Pdi_286 24.33

Pediatric Diabetes 2007: 8 (Suppl. 6): 24–33 Journal compilation # 2007 Blackwell Munksgaard Can we prevent diabetic ketoacidosis inchildren? Bismuth E, Laffel L. Can we prevent diabetic ketoacidosis in children? Pediatric Diabetes 2007: 8 (Suppl. 6): 24–33.
Abstract: Diabetic ketoacidosis (DKA) is an acute potentially life- threatening complication of diabetes affecting more than 100,000 persons annually in the United States. Although major advances have improved diabetes care, DKA remains the leading cause of hospitalization,morbidity, and death in youth with type 1 diabetes (T1D). As the majority of patients presenting with DKA have established diabetes, it is important to address outpatient educational approaches directed at sick-daymanagement and early identification and treatment of impending DKA.
Teaching and reinforcement of sick-day rules involves improved self-care Lori Laffel, MD, MPHSection on Genetics and Epidemiology with consistent self-monitoring of blood glucose and ketones, and timely administration of supplemental insulin and fluids. DKA as an initial manifestation of T1D may be less amendable to prevention except with an increased awareness by the lay and medical communities of the symptoms of diabetes and surveillance in high-risk populations potentially identified by family history or genetic susceptibility. New technologies that can detect the blood ketone 3b-hydroxybutyrate (3b-OHB) instead of traditional urine ketones appears to provide opportunity for early identification and treatment of impending DKA leading to reduced need for hospitalization Despite major advances in the care of diabetes, mately 50% of all deaths in individuals with diabetes diabetic ketoacidosis (DKA) remains a leading cause younger than 24 yr old (5, 6). Cerebral edema is an of hospitalization and the leading cause of morbidity uncommon but serious complication of DKA associ- and death in children and adolescents with type 1 dia- ated with morbidity and mortality. Estimates of the betes (T1D). There are over 150 000 annual episodes incidence of cerebral edema in DKA range from 0.4 to of DKA in the USA. The average cost of treating 3.1% (5), and recent reports from Britain and the USA a single episode of DKA in the USA is about $11 000; have shown a mortality rate of 21–24%, and sig- the cost of all episodes combined represents about 25% nificant neurologic sequelae in 21–35% (6, 7). Cerebral of the total spent on the care of patients with T1D (1).
edema accounts for 60–90% of all DKA-related The overall incidence of DKA varies with definition, age, and sex, ranging from 4.6 to 13.4 episodes per It is important to try to prevent DKA in order to 1000 persons with diabetes per year in the USA (2).
reduce morbidity and mortality associated with severe The majority of DKA cases have previously diag- metabolic decompensation. This prevention can be ac- nosed diabetes and it is estimated that 50% of hospital complished through appropriate education, improved admissions for DKA could be prevented with im- self-care and adherence, and consistent self-monitoring proved outpatient treatment and better adherence to of blood glucose and ketones. DKA as an initial manifestation of T1D is less amenable to prevention, Complications related to DKA are the most other than through surveillance in youth with a common cause of death in children, teenagers, and positive family history of diabetes, and increased young adults with diabetes, accounting for approxi- public awareness of the symptoms of diabetes.
T1D in youth worldwide, especially in the very young(19). In a recent study of incidence trends in childhood DKA is a serious complication of diabetes that is diabetes across Europe, the 0- to 4-yr age group associated with considerable mortality and morbidity.
displayed the highest annual increase (4.8%) (20).
Although the mortality rate in adults diminished from44% in the 1930s to 16% in the 1970s (9) and to 3–5%in the 1980–1990s (10), the recent rates in pediatric patients have been relatively stable at about 1% or less The incidence of DKA in adolescent patients enrolled (6, 11). In the context of evolving T1D, DKA is in the DCCT was 2.8 per 100 patient-years in the frequently an indicator of a delay in the recognition of intensive treatment group (n ¼ 92) vs. 4.7 per 100 the symptoms of diabetes, whereas in the context of patient-years in the conventional therapy group established diabetes, DKA is often indicative of either (n ¼ 103) (21). In a more recent study, incidence of insulin omission or suboptimally managed intercur- DKA in children and adolescents with T1D was found rent stress episodes. The latter offer opportunities for to be 8 per 100 patient-years (22). The incidence of enhanced diabetes education toward DKA preven- DKA in established diabetes is higher in females, tion. Given the intensity of public awareness cam- peaks in early teenage years, and rarely occurs in paigns as to the significance of diabetic symptoms and anyone diagnosed for less than 2 yr. Individuals with the increased effort expended in improving glycemic earlier age of onset and lower socioeconomic back- control after the Diabetes Control and Complications grounds seem to be at increased risk, along with Trial (DCCT), it might be anticipated that episodes of individuals who had psychopathology before diabetes DKA in children and adolescents with T1D would onset. Research reveals several consistent themes that have decreased over the last two decades. However, enable us to identify individuals at potential risk for data from Ontario between 1991 and 1999, revealed recurrent DKA, with about 20% of individuals that DKA admissions remained stable (4).
accounting for 80% of the hospital admissions forDKA in one report (23).
