- Email: [email protected]
Incidence and predictive criteria of nocturnal hypoglycemia in young children with insulin-dependent diabetes mellitus
Incidence and predictive criteria of nocturnal hypoglycemia in young children with insulin-dependent diabetes mellitus Paul A. Porter, FRACP,Barbara Keating, GradDipDiet, Geoffrey Byrne, FRACP, and Timothy W. Jones, FRACP Fromthe Department of Paediatric Endocrinology, PrincessMargaret Hospital for Children, Perth, Australia
Objective: To determine the incidence of significant nocturnal hypoglycemia occurring at home in young children with insulin-dependent diabetes mellitus using conventional therapy. Design: Sixty-one children (aged 2.6 to 8.5 years) were studied on one night, at home, with blood collection occurring at dinner, bedtime/supper, 11 PM,2 AM,and breakfast, with subsequent laboratory analysis. Results: The proportion of subjects with blood glucose levels less than 64, 55, 45, and 36 mg/dl (3.5, 3.0, 2.5, and 2.0 mmol/L) was 37.8%, 17%, 13%, and 8%, respectively. Nocturnal hypoglycemia was associated with younger age (<5 years 57% vs 5 to 8.5 years 36%; p <0.00I) and lowered glycosylated hemoglobin levels (HbAIc) with a greater than 50% incidence of hypoglycemia seen in subjects with HbAIc levels of less than 8.5%. The average HbAIc concentration was lower in the hypoglycemic group than in the nonhypoglycemic group (7.8 vs 8.3%; p <0.02). Nocturnal hypoglycemia occurred with increasing frequency throughout the night in subjects less than 5 years of age (dinner, supper, 11 PM,2 AM, and breakfast incidences being 0%, 12.5%, 26%, 33%, and 30%, respectively) but not in those older than 5 years. Carbohydrate intake at supper did not prevent subsequent hypoglycemia. Blood glucose levels at 11 PMwere poor predictors of subsequent hypoglycemia at 2 AMin either the group as a whole or in the children less than 5 years of age. Symptom recognition of nocturnal hypoglycemia was decreased in younger children (<5 years [36%] >5 years [58%]), in those with a lower HbAIc, and when hypoglycemia occurred at breakfast rather than at dinner (0% vs 50%). Conclusions: The incidence of nocturnal hypoglycemia in young children with insulin-dependent diabetes mellitus receiving conventional therapy is unacceptably high and is increased with lowered age and HbA1c concentration; the condition is often asymptomatic. Early-morning hypoglycemia is poorly predicted by a blood glucose determination at II PMand is not prevented by carbohydrate intake at supper. In younger children, blood glucose profiles should include earlymorning measurements. (J Pediatr 1997;130:366-72)
Submitted for publication Jan. 23, 1996; accepted Aug. 9, 1996. Reprint requests: Paul A. Porter, FRACP, Paediatrician, Princess Margaret Hospital for Children, PO Box D184, Perth 6001, Western Australia. Copyright © 1997 by Mosby-Year Book, Inc. 0022-3476/97/$5.00 + 0 9/21/77290 366
Hypoglycemia remains the most frequent acute complication of insulin-dependent diabetes mellitus in children and often limits attempts to improve glycemic control. 14 Severe hypoglycemia, as evidenced by coma and seizure activity, is more commonly seen during the night than during the day. 4-6 The incidence of less severe but significant nocturnal
The Journal of Pediatrics Volume 130, Number 3
See commentary, p. 339.
BGL HbAIc IDDM NPH
Blood glucose level Glycoslyatedhemoglobin Insulin-dependent diabetes mellitus [Neutral protamine Hagedorn]; isophane insulin suspension
hypoglycemia has been previously reported to be between 14% and 35% in pediatric populations with moderate or poor glycemic control; however, these studies have been limited by a number of factors, including a reliance on the reporting of symptoms;, the use of reflectance meter analysis of blood glucose, and conducting surveys in a hospital setting rather than at home. 2,4' 7-10 Younger children with IDDM are reported to have a higher incide,nce of nocturnal seizures than their older counterparts. Neurodevelopmental difficukies, which may be due to recurrent hypoglycemia, have been reported in children given a diagnosis before the age of 5 years and in those who have had hypoglycemic convulsions.11,12 With greater emphasis on improved glycemic control as a therapeutic goal, there is a concern that such adverse effects may increase. This study was designed to characterize accurately the frequency of nocturnal hypoglycemia in a large group of young children at home who were receiving conventional therapy (twice-daily insulin) and to determine whether there were any predictive criteria for such events. METHODS Subjects. All children less than 8fi years of age who were attending the diabetes clinic at Princess Margaret Hospital for Children were invited to participate in the study and were enrolled after agreement from their parents. Patients with a duration of diabetes of less than 6 months were excluded. Each subject was studied once on a night directly preceding a routine clinic visit. All subjects were on twice-daily subcutaneous insulin regimens using a rnixmre of isophane insulin suspension (NPH insulin) and regular insulin. Parents were requested not to change the subjects' usual routine or insulin dosage before the study night. Seventy patients in this age group were available for enrollment. Two were excluded because of needle (fingerprick) phobia, three because of significant Ongoing psychosocial stressors, and four because they did not return the blood samples. A total of 61 subjects, 31 male and 30 female, aged between 2.6 and 8.5 years (mean, 5.36, SD 1.7) were studied. The mean HbAI~ value of the group was 8.1% (5.8% to 10.3%, SD 0.95%) and the mean duration of diabetes 2.6 years (SD 1.7 years). Procedm'es. Fox" accurate measurement of nocturnal blood glucose levels, laboratory analysis is required. Home
Porter et at.
