Clinico-Epidemiological Profile of Diabetic Ketoacidosis in Hospitalised Children at a Tertiary Health Care Centre

Introduction: Diabetic Ketoacidosis (DKA) presents with spectrum of clinical manifestations and awareness regarding this amongst physicians is crucial. The study aimed to chronicle the clinical profile of DKA in type 1 diabetes mellitus (T1DM) children. Methods: The study was conducted at a tertiary care hospital including patients with signs and symptoms of DKA with either debut or established T1DM (n = 38). Detailed clinical history, examination and laboratory investigations were carried out. Differences in frequency distribution concerning demographics and clinical data were analysed in R-studio software (v.1.2.5001). Results: Kussmaul breathing (49.97%) was the commonest presentation. Infection (39.46%) and poor compliance due to inadvertent omission of insulin therapy (50%) were the commonest precipitating factors. Mild, moderate and severe dehydration was present in 44.74%, 39.47% and 15.79% respectively. Forty-five percent patients had moderate while 34.21% and 21.05% had mild and severe DKA respectively. Seventy-five percent and 52% patients were below 3rd percentile for height and weight respectively. High number of patients had past history of viral infections, were diagnosed in winter, belonged to 2nd or 3rd birth order and from lower middle class. Patients with poor compliance to insulin belonged to lower middle (40%), upper lower (53.33%) and lower class (6.67%). Conclusions: DKA can be diagnosed early by identifying dehydration, Kussmaul breathing, polyuria and altered sensorium. Infection and poor compliance due to omission of insulin therapy were the commonest and preventable precipitating factor for DKA. Poor compliance to insulin can be attributed to lower socioeconomic class. Awareness among physicians concerning clinical profile of DKA is crucial.


INTRODUCTION
Type 1 diabetes mellitus (T1DM) is among the common auto-immune, endocrinal disorder in children caused due to destruction of β cell. 1 T1DM mostly presents as diabetic ketoacidosis (DKA) and is the commonest paediatric endocrine emergency. 2 The prevalence of DKA at diagnosis of T1DM and in established T1DM varies widely and ranges from 13 -80% and from six to 28% respectively in different countries. 3 The clinical signs of DKA includes dehydration, tachycardia, tachypnoea, deep sighing respiration, breath smell of acetone, nausea and / or vomiting, abdominal pain, blurry vision, confusion, drowsiness, progressive decrease in level of consciousness and eventually, loss of consciousness (coma) etc. 4 The primary risk factors for DKA is especially more in poorer countries due to delayed diagnosis, lower prevalence, poor compliance, limited access to medical services etc. 4 DKA mimics infections and other acute medical conditions due to its varied clinical presentation. Therefore, it is associated with high mortality if misdiagnosed and mismanaged. 5 Early diagnosis and management of DKA prevents short-term and long-term complications. Therefore, awareness on the part of the physician accompanied by accessibility and availability of primary health-care facilities widely contributes to better management. 6 Additionally, there is overwhelming evidence that lack of awareness regarding symptoms, signs, and clinical profile of DKA in T1DM among individuals and physicians have resulted in increased mortality in developing countries. 7 Studies that deal exclusively with clinical profile of DKA in children of the Indian subcontinent are rather sparse. This study aimed to evaluate the clinico-epidemiological profile of DKA at presentation in hospitalised children with T1DM at a tertiary health care centre in Western part of India.