In Europe, Australia and North America, some 15– 70% of all newly diagnosed children with diabetes present with DKA (12, 13). Thus, there is widegeographic variation in the frequency of DKA at DKA at diagnosis is more common in younger onset of diabetes; rates inversely correlate with the children (,5 yr of age) and in children whose families regional incidence of T1D, likely as a result of more do not have ready access to medical care for social or experienced physicians recognizing symptoms of dia- economic reasons (22, 24). Lack of health insurance is betes (8). Most commonly, rates of DKA at diagnosis associated with higher rates (and greater severity) of are 25–30% (14–16). In addition, rates of DKA vary DKA at diagnosis, presumably because uninsured according to age at diagnosis of diabetes, with as persons delay seeking timely medical care (22). Lower many as 44% of youth ,6 yr of age presenting in income and lower parental educational achievement DKA in one study (17) and 30% presenting with (father’s work, education of parents) were also acidosis and/or coma in another report (18). Recently, associated with higher risk of DKA (25). History of DKA, defined by blood bicarbonate ,15 mmol/L parental depression and diminished parental anxiety, and/or pH , 7.25 (7.3 if arterial or capillary), was possibly due to lower parental awareness of symptoms found in 23.3% of a carefully analyzed US cohort (16).
in offspring, may also increase DKA risk at onset (26).
The prevalence of DKA decreased significantly with As expected, a lower frequency of DKA at T1D age from 36% in children ,5 yr of age to 16% in those diagnosis was evident in children with a family history .14 yr, but it did not differ significantly by sex or ethnicity (16). Indeed, in the younger child who mayalso have a more rapid rate of beta cell loss, it is more difficult to obtain a classic history of polyuria, poly-dipsia, and nocturia. Thus, infants and toddlers with Children whose insulin is administrated by a respon- impending DKA may go undiagnosed, thereby in- sible adult rarely have episodes of DKA; 75% of creasing their duration of symptoms, leading to more episodes of DKA after diagnosis are associated with severe dehydration and acidosis at presentation. For insulin omission or treatment error. The remainder example, a dramatic 220% increase in DKA admis- are a result of inadequate insulin therapy during sions was observed in the 0- to 4-yr age group between intercurrent illness (11). In a cohort of 1243 children 1991 and 1999 in Ontario (4). This trend is concern- with T1D followed prospectively for 4 yr in the ing, and supports the increase in the occurrence of Denver area, the incidence of DKA was 8 per 100 Pediatric Diabetes 2007: 8 (Suppl. 6): 24–33 person-years and increased with age in girls. In area, the incidence of DKA in new-onset cases aged younger children, the risk of DKA increased with 6–14 yr decreased from 78% in 1987–1991 to 12.5% higher A1c and higher reported insulin dose. In older in 1991–1997 with no cases reported after 1992. In the children, the risk of DKA increased with higher A1c, control region nearby, which received no information, higher reported insulin dose, underinsurance, and the 83% of new cases presented with DKA (33). This de- presence of psychiatric disorders (22). Eating disor- monstrates that by means of an aggressive but rela- ders, relatively common in young women with T1D, tively inexpensive information campaign, it is possible also contribute to impaired metabolic control, leading to reduce the incidence of DKA at onset. In another center, where tips about early symptoms of diabetes There is consistent evidence for psychosocial risk were provided to local pediatricians, a reduction from factors as predictive of recurrent DKA. Individuals 86 to 26% of DKA at diagnosis was observed (3).
from families low in warmth and support, where there A diagnosis of DKA may be delayed in new-onset are high levels of unresolved family conflict and a cases, especially in younger children, who may first be distinct lack of parental involvement in the adolescent’s diagnosed with pneumonia, reactive airway disease, or diabetes care, seem to be typical of this population (28).
bronchiolitis. Indeed, a recent study revealed that Objective assessment of insulin management behavior among youth diagnosed with T1D, children ,3 yr old through prescription data from Scotland indicated that and those presenting in DKA had more medical 28% of young adults (aged 15–25 yr) with T1D do not encounters in the week prior to diagnosis compared refill a sufficient amount of prescribed insulin to follow with those without DKA, suggesting missed opportu- their treatment regimen. This behavior, indicating nities to prevent DKA (34). Follow-up studies suggest under-insulinization, predicts admission for DKA (29).
that DKA is more frequent (24, 35), and more severe Individuals using continuous subcutaneous insulin (24), when missed at initial patient encounters. In New infusion (CSII) are potentially at increased risk for England, anecdotal reports include three deaths in DKA following unrecognized interruption in insulin youth from undiagnosed diabetes in 2003–2004.
delivery and inadequate monitoring (30). Population- Earlier diagnosis through immunologic and genetic based and retrospective clinical studies report a rela- screening of high-risk children, such as in the Diabetes tively low rate of DKA with pump therapy, but Prevention Trial (DPT)-1, decreased DKA incidence at a higher rate with CSII compared with injection diabetes onset (36). A high level of awareness in those therapy, at least in some countries (30). Recent data with positive family history of T1D also reduces the among 1041 pediatric patients using insulin pumps in occurrence of DKA at diagnosis (24). In a recent study, Europe revealed a risk of DKA of 6.6 per 100 patient- children with a positive family history of T1D presented years (31). In a Swedish population-based study, the 50% less often in DKA at diagnosis than those without incidence of DKA among pump users was 3.5 per 100 a family history of T1D (29 vs. 60%) (24). In addition, patient-years compared with 1.7 per 100 patient-years the DPT-1 revealed that a rising A1c even within the among patients treated with insulin injections (32).
reference range, may predict the diagnosis of T1D, thus In general, the risk of DKA appears relatively low in offering another means to avert an episode of DKA at research settings, among adherent patients, and in patients with sufficient family support (30).
