Introduction

Diabetic ketoacidosis (DKA) is a common, dangerous and costly acute complication in patients who suffer from Diabetes Mellitus (DM). Despite DKA being less frequent than severe hypoglycemia, it has a higher mortality rate and is the leading cause of death related to DM in pediatric patients [1].

DKA is defined biochemically, by glycemia levels higher than 250 mg/dl, with positive ketones in urine and blood, metabolic acidosis with pH equal or lower than 7.30, increased anion gap (equal or higher than 10), and decreased plasmatic bicarbonate (equal or lower than 18 mEq/l) [2].

Around the world, DKA hospitalizations have increased during the last decades, according to epidemiological studies [3,4].

DKA, as an acute complication of DM, requires urgent and expert management based on guidelines [5]. Management of DKA is considerably expensive. For example, in the United Kingdom, the cost to treat a single episode of DKA is estimated to be GBP 1,387 (around US$ 1,750) [6].

Despite DKA being the most severe acute complication of DM, this study had not been previously conducted in the Emergency Department (ED) of Thames Hospital, nor in other ED in New Zealand. In this study, we will describe the profile of the patient who attends the ED of Thames Hospital, with DKA.

Objectives

The main objective is to establish the social demographic data and the relationship between DKA and some variables related to DM patients attended in the ED of Thames Hospital, New Zealand, from the 7^th November 2013 to the 7^th November 2017.

Secondary objectives were to establish the relationship between DKA and clinical features, signs on physical examination and with other variables related to chronic complications of DM.

Patients and Methods

Study design

This is a descriptive, retrospective, transverse study.

This study was approved by the Clinical Director of Thames Hospital and by the Team Leader of Clinical Effectiveness of the Waikato District Health Board, New Zealand.

Study population

The universe of the study was all patients diagnosed with DKA in the ED of Thames Hospital, New Zealand, from the 7^th November 2013 to the 7^th November 2017.

Variables

The variables were divided in social demographic variables, variables related to DM, variables related to clinical background or past medical history, variables related to simultaneous diagnosis and variables related to history of chronic and acute complications of DM.

Method of data registry/gathering

Inclusion criteria: patients with clinical diagnosis of DKA in the ED, with biochemical confirmation. Exclusion criteria: no inclusion criteria.

Three templates were completed for each patient: for social demographic data, for clinical history, and another one for data related to physical exam and other diagnoses (see annexes 1, 2 and 3).

Analysis and data processing

Variables were described with their absolute frequency (number) and relative frequency (percentage), and depending on their distribution (normal or not normal), mean and standard deviation were used, or the median and percentiles 5 and 95. We studied 63 different variables, however, only the relevant ones are mentioned in results, discussion and conclusions.

A result is considered statistically significant when the “p” level is equal or lower than 0.05. For the association among variables, we used the coefficient of correlation of Pearson (r), Spearman (rho), or Kendall (tau). The tests of joint distributions of homogeneity / heterogeneity of Pearson (chi-square) or Exact Fisher test (F), and meaning of difference of Mann-Whitney (U) or Student (t), according to the characteristics, scales and distributions of the variables involved in the comparisons. All the statistical tests are two sided with a significance level of 0.05 (or less).

We used the statistical program SPSS 17.0 for Windows©, SPSS Co, Chicago, Illinois, USA.

Results

We recorded the frequency, percentage and simple comparisons of gender, age range, family history of DM, treatment of DM, treatment compliance/diet/control of DM. They are summarized in Table 1.

In Table 2, we summarize the frequency, percentage and simple comparisons of variables related to acute and chronic complications of DM.

In Table 3, we summarize the frequency, percentage and simple comparisons of past or current alcohol drinker, and past or current drug user.

Non-classifiable (NC): patient younger than 15 (arbitrary value).

Non alcohol drinker: every patient that drinks or drank up to: 2 standard drinks a day for women and no more than 10 standard drinks a week, 3 standard drinks a day for men and no more than 15 standard drinks a week and at least 2 alcohol-free days every week (arbitrary value).

