Trends in diabetic ketoacidosis in childhood and adolescence: a 15-yr experience

Chia sẻ: Lê Na | Ngày: | Loại File: PDF | Số trang:7

Thêm vào BST

Báo xấu

21
lượt xem 1
download

Download Vui lòng tải xuống để xem tài liệu đầy đủ

The number of episodes of diabetic ketoacidosis (DKA) is a Parkville, Melbourne, Australia significant outcome measure for diabetes care. We ascertained patterns of admission due to DKA over 15yr to determine whether this indicator of diabetes care had improved in parallel with clinical practices. Between 1 January 1985 and 31 December 1999, 630 admissions were reviewed. We subanalyzed these admissions according to whether the patient was newly diagnosed, had infrequent episodes of DKA (non-relapsers) or had frequent (2/yr) episodes of DKA (relapsers). Overall there was a slight downward trend in the incidence of DKA admissions over the study period

Chủ đề:

Bình luận(0) Đăng nhập để gửi bình luận!

Lưu

Nội dung Text: Trends in diabetic ketoacidosis in childhood and adolescence: a 15-yr experience

Pediatric Diabetes 2002: 3: 82–88 Copyright C Blackwell Munksgaard 2002 Printed in Denmark. All rights reserved Pediatric Diabetes ISSN 1399-543X Original Article Trends in diabetic ketoacidosis in childhood and adolescence: a 15-yr experience Bui TP, Werther GA, Cameron FJ. Trends in diabetic ketoacidosis in Thao P Bui, George A Werther childhood and adolescence: a 15-yr experience. and Fergus J Cameron Pediatric Diabetes 2002: 3: 82–88. C Blackwell Munksgaard, 2002 Department of Endocrinology and Diabetes, Royal Children’s Hospital, Abstract: The number of episodes of diabetic ketoacidosis (DKA) is a Parkville, Melbourne, Australia significant outcome measure for diabetes care. We ascertained patterns of admission due to DKA over 15 yr to determine whether this indicator of diabetes care had improved in parallel with clinical practices. Between 1 January 1985 and 31 December 1999, 630 admissions were reviewed. We subanalyzed these admissions according to whether the patient was newly diagnosed, had infrequent episodes of DKA (non-relapsers) or had frequent (ⱖ 2/yr) episodes of DKA (relapsers). Overall there was a slight downward trend in the incidence of DKA admissions over the study period. There was a proportionate increase in the incidence of DKA amongst newly diagnosed patients, with a proportionate decrease in the Key words: adverse events – diabetic incidence of DKA seen in relapsers. DKA occurring in non-relapsers ketoacidosis – incidence remained relatively stable. Adverse clinical events during the admission were relatively uncommon and occurred in all three subgroups. There Corresponding author: Dr Fergus Cameron, Head of Diabetes Unit, was no significant difference in HbA1C prior to admission between the Department of Endocrinology and relapser and non-relapser groups and there was similarity in the degree of Diabetes, Royal Children’s Hospital, acidosis between all three subgroups. The frequency of significant Flemington Road, Parkville, 3052, complications associated with DKA remained unchanged over the study Melbourne, Australia. period. Slower rehydration policies were not associated with decreases in Tel: π613 9345 6972 either cerebral edema or death rates. DKA remains a significant Fax: π613 9347 7763 complication of type 1 diabetes associated with a variety of significant e-mail: adverse events. Our experience indicates that further efforts to reduce the cameronf/cryptic.rch.unimelb.edu.au occurrence of DKA must be focused upon earlier diagnosis and Submitted 25 May 2001. Accepted for intervention in newly diagnosed patients. publication 12 December 2001 In diabetes occurring during childhood and adoles- two decades. In this study we have reviewed the ex- cence, a significant clinical outcome measure is the perience of a major pediatric diabetes care service frequency of diabetic ketoacidosis (DKA) (1, 2). Epi- over the last 15 yr in respect of admissions for DKA, sodes of DKA caused by relative hypoinsulinism are etiology and adverse events. potentially life-threatening. In the context of evolving type 1 diabetes, DKA is frequently an indicator of Methods diagnostic delay, whereas in the context of established diabetes, DKA is indicative of either insulin omission The diabetes care service at the Royal Children’s Hos- or suboptimally managed intercurrent stress episodes pital, Melbourne (RCH), provides care