Accepted Manuscript

Retrospective evaluation of a national guideline to prevent neonatal hypoglycemia

A.H. Rasmussen, MD., S. Biostatistician Wehberg, PhD, J. Fenger-Grøn, MD., H.T.

Christesen, MD, PhD, Professor

PII: S1875-9572(17)30021-9
DOI: 10.1016/j.pedneo.2016.12.002
Reference: PEDN 635

To appear in: Pediatrics & Neonatology

Received Date: 18 August 2016

Revised Date: 10 November 2016
Accepted Date: 4 December 2016

Please cite this article as: Rasmussen A, Wehberg SB, Fenger-Grøn J, Christesen H, Retrospective
evaluation of a national guideline to prevent neonatal hypoglycemia, Pediatrics and Neonatology (2017),
doi: 10.1016/j.pedneo.2016.12.002.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

1 This material is original research.

2 Title:

3 Retrospective evaluation of a national guideline to prevent neonatal

PT
4 hypoglycemia

RI
5

SC
6 Running title:

7 Neonatal hypoglycemia

U
AN
8

9 Rasmussen, AH. MD. Hans Christian Andersen Children´s Hospital, Odense University Hospital, Denmark
M

10 and Institute of Clinical Research, University of Southern Denmark, Odense, Denmark.

11 Wehberg, S. Biostatistician, PhD., Odense University Hospital, Odense, Denmark.

D

12 Fenger-Grøn, J. MD. Dept. of Pediatrics, Lillebaelt Hospital, Kolding, Denmark.

TE

13 Christesen, HT. Professor, MD, PhD, Hans Christian Andersen Children´s Hospital, Odense University

14 Hospital and Institute of Clinical Research, University of Southern Denmark, Odense, Denmark.
EP

15
C

16 Corresponding author:
AC

17 Henrik T Christesen. Professor, MD, PhD, Hans Christian Andersen Children´s Hospital, Odense University

18 Hospital and Institute of Clinically Research, University of Southern Denmark, Odense, Denmark.

19 Telephone: +45 2137 0888

20 Email: henrik.christesen@rsyd.dk

21

22

1
 ACCEPTED MANUSCRIPT

23

24 This material is original research. No related papers from the authors have been published or are under

25 consideration for publication elsewhere, and this paper has not been submitted elsewhere.

26

PT
27

28

RI
29

SC
30

31

32
U
AN
33

34
M

35
D

36
TE
C EP
AC

2
 ACCEPTED MANUSCRIPT

37 Retrospective evaluation of a national guideline to prevent neonatal hypoglycemia

38

39

40 This material is original research. No related paper from the authors has been published or is

PT
41 under consideration for publication elsewhere, and this paper has not been submitted elsewhere.

RI
42

43 Keywords: Hypoglycemia; Neonate; Prevention.

SC
44

45 Abstract

46
U
Background. Hypoglycemia is common in neonates and may cause adverse neurological
AN
47 outcomes. Guidelines should aim to prevent repeated hypoglycemic episodes in risk groups,
M

48 but they are not usually stratified according to the severity of hypoglycemia risk, which may

49 lead to the implementation of inappropriate and redundant interventions.

D

50 Aim. To evaluate the effect of a national prevention guideline stratified according to mild,
TE

51 moderate, and severe risks of hypoglycemia.

52 Methods. From national registers, a population cohort of 22,725 neonates was identified
EP

53 retrospectively before and after implementation of a national guideline. Of these, 1900 had
C

54 World Health Organization International Classification of Diseases 10 discharge diagnoses of

AC

55 hypoglycemia. Diagnoses indicating hypoglycemia risk [small/large for gestational age

56 (SGA/LGA), asphyxia, prematurity, maternal insulin-treated diabetes mellitus (mIDM)] were

57 recorded. Neonatal ward files were evaluated to validate hypoglycemia diagnoses. Adjusted

58 odds ratios (aORs) were calculated, adjusting for sex, parity, SGA, LGA, preterm birth, and

59 asphyxia, where relevant.

3
 ACCEPTED MANUSCRIPT

60 Results. Primiparity and male sex were associated independently with hypoglycemia

61 diagnosis [aORs, 1.29 (1.17–1.42) and 1.14 (1.03–1.26), respectively]. The overall incidence of

62 hypoglycemia diagnosis at discharge decreased from 9.4% to 5.5% after guideline

63 implementation [aORchange, 0.57 (0.50–0.64)]. The overall incidence of validated hypoglycemia

PT
64 decreased from 2.1% to 1.2% [aOR 0.59 (0.46–0.77), p<0.001]. By risk group, the

65 hypoglycemia incidence decreased from 30.5% to 18.6% [aOR 0.52 (0.36–0.75)] among SGA

RI
66 neonates, from 25.8% to 16.4% [aOR 0.57 (0.42–0.76)] among preterm infants, and from

SC
67 27.4% to 16.6% [aOR 0.63 (0.34–0.83)] among those with asphyxia. LGA neonates showed a

68 decreased incidence in obstetric wards only. No significant change was observed for the

69 mIDM group.
U
AN
70 Conclusion. Stratification of hypoglycemia risk in a hypoglycemia prevention guideline was

71 followed by decreased estimated hypoglycemia incidences, but no causative conclusion could

M

72 be drawn. Prospective studies with risk stratification for hypoglycemia prevention are
D

73 encouraged.
TE

74
C EP
AC

4
 ACCEPTED MANUSCRIPT

75 Background

76 Glucose is an essential fuel for brain metabolism and, depending on its severity and duration,

77 hypoglycemia may result in adverse neurological outcomes.(1-4) The key to hypoglycemia

78 prevention is to identify neonates at risk, as this condition may be difficult to detect clinically.