367
monitoring reflectance-based monitors are not accurate at the extremes of glucose readings expected in patients with diabetes and were considered inadequate for this study. Blood collection was performed at home by the subjects' parents, who used finger pricks and fluoride oxalate tubes. Fifty microliters of blood was collected at each sampling. Samples were collected before the evening meal, before supper or at bedtime (whichever was applicable), at 11 PM, at 2 AM, and before breakfast. The collected samples were then returned to the clinic the following day and analyzed in the department of biochemistry at Princess Margaret Hospital with the use of a Yellow Springs Industries glucose analyzer. This method of analysis was tested and verified for accuracy by a previous pilot study by the authors. This pilot analyzed 30 samples paired with venous blood analyses immediately after collection. The coefficient of variance at a BGL of 64 mg/dl was 4.2%, at 91 mg/dl it was 2.8%, and at 182 mg/dl it was 1.3% The parents of the subjects were asked not to measure the BGLs during the study unless there was evidence of symptomatic hypoglycemia. All episodes of hypoglycemia were treated with the administration of carbohydrate. On the day of the study a questionnaire was completed by the subjects' parents to determine the presence or absence of hypoglycemic symptoms (e.g., tremulousness, sweating, pallor, behavioral and neurocognitive changes) during the study period or the preceding 48 hours and the normality or otherwise of the day's carbohydrate intake, exercise, and insulin dosage. The presence or absence of supper was also determined, and the carbohydrate content was determined by analysis of a food record. Anthropometric data collected the following day included height and weight, and clinical information collected included duration of diabetes and insulin dose. Glycosylated hemoglobin levels were determined with a DCA 2000 (Bayer Diagnostics) analyzer. The HbAlc reference range for subjects without diabetes in our laboratory is less than 6.2%. Statistical analysis. Data analysis used means and standard deviations for descriptive statistics. Populations were compared with the chi-square, Student t, and Mann-Whitney tests. Specificity, sensitivity, and predictive values were calculated as described by Vechio. 13
RESULTS Of the 61 subjects studied, 23 (37.8%) recorded a BGL of less than 64 mg/dl (3.5 mmolFL), 17 (27.8%) less than 55 mg/dl (3.0 mmol/L), 8 (13.1%) less than 45 mg/dl (2.5 mmol/L), and 5 (8%) less than 36 mg/dl (2 mmol/L) at one or more sampling event. Eleven (47%) of those with BGLs less than 64 mg/dl had symptoms associated with their hypoglycemia. Hypoglycemia (arbitrarily defined as a BGL <64 mg/dl [3.5 mmol/L]) was significantly more common in the
368
Porter et al.
The Journal of Pediatrics March 1997
Table I. Nocturnal hypoglycemia (BGL <64 mg/dl [3.5 mmol/L]) profile by age in children with IDDM Age group All subjects (n = 61)
<5 yr (n = 21)
5-8.5 yr (n = 40)
p
8.1 (-+0.95) 23 (37%) 11 (47%)
7.9 (-~0.8) 11 (57%) 4 (36%)
8.2 (-+0.9) 12 (28%) 7 (58%)
NS <0.001" <0.001"
Mean HbAlc (% -+ SD) Hypoglycemic:No. (% of total) Symptomatic:No. (% of total) NS, Not significant. *Chi-square analysis.
Table II. Influence of age on the incidence of
Table Ill. Incidence of nocturnal hypoglycemia according
hypoglycemic episodes (BGL <64 mg/dl [3.5 mmol/L]) at each sample time (% of total samples)
to HbAlc concentration
Age group Time
All subjects (%)
<5 yr (%)
5-8.5 yr (%)
p
Dinner Bed/supper 11PM 2AM
ll 13 16 18
0 12.5 26 33
16 11 8.5 11.4
<0.00l* NS <0.00l* <0.001"
Breakfast
11
30
5.2
HbAIe (%)
Total subjects
<7.5 7.5-8.5 >8.5
16 24 21
Subjects with hypoglycaemia No.
(%)
Subjects with symptoms
8 12 3*
(50) (50) (15)
1 8 1
*p <0.01 by Mann-Whitney test.
<0.001"
NS, *Chi-square analysis.
younger children, with those aged less than 5 years having double the incidence of hypoglycemia compared with those more than 5 years (Table I). Analysis of the distribution of hypoglycemia during the night revealed an even occurrence at each monitoring time, with an incidence of 11% at dinner, 13% at bedfirne/supper, 16% at 11 PM, 18% at 2 AM, and 11% at breakfast. The pattern of hypoglycemia, however, differed according to age. Subjects less than 5 years of age were more fikely to be hypoglycemic at 11 PM, 2 AM, and before breakfast, whereas the subjects aged 5 to 8k1 years had a similar incidence at all sampling times (Table II). The average BGL of the subject group was not significantly different between any sampling period. The mean HbAlc value of the group who became hypoglycemic was lower than the group who did not (7.8% vs 8.3%; p <0.02). The incidence of hypoglycemia remained constant at HbAlc levels of less than 8.5% but fell significantly above this level (Table III). The predictive value of a BGL taken at 11PM for hypoglycemia at 2 AM was determined for the group as a whole and for those subjects less than 5 years of age (Tables IV and V). No single BGL achieved adequate results in all four tests (sensitivity, specificity, positive and negative predictive values). For the group as a whole there was little difference in sensitivity at BGLs between 90 mg/dl (5 mmol/L) and 181 mg/dl (10 mmol/L), whereas specificity was considered adequate at less than 128 mg/dl (7 mmol/L); however, the pos-
itive predictive value was acceptable only at a BGL less than 73 mg/dl (4 retool/L). In the younger subjects there was a similar pattern, with tittle variation in sensitivity between 90 and 181 mg/dl, and no BGL provided an adequate positive predictive value. A BGL o f 100 rag/d1 at supper (bedtime snack) was not helpful in predicting subsequent hypoglycemia (Fig. 1). When only children who did not have symptoms of hypoglycemia, and hence did not receive extra carbohydrate, were analyzed, only two of eight subjects with a BGL of less than 100 mg/dl at supper had values of less than 64 mg/dl at either 11 PN or 2 AM, whereas three were more than 145 mg/dl at 2 AM. In the group of subjects with a BGL of less than 64 mg/dl at 2 AM, 0nly three of eight had BGLs of less than 100 mg/dl at supper (including one of 52 mg/dl), whereas three had BGLs of more than 182 mg/dl (Fig. 2). The duration of hypoglycemia in the symptom-free group was measured by analyzing how many sample times were passed during the hypoglycemic period. Of the 23 subjects with hypoglycemia that began during the study, eight continued to have hypoglycemia for more than one sampling period, with five continuing to have hypoglycemia for more than two periods. Symptomatic awareness of hypoglycemia became less common throughout the night, with 50% of the hypoglycemic episodes being asymptomatic before dinner, 60% at 11 PM, 71% at 2 AM, and 100% at breakfast. There was no relationship between gender, height, weight, duration of diabetes, or insulin dose (measured in units per kilogram) and the development of subsequent hypoglyce-
The Journal of Pediatrics Volume 130, Number 3
Porter et aI.