for approxi- (3, 4). Given the intensity of public awareness cam- mately 880 children and adolescents with type 1 dia- paigns as to the significance of diabetic symptoms betes mellitus. The clinic population originates from and the increased effort expended in improving gly- all socioeconomic strata within Victoria. At our hos- cemic control post-Diabetes Control and Compli- pital one to two children or adolescents per week are cation Trial (DCCT) (5), it might be anticipated that newly diagnosed with diabetes. This study is a retro- episodes of DKA in children and adolescents with spective analysis of all episodes of admission caused type 1 diabetes mellitus have decreased over the last by DKA between 1 January 1985 and 31 December 82 DKA admissions: trends over 15 yr 1999. All notes were reviewed by one of the authors (non-relapsers) and those with established diabetes (TB). DKA was defined as an admission pH ⬍ 7.30, and having at least two episodes of DKA per calen- and ketonuria in the presence of type 1 diabetes mel- dar year (relapsers). litus (defined according to World Health Organiza- Precipitants of DKA were those documented tion criteria) (6, 7). Patient records were reviewed to within the case notes. In patients with established dia- ascertain age, sex, documented etiology of DKA, betes, metabolic control was calculated according to treatment details (duration of insulin infusion, dur- the mean value of the preceding three 3-monthly ation of treatment until ketone-free) and adverse HbA1C values. HbA1C data from 1985 to 1994 were events. Patients were subclassified into those with assessed using the BIO-RAD (Hercules, CA) affinity newly diagnosed type 1 diabetes mellitus (newly diag- column chromatography method and between 1994 nosed patients), those with established diabetes and 1999 using a Bayer (Calabria, Barcelona) DCA having only one episode of DKA per calendar year 2000 latex immunoagglutination method. Equili- bration between the two methods was achieved using the following equation (8): HbA1C (measured by immunoagglutination) Ω 1.2 π 1.07 ¿ HbA1C (measured by column chromatogra- phy)In patients with more than two episodes of DKA per year (relapsers), only the first admission per calen- dar year was assessed in the HbA1C analysis. The degree of acidosis, as measured by serum pH value (capillary or arterial), was analyzed, as were duration of symptoms and insulin infusion (con- tinued from diagnosis until pH was greater than or equal to 7.30). Significant adverse events, including episodes of hypoglycemia and electrolyte disturbance, cerebral edema and death, were analyzed. Cerebral edema was defined as having occurred if a clinical diagnosis was made and the patient treated with in- travenous mannitol, with or without respiratory sup- port and radiologic (CT) confirmation. Where outcome variables were found to be distrib- uted in a normal fashion, results have been reported as mean ∫ SD. When non-normal distribution oc- curred, median and interquartile ranges have been re- ported. Weighted linear regression was used to fit a straight line to percentages against the year for each plot. A p-value of less than 0.05 was considered to be significant. Fig. 1. Diabetic admissions by year. Results Total episodes of DKA Between 1 January 1985 and 31 December 1999, there were 4458 diabetes-related admissions to RCH with a range of 248–373/yr (Fig. 1). The number of newly diagnosed patients ranged from 80 to 110 cases/yr, showing minor variation on a year-to-year basis (Fig. 1). Over this same time period there were 630 admis- sions due to DKA in 191 female and 141 male pa- tients. The ratio of female to male DKA admissions over the study period was 388:242. The number of DKA admissions has remained between 26 and 59/yr over the period, with a slight downward trend. Newly diagnosed patients Fig. 2. Diabetic ketoacidosis (DKA) admissions by diabetic status There were 164 DKA admissions in newly diagnosed and by year. patients, accounting for 26.0% of all DKA admis- Pediatric Diabetes 2002: 3: 82–88 83 Bui et al. Fig. 5. pH on admission by diabetic status for diabetic ketoacidosis Fig. 3. Number of diabetic ketoacidosis (DKA) admissions in newly (DKA) admissions. diagnosed diabetics as a percentage of the number of newly diag- nosed diabetics per year and its regression line. (4.3%) episodes of cerebral edema and two (1.2%) deaths in this group (Table 2). The two deaths (1994 and 1997) were in patients with methylmalonic acide- mia and trichothiodystrophy/asthma. Two patients who developed acute renal failure as a consequence of dehydration later recovered. Non-relapsers Admissions for DKA among non-relapsers num- bered 238 over the study period, accounting for 37.8% of total cases of DKA (Fig. 2). The absolute number of DKA admissions in the non-relapser group did not change significantly over the study period (Fig. 2) with the female:male ratio being 1.6:1.0. The mean age at time of admission for DKA in this group was 14.0 ∫ 3.4 yr, with a median pH at Fig. 4. Number of diabetic ketoacidosis (DKA) admissions in newly presentation of 7.15 (range, 7.06–7.22) (Fig. 5). The diagnosed diabetics as a percentage of the total number of admis- median duration polyuria and/or polydipsia of ⬍ 1 d sions for DKA per year and its regression line. (0–9), and the median duration of vomiting was 1 d (0–14). The mean HbA1C was 11.3 ∫ 2.1% (Fig. 6), with preceding HbA1C data unavailable in 26 pa- sions over the study period (Fig. 2) and having the tients. In the non-relapsing group where etiology was female:male ratio of 1.3:1.0. The proportion of newly recorded, 67/127 admissions were precipitated by in- diagnosed patients with DKA:total newly diagnosed fection, and 49/127 episodes were caused by insulin patients/yr increased by 0.9% for each yr (p ⬍0.001, omission. No cause was reported in 111 admissions 95%CI Ω 0.5–1.4) over the study period (Fig. 3). Over (Table 1). The median duration of insulin infusion the same study period a significant increase was seen was 23 h (range, 19–30). Nineteen episodes of hyper- in the proportion of newly diagnosed patients with kalemia (8.0%) and seven (2.9%) of hypoglycemia DKA:total episodes of DKA (Fig. 4). were documented. There were three (1.3%) episodes The mean age at the time of DKA was 7.8 ∫ 4.9 yr. of cerebral edema with one associated death (0.4%) The median pH at presentation in this group was 7.12 (Table 2). (range, 7.01–7.20) (Fig. 5), with the median duration of polyuria and/or polydipsia of 14 d (7–21). The me- Relapsers dian duration of vomiting was 1 d (0–2). Amongst newly diagnosed patients with DKA with docu- There were 228 admissions for DKA among relapsers mentation of etiology, 46/46 episodes were precipi- over the 15 yr studied (Fig. 2). This accounted for tated by infection, with no cause being documented 36.2% of all DKA admissions. The female:male ratio in 118 of the admissions (Table 1). The median dur- of these admissions was 1.9:1.0, with a downward ation of insulin infusion was 29 h (range, 22–40). trend seen over the study period (p Ω0.09) (Fig. 2). There were 10 (6.0%) episodes of hypokalemia, seven The mean age at the time of admission was 14.0 ∫ 2.6 (4.3%) episodes of significant hypoglycemia, seven yr. The median pH at the time of presentation was 84 Pediatric Diabetes 2002: 3: 82–88 DKA admissions: trends over 15 yr Table 1. Precipitants of diabetic ketoacidosis (DKA) admissions Newly diagnosed Non-relapsers Relapsers Total diabetics Infection 46 67 43 156 Insulin misuse 0 49 35 84 Unknown 118 111 147 376 Other 0 11 3 14 Table 2. Adverse events during diabetic ketoacidosis (DKA) admissions. Data are number (%) Newly diagnosed Non-relapsers Relapsers Total diabetics Death 2 (1.2) 1 (0.8) 1 (0.4) 4 (0.6) Cerebral edema 7 (3.4) 3 (1.3) 3 (1.3) 13 (1.2) Hypoglycemia* 7 (3.4) 7 (2.9) 12 (3.5) 26 (1.4) (Glc ⬍ 2.6 mmol/L) Hypokalemia* 10 (6.0) 2 (0.8) 2 (0.9) 14 (2.2) (Kπ ⬍ 3.0 mmol/L) Hyperkalemia* 5 (3.0) 19 (8.0) 12 (3.5) 36 (5.7) (Kπ ⬎ 6.0 mmol/L) Hypernatremia* 2 (1.2) 3 (1.3) 2 (0.9) 7 (1.1) (Naπ ⬎ 160 mmol/L) Hypothermia* 1 (0.6) 0 4 (1.8) 5 (0.8) (temperature ⬍ 36.0æC) Renal failure 2 (1.2) 0 0 2 (0.3) Insulin edema 1 (0.6) 0 0 1 (0.2) Nosocomial 9 (5.5) 2 (0.8) 5 (2.2) 16 (2.5) infection *Occurring within 48 h of presentation. 