DKA in established diabetes is most often the result ofinappropriate management of intercurrent illness/ stress or accidental or deliberate omission of insulin Professional and public awareness of early signs and (Table 1). Patient and family training of sick-day symptoms of diabetes in children and adolescents is management can provide the education needed to required to decrease the incidence of DKA in patients prevent or treat severe hyperglycemia and ketosis.
with new-onset diabetes; a high index of suspicion Studies of the effects of such comprehensive diabetes of symptomatology may lead to earlier diagnosis.
management programs and telephone help lines report Although such strategies are intuitively obvious, pro- a reduction in the rates of DKA from 15–60 to 5–5.9 grams to decrease DKA at onset need to be designed per 100 patient-years (11). It is likely that both and evaluated in diverse populations and age groups diabetes sick-day education and 24-h telephone help (11). In a program in Parma, Italy, schools and doctors’ lines reduce the occurrence of DKA; however, studies offices were provided with colorful posters with have not evaluated the unique contribution of each practical messages about diabetes, and local pediatri- approach to decreasing the rates of DKA. Therefore, cians were instructed on the use of glucose meters.
episodes of DKA after diagnosis could be reduced if Parents and pediatricians received a toll-free number to all children with diabetes and their families receive facilitate contact with the diabetes unit. In the study comprehensive, ongoing diabetes education along Pediatric Diabetes 2007: 8 (Suppl. 6): 24–33 Table 1. Triggers for hyperglycemia, ketosis, and diabetic therapy involving intensive home-based psychother- apy appears to reduce DKA admissions compared tostandard care for youth in poor control (41, 42).
Emotional stressErrors in insulin administration Testing for ketones remains critical to the prevention of DKA, yet knowledge and practice of ketone testing Intentional manipulation of insulin dosing are usually deficient. Diabetes education requires not only initial teaching of diabetes management to Myocardial infarctionMedications (e.g. steroids) prevent metabolic decompensation but also ongoing reinforcement. Unfortunately, sick-day management is usually taught in a state of good health when practice of these principles is not needed. Thus,patients and families may have forgotten sick-day man-agement by the time illness ensues. Sick-day manage- with access to a 24-h diabetes telephone help line (3, ment requires the patient to check blood glucose and 11). Sick-day rules should be reinforced periodically, especially at the start of the school year and during fluseason when illness is more common. For patientsreceiving CSII, DKA may be avoidable with frequent monitoring of blood glucose along with urine/bloodketones, followed by appropriate intervention when Although glucose is the preferred metabolic fuel, needed (8). In Norway, the nationwide incidence of alternative fuel sources, such as free fatty acids DKA (approximately 4 per 100 patient-years) did not (FFAs) and ketones from fat breakdown, can be change despite an increase in CSII use from 5% in used if glucose is unavailable. In DKA, because of an absolute or relative lack of insulin, the insulin- Patients who experience multiple episodes or Ôre- dependent tissues are unable to metabolize glucose current’ DKA are more problematic. Insulin omission normally. The regulation of fatty acid breakdown is has been identified as the major factor in most of these influenced by several hormones, in particular insulin, cases. There is often a trigger for insulin omission and glucagon, and epinephrine. Low insulin levels along a psychiatric social worker or clinical psychologist with high levels of counter-regulatory hormones (low should be consulted to investigate any psychosocial ratio of insulin/glucagon) cause an increase in lipo- reasons contributing to development of DKA. Insulin lysis, mobilizing FFAs, and promoting ketogenesis.
omission may be preventable with multidisciplinary After transport into hepatic mitochondria, FFAs are support providing education, psychosocial evaluation, converted into acetyl-coenzyme A (CoA), which can and treatment combined with adult supervision of be used for energy synthesis in the tricarboxylic acid insulin administration. When responsible adults ad- cycle if adequate oxaloacetate is present. When oxalo- ministered insulin, episodes of DKA dropped by 90% acetate is being used for gluconeogenesis (as in DKA), (38). The DKA prevention program in Los Angeles, the acetyl-CoA is instead used for synthesis of ketones.
which included teaching patients and families the Acetoacetate (AcAc) is the primary product of b- early warning signs of DKA, providing sick-day oxidation; 3b-hydroxybutyrate (3b-OHB) results from management guidelines, and availability of a 24-h the reduction of AcAc; and acetone results from the health-care team, reduced the rate of recurrent DKA spontaneous decarboxylation of AcAc. Ordinarily, from 12 events per 100 patient-years to 4 events per AcAc and 3b-OHB exist in equimolar concentrations 100 patient-years in 1998 (3). Again, this study did not in the blood. During DKA, however, the reduced evaluate the unique contribution of the individual redox potential in the hepatic mitochondria favors the components of the DKA prevention program. Dedi- formation of 3b-OHB and the 3b-OHB to AcAc ratio cated outpatient diabetes treatment teams directed at increases from 1:1 to 6:1, reaching as high as 10:1 adult patients can also result in significant decreases in in some individuals. In contrast, with recovery from DKA-associated readmissions and A1c values in DKA, the ratio again falls with conversion of 3b-OHB ketosis-prone patient populations (39). When these efforts fail, attempts to prevent recurrent DKAmandate more aggressive approaches such as out-of- home placement away from dysfunctional familiesand even use of insulin pump therapy to provide For several decades, the only way to measure ketones consistent basal insulin delivery (3, 40). Multisystemic was testing the urine with a dipstick test such as Pediatric Diabetes 2007: 8 (Suppl. 6): 24–33 KetostixÒ (Bayer Diagnostics, USA) or with acetest suggestive of DKA are present. A few years ago, the tablets. KetostixÒ measures urinary AcAc, which ADA statement noted that ÔCurrently available urine reacts with nitroprusside to produce a purple-colored ketone tests are not reliable and blood ketone testing complex. Nitroprusside does not react with 3b-OHB, methods are now reliable for diagnosis and monitor- the predominant ketone body in DKA. Urine strips ing treatment of ketoacidosis’ (45).