369
Table IV. Predictive values (%) for nocturnal hypoglycemia (BGL <64 mg/dl [3.5 mmol/L]) at 2 AM with different BGL at 11 PM in all ages
Blood glucose at 11 PM mg/dP
(mmol/L)
Sensitivity
Specificity
Positive predictive value
Negative predictive value
<72 <91 <109 <127 <145 <163 <182 <200 <218
(4) (5) (6) (7) (8) (9) (10) (11) (12)
37 62 62 62 62 62 62 87 100
83 73 71 62 52 50 45 40 35
60 27 36 35 26 25 22 29 29
83 86 86 84 85 84 83 89 93
Table V. Predictive values (%) for nocturnal hypoglycemia at 2 AM (BGL <64 mg/dl [3.5 mmol/L]) with different BGL at 11 PM in the: group less than 5 years of age
Blood glucose at 11 PM mg/dl
mmol/L
Sensitivity
Specificity
Positive predictive value
Negative predictive value
<72 <91 <109 <127 <145 <163 <182 <200
(4) (5) (6) (7) (8) (9) (10) (11)
50 75 75 75 75 75 75 100
75 66 41 41 41 41 33 33
40 42 42 30 27 27 27 30
83 90 90 85 83 83 83 100
mia. All subjects were analyzed as to the normality of the day preceding the study night. No subject recorded an alteration in insulin dose, carbohydrate intake, or timing of either meals or insulin administration. Two subjects had an intercurrent upper respiratory tract infection that did not require adjustment of their usual regimen. Neither of these subjects subsequently had hypoglycemia. Five subjects reported hypoglycemic symptoms during the 12 hours before the study. None of these had hypoglycemia during the study. The ingestion of a carbohydrate-containing meal at bedtime (supper) was not associated with a lower frequency of subsequent hypoglycernia, with 13 (27%) of 47 subjects who consumed an adequate carbohydrate-containing supper having subsequent hypoglYcemia compared with 3 (21%) of 14 who did not have supper. DISCUSSION Nocturnal hypoglycemia is a common and unpredictable complication in adults and older children with IDDM. In this study, designed to measure the incidence of nocturnal hypoglycernia in the young child with IDDM, in whom the adverse effects of such hypoglycemia may be more significant, we found that in a group of children receiving
conventional therapy there is an unacceptably high rate of nocturnal hypoglycemia. For example, on a single night, glucose values of less than 64 mg/dl (3.5 retool/L) occurred in 57% of the group less than 5 years of age and in 28.5% of the group aged 5 to 8fi years. Nocturnal hypoglycemia was more likely to occur in children younger than the age of 5 years and with better glycemic control. Because most episodes of hypoglycemia were asymptomatic, previous studies estimating the incidence of hypoglycemia by monitoring the frequency of symptomatic episodes have underestimated the mae frequency of the complication. 2, 4, 8 Previous studies investigating nocturnal hypoglycemia in pediatric patients with diabetes have mainly examined older children and with poorer glycemic control. Shalwitz et at.7 studied 135 patients (mean age, 14 years; mean HbAlc value, 9.4% to 10.6%) and found that hypoglycemia (BGL <60 mg/dl) occurred at an incidence of 14.4% at 2 AM and 6.7% at 6 AM, whereas Whincup and Milner9 found a 34% incidence of hypoglycemia at 2 AM (at age 3 to 16 years, HbAlc value 7.4% to 11.4%). Simell et at. 10 studied 23 children less than 8 years of age with good glycemic control and found a 14% incidence of hypoglycemia (<55 mg/dl) at 3 AM in the 2- to 4-year-old subjects and a 50% incidence in the 5- to
370
Porter et al.
The Journal of Pediatrics March 1997
290
218-
182-
/
I
.
.4 2300 TIME
B e d t i m e snack
(1915-1945)
0200
Fig. l. Blood glucose profiles for subjects with BGL of less than 100 mg/dl at supper (excluding those treated for symptomatic hypoglycemia at these times).
2
1
8
~
BGL (mg/dl) ~
36 L
0+Bedtime snack (1915-1945)
~
2300
/
0200
TIME Fig. 9. Blood glucose profiles for subjects with BGL of less than 64 mg/dl at 2 AM (excluding those treated for symptomatic hypoglycemia at these times).
8-year-old subjects. All three Of these studies, however, examined their subjects in the hospital, a setting that may not mimic the home environment. There are no previously pubfished data on the incidence of hypoglycemia occurring at home under normal conditions in the very Young diabetic group, so direct comparisons from the studies cited above cannot be drawn to the children included in the present study. By studying our subjects at home, we were able to remove the influence of confounding variables such as differences in exercise, carbohydrate intake, and anxiety levels that may occur in the hospital. It is possible that the very act of studying these children resulted in subtle changes in the daily regimens not detected by the questionnaire; however, no difference could be found in the major influences on BGLs,
including the timing and amount of insulin dose and carbohydrate intake. Nocturnal hypoglycemia was associated in those children with an HbAlc value of less than 8.5%. This is in keeping with the findings of other studies, such as the Diabetes Control and Complications Trial, which documented a close relationship between glycemic control and the incidence of significant hypoglycemia. 18The effect of fight control would seem to be as important in the younger child. Nocturnal hypoglycemia has been reported to be more frequent between 2 AM and 3 AM than at other times during the night in children using mixed NPH insulin and regular insulin, which may be explained on the basis of the dynamics of insulin action. 7, 9, 10 We did not find evidence to sup-
The Journal ~.f Pediatrics Volume 130, Number 3
port this view, particularly in the younger child, in whom a progressive increase in the incidence of hypoglycemia was found throughout the night, peaking before breakfast. The absence of an early-morning rise in BGLs raises the possibility that the dawn phenomenon is not as significant in the younger child as in the older age group. The dawn phenomenon has been related to the effect of growth hormone. Because growth hormone levels are lower in the younger child than in the adolescent, it is tempting to speculate whether this difference explains the lack of an early-morning rise in BGL in younger subjects. Studies in older children have suggested that BGLs obtained during the evening are predictive of subsequent hypoglycemia.7, 9 We found that although a BGL of less than 128 mg/dl (7 mmol/L) at 11 PM is reasonably sensitive and specific for hypoglycemia at 2 aM, the positive predictive value of any particular BGL at this time is poor. A BGL of less than 100 mg/dl at supper has been used in predicting the likelihood of subsequent hypoglycemia and as an indication to consume more carbohydrate as a preventive measure. We have not been able to support this notion. Of the subjects whose BGL was less than 64 mg/dl at 2 aM, more had BGLs of greater than 100 mg/dl at both supper and 11 PM, with one subject falling from 218 mg/dl despite normal insulin and carbohydrate intake. In combination with the decreased awareness of symptoms that occurs during early-morning hypoglycemia, the observations noted above indicate that direct blood glucose monitoring is needed to exclude hypoglycemia at this time. This reinforces the recommendation to moifitor regularly the early-morning BGLs in the young child[ with IDDM. The use of mixed NPH insulin and regular insulin at night may be contributing to the incidence of nocturnal hypoglycemia because of a combined action occurring during the early hours of the morning. A possible solution is the introduction of a split insulin regimen at night (i.e., regular insulin at dinner and NPH insulin at supper) in the younger child, as is recommended in older patients. Many children appeared to their caregivers to be free of symptoms. This was especially so in younger children and was surprising because children without diabetes often have symptoms of hypoglycemia at BGLs of 64 to 73 mg/dl (3.5 to 4 mmol/L), t4, 15 The lack of signs of hypoglycemia suggests that children, like adults, may have hypoglycemia in association with autonomic failure. Whereas the hormonal counterregulation to diurnal hypoglycemia has been found to be maintained in the young child with diabetes, responses to nocturnal hypoglycemia have not been investigated. 16 Hypoglycemia per se may decrease the response to and recognition of s~absequent hypoglycemia. 17 No subject in this study was known to have hypoglycemia in the 12 hours preceding the monitored period; however, asymptomatic hy-
Porter et al.