PT
79 (1, 5-8)

80 The threshold for intervention in neonatal hypoglycemia remains controversial.(1, 9) After

RI
81 the physiological nadir, the mean plasma glucose level in 3–47-h-old, healthy, non-exclusively

SC
82 breastfed term neonates is 3.5–3.7 mmol/L, with a fifth percentile of 2.2–2.3 mmol/L.(10)

83 This value is lower in healthy, exclusively breastfed term neonates (11), but higher ketone

84
U
bodies provide alternative fuel in this group.(5)
AN
85 In risk groups, the clinical cut-off value for treatment of hypoglycemia is usually set around

86 2.5 mmol/L (45 mg/dL) after the first hours of life. This threshold is used widely in clinical
M

87 practice, despite differences in the methods and devices used for determination of the glucose
D

88 concentration (12), and is supported by the recent guideline from The American Academy of
TE

89 Pediatrics.(13) Others have stated that asymptomatic neonates at risk should not maintain

90 glucose levels < 2.0 mmol/L after feeding (1, 14), and treatment thresholds of 1.8 mmol/L on
EP

91 one occasion and 2.6 mmol/L on multiple occasions have been proposed.(15)
C

92 The incidence of neonatal hypoglycemia ranges from 5% to 15% (1, 16, 17), possibly
AC

93 reflecting differences in study populations, preventive measures, and hypoglycemia

94 definitions and measuring techniques (18), in addition to normal biological variation.

95 According to the Institute of Medicine’s definition, clinical guidelines are “ systematically

96 developed statements to assist practitioner and patient decisions about appropriate health

97 care for specific clinical circumstances” (19) , and guidelines intend to improve efficiency and

98 quality of care and reduce inappropriate practice.(20, 21) Hypoglycemia prevention protocols

5
 ACCEPTED MANUSCRIPT

99 should be individualized according to risk stratification (22), and evidence-based guidelines

100 can improve the standard of care for neonates with hypoglycemia.(23) Although guidelines

101 have the potential to have harmful effects as well as benefits if not evaluated, (20, 21) the

102 consequences of guideline implementation have not been well studied. We are not aware of

PT
103 any previous study evaluating the effects of the implementation of a systematic, stratified

104 guideline on the prevention and treatment of neonatal hypoglycemia.

RI
105 In Denmark, a national guideline for the prevention of neonatal hypoglycemia was introduced

SC
106 in 2010 with the primary aims of avoiding repeated hypoglycemic episodes in risk groups and

107 enhancing the quality of care. The guideline stratified hypoglycemia risk as mild, moderate,

108
U
and severe, allowing for the differentiation of preventive actions according to severity.
AN
109 In this study, we aimed to evaluate the prevention effect of the national prevention guideline

110 by estimating the incidence proportions of neonatal hypoglycemia overall and for major risk
M

111 groups, before and after the guideline implementation.

D

112
TE

113 Methods

114 Data sources and uptake area

EP

115 This retrospective, observational cohort study was based on data from the Danish National
C

116 Patient Registry (DNPR) and the Danish Medical Birth Registry (DMBR). The DNPR contains
AC

117 data for all hospitalized patients and newborns in Denmark. The DMBR has more detailed

118 information on pregnancies and deliveries; it contains highly reliable and valid data on 99.8%

119 of all Danish deliveries. (24) The following data were extracted from the DNPR and DMBR for

120 each subject included in the study: personal identification number, postal code, parity,

121 gestational age (GA), sex, birth weight, hospital/department, and all discharge diagnoses

122 according to the World Health Organization’s International Classification of Diseases 10

6
 ACCEPTED MANUSCRIPT

123 (ICD10). Additional data were collected from the medical records of three hospitals in the

124 Region of Southern Denmark: Odense University Hospital (OUH; Department of Obstetrics

125 and Hans Christian Andersen Children’s Hospital) and the two adjacent regional hospitals

126 Lillebaelt Hospital, Kolding (Departments of Obstetrics and Pediatrics) and Svendborg

PT
127 Hospital (Department of Obstetrics). The OUH receives at-risk and high-risk mothers and

128 neonates from the Svendborg and Kolding uptake areas, including all mothers with insulin-

RI
129 treated diabetes mellitus (mIDM). After delivery, all mother–infant pairs with mIDM were

SC
130 referred to the neonatal ward.