containing glycine in addition to nitroprusside can Published reports on blood ketone monitoring have appeared from clinical research and from inpatient Urine ketone testing has limitations for several and outpatient settings. Elevations in blood 3b-OHB reasons. Urine ketone measurements do not accu- and urine ketones have been monitored prospectively rately reflect current conditions if the urine has been during investigations in which CSII with lispro or in the bladder for several hours (e.g. overnight) before regular insulin was discontinued under controlled testing. Similarly, if bottled strips are used and the conditions (46–48). After 5–6 h following discontin- bottle was first opened more than 6 months earlier (it uation of CSII, moderate ketones generally appeared is recommended that the date be written on the bottle in the urine and blood 3b-OHB rose from 0.1– when first opened), the strips can lose their accuracy.
0.2 mmol/L to 1–1.2 mmol/L. Monitoring blood Urine ketone results can also be affected by medi- ketones would likely allow the earlier detection of cations, giving false-positive results in the presence of ketosis when levels exceed 0.5 mmol/L, before keto- drugs containing sulfhydryl groups, like captopril.
nuria becomes evident, allowing for earlier interven- Also, obtaining a urine sample is sometimes problem- tion. Patients on CSII, in whom ketosis can appear atic with very young children, teenagers, and people rapidly in cases of pump failure or catheter occlusion, who are unable to void or who are too ill or exhausted Measuring 3b-OHB in a hyperglycemic patient in The inability of the nitroprusside test to detect the acute setting can expedite diagnosis and treatment 3b-OHB and an increasing belief that blood levels of as a point-of-care tool in an emergency department 3b-OHB might prove useful in the management (50). In this setting, sensitivity and specificity of the and prevention of DKA prompted the development capillary 3b-OHB measurements in determining DKA of rapid enzymatic methods for the quantification of were 72 and 82% (vs. 66 and 78% for urine ketone 3b-OHB in small-volume blood samples. The first of dipstick), respectively (51). In another study, among these systems was marketed by GDS Diagnostic 50 patients who had 3b-OHB measurements in an (Elkart, IN, USA), which provided a bench-top emergency room when fingerstick glucose exceeded analyzer for use in clinical laboratories and physician 11 mmol/L (200 mg/dL), 3b-OHB level of .3 mmol/L offices. The GDS System determines 3b-OHB levels had a sensitivity of 100% and specificity of 88% for on a drop of blood (25 mL) in about 2 min, with a detection range between 0 and 2 mmol/L (44). More In the outpatient setting, blood ketone monitoring recently, a hand-held device has been developed that may improve self-care management during hypergly- allows the determination of 3b-OHB from capillary cemia, sick days, and for symptoms suggestive of blood in 10 s (initially 30 s) at home or at the patient’s DKA; however, patient adherence needs to be bedside. This system for the precise quantification of evaluated. Among healthy outpatients in suboptimal 3b-OHB levels on a fingerstick blood specimen (2 mL) glycemic control, measurements of blood 3b-OHB can has been introduced by MediSense/Abbott Labora- help distinguish ketotic patients from those with tories (Precision XtraÒ/OptiumÒ) (Abbott Diabetes hyperglycemia alone (53). For sick-day management, Care, Alameda, CA) and is available for clinical the efficacy of blood 3b-OHB monitoring was practice. Accuracy testing against a reference labora- evaluated in a 6-month two-centre, prospective, tory instrument demonstrates a high correlation randomized clinical trial comparing blood 3b-OHB (r ¼ 0.94) with a measurement range between 0 and monitoring (Precision XtraTM) with traditional urine 6 mmol/L. Elevated ketones levels, called hyperketo- ketone testing. We enrolled 123 children and adoles- cents who were randomized, according to pump statusand A1c, to receive the Precision XtraTM system orurine ketone strips for ketone monitoring. Partic- ipants continued routine diabetes care throughout thestudy, which included 24-h access to an on-call physician (54). Participants received sick-day guide- The ADA previously recommended that all people lines within logbooks specific for each group. Rec- with diabetes should test their urine for ketones ommendations were based on blood glucose results during periods of acute illness or stress, when blood and either blood 3b-OHB or urine ketone measure- glucose levels are consistently in excess of 16.6 mmol/L ments (see Tables 2 and 3). Adherence to ketone (300 mg/dL), during pregnancy, or when symptoms monitoring during sick days was 90.8% for partic- Pediatric Diabetes 2007: 8 (Suppl. 6): 24–33 Table 2. Supplemental insulin dosages based on blood glucose and urine ketone results Omit fast-acting analog or regular insulin Decrease intermediate or long-acting insulin by 20% Decrease intermediate or long-acting insulin by 20%; contact health care team,especially with vomiting Note: % refers to percentage of total daily dosage (TDD) given as fast-acting analog or regular insulin. TDD is calculatedby adding up all the insulin administered on a usual day, including the fast-acting analog or regular insulin and theintermediate/long-acting insulin. Do not include supplements added to the usual dose because of unexpectedhyperglycemia. In calculating the TDD when sliding scales are used, select the sliding scale dose for blood glucose ofabout 8.3 mmol/L (150 mg/dL). Blood glucose and urine ketones should be monitored every 2–4 h. Supplemental insulinboosters are repeated every 2–3 h with the fast-acting analog or every 3–4 h with regular insulin. If hyperglycemia or urineketones do not improve after two supplemental dosages, the health-care team should be contacted. Pump basal ratesshould be increased by 20–50% during illness along with additional bolus doses. If blood glucose level is ,4.4 mmol/Land there is decreased po intake, omit the fast-acting analog or regular insulin and decrease intermediate/long-actinginsulin by 20%. Contact health-care team, especially if there is vomiting. The range of dosage adjustments accounts forthe need for clinical judgment based upon clinical status and anticipated oral intake of food and fluids.