37 1
poglycemia could not be excluded. No subject had symptoms at breakfast, a circumstance that again emphasizes the need for frequent routine glucose monitoring in this age group. It has been suggested that early onset of IDDM is associated with neurocognitive deficits as a result of hypoglycemia) 1 Hypoglycemic convulsions have been shown to be associated with lowered mean performance IQ scores in boys. 12 In view of the frequency of severe hypoglycemia found in this study, including very low blood glucose readings (18 mg/dl [1 retool/L]), questions are raised as to the possible neurocognitive sequelae. Longitudinal studies are needed to delineate clearly what level of hypoglycemia is associated with cognitive changes in this age group and to determine the effects of :frequent, ongoing hypoglycemia on the developing brain during an extended period. As diabetologists continue to pursue improved glycemic control in an attempt to decrease the incidence of long-term complications, care needs to be taken, especially in younger children, to prevent a new range of potentially disabling neurocognitire deficits. We thank Dr. Peter O'Leary (Biochemist, Princess Margaret Hospital) for his invaluable help and advice during this study. REFERENCES
1. Cryer P, Gefich J. Glucose counterregulation, hypoglycemia and intensive insulin therapy in diabetes mellitus. N Engl J Meal 1985;313:232-41. 2. Maefarlane PI, Wakers M, Stutchfield P, Smith CS. A prospective study of symptomatic hypoglycemia in childhood diabetes. Diabetic Medicine 1989;6:62%30. 3. Soltesz G. Hypoglycaemia in the diabetic child. Bailli~res Clin Endocrinol Metab 1993;7:741-55. 4. Ward CM, Stewart AW, Cntfield RG. Hypoglycaemia in insulin- dependent diabetic patients attending an outpatients' clinic. N Z Med J 1990;103:339-41. 5. Bendtson I, Rosenfalck AM, Binder C. Nocturnal versus diurnal hormonal counter-regulation to hypoglycaemia in type 1 (insulin-dependent) diabetic patients. Acta Endocrinol 1993; 128:109-15. 6. Fisher BM, Frier BM. Nocturnal convulsions and insulininduced hypoglycaemia in diabetic patients. Postgrad Med J 1987;63:673-6. 7. Shalwitz RA, Farkas-Hirsch R, White NH, Santiago JV. Prevalence and consequences of nocturnal hypoglycemia among conventionally treated children with diabetes mellitus. J Pediatr 1990;116:685-9. 8. Daneman D, Frank M, Peflman K, Tamm J, Ehrlich R. Severe hypoglycemiain children with insulin dependent diabetes mellitus: frequency and predisposing factors. J Pediatr 1989;115: 681-5. 9. Whincup G, Milner R. Prediction and management of nocturnal hypoglycaemia in diabetes. Arch Dis Child 1987;62:333-7. 10. Simell T, Simell O, Cammi E, Caprio E, Hakulinen A, Haxalax J, et al. Glucose profiles in children two years after onset of type-1 diabetes. Diabetic Medicine 1993;10:524-9. 11. Ryan C, Vega A, Drash A. Cognitive deficits in adolescents
372
Porter et al.
The Journal of Pediatrics March 1997
who developed diabetes early in life. Pediatrics 1985;75:921-7. 12. Rovet JF, Ehrlich RM, Hoppe M. Intellectual deficits associated with early onset of IDDM in children. Diabetes Care 1987;10:510-5. 13. VechioTJ. Predictive value of a single diagnostic test in unselected populations. N Engl J Med 1966;3:206-7. 14. Jones TW, Boulware SD, Kramer DK, Caprio S, Sherwin RS, Tamborlane WV. Independent effects of youth and poor diabetes control on responses to hypoglycaemia in children. Dia, betes 1991;40:358-61. 15. Jones TW, Boulware SD, McCarthy G, Sherwin RS, Tamborlane WV. Enhanced adrenomedullaryresponse and increased susceptibility to neuroglycopenia:mechanisms underlying the
adverse effects of sugar ingestion in healthy children. J Pediatr 1995;126:171-7. 16. Brambilla P, Bougneres PF, Santiago JV, Chaussain JL, Pouplard A, Castano L. Glucose counterregulationin pre-schoolaged diabetic children with recurrent hypoglycaemia during conventional treatment. Diabetes 1987;36:300-4. 17. Veneman T, Mitrakou A, Mokan M, Cryer P, Gerich J. Induction of hypoglycaemic unawareness by asymptomatic noctttrnal hypoglycaemia. Diabetes 1993;42:1233-7. 18. The Diabetes Control and Complications Trial. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 1993;329:977-86.
50 Years Ago in The Journal of Pediatrics Children born during the siege of Leningrad in 1942 Antonov AN. J Pediatr 1947;30:250-9. From August 1941 through January 1943 the city of Leningrad was under almost continous siege fr the German army. The winter of 1941-1942 was particularly harsh, and food resources were scant The records of the newborn department at the Leningrad State Pediatric Institute were studied to tablish the effect of this military operation on pregnancy and newborn health. At the height of the sic the stillbirth rate doubled to 5.6%, and the rate of prematurity reached 41.2%. The term infant morta rate reached 9%, and 30.8% of prematurely born infants died. Even among term infants, birth wei8 were compromised. In early 1942, for example, the average birth weight of term infants was 500 to gm less than the previously established norms for Leningrad. Other than the stress of continual warfi two factors were impficated in this tragedy. First, the food ration was severely limited, amount to between 150 and 500 gm of poor-quality bread daily. Second, fuel shortages resulted in signific cold stress to infants. At one point the temperature in the newborn unit was 40 ° F and the staff worl in overcoats. Comment: During the past 50 years, the pages of The Journal have chronicled a host of dramatic vances in newborn care. Little has changed, however, with the politics of warfare. Infants continue suffer disproportionately. One has to wonder if comparable data are available for, as an example, the lq kan peninsula or Iraq.