131

132 Guideline
U
AN
133 Before 2010, no uniform hypoglycemia prevention program was established in Denmark. The

134 national guideline was introduced in 2010 (Supplementary Table 1). In brief, infants were
M

135 classified into no-, low-, moderate-, and high-risk groups at birth, depending on the risk
D

136 factors present. Prevention measures ranged from breast contact to intravenous glucose
TE

137 administration within the first 30 min of life. If the first two glucose measurements were ≥2.5

138 mmol/L, no further glucose determination was required.

EP

139
C

140 Time periods

AC

141 The pre-guideline period extended from August 1, 2006 to July 31, 2008; the post-guideline

142 period extended from January 1, 2011 to December 31, 2011, or from July 1, 2012 to

143 December 31, 2012 (for Lillebaelt Hospital, Kolding). Data collected between these periods

144 were not included due to use of pilot versions of the guideline, and the difference in the post-

145 guideline period is due to a delay in guideline implementation at Lillebaelt Hospital, Kolding.

146

7
 ACCEPTED MANUSCRIPT

147 Risk groups

148 The risk groups were defined retrospectively according to small for gestational age (SGA) or

149 large for gestational age (LGA) status [birth weight exceeding ±2 standard deviations (SDs) of

150 the reference value], prematurity (GA<37+0 weeks), asphyxia (umbilical cord pH<7.1 and/or

PT
151 base excess≤10), and mIDM. Neonates with ICD10 diagnoses of hypoglycemia (P70.0–70.9)

152 were identified. Mothers with mIDM attending the neonatal ward of OUH could be identified,

RI
153 but maternal ICD10 diabetes diagnosis codes were generally not available in the national

SC
154 registries used for the present study. Neonates with more than one hypoglycemia risk factor

155 were assigned to more than one risk group.

156
U
AN
157 Blood glucose measurement

158 Glucose concentrations were measured using different devices. In the neonatal ward of OUH,
M

159 capillary blood glucose was tested using the HemoCue 201/201+ point-of-care testing (POCT)
D

160 device until mid-April 2011, where after the HemoCue 201 DM RT and 201 DM were used.
TE

161 The Roche Accu-Chek Inform II system was used in the neonatal ward of Lillebaelt Hospital,

162 Kolding, and Radiometer ABL 800 flex analyzers were used in the obstetric departments
EP

163 throughout the study period.

C

164 All HemoCue devices measured glucose concentrations in whole blood, but the HemoCue 201
AC

165 RT/201 DM device automatically presented calculated plasma glucose values after

166 multiplying by 1.11.

167

168 Populations and outcome measures

169 We studied the incidence proportions of neonatal hypoglycemia in the hospitals’ uptake area

170 for each of the two time periods in four different patient populations, overall and separately

8
 ACCEPTED MANUSCRIPT

171 for risk groups (SGA, LGA, preterm, asphyxia, and mIDM). For national registry data, the

172 denominator was the total birth count extracted from the DNPR/DMBR for the uptake area,

173 and the numerator (count of neonates with hypoglycemia within the denominator group) was

174 based on ICD10 diagnoses of hypoglycemia (P70.0–70.9) from the DNPR/DMBR in the

PT
175 neonatal period. For obstetric ward neonatal data, the denominator consisted of all neonates

176 discharged from one of the three obstetric wards to home, and the numerator was based on

RI
177 recording of P70.0–70.9 diagnoses in the medical records. Bedside blood glucose values were

SC
178 not recorded systematically in the medical records of the obstetric departments, disallowing

179 blood sample–based validation of neonatal hypoglycemia diagnoses in this population. For

180
U
neonatal ward data, the denominator consisted of all neonates referred to one of the two
AN
181 neonatal wards. The numerator was based on recording of P70.0–70.9 diagnoses in the

182 medical records. For neonatal ward data on validated hypoglycemia, the denominator was the
M

183 same and the numerator was based on recording of P70.0–70.9 diagnoses in the medical
D

184 records in cases for which data on blood glucose level and age (in hours) at the time of blood
TE

185 glucose determination were also available. These data were analyzed to validate diagnoses of

186 hypoglycemia. Hypoglycemia was defined as glucose value <1.7 mmol/L at 0–2 h of life and
EP

187 <2.5 mmol/L after 2 h of life.

C

188
AC

189 Methodological bias (sensitivity analysis)

190 We reviewed a random sample of data from 400 neonates without ICD10 hypoglycemia

191 diagnoses in the neonatal ward of OUH within the asphyxia and SGA hypoglycemia risk factor

192 groups. Hypoglycemic neonates identified by this search, but not identified through the

193 original registry data search, were not included in the incidence estimates.