ipants checking blood ketones, compared with 60.3% 13.9 mmol/l), when ketone monitoring occurred in for those checking urine ketones. There was no only 34% of each group. Thus, the common occur- difference in monitoring frequency between groups rence of hyperglycemia in youth with diabetes was during periods of hyperglycemia alone (blood glucose often an insufficient motivation to test for ketones inthe absence of illness. In this study, the need foremergency department assessment and treatment or Table 3. Algorithms for supplemental fast-acting analogor regular insulin dosages incorporating blood 3b-hydroxy- urgent hospitalization was 50% lower in the blood ketone group compared with the urine ketone group,38 episodes per 100 patient-years compared to 75 per 100 patient-years, respectively. At study’s end, 70% of those checking ketones reported that they preferred to check blood versus urine ketones (54). Thus, blood ketone monitoring seems to be well accepted in Hyperketonemia appears to be common in the setting of uncontrolled diabetes (55). Further studies are needed to confirm the impact of blood ketone Note: % refers to percentage of total daily dosage (TDD) given as fast-acting analog or regular insulin. TDD iscalculated by adding up all the insulin administered on a usual day, including the fast-acting analog or regularinsulin and the intermediate/long-acting insulin. Do not Patient and parent education remains the cornerstone include supplements added to the usual dose because ofunexpected hyperglycemia. In calculating the TDD when for sick-day management and prevention of metabolic sliding scales are used, select the sliding scale dose for decompensation in youth with T1D. New technologies blood glucose of about 8.3 mmol/L (150 mg/dL). Blood like a handheld blood ketone meter can improve self- glucose and ketones should be monitored every 2–4 h.
care management. It provides a method to detect Supplemental insulin boosters are repeated every 2–3 hwith the fast-acting analog or every 3–4 h with regular metabolic disturbance and correct it if appropriate insulin. If hyperglycemia or blood ketones do not improve guidelines are followed. First, patients, family, and after two supplemental dosages, the health-care team general practitioners should be aware of the triggers should be contacted. Pump basal rates increased by 20– for hyperglycemia, ketosis, and DKA (Table 1). For 50% during illness along with additional bolus doses. If example, substance abuse may trigger DKA through blood glucose level should be ,4.4 mmol/L and there isdecreased po intake, omit the fast-acting analog or regular non-adherence and comorbidities such as reactive insulin and decrease intermediate/long-acting insulin by airway disease and inflammatory bowel disease may 20%. Contact health-care team, especially if there is warrant systemic steroid treatment, which may pro- vomiting. The range of dosage adjustments accounts for mote metabolic decompensation in the absence of the need for clinical judgment based upon the clinicalstatus and anticipated oral intake of food and fluids.
supplemental insulin. Next, algorithms for management Pediatric Diabetes 2007: 8 (Suppl. 6): 24–33 of hyperglycemia and sick days should be explained children and adolescents with T1D, usually at diagnosis and reinforced repeatedly. Before routine acceptance with reinforcement at the time of illness. These of ketone monitoring is realized, it is necessary that all guidelines should be instituted with illness, ketonuria, patients with T1D check their glucose levels fre- or hyperketonemia, when blood glucose values are quently. In a 1-year prospective study enrolling 300 elevated (.13.8 mmol/L – 250 mg/dL – on two con- youth with T1D (7–16 yr old) in our centre, we found secutive readings), or with symptoms of DKA, such as that 24% of the patients checked their blood glucose nausea, vomiting, or abdominal pain (54, 57) (Fig. 1).
twice daily or less often (56). Further, glycemic The cornerstone of sick-day management includes: control improves significantly as the frequency ofblood glucose monitoring increases. Indeed, blood (i) Never omit insulin: Insulin must always be glucose monitoring frequency was the sole modifiable administered during illness, even at times when predictor of glycemic control in this study (56). Thus, eating is markedly diminished as infection induces through increased blood monitoring of glucose and insulin resistance, often necessitating increased or ketones, we have significant opportunity to improve supplemental doses of insulin. The additional or supplemental dose is needed to manage thehyperglycemia and ketosis. Supplemental insulindosages generally consist of 10–20% of the total daily insulin dose administrated every 2–3 h if The most common reason for families to report given as fast-acting analog insulin or every 3–4 h if elevated ketone levels is an infection or illness; another given as regular insulin, based on both the blood frequent cause is failure of insulin delivery resulting glucose and ketone results, using algorithms.
from a pump mishap (pump failure, catheter kinking (ii) Ongoing self-blood-glucose monitoring with adult or slipping out, etc.). Insulin omission is also supervision at least every 2–4 h, occasionally every a common cause for ketosis. Ketones are checked 1–2 h, and with results recorded in a log book.
more commonly in association with illness than with (iii) Monitoring for ketosis every 2–4 h with results hyperglycemia alone (54). The objective of sick-day management in T1D is to minimize metabolic (iv) Continuation of monitoring and supplemental imbalance, avoid severe hypoglycemia with gastroin- testinal illness, and prevent unchecked hyperglycemia (v) Increased intake of salty fluids to combat and ketosis leading to DKA. Sick-day guidelines are dehydration associated with hyperglycemia and taught to all families and, as age appropriate, to possible fever. The blood glucose level determines - Call the health care team foradvice when the blood ketone level until ketone levels are < 0.6mmol/L(see table 2) Fig. 1. Flow chart for ketone checking and treatment of illness or infection [adapted from Burdick et al., Practical Diabetology, 2004 (43)].