Thomas R. Welch, Division of Pediatric Nephroh Children's Hospital Medical Cen Cincinnati, OH 45229-3(
Objective: To determine the incidence of significant nocturnal hypoglycemia occurring at home in young children with insulin-dependent diabetes mellitus using conventional therapy. Design: Sixty-one children (aged 2.6 to 8.5 years) were studied on one night, at home, with blood collection occurring at dinner, bedtime/supper, 11 PM,2 AM,and breakfast, with subsequent laboratory analysis. Results: The proportion of subjects with blood glucose levels less than 64, 55, 45, and 36 mg/dl (3.5, 3.0, 2.5, and 2.0 mmol/L) was 37.8%, 17%, 13%, and 8%, respectively. Nocturnal hypoglycemia was associated with younger age (<5 years 57% vs 5 to 8.5 years 36%; p <0.00I) and lowered glycosylated hemoglobin levels (HbAIc) with a greater than 50% incidence of hypoglycemia seen in subjects with HbAIc levels of less than 8.5%. The average HbAIc concentration was lower in the hypoglycemic group than in the nonhypoglycemic group (7.8 vs 8.3%; p <0.02). Nocturnal hypoglycemia occurred with increasing frequency throughout the night in subjects less than 5 years of age (dinner, supper, 11 PM,2 AM, and breakfast incidences being 0%, 12.5%, 26%, 33%, and 30%, respectively) but not in those older than 5 years. Carbohydrate intake at supper did not prevent subsequent hypoglycemia. Blood glucose levels at 11 PMwere poor predictors of subsequent hypoglycemia at 2 AMin either the group as a whole or in the children less than 5 years of age. Symptom recognition of nocturnal hypoglycemia was decreased in younger children (<5 years [36%] >5 years [58%]), in those with a lower HbAIc, and when hypoglycemia occurred at breakfast rather than at dinner (0% vs 50%). Conclusions: The incidence of nocturnal hypoglycemia in young children with insulin-dependent diabetes mellitus receiving conventional therapy is unacceptably high and is increased with lowered age and HbA1c concentration; the condition is often asymptomatic. Early-morning hypoglycemia is poorly predicted by a blood glucose determination at II PMand is not prevented by carbohydrate intake at supper. In younger children, blood glucose profiles should include earlymorning measurements. (J Pediatr 1997;130:366-72)
Submitted for publication Jan. 23, 1996; accepted Aug. 9, 1996. Reprint requests: Paul A. Porter, FRACP, Paediatrician, Princess Margaret Hospital for Children, PO Box D184, Perth 6001, Western Australia. Copyright © 1997 by Mosby-Year Book, Inc. 0022-3476/97/$5.00 + 0 9/21/77290 366
Hypoglycemia remains the most frequent acute complication of insulin-dependent diabetes mellitus in children and often limits attempts to improve glycemic control. 14 Severe hypoglycemia, as evidenced by coma and seizure activity, is more commonly seen during the night than during the day. 4-6 The incidence of less severe but significant nocturnal
The Journal of Pediatrics Volume 130, Number 3
See commentary, p. 339.
BGL HbAIc IDDM NPH
Blood glucose level Glycoslyatedhemoglobin Insulin-dependent diabetes mellitus [Neutral protamine Hagedorn]; isophane insulin suspension
hypoglycemia has been previously reported to be between 14% and 35% in pediatric populations with moderate or poor glycemic control; however, these studies have been limited by a number of factors, including a reliance on the reporting of symptoms;, the use of reflectance meter analysis of blood glucose, and conducting surveys in a hospital setting rather than at home. 2,4' 7-10 Younger children with IDDM are reported to have a higher incide,nce of nocturnal seizures than their older counterparts. Neurodevelopmental difficukies, which may be due to recurrent hypoglycemia, have been reported in children given a diagnosis before the age of 5 years and in those who have had hypoglycemic convulsions.11,12 With greater emphasis on improved glycemic control as a therapeutic goal, there is a concern that such adverse effects may increase. This study was designed to characterize accurately the frequency of nocturnal hypoglycemia in a large group of young children at home who were receiving conventional therapy (twice-daily insulin) and to determine whether there were any predictive criteria for such events. METHODS Subjects. All children less than 8fi years of age who were attending the diabetes clinic at Princess Margaret Hospital for Children were invited to participate in the study and were enrolled after agreement from their parents. Patients with a duration of diabetes of less than 6 months were excluded. Each subject was studied once on a night directly preceding a routine clinic visit. All subjects were on twice-daily subcutaneous insulin regimens using a rnixmre of isophane insulin suspension (NPH insulin) and regular insulin. Parents were requested not to change the subjects' usual routine or insulin dosage before the study night. Seventy patients in this age group were available for enrollment. Two were excluded because of needle (fingerprick) phobia, three because of significant Ongoing psychosocial stressors, and four because they did not return the blood samples. A total of 61 subjects, 31 male and 30 female, aged between 2.6 and 8.5 years (mean, 5.36, SD 1.7) were studied. The mean HbAI~ value of the group was 8.1% (5.8% to 10.3%, SD 0.95%) and the mean duration of diabetes 2.6 years (SD 1.7 years). Procedm'es. Fox" accurate measurement of nocturnal blood glucose levels, laboratory analysis is required. Home
Porter et at.