Drug user: we defined this as a behavior alteration that leads (or led) to an inability to control the use of a legal or illegal drug or medication.

Only the following variables had a relevant result and will be consider in discussion and conclusions: obesity (BMI of 25 or more), appendectomy, relevant weight loss (≤5 kg in the last 3 months; arbitrary value), polyuria and / or polydipsia, asthenia and / or feeling generally unwell, consciousness and / or balance impairment, simultaneous diagnosis of respiratory infection and simultaneous diagnosis of urinary infection.

In Table 4 are summarized the frequency, percentage and simple comparisons of the variables.

Discussion

We cannot compare our data with similar studies in New Zealand´s ED, because to our knowledge, no previous data had been published before. Therefore, we had to compare our data with other studies out of New Zealand.

The main social demographic variables of our patients do not differ in essence from other patients published in the literature.

Regarding the variables related to signs on physical examination, no data was found relevantly different from other studies.

About not expected findings, previous history of DKA, history of obesity, simultaneous diagnosis of respiratory infection and of urinary infection were the most relevant.

Regarding DM patients in previous published studies, the great majority of authors do not specify if DM was T1DM or T2DM. Furthermore, no author specified if DM was previously known or not.

When mentioning all patients, we refer to patients with and without previous diagnosis of DM.

In our study, 54% of patients with previous diagnosis of T1DM had history of DKA, and 42% of T2DM had history of DKA. In the literature, Shera, et al. [7] report 21% of the history of DKA in T1DM patients, and Javor, et al. [8] report 36%. According to Gill, et al. [9], 59% of patients with a previous diagnosis of DM had a history of DKA. According to Shibeshi, et al. [10], 25% of patients with a previous diagnosis of DM had a history of DKA.

It is expected to find a higher percentage of DKA history among patients with previous diagnosis of T1DM than T2DM. This is explained because of the pathogenesis of the illness, and because an acute decompensation of T1DM is more probable to happen than in T2DM.

It is worth mentioning that the history of DKA in the present study is higher compared to the published studies, among patients with previous diagnosis of T1DM, and T2DM.

If a patient had previous DKA history, it was more probable to have another episode of DKA in the future, because of the general condition of the patient, lifestyle, and the level of commitment to follow treatment and recommendations about DM.

In our study, 30% of patients had hypoglycemia history. In the group with no previous diagnosis of DM, 3% had previous hypoglycemia history, 42% in the T1DM and 73% in the T2DM. In the literature, Cosmescu, et al. [11] published 27% of hypoglycemia history in pediatric patients with previous diagnosis of T1DM, Soliman, et al. [12] reported 28%, Gill, et al. [13] 12%. This result is remarkable. It was expected to have a higher percentage of hypoglycemia history among patients with previous T1DM, but the percentage among T2DM patients was higher than among T1DM patients. Up to 24% of patients with T2DM were on treatment with oral hypoglycemiant drugs and insulin.

One reason for this difference could be the treatment with insulin and fast acting sulfonylureas. Another plausible explanation could be that T1DM patients are usually younger and can tolerate better hypoglycemias, while T2DM patients are usually older and tolerate worse the episodes of hypoglycemia.

About history of nonketotic hyperosmolar coma, we found it in 5% of the patients with no previous diagnosis of DM, 0 patients with previous diagnosis of T1DM, and 14% of T2DM. We could not find any studies in the literature. Because of pathogenesis, nonketotic hyperosmolar coma is exceptional among T1DM patients. It has a high mortality rate, and is more frequent among elderly people. Despite nonketotic hyperosmolar coma is quite rare among T1DM, in these patients, the mortality rate is even higher.

In our study, 0 patients with T1DM had diabetic foot history, and 16% of T2DM patients did. Regarding diabetic foot history in previous studies, among patients with previous diagnosis of DM, Feleke, et al. published 35% [14], and Pankaj, et al. 9%. [15] These authors did not specify if patients were T1DM or T2DM.