7.13 (range, 7.05–7.22) (Fig. 5), the median duration episodes of hypoglycemia (5.3%), 12 episodes of of polyuria and/or polydypsia ⬍ 1 d (0–7) and the hyperkalemia (5.3%), three cases of cerebral edema median duration of vomiting ⬍ 1 d (0–4). In each re- (1.3%). One death (0.4%) resulted from a combi- lapsing patient per calendar year, the mean HbA1C nation of cerebral edema and pulmonary edema/ was 11.3 ∫ 1.8% (preceding HbA1C data unavailable pneumothorax (Table 2). in 32 patients), and was not significantly different from that in non-relapsers (pΩ0.998) (Fig. 6). Among DKA admissions in this group where etiol- Discussion ogy was documented, 43/81 admissions were precipi- There has been a younger age of onset and an in- tated by infection and 35/81 by insulin omission. In creased incidence of type 1 diabetes mellitus docu- 147 (65%) admissions no cause was documented mented worldwide (9–11). However, when the mean (Table 1). The median duration of insulin infusion admission rate per year from 1985 to 1999 (297.2/900 was 24 h (range, 19–31). Adverse events included 12 patients/yr) is compared with previous data published from RCH from 1973 to 1981 (348.2/500 patients/ yr) (12), there appears to have been a downward trend in the overall number of diabetic admissions in our hospital. Between 1985 and 1999, the number of DKA admissions in established diabetes cases de- creased from 52 cases to 16 cases. The proportion of DKA admissions attributable to newly diagnosed patients was 26.0% over the entire study period. This figure is comparable to previous cross-sectional studies reporting rates of 20–36% (4, 13, 14). There are no published reports showing longitudinal trends of this ratio. In this study we found a longitudinal trend of a 2.7%/yr increase in DKA admissions in newly diagnosed patients as a percentage of total Fig. 6. Mean HbA1C values by diabetic status. DKA admissions. Over the same time-frame, the pro- Pediatric Diabetes 2002: 3: 82–88 85 Bui et al. portion of newly diagnosed patients with DKA com- relapsers and non-relapsers were significantly worse pared with the total newly diagnosed patients per year than cross-sectional mean HbA1C reports of 8.6% also increased (by 0.9%/yr). Given the constancy or seen in other Australian and international centres slight increase in the rate of newly diagnosed patients (26–28) and it is likely that rates of insulin omission and the slight decline in overall DKA admission rate leading to DKA were higher than those recorded. seen in this study, the overall decrease in diabetes- It is noteworthy that despite various precipitants related admissions presumably reflects better stan- and variable duration of symptoms prior to presen- dards of care for patients with established diabetes tation, the degree of acidosis (as measured by pH) and fewer readmissions. was remarkably concordant between the newly diag- In our experience, patients who are admitted with nosed, non-relapsing and relapsing groups. This DKA are more likely to be female with an overall would seem to imply that regardless of speed of onset, female:male ratio of 1.6:1.0. Whilst the female:male once a metabolic threshold had been reached our pa- ratio amongst newly diagnosed patients with DKA tients sought medical attention in a fairly consistent was closer to an equal sex distribution, the number time-frame. It is our practice to maintain all patients of admissions involving females with recurrent epi- with DKA on insulin infusions (0.05–0.15 units/kg/ sodes of DKA was almost double that of males. This h) until such time as their pH is greater than, or equal phenomenon has been found in other diabetes care to, 7.30. After this time we change to preprandial sub- centres (15, 16) but not universally (3, 17, 18). Pos- cutaneous insulin. Despite the similarity in degree of sible reasons for a female preponderance in DKA ad- acidosis, the duration of insulin infusions was much missions include a greater likelihood of family con- greater for those patients who were newly diagnosed flict and behavioral problems (19) and a greater likeli- as compared with those with established diabetes (29 hood of weight reduction associated with insulin vs. 23 and 24 h). In our experience, then, it appears omission occurring in females with type 1 diabetes that the duration of preceding symptoms, rather than (20). the degree of acidosis on presentation, is most predic- Whilst patients with established diabetes appeared tive of the time taken to normalize blood pH. to have a very short duration of hyperglycemic symp- Patients who were admitted and treated for DKA toms prior to admission for DKA, approximately had significant rates of electrolyte and metabolic dis- 50% of newly diagnosed patients had experienced turbance during the course of their resuscitation. In polyuria and/or polydipsia for greater than 2 wk prior 26 admissions (4.1%), significant hypoglycemia was to