367
monitoring reflectance-based monitors are not accurate at the extremes of glucose readings expected in patients with diabetes and were considered inadequate for this study. Blood collection was performed at home by the subjects' parents, who used finger pricks and fluoride oxalate tubes. Fifty microliters of blood was collected at each sampling. Samples were collected before the evening meal, before supper or at bedtime (whichever was applicable), at 11 PM, at 2 AM, and before breakfast. The collected samples were then returned to the clinic the following day and analyzed in the department of biochemistry at Princess Margaret Hospital with the use of a Yellow Springs Industries glucose analyzer. This method of analysis was tested and verified for accuracy by a previous pilot study by the authors. This pilot analyzed 30 samples paired with venous blood analyses immediately after collection. The coefficient of variance at a BGL of 64 mg/dl was 4.2%, at 91 mg/dl it was 2.8%, and at 182 mg/dl it was 1.3% The parents of the subjects were asked not to measure the BGLs during the study unless there was evidence of symptomatic hypoglycemia. All episodes of hypoglycemia were treated with the administration of carbohydrate. On the day of the study a questionnaire was completed by the subjects' parents to determine the presence or absence of hypoglycemic symptoms (e.g., tremulousness, sweating, pallor, behavioral and neurocognitive changes) during the study period or the preceding 48 hours and the normality or otherwise of the day's carbohydrate intake, exercise, and insulin dosage. The presence or absence of supper was also determined, and the carbohydrate content was determined by analysis of a food record. Anthropometric data collected the following day included height and weight, and clinical information collected included duration of diabetes and insulin dose. Glycosylated hemoglobin levels were determined with a DCA 2000 (Bayer Diagnostics) analyzer. The HbAlc reference range for subjects without diabetes in our laboratory is less than 6.2%. Statistical analysis. Data analysis used means and standard deviations for descriptive statistics. Populations were compared with the chi-square, Student t, and Mann-Whitney tests. Specificity, sensitivity, and predictive values were calculated as described by Vechio. 13
RESULTS Of the 61 subjects studied, 23 (37.8%) recorded a BGL of less than 64 mg/dl (3.5 mmolFL), 17 (27.8%) less than 55 mg/dl (3.0 mmol/L), 8 (13.1%) less than 45 mg/dl (2.5 mmol/L), and 5 (8%) less than 36 mg/dl (2 mmol/L) at one or more sampling event. Eleven (47%) of those with BGLs less than 64 mg/dl had symptoms associated with their hypoglycemia. Hypoglycemia (arbitrarily defined as a BGL <64 mg/dl [3.5 mmol/L]) was significantly more common in the
368
Porter et al.
The Journal of Pediatrics March 1997
Table I. Nocturnal hypoglycemia (BGL <64 mg/dl [3.5 mmol/L]) profile by age in children with IDDM Age group All subjects (n = 61)
<5 yr (n = 21)
5-8.5 yr (n = 40)
p
8.1 (-+0.95) 23 (37%) 11 (47%)
7.9 (-~0.8) 11 (57%) 4 (36%)
8.2 (-+0.9) 12 (28%) 7 (58%)
NS <0.001" <0.001"
Mean HbAlc (% -+ SD) Hypoglycemic:No. (% of total) Symptomatic:No. (% of total) NS, Not significant. *Chi-square analysis.
Table II. Influence of age on the incidence of
Table Ill. Incidence of nocturnal hypoglycemia according
hypoglycemic episodes (BGL <64 mg/dl [3.5 mmol/L]) at each sample time (% of total samples)
to HbAlc concentration
Age group Time
All subjects (%)
<5 yr (%)
5-8.5 yr (%)
p
Dinner Bed/supper 11PM 2AM
ll 13 16 18
0 12.5 26 33
16 11 8.5 11.4
<0.00l* NS <0.00l* <0.001"
Breakfast
11
30
5.2
HbAIe (%)
Total subjects
<7.5 7.5-8.5 >8.5
16 24 21
Subjects with hypoglycaemia No.
(%)
Subjects with symptoms
8 12 3*
(50) (50) (15)
1 8 1
*p <0.01 by Mann-Whitney test.
<0.001"
NS, *Chi-square analysis.
younger children, with those aged less than 5 years having double the incidence of hypoglycemia compared with those more than 5 years (Table I). Analysis of the distribution of hypoglycemia during the night revealed an even occurrence at each monitoring time, with an incidence of 11% at dinner, 13% at bedfirne/supper, 16% at 11 PM, 18% at 2 AM, and 11% at breakfast. The pattern of hypoglycemia, however, differed according to age. Subjects less than 5 years of age were more fikely to be hypoglycemic at 11 PM, 2 AM, and before breakfast, whereas the subjects aged 5 to 8k1 years had a similar incidence at all sampling times (Table II). The average BGL of the subject group was not significantly different between any sampling period. The mean HbAlc value of the group who became hypoglycemic was lower than the group who did not (7.8% vs 8.3%; p <0.02). The incidence of hypoglycemia remained constant at HbAlc levels of less than 8.5% but fell significantly above this level (Table III). The predictive value of a BGL taken at 11PM for hypoglycemia at 2 AM was determined for the group as a whole and for those subjects less than 5 years of age (Tables IV and V). No single BGL achieved adequate results in all four tests (sensitivity, specificity, positive and negative predictive values). For the group as a whole there was little difference in sensitivity at BGLs between 90 mg/dl (5 mmol/L) and 181 mg/dl (10 mmol/L), whereas specificity was considered adequate at less than 128 mg/dl (7 mmol/L); however, the pos-
itive predictive value was acceptable only at a BGL less than 73 mg/dl (4 retool/L). In the younger subjects there was a similar pattern, with tittle variation in sensitivity between 90 and 181 mg/dl, and no BGL provided an adequate positive predictive value. A BGL o f 100 rag/d1 at supper (bedtime snack) was not helpful in predicting subsequent hypoglycemia (Fig. 1). When only children who did not have symptoms of hypoglycemia, and hence did not receive extra carbohydrate, were analyzed, only two of eight subjects with a BGL of less than 100 mg/dl at supper had values of less than 64 mg/dl at either 11 PN or 2 AM, whereas three were more than 145 mg/dl at 2 AM. In the group of subjects with a BGL of less than 64 mg/dl at 2 AM, 0nly three of eight had BGLs of less than 100 mg/dl at supper (including one of 52 mg/dl), whereas three had BGLs of more than 182 mg/dl (Fig. 2). The duration of hypoglycemia in the symptom-free group was measured by analyzing how many sample times were passed during the hypoglycemic period. Of the 23 subjects with hypoglycemia that began during the study, eight continued to have hypoglycemia for more than one sampling period, with five continuing to have hypoglycemia for more than two periods. Symptomatic awareness of hypoglycemia became less common throughout the night, with 50% of the hypoglycemic episodes being asymptomatic before dinner, 60% at 11 PM, 71% at 2 AM, and 100% at breakfast. There was no relationship between gender, height, weight, duration of diabetes, or insulin dose (measured in units per kilogram) and the development of subsequent hypoglyce-
The Journal of Pediatrics Volume 130, Number 3
Porter et aI.