About diabetic neuropathy history, it was found in 9% of T1DM patients and 31% of T2DM. Similar percentages were published by Jaiswal, et al. [16] among adolescent patients (8% for T1DM and 26% for T2DM), showing that the prevalence of diabetic neuropathy is more frequent in T2DM than in T1DM. In other studies, Shera, et al. [7] published 3% and Gibb, et al. [17] 33% in T1DM. Among DM patients, Harzallah, et al. [18] reported 24% and Lundman, et al. [19] 27%.

In our study, there was a diabetic retinopathy history in 5% of patients with T1DM, and 31% of T2DM. Regarding diabetic retinopathy history in the literature, among patients with DM, Harzallah, et al. [18] published 8% of patients and Lundman, et al. [19] 27%; none of the authors specified if their patients were T1DM or T2DM. Among patients with T1DM, Shera, et al. [7] reported 7%, and Gibb, et al. [17] 42%. This can be explained like the previous variable. This is a consequence of microangiopathy; it takes over 20 years to develop it that is the reason why it is more common among patients with previous diagnosis of T2DM than T1DM.

No diabetic nephropathy history was found in T1DM in our study, but it was found in 9% of T2DM patients. In the literature, Lundman, et al. [19] reported 13% of patients and El Mahdi, et al. [20], 11% (they did not specify the proportion of T1DM or T2DM). Among T1DM patients, Shera, et al. [7] published 5% of patients, and Gibb, et al. [17] 11%. This could be explained like the previous variable. It is a consequence of the microangiopathy, it takes over 20 years to develop it; that is the main reason why it is more common among T2DM than T1DM.

Regarding to past or current alcohol drinkers, we had a lower percentage compared to published studies: 26%. Literature shows higher percentages: Kelly, et al. 35% [21], Umpierrez, et al. 51% [22], and Wilson, et al. 94% [23]. It is possible that alcohol intake in our general population might be also lower, but we lack comparative data among alcohol drinkers in our population, related with others.

We reported that 3.5% of our patients were past or current drug users; this incidence is significantly lower than other studies. In the literature, data rank from about the 13-14%, of Umpierrez, et al. [22], Kelly, et al. [21], and Warner, et al. [24], the 45% of Nyenwe, et al. [25], and finally Isidro, et al. [26], who reported a 70% of all patients were current or past drug users. We suppose this has something to do with the drug intake in the general population, but again, we lack comparative data among drug intake with other populations.

We reported 12% of total patients with obesity history. 25% of patients had no previous diagnosis of DM, 0 patients had a history of T1DM, and only 1 patient (2%) T2DM. Our data matches with other T1DM results, but not with T2DM. For example, Kelly, et al. [21] reported that 56% of DKA patients with previous T2DM diagnosis had also an obesity history. Newton, et al. [27], reported up to 70%. We believe that this huge discrepancy can only be explained due to a lack of previous diagnosis of obesity in T2DM (despite being obese). As these patients already did have a previous diagnosis of DM, usually an additional obesity diagnosis was not added. As these patients are already diagnosed with DM, they usually had a treatment, diet and recommendations about the illness. This could explain the higher percentage (24 patients) of obesity history among patients with no previous diagnosis of DM, as they were not under any treatment for DM.

In our study, 10% patients had appendectomy: 13% in the group of patients with no previous diagnosis of DM, 11% of T1DM, and 0 patients of T2DM. Pawlowicz, et al. [28], reported a 4% of appendectomy in their DKA patients. The clinical significance of this feature is unknown to us, and we cannot find a plausible explanation for this. We can speculate that previous episodes of hyperglycemia or DKA produced abdominal pain, and this could mimic acute abdomen, including appendicitis.

About the clinical feature of recent weight loss, it was found in 88% of the patients with no previous diagnosis of DM, in 9% of T1DM, and in 45% of T2DM. In the literature, among all patients, Iddi, et al. [29] reported 7% of weight loss, and Westerberg, et al. [30] 81%. Among DM patients, Harzallah, et al. [18] published 76% (none of these authors specified what type of DM were their patients). Among patients with T1DM, Shera, et al. [7] reported 79%. In this variable, it is worth mentioning that from the group of patients with no previous diagnosis of DM, there were 81% onset of T1DM, and 19% onset of T2DM. We cannot give a convincing explanation about the difference of percentage between patients with previous diagnosis of T1DM and T2DM in this study. Besides this, we have no explanation about why the percentage is higher among the patients with no previous diagnosis of DM.