METHODS
The prospective study was conducted at the Department of Paediatrics at a tertiary health care centre in Kolhapur, Maharashtra, Western India between January 2018 to December 2019. The study was approved by Institutional Ethics committee. Informed and written consent from parents / guardian was obtained. Sample size was calculated using standard formula [n = Z 2 P(1-P)/d 2 to estimate proportion considering level of confidence (Z) = 1.96, expected prevalence or proportion (P) = 0.02 and level of precision (d) = 0.05. Thirty-eight children below 18 years of age with either debut or established T1DM admitted in the PICU with signs and symptoms and laboratory parameters of DKA were included. The demographic data (age, sex, height, weight), clinical data (presenting complaints, duration of symptoms, associated symptoms and precipitating factors for DKA episodes) and other clinical history including history of past infections, season at which they were first diagnosed, family diabetic history, birth weight in kilograms (High > 2.7, Normal 2.5 -2.7, Low < 2.5), birth order, diet history specifically feeding in the first six months were collected. Details of Socioeconomic status as per modified Kuppuswamy's scale were also collected. 8 Degree of dehydration was assessed clinically as mild (5%), moderate (5 -10%) and severe (> 10%). 4 Growth parameters like height and weight of the children were assessed according to World Health Organisation 2006 and Indian Academy of Paediatrics 2015 combined height and weight growth charts for 0 to 18 years. 9,10 In children with established T1DM, additional data regarding duration of diabetes and compliance for insulin therapy either good or poor were collected. The poor compliance for insulin therapy was categorised as omission of insulin (either inadvertently or deliberately due to psychosocial issue) or due to limited access to medical services.
Along with detailed clinical history and examination, laboratory investigations including total blood count, serum and urine sugar, venous blood gases (VBG), serum electrolytes, urine ketone bodies, urine culture and blood cultures were carried out to confirm and manage DKA. Radiological investigations like x-ray chest and USG abdomen were done wherever required. DKA was defined as the presence of blood sugar > 200 mg/dl with a venous pH < 7.3 and or bicarbonate < 15 mmol/L with associated glycosuria, ketonuria and ketonemia. 5  mmol/L and Severe: pH < 7.1, serum bicarbonate < 5 mmol/L). 4 All the patients diagnosed with DKA were managed as per International Society for Paediatric and Adolescent Diabetes (ISPAD) clinical practice consensus guidelines 2018 protocol. 4 During treatment other possible complications like cerebral oedema, electrolyte abnormalities and acute renal failure, to name few were also evaluated and managed accordingly. After successful management of acute DKA episode, at the time of discharge all patients were advised to undergo treatment with basal bolus Insulin therapy with appropriate counselling and regular follow-up.
The data collected was organised in MS Excel (2016). Frequency distribution concerning demographics, history and clinical data were studied. Data is represented as frequency, percentage and mean ± SD where relevant. Differences in proportion of population based on various factors were analysed in R studio software (v.1.2.5001).

RESULTS
Of the total 38 patients, 21.1% (n = 8) were debut and 78.9% (n = 30) were established DKA. The age of the patients ranged between five months to 12 years with the mean age of 7.4 ± 3.7 years. High number of patients (52.6%; n = 20) were diagnosed within first four years of life followed by 34.2% (n = 13) and 13.2% (n = 5) at the age of 4 -8 years and 8 -12 years respectively. The gender distribution was balanced with 47% (n = 18) males and 53% (n = 20) females.
Clinical data of the patients is represented in Among 15 patients with poor compliance, 40% (n = 6) patients belonged to lower middle class family, 53.33% (n = 8) and 6.67% (n = 1) patients belonged to upper lower class and lower class family respectively. Assessment of growth concerning height and weight revealed that 75% (n = 25) and 52% (n = 20) children were below 3 rd percentile for height and weight respectively.
Among the patients with established T1DM (n = 30), 43.33% (n = 13) patients had diabetic history of up to two years followed by 26.66% (n = 8), 23.33% (n = 7) and 6.66% (n = 2) patients with the diabetic history of two to four years, four to six years and six to eight years respectively. Forty-five percent of patients had moderate DKA (n = 17) followed by 34.21% (n = 13) with mild DKA and 21.05% (n = 8) with severe DKA.
Leucocytosis was observed in 73.68% patients (n = 28). Hyponatremia and hypokalaemia were seen in 78.95% (n = 30) and 65.79% (n = 25) patients respectively. Table 3 represents the other laboratory profiles of DKA patients. The average duration of stay in hospital was 8.23 ± 5.2 days. There was no mortality in DKA patients.  In this study, 38 children below the age of 18 years were studied and the mean age of the patients was 7.4 ± 3.7 years with male:female ratio of 0.9:1. This is similar to the previous study conducted by Al-Eisa A et al. 12 Osmotic symptoms including polyuria, polydipsia, polyphagia, vomiting and fatigue are usually observed in initial stages of DKA. With the progression of acidosis, the pattern of breathing become more fast, deep and agonal, called Kussmaul breathing. 13 The commonest clinical presentation along with dehydration reported in our study were Kussmaul breathing, vomiting, fever, polyuria and altered sensorium.

