK et al. (2012) reported polyuria, polydipsia as the commonest presenting features followed by vomiting, altered sensorium and abdominal pain [15].

Dehydration was the primary presenting sign in 83.3% cases, followed by Kussmaul breathing in 72.2% cases, altered sensorium in 50%, tachycardia was found in 24% cases, shock in 20.3% and 5.55% cases were comatose. This was comparable to the study by Neu A et al. (2003), where almost 53% had altered levels of consciousness, with 10.9% of them being unconscious [22]. Islam R et al. (2014) also found that Kussmaul's breathing and dehydration were the commonest clinical feature of DKA [23]. This is unlike the experience from Saudi Arabia, by Satti SA et al. (2013), where only

Varshney GA et al. (2015) they reported the median time for the arterial blood gases to become normal was 26 hrs. The average length of the hospital was 7.8 days [26].

In our study, among the demographic parameters, gender, area of residence, socioeconomic status, and B.M.I. of the patient were significantly associated with the severity of DKA. (P-value < 0.05 for each). Similarly, among clinical and biochemical parameters, presence of infection, insulin omission, DKA 1st episode, presence of diarrhoea, presence of shock, poor G.C.S. on admission, fever on admission, and time required for resolution of Diabetic ketoacidosis was significantly associated with severity of DKA. (P-value <0.05 for each). This was supported by Syed M et al. (2011), they observed that severity of diabetic ketoacidosis was significantly associated with the presence of infection, history of omission of insulin, poor compliance, presence of shock at the time of presentation, length of stay in the hospital, outcome (p < 0.01 for each of these associations) and Glasgow Coma Scale score (p=0.02) [10].

In multivariate logistic regression analysis done between survivors and deaths using different parameters in our study suggest that likelihood of death was significantly higher among the patients who had poor G.C.S. on admission <8 (OR=34.5, P=0.05), cerebral edema (OR=11.5, p=0.03), age <5years (OR=6.09, p=0.015), hyponatremia (serum sodium <130) (OR=4.14, p=0.048) and requirement of insulin infusion >72hrs (OR=4.04, p=0.017). This was in concordance with a study by Syed M et al. (2011), they observed the presence of cerebral edema, need for mechanical ventilation, low socioeconomic status, low serum sodium level, low serum potassium level, low arterial blood pH, presence of infection and stay in the pediatric intensive care unit were factors that significantly impacted the outcome of the patient [10]. Hence, children with these characteristics should be more intensively monitored.

Out of 54 cases admitted with DKA, 50 improved and were discharged, 4 (7.4%) patients died during treatment. This was comparable with the reported mortality rate from developing countries like India, Pakistan, and Bangladesh (3.4% to 13.4%) [9-12].