In our study, 57% of total patients had clinical features of polyuria and/or polydipsia, 94% in the group with no previous diagnosis of DM, 14% of T1DM, and 25% of T2DM. Literature show similar results: Rajasoorva, et al. [31] reported 58%, Nazneen, et al. [32] 58%, and Iddi, et al. [29] 41%. Among previously diagnosed DM patients, Pankaj, et al. [15] published 26% and Harzallah, et al. [18] 87%. We cannot find a reasonable explanation of this high percentage in our study of patients with no history of DM that presented polyuria and / or polydipsia, compared to patients with previous diagnosis of DM.

Regarding asthenia and / or feeling general unwell, 23% had them, 14% of T1DM patients, and 40% of the T2DM. Literature shows that among all patients, Alourfi, et al. [33] reported 23% and Westerberg, et al. [30] reported 62%. Among DM patients, Pankaj, et al. [15] published 16% (but they do not specify what type of DM). This could be explained because of the older age. In our study the proportion of this feature in T1DM is much lower than in T2DM. The cause of feeling asthenic and unwell is multifactorial, and because of this, under many variables, and very difficult to explain.

An alteration of consciousness and / or balance was found in 4.8% of T1DM, and 16% of T2DM. In similar studies, among all patients, Rajasoorva, et al. [31] reported 61%, Pankaj, et al. [15] 30%, Kakusa, et al. [34] 64%, and Hamed, et al. [35] 60%. We do believe that the older the patient, the worse the general condition on arrival to the emergency department, and this could explain why it is more frequent among T2DM. However, in the published studies the situation is the opposite: the younger the patient, the higher the clinical feature of alteration of consciousness and / or balance (especially in pediatric patients). Once again, we cannot find a good explanation about our results and the different results in the literature.

No T1DM patients in our study had simultaneous respiratory infection, and 5% of T2DM did. In other studies, among all patients (i.e.: with and without previous DM diagnosis), Kakusa, et al. [34] reported 16%, Hamed, et al. [35] 27%, and Satti, et al. [36] 36%. Among previous known DM patients (but once again, T1DM and T2DM are together), Feleke, et al. [14] published 37%, Pankaj, et al. [15] 37%. We cannot find a plausible explanation of why this is the most common simultaneous diagnosis in the published studies, and in comparison, there is a very low percentage of this variable in the present study.

In our study, 17% of all patients presented simultaneous diagnosis of urinary infection; this was found in 25% of T1DM patients, and 28% of T2DM. In other studies, among DM patients, Feleke, et al. [14] published 14%, Pankaj, et al. [15] 37%, and Hamed, et al. [35] 31%. We cannot find a reasonable explanation of these differences among patients with and without previous diagnosis of DM, nor for the difference with other published studies.

The main limitation of our study is that it is a retrospective study, with the underlying inconveniences. Despite the fact that there are treatment guidelines about DKA in our hospital, there is no specific form (as in other hospitals) for fulfilling all standardized data. It is probable that some (or many) data were not taken, as in free text the busy emergency doctor could had forgotten.

Conclusion

The main characteristics of our patients do not differ in essence from other patients published in the literature.

Nonetheless, in this study, compared with the literature data, there is a significantly lower percentage of patients who were past or current alcohol drinkers, of past or current drug users, of patients with a previous diagnosis of T2DM who had a history of obesity, of patients with respiratory infection, and less asthenia among patients with previous diagnosis of T1DM.

On the contrary, we had a considerably higher percentage of patients with a previous episode of DKA, a higher percentage of patients with history of weight loss, and polyuria and / or polydipsia among patients with no previous diagnosis of DM.