Pediatric Diabetes 2007: 8 (Suppl. 6): 24–33 whether sugar-containing or sugar-free fluids mean cost per admission was lowest for DKA should be consumed (usually sugar-free fluids if precipitated by non-compliance in established T1D blood glucose .10 mmol/L (180 mg/dL), sugar- patients than for patients with new onset diabetes containing fluids if blood glucose 10 mmol/L).
or acute illness; nonetheless, this category remains re- (vi) Treatment of any underlying illness.
sponsible for the greatest portion of the economic (vii) Anti-emetics if severe vomiting prevents ad- burden of DKA (1). Furthermore, patients with DKA equate fluid intake, and mental status does not resulting from insulin omission or those with recurrent DKA could potentially increase the costs of hospital- (viii) Frequent contact with the health-care team to ization if length of stay were increased for ongoing psychosocial assessment and management. New tech-nologies like blood ketone monitoring and real-time If illness persists more than a few hours, emergency continuous glucose sensing may provide opportunities to prevent or reduce the occurrence of DKA with Tables 2 and 3 provide specific guidelines based on potential cost savings in those patients and families urine ketone or blood ketone levels. Patient/families willing to utilize such new approaches.
should contact their health-care team if blood ketonelevels persist above 1 mmol/L or urine ketones remain LL has acted as a paid consultant and has receivedinvestigator-initiated funding from Abbott Diabetes Care. EB has declared no conflicts of interest.
Diabetic ketoacidosis is a serious complication ofdiabetes and the leading cause of death in children with diabetes. Delay in the diagnosis and treatment of DKA increases morbidity and can lead to mortality, BALASUBRAMANYAM A. Economic impact of diabetic usually from cerebral edema. The initial diagnosis ketoacidosis in a multiethnic indigent population: of diabetes or DKA may not be straightforward.
analysis of costs based on the precipitating cause.
Some children present with flu-like symptoms, and Diabetes Care 2003: 26: 1265–1269.
2. FISHBEIN H, PALUMBO P. Acute metabolic complications the diagnosis can be missed, especially during flu sea- in diabetes. In: Harris MI, Cowie CC, Stern MP, Boyco son. Prevention of severe metabolic decompensation EJ, Reiber GE, Bennett PH, eds. Diabetes in America.
through sick-day management remains a cornerstone Bethesda, MD: NIH/NIDDK; 1995: 283–292.
of comprehensive diabetes treatment. Even in developed 3. KAUFMAN FR, HALVORSON M. The treatment and prevention of diabetic ketoacidosis in children and countries, children have died from DKA and undiag- adolescents with type I diabetes mellitus. Pediatr Ann nosed diabetes in recent years. Increased awareness of diabetes symptoms should be reinforced among 4. CURTIS JR, TO T, MUIRHEAD S, CUMMINGS E, DANEMAN general practitioners, pediatricians, and in schools.
D. Recent trends in hospitalization for diabetic keto- Simple interventions like posters and educational acidosis in ontario children. Diabetes Care 2002: 25:1591–1596.
leaflets circulated to physician offices and schools 5. SCIBILIA J, FINEGOLD D, DORMAN J, BECKER D, DRASH A.
may help reduce DKA at diagnosis through earlier Why do children with diabetes die? Acta Endocrinol recognition of symptoms. In children with established Suppl (Copenh) 1986: 279: 326–333.
diabetes, education remains the most powerful tool to 6. EDGE JA, FORD-ADAMS ME, DUNGER DB. Causes of death in children with insulin dependent diabetes 1990- prevent DKA. Identification of triggers for metabolic 96. Arch Dis Child 1999: 81: 318–323.
decompensation and knowledge of implementation of 7. GLASER N, BARNETT P, MCCASLIN I et al. Risk factors for sick-day management rules should be reinforced cerebral edema in children with diabetic ketoacidosis.
repeatedly, especially during flu season. Screening The Pediatric Emergency Medicine CollaborativeResearch Committee of the American Academy of for the risk factors for recurrent DKA may identify Pediatrics. N Engl J Med 2001: 344: 264–269.
the child/family in need of specific interventions and 8. WOLFSDORF J, GLASER N, SPERLING MA. Diabetic attention to psychosocial factors. Monitoring 3b- ketoacidosis in infants, children, and adolescents: A OHB levels in patients with DKA can reduce the consensus statement from the American DiabetesAssociation. Diabetes Care 2006: 29: 1150–1159.
length of stay and costs (58). While blood ketone 9. TUNBRIDGE W. Deaths due to diabetic ketoacidosis.
monitoring strips are costlier than urine ketone strips, there remains opportunity for overall cost savings 10. BASU A, CLOSE CF, JENKINS D, KRENTZ AJ, NATTRASS with routine use of blood ketone testing in the M, WRIGHT AD. Persisting mortality in diabeticketoacidosis. Diabet Med 1993: 10: 282–284.
ambulatory and hospitalization settings because of 11. DUNGER DB, SPERLING MA, ACERINI CL et al. European reduced rates of hospitalization and emergency assess- Society for Paediatric Endocrinology/Lawson Wilkins ments and shorter lengths of stay. In one study, the Pediatric Endocrine Society consensus statement on Pediatric Diabetes 2007: 8 (Suppl. 6): 24–33 diabetic ketoacidosis in children and adolescents.