194

9
 ACCEPTED MANUSCRIPT

195 Statistical methods

196 First, odds ratios (ORs) for hypoglycemia diagnosis with 95% confidence intervals (CIs) were

197 calculated based on logistic regression models for the following risk factors: SGA, LGA,

198 preterm birth, asphyxia, and mIDM (where relevant), as well as parity and sex. This analysis

PT
199 was conducted for the overall patient population, across both time periods. Second, neonatal

200 characteristics were compared between neonates with and without hypoglycemia diagnoses

RI
201 for each time period using the Mann–Whitney U, Student’s t, and chi-squared tests, where

SC
202 appropriate. Third, incidence proportions of hypoglycemia with 95% CIs were calculated for

203 five patient populations (overall and separately according to risk groups). In addition, ORs for

204
U
change with 95% CIs were calculated based on logistic regression models (univariate and
AN
205 adjusted for sex, parity, SGA, LGA, preterm birth, and asphyxia, where relevant). No

206 adjustment for multiple testing was performed. Data were analyzed using Stata software
M

207 (StataCorp, College Station, TX, USA).

D

208
TE

209 Ethics

210 The regional ethics committee approved the study (protocol no. S-20120119).
EP

211
C

212 Results
AC

213 Overall, the study population included 22,725 neonates (48.4% girls, 51.6% boys). The mean

214 (±SD) birth weights were 3400 (±600) g for girls and 3506 (±644) g for boys. The mean GA

215 was 39 (±2.2) weeks. Prematurity was seen in 1693 (7.4%), asphyxia in 902 (3.9%), SGA in

216 848 (3.7%), LGA in 822 (3.6%), and mIDM in 245 (1.1%) neonates. In total, 3949 (17.3%)

217 neonates fell into risk groups, of which 406 (10.3%) had more than one risk factor.

10
 ACCEPTED MANUSCRIPT

218 In univariate and adjusted analyses, SGA, LGA, preterm birth, and asphyxia were associated

219 strongly with increased odds of hypoglycemia diagnosis (Table 1). Surprisingly, primiparity

220 status and male sex were also associated strongly with increased adjusted odds ratios (aORs)

221 for hypoglycemia. Neonates belonging to the risk groups defined in this study accounted for

PT
222 1289/1900 (67.8%) of all neonates with hypoglycemia diagnoses. Accordingly, almost one-

223 third of neonates with such diagnoses could not be assigned to any risk group.

RI
224 The baseline characteristics of all neonates and those with hypoglycemia, defined by

SC
225 discharge diagnoses, in the two study periods are shown in Figure 1.

226 The estimated incidence proportion of hypoglycemia by discharge diagnosis was 8.4% (Table

227 2).
U
AN
228 National registry data showed that the incidence of hypoglycemia diagnoses in the SGA, LGA,

229 preterm, and asphyxia risk groups was 26.2% (22.7–27.0%), with no significant difference
M

230 among risk group estimates. In the neonatal ward, the incidence of validated hypoglycemia
D

231 ranged from 5.5% to 10.5% among risk groups, but was much higher (32.2%) in the mIDM
TE

232 group.

233 The implementation of the hypoglycemia guideline was associated with a highly significant
EP

234 decrease in the incidence proportion of this condition in the overall population in all analyses
C

235 (Table 2).

AC

236 For neonates in the neonatal ward, the incidence of validated hypoglycemia also decreased,

237 from 2.1% to 1.2% [crude OR for change, 0.55 (0.43–0.71), p<0.001; aOR, 0.59 (0.46–0.77)].

238 All analyses showed decreased incidence proportions in the SGA and preterm birth groups.

239 For the asphyxia group, a highly significant decrease was seen in the national registry and in

240 the obstetric wards, but no change was observed in the neonatal wards. For the mIDM and

241 LGA risk groups, no significant change was observed in any analysis.

11
 ACCEPTED MANUSCRIPT

242 The change in glucose measurement technique during the study period could theoretically

243 have affected the incidence of mild hypoglycemia (2.3–2.4 mmol/L) diagnoses within the

244 post-implementation period. To compensate for this bias, we conducted a subanalysis

245 including only OUH neonatal ward patients with validated hypoglycemia (<2.3 mmol/L)

PT
246 through April 15, 2011. This methodological bias correction did not change the incidence

247 estimates.

RI
248 In the review of 400 neonatal ward files with ICD10 codes for asphyxia and SGA, without

SC
249 discharge diagnosis codes for hypoglycemia, we identified unreported hypoglycemia in 16.5%

250 of neonates. Underreporting showed a decreasing trend between the pre- and post-

251
U
implementation periods, especially for SGA (overall, from 20% to 13%, p=0.18; asphyxia, from
AN
252 16% to 12%, p=0.42; SGA, from 24% to 14%, p=0.07).

253
M

254 Discussion
D

255 In Danish neonatal departments, various initiatives to prevent and treat neonatal
TE

256 hypoglycemia have existed for decades. Until the introduction of the national guideline,

257 however, practices were heterogeneous, differing within as well as between departments. The
EP

258 implementation of the national guideline instituted a homogenous, consistent, and systematic
C

259 approach, and the current study was conducted to evaluate its effect on the risk of
AC

260 hypoglycemia. The retrospective design of our study and methodological constraints,

261 including the lack of a uniform blood glucose recording protocol and the quality dependence

262 of discharge diagnoses, except for the validation substudy that was conducted using data from

263 the neonatal ward, do not allow the calculation of exact incidence estimates or drawing of

264 definitive conclusions on the causes of changes. However, the estimated overall hypoglycemia

265 incidence fell significantly between the two study periods; this effect was most pronounced in

12
 ACCEPTED MANUSCRIPT

266 the obstetric wards, but was also seen in neonatal wards among cases of validated

267 hypoglycemia discharge diagnoses. No evidence of an increase in hypoglycemia was observed

268 after the guideline introduction, suggesting that the uniform incidence decrease was not due

269 to methodological bias.