369
Table IV. Predictive values (%) for nocturnal hypoglycemia (BGL <64 mg/dl [3.5 mmol/L]) at 2 AM with different BGL at 11 PM in all ages
Blood glucose at 11 PM mg/dP
(mmol/L)
Sensitivity
Specificity
Positive predictive value
Negative predictive value
<72 <91 <109 <127 <145 <163 <182 <200 <218
(4) (5) (6) (7) (8) (9) (10) (11) (12)
37 62 62 62 62 62 62 87 100
83 73 71 62 52 50 45 40 35
60 27 36 35 26 25 22 29 29
83 86 86 84 85 84 83 89 93
Table V. Predictive values (%) for nocturnal hypoglycemia at 2 AM (BGL <64 mg/dl [3.5 mmol/L]) with different BGL at 11 PM in the: group less than 5 years of age
Blood glucose at 11 PM mg/dl
mmol/L
Sensitivity
Specificity
Positive predictive value
Negative predictive value
<72 <91 <109 <127 <145 <163 <182 <200
(4) (5) (6) (7) (8) (9) (10) (11)
50 75 75 75 75 75 75 100
75 66 41 41 41 41 33 33
40 42 42 30 27 27 27 30
83 90 90 85 83 83 83 100
mia. All subjects were analyzed as to the normality of the day preceding the study night. No subject recorded an alteration in insulin dose, carbohydrate intake, or timing of either meals or insulin administration. Two subjects had an intercurrent upper respiratory tract infection that did not require adjustment of their usual regimen. Neither of these subjects subsequently had hypoglycemia. Five subjects reported hypoglycemic symptoms during the 12 hours before the study. None of these had hypoglycemia during the study. The ingestion of a carbohydrate-containing meal at bedtime (supper) was not associated with a lower frequency of subsequent hypoglycernia, with 13 (27%) of 47 subjects who consumed an adequate carbohydrate-containing supper having subsequent hypoglYcemia compared with 3 (21%) of 14 who did not have supper. DISCUSSION Nocturnal hypoglycemia is a common and unpredictable complication in adults and older children with IDDM. In this study, designed to measure the incidence of nocturnal hypoglycernia in the young child with IDDM, in whom the adverse effects of such hypoglycemia may be more significant, we found that in a group of children receiving
conventional therapy there is an unacceptably high rate of nocturnal hypoglycemia. For example, on a single night, glucose values of less than 64 mg/dl (3.5 retool/L) occurred in 57% of the group less than 5 years of age and in 28.5% of the group aged 5 to 8fi years. Nocturnal hypoglycemia was more likely to occur in children younger than the age of 5 years and with better glycemic control. Because most episodes of hypoglycemia were asymptomatic, previous studies estimating the incidence of hypoglycemia by monitoring the frequency of symptomatic episodes have underestimated the mae frequency of the complication. 2, 4, 8 Previous studies investigating nocturnal hypoglycemia in pediatric patients with diabetes have mainly examined older children and with poorer glycemic control. Shalwitz et at.7 studied 135 patients (mean age, 14 years; mean HbAlc value, 9.4% to 10.6%) and found that hypoglycemia (BGL <60 mg/dl) occurred at an incidence of 14.4% at 2 AM and 6.7% at 6 AM, whereas Whincup and Milner9 found a 34% incidence of hypoglycemia at 2 AM (at age 3 to 16 years, HbAlc value 7.4% to 11.4%). Simell et at. 10 studied 23 children less than 8 years of age with good glycemic control and found a 14% incidence of hypoglycemia (<55 mg/dl) at 3 AM in the 2- to 4-year-old subjects and a 50% incidence in the 5- to
370
Porter et al.
The Journal of Pediatrics March 1997
290
218-
182-
/
I
.
.4 2300 TIME
B e d t i m e snack
(1915-1945)
0200
Fig. l. Blood glucose profiles for subjects with BGL of less than 100 mg/dl at supper (excluding those treated for symptomatic hypoglycemia at these times).
2
1
8
~
BGL (mg/dl) ~
36 L
0+Bedtime snack (1915-1945)
~
2300
/
0200
TIME Fig. 9. Blood glucose profiles for subjects with BGL of less than 64 mg/dl at 2 AM (excluding those treated for symptomatic hypoglycemia at these times).
8-year-old subjects. All three Of these studies, however, examined their subjects in the hospital, a setting that may not mimic the home environment. There are no previously pubfished data on the incidence of hypoglycemia occurring at home under normal conditions in the very Young diabetic group, so direct comparisons from the studies cited above cannot be drawn to the children included in the present study. By studying our subjects at home, we were able to remove the influence of confounding variables such as differences in exercise, carbohydrate intake, and anxiety levels that may occur in the hospital. It is possible that the very act of studying these children resulted in subtle changes in the daily regimens not detected by the questionnaire; however, no difference could be found in the major influences on BGLs,
including the timing and amount of insulin dose and carbohydrate intake. Nocturnal hypoglycemia was associated in those children with an HbAlc value of less than 8.5%. This is in keeping with the findings of other studies, such as the Diabetes Control and Complications Trial, which documented a close relationship between glycemic control and the incidence of significant hypoglycemia. 18The effect of fight control would seem to be as important in the younger child. Nocturnal hypoglycemia has been reported to be more frequent between 2 AM and 3 AM than at other times during the night in children using mixed NPH insulin and regular insulin, which may be explained on the basis of the dynamics of insulin action. 7, 9, 10 We did not find evidence to sup-
The Journal ~.f Pediatrics Volume 130, Number 3
port this view, particularly in the younger child, in whom a progressive increase in the incidence of hypoglycemia was found throughout the night, peaking before breakfast. The absence of an early-morning rise in BGLs raises the possibility that the dawn phenomenon is not as significant in the younger child as in the older age group. The dawn phenomenon has been related to the effect of growth hormone. Because growth hormone levels are lower in the younger child than in the adolescent, it is tempting to speculate whether this difference explains the lack of an early-morning rise in BGL in younger subjects. Studies in older children have suggested that BGLs obtained during the evening are predictive of subsequent hypoglycemia.7, 9 We found that although a BGL of less than 128 mg/dl (7 mmol/L) at 11 PM is reasonably sensitive and specific for hypoglycemia at 2 aM, the positive predictive value of any particular BGL at this time is poor. A BGL of less than 100 mg/dl at supper has been used in predicting the likelihood of subsequent hypoglycemia and as an indication to consume more carbohydrate as a preventive measure. We have not been able to support this notion. Of the subjects whose BGL was less than 64 mg/dl at 2 aM, more had BGLs of greater than 100 mg/dl at both supper and 11 PM, with one subject falling from 218 mg/dl despite normal insulin and carbohydrate intake. In combination with the decreased awareness of symptoms that occurs during early-morning hypoglycemia, the observations noted above indicate that direct blood glucose monitoring is needed to exclude hypoglycemia at this time. This reinforces the recommendation to moifitor regularly the early-morning BGLs in the young child[ with IDDM. The use of mixed NPH insulin and regular insulin at night may be contributing to the incidence of nocturnal hypoglycemia because of a combined action occurring during the early hours of the morning. A possible solution is the introduction of a split insulin regimen at night (i.e., regular insulin at dinner and NPH insulin at supper) in the younger child, as is recommended in older patients. Many children appeared to their caregivers to be free of symptoms. This was especially so in younger children and was surprising because children without diabetes often have symptoms of hypoglycemia at BGLs of 64 to 73 mg/dl (3.5 to 4 mmol/L), t4, 15 The lack of signs of hypoglycemia suggests that children, like adults, may have hypoglycemia in association with autonomic failure. Whereas the hormonal counterregulation to diurnal hypoglycemia has been found to be maintained in the young child with diabetes, responses to nocturnal hypoglycemia have not been investigated. 16 Hypoglycemia per se may decrease the response to and recognition of s~absequent hypoglycemia. 17 No subject in this study was known to have hypoglycemia in the 12 hours preceding the monitored period; however, asymptomatic hy-
Porter et al.