admission. In one retrospective study of 66 newly recorded within the first 48 h. This compares with diagnosed children with diabetes, 31 had been seen other reported figures of 12.7% within the first 48 h by their family doctor at least once without the diag- and 30% within the first 14 d after admission (29). nosis being made (21). Seventeen patients developed The pathogenesis of cerebral edema remains enig- DKA with 11 of the 17 having been seen on multiple matic (12, 30). Some authors have reported reduced occasions by family doctors prior to admission (21). incidence of cerebral edema with slower rehydration Diabetes usually develops slowly with easily recogniz- policies (31–33). In 1992, midway during the study able symptoms prior to deteriorating to DKA (22). period, we changed our rehydration rate from 24 to Earlier diagnosis in newly diagnosed patients appears 48 h. Rehydration fluids (normal saline or normal sa- to be feasible, with public health education cam- line and dextrose) have remained constant over the paigns having been shown to radically reduce epi- entire study period. Between 1985 and 1992 there sodes of DKA in other centres (23). were 6/345 (1.7%) cases of clinically diagnosed cer- In this study there were relatively high rates of pre- ebral edema and between 1992 and 1999 there were cipitants to DKA not having been recorded and it is 7/285 (2.4%) cases of cerebral edema. The lack of re- difficult to be sure that the etiologic factors seen are sponse in cerebral edema prevention to slower rehy- truly representative. Insulin omission has been re- dration in our experience presumably reflects the ported as accounting for anywhere between 27 and multifactorial and unpredictable nature of this most 67% of episodes of DKA in previous studies (3, 24, devastating complication of DKA in childhood and 25). Certainly the patients reported in this study, pre- adolescence (34). senting with either recurrent or non-recurrent DKA, In summary, in our experience there has been a had evidence of long-standing previous poor control change in the nature of DKA admissions over the last with mean yearly HbA1C values approximating 11.0– 15 yr. DKA admission rates have slightly fallen, with 11.5%. The similarity in mean HbA1C values be- newly diagnosed patients contributing a relatively tween relapsers and non-relapsers may potentially greater proportion to the overall DKA rate. Most es- have been due to individual patients being classified tablished patients with diabetes who present with as relapsers in one year and non-relapsers in a sub- either recurrent or non-recurrent DKA have evidence sequent year. This, however, was found to occur in of long-standing preceding poor metabolic control. only seven patients. The levels of control seen in both We found no difference in previous metabolic control 86 Pediatric Diabetes 2002: 3: 82–88 DKA admissions: trends over 15 yr amongst those patients who had either recurrent or 14. E K, S JN, P L, E B, A OO. Epidemiology and treatment of diabetic non-recurrent episodes of DKA in any given year. ketoacidosis in a community population. Diabetes Care The frequency of significant complications associated 1984: 7: 528–532. with DKA remains unchanged despite clinical efforts 15. G GV, L S, K LA. Prevalence and character- to the contrary. In particular, a slower rehydration istics of brittle diabetes in Britain. Q J Med 1996: 89: 839– policy in this study was not associated with a decrease 843. in cerebral edema rates or a decrease in rates of 16. T CJ, C F, C J, N RW. Abnormal insulin treatment behaviour: a major cause of death. Thus DKA remains a significant complication ketoacidosis in the young adult. Diabet Med 1995: 12: of type 1 diabetes associated with adverse metabolic 429–432. events, cerebral edema and death. Given the increas- 17. W AD, H PJ, S BM, K JA, N ing contribution by newly diagnosed patients pre- M, FG MG. Changing sex ratio in diabetic ke- senting with DKA, our experience indicates that toacidosis. Diabet Med 1990: 7: 628–632. 18. G MP, H AJ, O DP. An approach to pre- further efforts to reduce the occurrence of DKA must vention of recurrent diabetic ketoacidosis in the pediatric focus upon earlier diagnosis and intervention. population. J Pediatr 1985: 107: 195–200. 19. D RH, J AM, C C et al. Psychosocial predictors of acute complications of diabetes in youth. Di- References abet Med 1995: 12: 612–618. 