PT
270 The overall population incidence of neonatal hypoglycemia, according to discharge diagnosis,

271 was 8.4%, which was within previously reported incidences of 5–15%. (2, 7, 8, 16, 17, 25, 26)

RI
272 Our estimate had the advantage of being based on a large population birth cohort, but the

SC
273 disadvantage of being based predominantly on discharge diagnoses and glucose

274 measurements obtained with different methods. Validation exercises for these diagnoses

275
U
showed both overreporting and underreporting of hypoglycemia compared with medical
AN
276 records reviews.

277 In a multivariable analysis, we identified SGA, LGA, preterm birth, and asphyxia as
M

278 independent risk factors for hypoglycemia. These risk groups, as well as the risk faced by
D

279 infants of diabetic mothers, are well known.(5, 7, 25, 27) Almost one-third of our neonates
TE

280 with hypoglycemia diagnoses were not assigned to a risk group. One explanation for this

281 finding could be missing discharge diagnoses of gestational diabetes, which in the absence of
EP

282 hypoglycemia could have been reported in the mothers’ hospital records only, and hence not
C

283 included in the current study. Surprisingly, the analysis based on the identified risk groups
AC

284 showed that primiparity and male sex were independent risk factors for hypoglycemia.

285 Explanatory factors may include delayed onset of lactogenesis in breastfeeding primipara

286 (28) and increased male representation in neonates below the LGA cut-offs, but >4000 g,

287 above which the risk of hypoglycemia also may be increased.(29)

288 The overall incidence of hypoglycemia according to discharge diagnosis among the risk

289 groups of SGA, LGA, preterm birth, and asphyxia was 26.2%. Others have reported

13
 ACCEPTED MANUSCRIPT

290 hypoglycemia incidences among at-risk neonates of 11.7–51%.(18, 26, 30) In the analysis

291 restricted to neonatal ward data for newborns with validated hypoglycemia, the incidence

292 was at the lower end of this reported range (12.4%).

293 The estimated incidence of hypoglycemia diagnosis fell significantly after the implementation

PT
294 of the prevention guideline. This change may be attributed to a preventive effect of the

295 guideline, but the contribution of methodological bias, including errors in discharge diagnoses

RI
296 and variations in blood glucose measurement counts and methods, cannot be ruled out. In

SC
297 support of a true decrease in the incidence of hypoglycemia, a decreased incidence was also

298 seen in the validation exercise. Moreover, our subanalysis of neonatal ward cases of glucose

299
U
values <2.3 mmol/L also showed a highly significant decrease in the incidence.
AN
300 Decreases in the estimated incidence of hypoglycemia were seen in the SGA and preterm risk

301 groups, regardless of data source. The decrease in preterm neonates occurred despite a
M

302 decrease in the mean GA and birth weight, which would tend to increase the risk of
D

303 hypoglycemia, among all such neonates in the post-guideline period.

TE

304 For neonates with asphyxia, a decrease was seen only in the obstetric ward, suggesting that

305 hypoglycemia prevention in the neonatal wards was effective before guideline
EP

306 implementation, or that detection of hypoglycemia among those with milder asphyxia in the
C

307 obstetric ward who were not referred to the neonatal ward was more complete after
AC

308 guideline implementation.

309 The LGA and mIDM risk groups showed no decrease in hypoglycemia, regardless of the data

310 source. The unchanged incidence in the LGA group may reflect the ease of hypoglycemia

311 prevention in this group by early feeding, a prevention strategy which was carried out before

312 guideline implementation. For mIDM, the avoidance of hypoglycemia in this patient group

313 received much attention before guideline implementation.

14
 ACCEPTED MANUSCRIPT

314 Ongoing controversy regarding the definition and consequences of hypoglycemia, and the

315 methodological problems associated with POCT device use, raises open questions about the

316 appropriateness of the prevention strategy examined here and others. At the least, the

317 national prevention program introduced less-intrusive interventions for low-risk neonates

PT
318 without increasing the incidence of hypoglycemia in any risk group.

319 The strengths of our study included the examination of a large, population-based birth cohort;

RI
320 the inclusion of data from several sources; the validation of hypoglycemia diagnoses; and the

SC
321 adjustment of analyses of incidence changes in risk groups. Limitations included the

322 retrospective design, with the risk of inaccuracy of discharge diagnoses and medical records;

323
U
the methodological problems associated with glucose measurement using different POCT
AN
324 devices; the non-inclusion of neonates with diet-treated maternal diabetes; and potential

325 undiscovered reporting bias from unidentified changes in practice.