DISCUSSION
Other studies conducted in other parts of India also showed vomiting, polyuria and polydipsia as the commonest symptoms. 14 DKA is characterised by severe depletion of water and electrolytes from both intra and extracellular fluid compartments. 15 We observed dehydration in all the patients of which mild, moderate and severe dehydration was present in 44.74%, 39.47% and 15. Management of T1DM includes monitoring of blood glucose levels and life-long insulin therapy. However, poor compliance due to inadvertent omission with insulin therapy among 50% patients in established T1DM can solely be attributed to either financial constrain or lack of education and parental knowledge. There was not a single patient who omitted insulin deliberately as a psychological cause or due to limited access to medical services. In the study conducted by Del Pozo P et al., poor compliance was observed in 66.6% patients. 2 I n s u l i n o m i s s i o n c a n b e p r e v e n t e d b y comprehensive programs that provide education, psychosocial evaluation and treatment combined with adult supervision of the entire process of insulin administration. 17 The infections, febrile illnesses, trauma change the body's metabolism and homeostasis resulting in drop in insulin level and increased glucagon and adrenaline levels. The liver also tends to release more glucose ensuring that sufficient energy is available. Concurrently, growth hormone and cortisol levels rise, reducing the sensitivity of tissues towards insulin and increasing glucose level. 18 This triggers episodes of DKA. We observed UTI as the commonest trigger. Several studies in the past also have reported infections as the commonest precipitants. 19,20 Certain viruses like Coxsackie virus B, CMV, adenovirus, mumps, rubella have been reported to target exocrine pancreas and are known to be associated with T1DM. 21,22 This is reflected in this study too as 36.8% had the history of viral fever. However, the mechanism behind development of autoantibodies due to viremia is not clear. We also noted significantly high number of patients (94%) being diagnosed in the winter. This was also reported Baruah MP et al. who attributed this to sudden environmental changes and the pattern of eating. 23 In our study, 29% had parental history of diabetes. The risk of developing diabetes increases if both the parents are diabetic. 24 Paternal diabetes is found to show a stronger association with diabetes in the offspring than maternal. This is also reflected in this study where 26.3% of patients gave a history of paternal diabetes while only 2.6% patients gave a history of maternal diabetes. The association of certain combination of HLA with significantly higher risk of developing T1DM have been reported, especially the combinations of HLA-DQA1, HLA-DQB1, and HLA-DRB1 gene variations. This combination of HLA genes seems to escalate the risk of beta cell destruction. 25 High birth weight and rapid weight gain in initial years of life are associated with risk of T1DM. 26,27 We also found a weak association between birth weight and the risk of diabetes as 76% T1DM children had normal birth weight. Stene LC et al. had also reported similar findings. 28 Effect of birth order on risk of T1DM is still a topic of debate. The finding of this study reflects that the children born later are less prone to T1DM but further research in this arena is required. Breast milk is thought to be the best suited diet during the initial days of life due to the presence of immunoglobulins. The majority of the patients in this study were exclusively breast milk fed which reflects that onset of T1DM was due to other factors.
Growth is usually influenced by duration of diabetes, metabolic control, genetic endowment, puberty, and growth hormone level including IGF binding proteins (IGFBPs) and insulin-like growth factors (IGFs). 29 We also noticed that 75% and 52% children were below 3 rd percentile for height and weight respectively. This study attempts to elaborate on the clinical profile of DKA in children. Although it is single centric, small study, it is hoped that this study would provide insight into the clinical profile of DKA children and stimulate further multi-centric, larger study in the future.

CONCLUSIONS
DKA can be diagnosed early by identifying certain clinical features like dehydration, Kussmaul breathing, vomiting, fever, polyurea and altered sensorium. Infection and poor compliance due to omission of insulin therapy were the commonest precipitating factors for DKA in Indian children.