Conflict of interest statement

The authors declare no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Statement of funding

This research received no specific grant nor financial support from any funding agency in the public, commercial, or not-for-profit sectors.

References

1. Realsen J, Goettle H, Chase HP (2012) Morbidity and mortality of diabetic ketoacidosis with and without insulin pump care. Diabetes Technol Ther 14: 1149-1154.
2. Gosmanov AR, Kitabchi AE (2000) Diabetic Ketoacidosis. In: Feingold KR, Anawalt B, Boyce A et al. eds. Endotext. South Dartmouth (MA): MDText.com.
3. Liu CC, Chen KR, Chen HF, Huang HL, Ko MC, et al. (2010) Trends in hospitalization for diabetic ketoacidosis in diabetic patients in Taiwan: analysis of national claims data, 1997-2005. J Formos Med Assoc 109: 725-734.
4. Venkatesh B, Pilcher D, Prins J, Bellomo R, Morgan TJ, et al. (2015) Incidence and outcome of adults with diabetic ketoacidosis admitted to ICUs in Australia and New Zealand. Crit Care 19: 451.
5. Wolfsdorf JI, Glaser N, Agus M, Fritsch M, Hanas R, et al. (2018) ISPAD Clinical Practice Consensus Guidelines 2018: Diabetic ketoacidosis and the hyperglycemic hyperosmolar state. Pediatr Diabetes 19: 155-177.
6. Dhatariya KK, Parsekar K, Skedgel C, Datta V, Hill P, et al. (2019) The cost of treating diabetic ketoacidosis in an adolescent population in the UK: a national survey of hospital resource use. Diabet Med 36: 982-987.
7. Shera AS, Miyan Z, Basit A, Maqsood A, Ahmadani MY, et al. (2008) Trends of type 1 diabetes in Karachi, Pakistan. Pediatr Diabetes 9: 401-406.
8. Javor KA, Kotsanos JG, McDonald RC, Baron AD, Kesterson JG, et al. (1997) Diabetic ketoacidosis charges relative to medical charges of adult patients with type I diabetes. Diabetes Care 20: 349-354.
9. Gill GV, Lucas S, Kent LA (1996) Prevalence and characteristics of brittle diabetes in Britain. QJM 89: 839-843.
10. Shibeshi MS, Fantahun B, Kebede T, Tilahun B (2016) Pediatric diabetic retinopathy: experience of a tertiary hospital in Ethiopia. BMC Res Notes 9: 116.
11. Cosmescu A, Felea D, Mătăsaru S (2008) Severe hypoglycemia in children and adolescents suffering from type 1 diabetes mellitus. Rev Med Chir Soc Med Nat Iasi 112: 955-958.
12. Soliman AT, al Salmi I, Asfour M (1997) Mode of presentation and progress of childhood diabetes mellitus in the Sultanate of Oman. J Trop Pediatr 43: 128-132.
13. Gill GV (1992) The spectrum of brittle diabetes. J R Soc Med 85: 259-261.
14. Feleke Y, Mengistu Y, Enquselassie F (2007) Diabetic infections: clinical and bacteriological study at Tikur Anbessa Specialized University Hospital, Addis Ababa, Ethiopia. Ethiop Med J 45: 171-179.
15. Pankaj S, Kaur H, Maneet Kaur M (2015) Clinical Profile of Diabetic Ketoacidosis: A Prospective Study in a Tertiary Care Hospital. J Clin Diagn Res 9: OC01-OC04.
16. Jaiswal M, Lauer A, Martin CL, Bell RA, Divers J, et al. (2013) Search for Diabetes in Youth Study Group. Diabetes Care 36: 3903-3908.
17. Gibb FW, Teoh WL, Graham J, Lockman KA (2016) Risk of death following admission to a UK hospital with diabetic ketoacidosis. Diabetologia 59: 2082-2087.