37 1
poglycemia could not be excluded. No subject had symptoms at breakfast, a circumstance that again emphasizes the need for frequent routine glucose monitoring in this age group. It has been suggested that early onset of IDDM is associated with neurocognitive deficits as a result of hypoglycemia) 1 Hypoglycemic convulsions have been shown to be associated with lowered mean performance IQ scores in boys. 12 In view of the frequency of severe hypoglycemia found in this study, including very low blood glucose readings (18 mg/dl [1 retool/L]), questions are raised as to the possible neurocognitive sequelae. Longitudinal studies are needed to delineate clearly what level of hypoglycemia is associated with cognitive changes in this age group and to determine the effects of :frequent, ongoing hypoglycemia on the developing brain during an extended period. As diabetologists continue to pursue improved glycemic control in an attempt to decrease the incidence of long-term complications, care needs to be taken, especially in younger children, to prevent a new range of potentially disabling neurocognitire deficits. We thank Dr. Peter O'Leary (Biochemist, Princess Margaret Hospital) for his invaluable help and advice during this study. REFERENCES
1. Cryer P, Gefich J. Glucose counterregulation, hypoglycemia and intensive insulin therapy in diabetes mellitus. N Engl J Meal 1985;313:232-41. 2. Maefarlane PI, Wakers M, Stutchfield P, Smith CS. A prospective study of symptomatic hypoglycemia in childhood diabetes. Diabetic Medicine 1989;6:62%30. 3. Soltesz G. Hypoglycaemia in the diabetic child. Bailli~res Clin Endocrinol Metab 1993;7:741-55. 4. Ward CM, Stewart AW, Cntfield RG. Hypoglycaemia in insulin- dependent diabetic patients attending an outpatients' clinic. N Z Med J 1990;103:339-41. 5. Bendtson I, Rosenfalck AM, Binder C. Nocturnal versus diurnal hormonal counter-regulation to hypoglycaemia in type 1 (insulin-dependent) diabetic patients. Acta Endocrinol 1993; 128:109-15. 6. Fisher BM, Frier BM. Nocturnal convulsions and insulininduced hypoglycaemia in diabetic patients. Postgrad Med J 1987;63:673-6. 7. Shalwitz RA, Farkas-Hirsch R, White NH, Santiago JV. Prevalence and consequences of nocturnal hypoglycemia among conventionally treated children with diabetes mellitus. J Pediatr 1990;116:685-9. 8. Daneman D, Frank M, Peflman K, Tamm J, Ehrlich R. Severe hypoglycemiain children with insulin dependent diabetes mellitus: frequency and predisposing factors. J Pediatr 1989;115: 681-5. 9. Whincup G, Milner R. Prediction and management of nocturnal hypoglycaemia in diabetes. Arch Dis Child 1987;62:333-7. 10. Simell T, Simell O, Cammi E, Caprio E, Hakulinen A, Haxalax J, et al. Glucose profiles in children two years after onset of type-1 diabetes. Diabetic Medicine 1993;10:524-9. 11. Ryan C, Vega A, Drash A. Cognitive deficits in adolescents
372
Porter et al.
The Journal of Pediatrics March 1997
who developed diabetes early in life. Pediatrics 1985;75:921-7. 12. Rovet JF, Ehrlich RM, Hoppe M. Intellectual deficits associated with early onset of IDDM in children. Diabetes Care 1987;10:510-5. 13. VechioTJ. Predictive value of a single diagnostic test in unselected populations. N Engl J Med 1966;3:206-7. 14. Jones TW, Boulware SD, Kramer DK, Caprio S, Sherwin RS, Tamborlane WV. Independent effects of youth and poor diabetes control on responses to hypoglycaemia in children. Dia, betes 1991;40:358-61. 15. Jones TW, Boulware SD, McCarthy G, Sherwin RS, Tamborlane WV. Enhanced adrenomedullaryresponse and increased susceptibility to neuroglycopenia:mechanisms underlying the
adverse effects of sugar ingestion in healthy children. J Pediatr 1995;126:171-7. 16. Brambilla P, Bougneres PF, Santiago JV, Chaussain JL, Pouplard A, Castano L. Glucose counterregulationin pre-schoolaged diabetic children with recurrent hypoglycaemia during conventional treatment. Diabetes 1987;36:300-4. 17. Veneman T, Mitrakou A, Mokan M, Cryer P, Gerich J. Induction of hypoglycaemic unawareness by asymptomatic noctttrnal hypoglycaemia. Diabetes 1993;42:1233-7. 18. The Diabetes Control and Complications Trial. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 1993;329:977-86.
50 Years Ago in The Journal of Pediatrics Children born during the siege of Leningrad in 1942 Antonov AN. J Pediatr 1947;30:250-9. From August 1941 through January 1943 the city of Leningrad was under almost continous siege fr the German army. The winter of 1941-1942 was particularly harsh, and food resources were scant The records of the newborn department at the Leningrad State Pediatric Institute were studied to tablish the effect of this military operation on pregnancy and newborn health. At the height of the sic the stillbirth rate doubled to 5.6%, and the rate of prematurity reached 41.2%. The term infant morta rate reached 9%, and 30.8% of prematurely born infants died. Even among term infants, birth wei8 were compromised. In early 1942, for example, the average birth weight of term infants was 500 to gm less than the previously established norms for Leningrad. Other than the stress of continual warfi two factors were impficated in this tragedy. First, the food ration was severely limited, amount to between 150 and 500 gm of poor-quality bread daily. Second, fuel shortages resulted in signific cold stress to infants. At one point the temperature in the newborn unit was 40 ° F and the staff worl in overcoats. Comment: During the past 50 years, the pages of The Journal have chronicled a host of dramatic vances in newborn care. Little has changed, however, with the politics of warfare. Infants continue suffer disproportionately. One has to wonder if comparable data are available for, as an example, the lq kan peninsula or Iraq.
Thomas R. Welch, Division of Pediatric Nephroh Children's Hospital Medical Cen Cincinnati, OH 45229-3(