M

326
D

327 Conclusion
TE

328 We conclude that the introduction of a standardized, stratified hypoglycemia prevention

329 program was followed by the use of a more systematic and consistent prevention strategy for
EP

330 hypoglycemia. The stratification of risk as mild, moderate, and severe, which enables the
C

331 implementation of differentiated prevention initiatives, was met not by an increased

AC

332 hypoglycemia incidence, but by strong suggestions of a decrease. These findings encourage

333 the further use of a stratified approach, which should be studied in other settings using a

334 prospective design.

335

15
 ACCEPTED MANUSCRIPT

336
337 Table 1. Factors associated with hypoglycemia diagnosis (ICD10 discharge diagnosis). Odds ratios with 95% confidence intervals based on
338 logistic regression analysis of 22,725 neonates in the national registry database are presented. Both time periods are included.
339
340
Factor Crude P-value Adjusted* P-value
OR (95% CI) OR (95% CI)
SGA vs. AGA 4.88 (4.15-5.72) <0.001 3.39 (2.86-4.03) <0.001
LGA vs. AGA 3.88 (3.27-4.61) <0.001 4.10 (3.43-4.90) <0.001
Preterm 4.12 (3.54-4.54) < 0.001 3.41 (3.04-3.96) <0.001

PT
Asphyxia 4.10 (3.50-4.81) < 0.001 4.20 (3.56-4.95) <0.001
Parity 1 vs. >1 1.47 (1.34-1.62) <0.001 1.29 (1.17-1.42) <0.001
Male vs. female 1.17 (1.06-1.28) 0.001 1.14 (1.03-1.26) 0.008
341
342 *OR adjusted for sex, parity, and SGA, LGA, preterm birth, and asphyxia, where relevant.

RI
343 Abbreviations: OR, odds ratio; CI, confidence interval; SGA, small for gestational age; LGA, large for gestational age; AGA, appropriate for
344 gestational age.
345
346

U SC
AN
M
D
TE
C EP
AC

16
 ACCEPTED MANUSCRIPT

347 Table 2. Incidence proportions of hypoglycemia before and after national guideline implementation, and odds ratios for change.
348
349
Incidence proportions (95% CI) in relation to guideline implementation Odds ratio (95% CI) for change
Both periods Before After Crude OR p Adjusted* OR p
National registry
Overall 8.4% (8.0-8.8) 9.4% (9.0-9.9) 5.5% (4.9-6.1) 0.55 (0.49-0.62) <0.001 0.57 (0.50-0.64) <0.001
SGA 27.0% (24.0-30.1) 30.5% (26.9-34.5) 18.6% (13.9-23.9) 0.51 (0.36-0.74) <0.001 0.52 (0.36-0.75) <0.001
LGA 23.6% (19.9-25.7) 25.3% (19.9-26.7) 21.1% (15.8-27.3) 0.89 (0.60-1.30) 0.554 0.96 (0.65-1.41) 0.839

PT
Preterm 23.6% (21.6-25.7) 25.8% (23.4-28.2) 16.4% (12.9-20.0) 0.56 (0.42-0.75) <0.001 0.57 (0.42-0.76) <0.001
Asphyxia 25.3% (22.6-28.3) 27.4% (24.2-30.8) 16.6% (11.3-23.0) 0.52 (0.34-0.81) 0.004 0.53 (0.34-0.83) 0.006
Obstetric ward neonates
Overall 5.3% (5.0-5.6) 6.1% (5.8-6.5) 3.0% (2.7-3.5) 0.48 (0.41-0.56) <0.001 0.51 (0.43-0.59) <0.001
SGA 13.2% (11.0-15.6) 15.0% (12.8-18.8) 9.1% (5.8-13.3) 0.56 (0.35-0.92) 0.022 0.55 (0.34-0.90) 0.018

RI
LGA 10.3% (8.3-12.6) 11.2% (8.8-13.9) 7.7% (4.5-12.1) 0.66 (0.38-1.17) 0.156 0.66 (0.38-1.18) 0.167
Preterm 11.1% (9.6-12.7) 12.2% (10.4-14.1) 7.7% (5.3-10.7) 0.60 (.040-0.90) 0.014 0.61 (0.48-0.78) 0.016
Asphyxia 18.2% (15.7-20.8) 20.3% (17.4-23.4) 8.9% (5.1-14.2) 0.38 (0.22-0.67) 0.001 0.37 (0.21-0.65) 0.001
Neonatal ward neonates

SC
Overall 3.0% (2.8-3.3) 3.3% (3.0-3.6) 2.4% (2.0-2.8) 0.72 (0.60-0.86) 0.001 0.76 (0.63-0.77) 0.001
SGA 13.7% (11.5-16.3) 15.6% (0.13-0.19) 9.3% (6.2-13.8) 0.56 (0.35-0.91) 0.019 0.55 (0.28-0.87) 0.018
LGA 12.4% (10.2-14.8) 12.0% (9.6-14.8) 13.5% (9.1-18.8) 1.14 (0.71-1.80) 0.573 1.30 (0.81-2.11) 0.286
Preterm 12.5% (10.9-14.1) 13.6% (11.8-15.6) 8.7% (6.1-11.9) 0.60 (0.41-0.88) 0.010 0.59 (0.40-0.88) 0.009
Asphyxia 7.2% (5.6-9.0) 7.1% (5.3-9.1) 7.7% (4.1-12.7) 1.09 (0.58-2.10) 0.579 1.27 (0.65-2.46) 0.482