MEMO Collaboration. Diabetes Audit and Research in Tayside Scotland. Medicines Monitoring Unit.
12. LEVY-MARCHAL C, PATTERSON CC, GREEN A. Geograph- ical variation of presentation at diagnosis of type I 30. HANAS R, LUDVIGSSON J. Hypoglycemia and ketoacido- diabetes in children: the EURODIAB study. European sis with insulin pump therapy in children and adoles- and Diabetes. Diabetologia 2001: 44 (Suppl. 3): B75–B80.
cents. Pediatr Diabetes 2006: 7 (Suppl. 4): 32–38.
13. BUI TP, WERTHER GA, CAMERON FJ. Trends in diabetic 31. DANNE T, BATTELINO T, JAROSZ-CHOBOT P. The Ped- ketoacidosis in childhood and adolescence: a 15-yr Pump Study: a low percentage of basal insulin and more experience. Pediatr Diabetes 2002: 3: 82–88.
than five daily boluses are associated with better 14. SMITH CP, FIRTH D, BENNETT S, HOWARD C, CHISHOLM P.
centralized HbA1c in 1041 children on CSII from 17 Ketoacidosis occurring in newly diagnosed and countries. Diabetes 2005: 54 (Suppl. 1): A453.
established diabetic children. Acta Paediatr 1998: 87: 32. HANAS, R, LINDBLAD B, LINDGREN F. A 2-year population study of pediatric ketoacidosis in Sweden: 15. PINKEY JH, BINGLEY PJ, SAWTELL PA, DUNGER DB, predisposing conditions and insulin pump use. Diabetes GALE EA. Presentation and progress of childhood diabetes mellitus: a prospective population-based study.
33. VANELLI M, CHIARI G, GHIZZONI L, COSTI G, GIACALONE T, The Bart’s-Oxford Study Group. Diabetologia 1994: 37: CHIARELLI F. Effectiveness of a prevention program for diabetic ketoacidosis in children. An 8-year study in 16. REWERS A, KLINGENSMITH G, DAVIS C et al. Diabetic schools and private practices. Diabetes Care 1999: 22: ketoacidosis at onset of diabetes: the SEARCH for Diabetes in Youth Study (Abstract). Diabetes 2005: 54 34. BUI H, STEIN R, FUNG K, TO T, DANEMAN D. DKA at diabetes onset may be caused by missed diagnosis especially 17. QUINN M, FLEISCHMAN A, ROSNER B, NIGRIN DJ, in children , 3y. Diabetes 2006: 55 (Suppl. 1): A54.
WOLFSDORF JI. Characteristics at diagnosis of type 1 35. MALLARE JT, CORDICE CC, RYAN BA, CAREY DE, diabetes in children younger than 6 years. J Pediatr KREITZER PM, FRANK GR. Identifying risk factors for the development of diabetic ketoacidosis in new onset 18. KOMULAINEN J, KULMALA P, SAVOLA K et al. Clinical, type 1 diabetes mellitus. Clin Pediatr (Phila) 2003: 42: autoimmune, and genetic characteristics of very young children with type 1 diabetes. Childhood Diabetes in 36. DIABETES PREVENTION TRIAL – TYPE 1 DIABETES STUDY Finland (DiMe) Study Group. Diabetes Care 1999: 22: GROUP. Effects of insulin in relatives of patients with type 1 diabetes mellitus. N Engl J Med 2002; 346: 1685–1691.
19. DIAMOND PROJECT GROUP. Incidence and trends of 37. MAGEIRSDOTTIR H, LARSEN J, BRUNBORG C, DAHL- childhood Type 1 diabetes worldwide 1990-1999. Diabet JORGENSEN K. Nationwide improvement in HBA1c and complication screening in a benchmarking project 20. GREEN A, PATTERSON CC. Trends in the incidence of in childhood diabetes. Pediatr Diabetes 2006: 7: 18.
childhood-onset diabetes in Europe 1989-1998. Diabe- 38. GOLDEN MP, HERROLD AJ, ORR DP. An approach to tologia 2001: 44 (Suppl. 3): B3–B8.
prevention of recurrent diabetic ketoacidosis in the 21. DIABETES CONTROL AND COMPLICATIONS TRIAL RESEARCH pediatric population. J Pediatr 1985: 107: 195–200.
GROUP. Effect of intensive diabetes treatment on the 39. MALDONADO MR, D’AMICO S, RODRIGUEZ L, IYER D, development and progression of long-term complica- BALASUBRAMANYAM A. Improved outcomes in indigent tions in adolescents with insulin-dependent diabetes patients with ketosis-prone diabetes: effect of a dedi- mellitus: J Pediatr 1994: 125: 177–188.
cated diabetes treatment unit. Endocr Pract 2003: 9: 22. REWERS A, CHASE HP, MACKENZIE T et al. Predictors of acute complications in children with type 1 diabetes.