20. B KS, N A, M RA, P RC, F 1. C FA, L JM. Diabetes mortality in persons CG, D DB. Eating habits, body weight, and insulin under 45 years of age. Am J Public Health 1983: 73: 1174– misuse. A longitudinal study of teenagers and young 1177. adults with type 1 diabetes. Diabetes Care 1999: 22: 1956– 2. E JA, F-A ME, D DB. Causes of death 1960. in children with insulin dependent diabetes 1990–96. Arch 21. H DV, M SS, L J. Mode of presen- Dis Child 1999: 81: 318–323. tation of juvenile diabetes. Br Med J 1976: 2: 211–212. 3. S O, E PC, V S, N J. Dia- 22. T AC, S CP, S KM, B GF, G betic ketoacidosis in Denmark: epidemiology, incidence EA. Type I (insulin dependent) diabetes: a disease of slow rates, precipitating factors and mortality rates. J Intern clinical onset? Br Med J 1987: 294: 342–345. Med 1989: 226: 223–228. 23. V M, C G, G L, C G, G 4. F GA, F HA, E SE. The epidemiology of T, C F. Effectiveness of a prevention program for diabetic acidosis: a population-based study. Am J Epide- diabetic ketoacidosis in children. An 8-year study in miol 1983: 117: 551–558. schools and private practices. Diabetes Care 1999: 22: 7– 5. T D C and C T R- 9.  G. The effect of intensive treatment of dia- 24. M VC, L JK, C R, K MA, MA- betes on the development and progression of long-term  D, P LS. Diabetes in urban African-Ameri- complications in insulin-dependent diabetes mellitus. N cans. I. Cessation of insulin therapy is the major precipit- Engl J Med 1993 September 30: 329: 977–986. ating cause of diabetic ketoacidosis. Diabetes Care 1995: 6. W H O. Diabetes mellitus. Report 18: 483–489. of a WHO Study Group. World Health Organization 25. S CP, F D, B S, H C, C P. Technical Report Series 727. WHO, Geneva, 1985, pp 1– Ketoacidosis occurring in newly diagnosed and estab- 113. lished diabetic children. Acta Paediatr 1998: 87: 537–541. 7. A KG, Z PZ. Definition, diagnosis and classi- 26. R M, C JB, W MC et al. Factors as- fication of diabetes mellitus and its complications. Part 1: sociated with glycemic control. A cross-sectional nation- diagnosis and classification of diabetes mellitus pro- wide study in 2,579 French children with type 1 diabetes. visional report of a WHO consultation. Diabet Med 1998: Diabetes Care 1998: 21: 1146–1153. 15: 539–553. 27. T M, S G, B J, C A. How well 8. D P, V S, C FJ. Deteriorating dia- are we doing? Metabolic control in patients with diabetes. betic control through adolescence-do the origins lie in J Pediatr Child Health 1999: 35: 479–482. childhood? Diabetic Med 2001: 18: 889–894. 28. M HB, R KJ, A HJ et al. In- 9. O P, V S, K M, T J. sulin management and metabolic control of type 1 dia- Worldwide increase in incidence of type I diabetes – the betes mellitus in childhood and adolescence in 18 coun- analysis of the data on published incidence trends. Diab- tries. Hvidore Study Group on Childhood Diabetes. Di- etologia 1999: 42: 1395–1403. abet Med 1998: 15: 752–759. 10. K M, P J, T J. The onset age 29. M ML, K SE, G VM, G JS. Fre- of type 1 diabetes in Finnish children has become younger. quent hypoglycemic episodes in the treatment of patients Diabetes Care 1999: 22: 1066–1070. with diabetic ketoacidosis. Arch Intern Med 1992: 152: 11. C ME, H NJ, S M, C A. The rising 2472–2477. incidence of childhood type 1 diabetes in New South 30. R AL, R WJ, W FT, M JI, D- Wales, Australia. J Pediatr Endocrinol Metab 2000: 13:  WH. Cerebral edema complicating diabetic ketoac- 363–372. idosis in childhood. J Pediatr 1980: 96(3 Part 1): 357–361. 12. M JM, W GA. Incidence and outcome of dia- 31. M CP, D V BWI, A M. Risk factors betic cerebral oedema in childhood: are there predictors? for developing brain herniation during diabetic ketoac- J Paediatr Child Health 1995: 31: 17–20. idosis. Pediatr Neurol 1999: 21: 721–727. 13. K RA. Diabetic ketoacidosis. new concepts and 32. E JA, D DB. Variations in the management of trends in pathogenesis and treatment. Ann Intern Med diabetic ketoacidosis in children. Diabet Med 1994: 11: 1978: 88: 681–695. 984–986. Pediatric Diabetes 2002: 3: 82–88 87 Bui et al. 33. H GD, F I. Physiologic management of 34. E JA. Cerebral oedema during treatment of diabetic diabetic ketoacidemia. A 5-year prospective pediatric ex- ketoacidosis: are we any nearer finding a cause? Diabetes perience in 231 episodes. Arch Pediatr Adolesc Med 1994: Metab Res Rev 2000: 16: 316–324. 148: 1046–1052. 88 Pediatric Diabetes 2002: 3: 82–88