U
Neonatal ward, validated hypoglycemia
Overall 1.8% (1.7-2.0) 2.1% (1.9-2.3) 1.2% (0.1-0.2) 0.55 (0.43-0.71) <0.001 0.59 (0.46-0.77) <0.001
AN
SGA 10.5% (8.5-12.7) 11.9% (9.4-14.8) 7.1% (4.2-11.0) 0.62 (0.31-0.93) 0.003 0.56 (0.34-0.94) 0.021
LGA 6.2% (4.6-8.1) 6.8% (5.0-9.1) 4.3% (2.0-8.0) 0.62 (0.29-1.29) 0.204 0.72 (0.33-1.55) 0.406
Preterm 7.9% (6.7-9.3) 9.4% (7.9-11.2) 2.9% (1.6-5.2) 0.29 (0.16-0.54) <0.001 0.28 (0.15-0.52) <0.001
Asphyxia 5.5% (4.2-7.4) 5.4% (3.9-7.4) 6.5% (3.3-11.3) 1.20 (0.61-2.40) 0.594 1.33 (0.65-2.73) 0.430
Maternal diabetes, 32.2% (26.4-38.4) 34.5% (27.3-42.3) 27.5% (18.1-38.6) 1.2 (0.40- (1.29) 0.270 0.83 (0.45-1.53) 0.559
M

insulin-treated
Bias-corrected 1.8% (1.6-1.9) 2.1 (1.9-2.3) 1.1% (0.8-1.4) 0.44 (0.30-0.55) <0.001 0.43 (0.31-0.58) <0.001
hypoglycemia
350 *Adjusted for sex, parity, SGA, LGA, preterm birth, and asphyxia, where relevant.
351 Abbreviations: CI, confidence interval; OR, odds ratio; SGA, small for gestational age; LGA, large for gestational age.
D

352
353
TE
C EP
AC

17
 ACCEPTED MANUSCRIPT

354 Figure 1. Baseline characteristics for all neonates, and neonates at risk of hypoglycemia, defined by discharge diagnosis, before and after

355 implementation of a national guideline for the prevention of hypoglycemia.

356 SGA, small for gestational age; LGA, large for gestational age; DM, diabetes mellitus.

357
358

PT
RI
U SC
AN
M
D
TE
C EP
AC

18
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC

359

19
 ACCEPTED MANUSCRIPT

360

361 COI statement

362 The authors have no conflict of interest relevant to this article.

363

PT
RI
U SC
AN
M
D
TE
C EP
AC

20
 ACCEPTED MANUSCRIPT

364 References

365 1.Cornblath M, Hawdon JM, Williams AF, Aynsley-Green A, Ward-Platt MP, Schwartz R, et al.

366 Controversies regarding definition of neonatal hypoglycemia: suggested operational

367 thresholds. Pediatrics. 2000;105(5):1141-5.

PT
368 2.Lucas A, Morley R, Cole TJ. Adverse neurodevelopmental outcome of moderate neonatal

369 hypoglycaemia. BMJ (Clinical research ed). 1988;297(6659):1304-8.

RI
370 3.Arya VB, Senniappan S, Guemes M, Hussain K. Neonatal Hypoglycemia. Indian J Pediatr.

SC
371 2013.

372 4.McKinlay CJ, Alsweiler JM, Ansell JM, Anstice NS, Chase JG, Gamble GD, et al. Neonatal

373
U
Glycemia and Neurodevelopmental Outcomes at 2 Years. The New England journal of
AN
374 medicine. 2015;373(16):1507-18.

375 5.Hawdon JM, Ward Platt MP, Aynsley-Green A. Patterns of metabolic adaptation for preterm
M

376 and term infants in the first neonatal week. Arch Dis Child. 1992;67(4 Spec No):357-65.
D

377 6.Altman M, Vanpee M, Cnattingius S, Norman M. Neonatal morbidity in moderately preterm

TE

378 infants: a Swedish national population-based study. The Journal of pediatrics.

379 2011;158(2):239-44.e1.
EP

380 7.Agrawal RK, Lui K, Gupta JM. Neonatal hypoglycaemia in infants of diabetic mothers. Journal

381 of paediatrics and child health. 2000;36(4):354-6.

C
AC

382 8.Maayan-Metzger A, Lubin D, Kuint J. Hypoglycemia rates in the first days of life among term

383 infants born to diabetic mothers. Neonatology. 2009;96(2):80-5.

384 9.Guemes M, Rahman SA, Hussain K. What is a normal blood glucose? Arch Dis Child. 2015.