40. STEINDEL BS, ROE TR, COSTIN G, CARLSON M, KAUFMAN FR. Continuous subcutaneous insulin infu- 23. KOVACS M, CHARRON-PROCHOWNIK D, OBROSKY DS. A sion (CSII) in children and adolescents with chronic longitudinal study of biomedical and psychosocial poorly controlled type 1 diabetes mellitus. Diabetes Res predictors of multiple hospitalizations among young people with insulin-dependent diabetes mellitus. Diabet 41. ELLIS DA, FREY MA, NAAR-KING S, TEMPLIN T, CUNNINGHAM P, CAKAN N. Use of multisystemic therapy 24. BLANC N, LUCIDARME N, TUBIANA-RUFI N. [Factors to improve regimen adherence among adolescents with associated with childhood diabetes manifesting as type 1 diabetes in chronic poor metabolic control: ketoacidosis and its severity]. Arch Pediatr 2003: 10: a randomized controlled trial. Diabetes Care 2005: 28: 25. FRANZESE A, VALERIO G, ARGENZIANO A et al. Social 42. ELLIS DA, TEMPLIN T, NAAR-KING S et al. Multisystemic deprivation influences illness onset in diabetic children.
therapy for adolescents with poorly controlled type I diabetes: Stability of treatment effects in a randomized 26. SVOREN B, HOOD K, VOLKENING L et al. Predictors of controlled trial. J Consult Clin Psychol 2007: 75: 168–74.
DKA at onset in pediatric type 1 diabetes. Diabetes 43. BURDICK J, HARRIS S, CHASE P. The importance of ketone testing. Pract Diabetol 2004: 3–11.
27. RODIN GM, DANEMAN D. Eating disorders and IDDM.
44. LAFFEL L. Ketone bodies: a review of physiology, A problematic association. Diabetes Care 1992: 15: pathophysiology and application of monitoring to diabetes. Diabetes Metab Res Rev 1999: 15: 412–426.
28. SKINNER TC. Recurrent diabetic ketoacidosis: causes, 45. GOLDSTEIN DE, LITTLE RR, LORENZ RA et al. Tests of prevention and management. Horm Res 2002: 57 glycemia in diabetes. Diabetes Care 2004: 27: 1761–1773.
46. ATTIA N, JONES TW, HOLCOMBE J, TAMBORLANE WV.
29. MORRIS AD, BOYLE DI, MCMAHON AD, GREENE SA, Comparison of human regular and lispro insulins after MACDONALD TM, NEWTON RW. Adherence to insulin interruption of continuous subcutaneous insulin infu- treatment, glycaemic control, and ketoacidosis in sion and in the treatment of acutely decompensated insulin-dependent diabetes mellitus. The DARTS/ IDDM. Diabetes Care 1998: 21: 817–821.
Pediatric Diabetes 2007: 8 (Suppl. 6): 24–33 47. GUERCI B, MEYER L, SALLE A et al. Comparison of 53. WALLACE TM, MATTHEWS DR. Recent advances in the metabolic deterioration between insulin analog and monitoring and management of diabetic ketoacidosis.
regular insulin after a 5-hour interruption of a con- tinuous subcutaneous insulin infusion in type 1 dia- 54. LAFFEL LM, WENTZELL K, LOUGHLIN C, TOVAR A, betic patients. J Clin Endocrinol Metab 1999: 84: MOLTZ K, BRINK S. Sick day management using blood 3-hydroxybutyrate (3-OHB) compared with urine 48. ORSINI-FEDERICI M, AKWI JA, CANONICO V et al. Early ketone monitoring reduces hospital visits in young detection of insulin deprivation in continuous sub- people with T1DM: a randomized clinical trial. Diabet cutaneous insulin infusion-treated patients with type 1 diabetes. Diabetes Technol Ther 2006: 8: 67–75.
55. LAFFEL L, BRINK S, KAUFMAN FR, BERGENSTAL R, 49. MATTA MP, MELKI V, BESSIERE-LACOMBE S, HANAIRE- FINEGOLD SE, JENKINS M. Frequency of elevation BROUTIN H. What are capillary blood ketone levels in in blood B-hydroxybutyrate (B-OHB) during home type 1 diabetic patients using CSII in normal conditions monitoring and association with glycemia in insulin- of insulin delivery? Diabetes Metab 2004: 30: 543–547.
treated children and adults. Diabetes 2000: 49 (Suppl. 1): 50. REWERS A, MCFANN K, CHASE HP. Bedside monitoring of blood beta-hydroxybutyrate levels in the manage- 56. LEVINE BS, ANDERSON BJ, BUTLER DA, ANTISDEL JE, ment of diabetic ketoacidosis in children. Diabetes BRACKETT J, LAFFEL LM. Predictors of glycemic control and short-term adverse outcomes in youth with type 1 51. BEKTAS F, ERAY O, SARI R, AKBAS H. Point of care diabetes. J Pediatr 2001: 139: 197–203.
blood ketone testing of diabetic patients in the 57. BRINK SJ. Diabetic ketoacidosis. Acta Paediatr 1999: 88 emergency department. Endocr Res 2004: 30: 395–402.
52. HARRIS S, NG R, SYED H, HILLSON R. Near patient 58. VANELLI M, CHIARI G, CAPUANO C. Cost effectiveness of blood ketone measurements and their utility in predict- the direct measurement of 3-beta-hydroxybutyrate in ing diabetic ketoacidosis. Diabet Med 2005: 22: 221– the management of diabetic ketoacidosis in children.
Pediatric Diabetes 2007: 8 (Suppl. 6): 24–33

Source: http://glucomen.co.uk/_assets/how_prevent_dka_2007_laffel.pdf