18. Harzallah F, Ncibi N, Alberti H, Ben Brahim A, Smadhi H, et al. (2006) Clinical and metabolic characteristics of newly diagnosed diabetes patients: experience of a university hospital in Tunis. Diabetes Metab 32: 632-635.
19. Lundman B, Engström L (1998) Diabetes and its complications in a Swedish county. Diabetes Res Clin Pract 39: 157-164.
20. El Mahdi EM, Abdel Rahman Iel M, Mukhtar Sel D (1989) Pattern of diabetes mellitus in the Sudan. Trop Geogr Med 41: 353-357.
21. Kelly JP, Umpierrez GE, Navarrete JE, Casals MM, Kitabchi AE (1997) Hyperglycemic crises in urban blacks. Arch Intern Med 157: 669-675.
22. Umpierrez G, Freire AX (2002) Abdominal pain in patients with hyperglycemic crises. J Crit Care 17: 63-67.
23. Wilson C, Krakof J, Gohdes D (1997) Ketoacidosis in Apache Indians with non-insulin-dependent diabetes mellitus. Arch Intern Med 157: 2098-2100.
24. Warner EA, Greene GS, Buchsbaum MS, Cooper DS, Robinson BE (1998) Diabetic ketoacidosis associated with cocaine use. Arch Intern Med 158: 1799-1802.
25. Nyenwe EA, Loganathan RS, Blum S, Ezuteh DO, Erani DM, et al. (2007) Active use of cocaine: an independent risk factor for recurrent diabetic ketoacidosis in a city hospital. Endocr Pract 13: 22-29.
26. Isidro ML, Jorge S (2013) Recreational drug abuse in patients hospitalized for diabetic ketosis or diabetic ketoacidosis. Acta Diabetol 50: 183-187.
27. Newton CA, Raskin P (2004) Diabetic ketoacidosis in type 1 and type 2 diabetes mellitus: clinical and biochemical differences. Arch Intern Med 164: 1925-1931.
28. Pawlowicz M, Birkholz D, Niedźwiecki M, Balcerska A (2008) Difficulties or mistakes in diagnosing type 1 diabetes mellitus in children? The consequences of delayed diagnosis. Pediatr Endocrinol Diabetes Metab 14: 7-12.
29. Iddi S, Brayson FA, Jaka H, Mirambo MM, Mushi MF (2017) Clinical presentation and precipitating factors of diabetic ketoacidosis among patients admitted to intensive care unit at a tertiary hospital in Mwanza, Tanzania. Tanzania J Hlth Res: 19.
30. Westerberg DP (2013) Diabetic ketoacidosis: evaluation and treatment. Am Fam Physician 87: 337-346.
31. Rajasoorva C, Wong SF, Chew LS (1993) Diabetic ketoacidosis - a study of 33 episodes. Singapore Med J 34: 381-384.
32. Nazneen S, Ahmed F, Ashrafuzzaman SM, Uddin KN, Ahsan ASM, et al. (2017) Clinical Presentation and Biochemical Abnormalities in Patients Presented with Diabetic Ketoacidosis in BIRDEM Hospital. Bangladesh Crit Care J 5: 7-10.
33. Alourfi Z, Homsi H (2015) Precipitating factors, outcomes, and recurrence of diabetic ketoacidosis at a university hospital in Damascus. Avicenna J Med 5: 11-15.
34. Kakusa M, Kamanga B, Ngalamika O, Nyirenda S (2016) Comatose and noncomatose adult diabetic ketoacidosis patients at the University Teaching Hospital, Zambia: Clinical profiles, risk factors, and mortality outcomes. Indian J Endocr Metab 20: 199-205.
35. Hamed ZS, Gawaly AM, Abbas KM, El Ahwal LM (2017) Epidemiology of infection as a precipitating factor for diabetic ketoacidosis at Tanta University Hospital. Tanta Med J 45: 68-72.
36. Satti SA, Saadeldin IY, Dammas AS (2013) Diabetic Ketoacidosis in children admitted to Pediatric Intensive Care Unit of King Fahad Hospital, Al-Baha, Saudi Arabia: Precipitating factors, epidemiological parameters and clinical presentation. Sudan J Paediatr 13: 24-30.