385 10.Alkalay AL, Sarnat HB, Flores-Sarnat L, Elashoff JD, Farber SJ, Simmons CF. Population

386 meta-analysis of low plasma glucose thresholds in full-term normal newborns. Am J Perinatol.

387 2006;23(2):115-9.

21
 ACCEPTED MANUSCRIPT

388 11.Swenne I, Ewald U, Gustafsson J, Sandberg E, Ostenson CG. Inter-relationship between

389 serum concentrations of glucose, glucagon and insulin during the first two days of life in

390 healthy newborns. Acta Paediatr. 1994;83(9):915-9.

391 12.D'Orazio P, Burnett RW, Fogh-Andersen N, Jacobs E, Kuwa K, Kulpmann WR, et al.

PT
392 Approved IFCC recommendation on reporting results for blood glucose (abbreviated). Clinical

393 chemistry. 2005;51(9):1573-6.

RI
394 13.Adamkin DH. Postnatal glucose homeostasis in late-preterm and term infants. Pediatrics.

SC
395 2011;127(3):575-9.

396 14.Hawdon JM. Definition of neonatal hypoglycaemia: time for a rethink? Archives of disease

397
U
in childhood Fetal and neonatal edition. 2013;98(5):F382-3.
AN
398 15.Screening guidelines for newborns at risk for low blood glucose. Paediatr Child Health.

399 2004;9(10):723-40.
M

400 16.Merenstein GB, Gardner SL. Handbook of neonatal intensive care. 5 ed. St. Louis, Mo.:
D

401 Mosby; 2002. xx, 863 s., illustreret (pbk) p.

TE

402 17.Hay WW, Jr., Raju TN, Higgins RD, Kalhan SC, Devaskar SU. Knowledge gaps and research

403 needs for understanding and treating neonatal hypoglycemia: workshop report from Eunice
EP

404 Kennedy Shriver National Institute of Child Health and Human Development. The Journal of
C

405 pediatrics. 2009;155(5):612-7.

AC

406 18.Harris DL, Weston PJ, Harding JE. Incidence of neonatal hypoglycemia in babies identified

407 as at risk. The Journal of pediatrics. 2012;161(5):787-91.

408 19.Institute of Medicine Committee to Advise the Public Health Service on Clinical Practice G.

409 In: Field MJ, Lohr KN, editors. Clinical Practice Guidelines: Directions for a New Program.

410 Washington (DC): National Academies Press (US)

411 Copyright (c) National Academy of Sciences.; 1990.

22
 ACCEPTED MANUSCRIPT

412 20.Grimshaw JM, Russell IT. Effect of clinical guidelines on medical practice: a systematic

413 review of rigorous evaluations. Lancet (London, England). 1993;342(8883):1317-22.

414 21.Woolf SH, Grol R, Hutchinson A, Eccles M, Grimshaw J. Clinical guidelines: potential

415 benefits, limitations, and harms of clinical guidelines. BMJ (Clinical research ed).

PT
416 1999;318(7182):527-30.

417 22.Rozance PJ, Hay WW, Jr. Describing hypoglycemia--definition or operational threshold?

RI
418 Early Hum Dev. 2010;86(5):275-80.

SC
419 23.Stewart CE, Sage EL, Reynolds P. Supporting 'Baby Friendly': a quality improvement

420 initiative for the management of transitional neonatal hypoglycaemia. Archives of disease in

421 childhood Fetal and neonatal edition. 2015.

U
AN
422 24.Knudsen LB, Olsen J. The Danish Medical Birth Registry. Dan Med Bull. 1998;45(3):320-3.

423 25.Hume R, McGeechan A, Burchell A. Failure to detect preterm infants at risk of

M

424 hypoglycemia before discharge. The Journal of pediatrics. 1999;134(4):499-502.

D

425 26.Sexson WR. Incidence of neonatal hypoglycemia: a matter of definition. The Journal of
TE

426 pediatrics. 1984;105(1):149-50.

427 27.Rozance PJ. Update on neonatal hypoglycemia. Curr Opin Endocrinol Diabetes Obes.
EP

428 2014;21(1):45-50.
C

429 28.Nommsen-Rivers LA, Chantry CJ, Peerson JM, Cohen RJ, Dewey KG. Delayed onset of
AC

430 lactogenesis among first-time mothers is related to maternal obesity and factors associated

431 with ineffective breastfeeding. Am J Clin Nutr. 2010;92(3):574-84.

432 29.Linder N, Lahat Y, Kogan A, Fridman E, Kouadio F, Melamed N, et al. Macrosomic newborns

433 of non-diabetic mothers: anthropometric measurements and neonatal complications.

434 Archives of disease in childhood Fetal and neonatal edition. 2014;99(5):F353-8.

23
 ACCEPTED MANUSCRIPT

435 30.Jonas D, Dietz W, Simma B. Hypoglycemia in Newborn Infants at Risk. Klinische Padiatrie.

436 2014.

437

PT
RI
U SC
AN
M
D
TE
C EP
AC

24