Infants of Diabetic Mothers: Data Review and Clinical Practice Change

1
Delivery Room Triage of Infants of Medication
Dependant Diabetic Mothers Validation of a Risk
Score for Hypoglycemia
May 19, 2009 Finger Lakes Region Perinatal Forum
Meeting Andrea Scheurer MD
2
Disclosures

• This research was funded in part by the Finger
  Lakes Region Perinatal Forum Mini-Grant Program.

3
Infants of Diabetic Mothers

• What we know
• Increased risk for hypoglycemia.
• May have hypoglycemia but be symptom free.
• Frequent determination of blood glucose level is
  necessary.

• What we dont know
• Which infants will actually develop hypoglycemia.
• Which infants will need IV therapy.
• Best place to admit these infants.
• Different protocols in different hospitals.

4
Previous Protocol at Strong/Highland
Transfer to NBN
Remain asymptomatic
Develop symptoms or need IV glucose
Stay in NICU/SCN
Admit to NICU or SCN
IMDDM
IMDDM Infant of Medication Dependant Diabetic
Mother (Medication insulin, glyburide, or other
oral hypoglycemic)
5
Infants of Diabetic MothersImplications of the
Previous Protocol

• Separation of mothers and infant.
• Impacts breastfeeding initiatives
• Interferes with initiation of breastfeeding
  within 1 hour of birth
• Delays rooming-in 24 hours a day
• Makes breastfeeding on demand a challenge
• Impacts mother-infant bonding
• Increased monetary cost

Steps 4,7,and 8 of Ten Steps To Successful
Breastfeeding, The Academy of Breastfeeding
Medicine, Protocol 7
6
Background

• Haidar-Ahmad developed and tested a 4-component
  hypoglycemic risk score in IMDDM.
• Risk score helped predict which IMDDM were at
  risk for development of hypoglycemia requiring IV
  dextrose.

7
Objective

• To validate this hypoglycemia risk score in a
  population of IMDDM at a level I community
  hospital.

8
Design/Methods

• Eligible infants were IMDDM (insulin or
  glyburide).
• 35 wks and without other indications for SCN
  care.
• Risk score components were collected
  retrospectively.
• 1 point
• Maternal age 35 yrs (AMA).
• Maternal blood glucose before delivery 120mg/dl
  (MBG).
• Infant size for dates (LGA/SGA).
• 2 points
• Infant delivery room blood glucose lt40 or
  120mg/dl (DRBG).
• Total risk score was calculated
• 0 to 5 point total.

9
Design/Methods

• Primary outcome was hypoglycemia requiring IV
  dextrose.
• The total risk score for each infant was
  calculated.
• The infants were divided into 2 groups
• Low risk
• Risk score of 1.
• High risk
• Risk score of 2.

10
Design/Methods

• The score was considered valid if an infant with
  a low risk score was at least 90 likely NOT to
  develop hypoglycemia requiring IV dextrose.
• (95 CI of NPV 0.9-1.0).
• To achieve this with a 0.05 and power 0.8, a
  sample size of 71 patients was required.

11
Results

12
Results

13
Conclusion

• Conclusion
• Validated a 4-component risk score as a strong
  and reliable predictor of hypoglycemia requiring
  IV dextrose among IMDDM at SMH and HH.

14
Outcome

• Change in clinical practice
• The risk score was implemented at SMH and HH to
  triage IMDDMs from delivery room to appropriate
  level of care for glucose monitoring.
• A blood glucose monitoring protocol was followed
  for all infants regardless of admit location.

15
Is this infant symptomatic or lt35 weeks?
Is this an infant of a diabetic mother?
No
No
Birth Center/NBN Routine Newborn Care
Yes
Yes
Was mother taking insulin or an oral
hypoglycemic agent during pregnancy?
Birth Center/NBN Newborn of Diabetic Mother
Protocol
No
Admit to NICU
Glyburide, Metformin, or other
Yes
Assign Risk Score
Score Components 0 points 1 point 2 points
Maternal pre-delivery BG lt 120 120
Maternal age lt 35 35
Neonatal weight for age AGA SGA, LGA
Neonatal BG 40-120 lt 40 or 120
If maternal BG unknown use total score of
remaining components
Total Score 0 to 1
Birth Center/NBN Newborn of Diabetic Mother
Protocol
Admit to NICU
Total Score 2 to 5
16
Hypothesis 2

• By keeping low risk infants with their mothers
  after birth we hope to improve maternal-infant
  bonding, early breastfeeding success, and overall
  maternal satisfaction.

17
Time to First Breast-Feed
18
Thank you!
19
Goal

• Examine the impact of the change in practice on
  breastfeeding initiation/success and maternal
  satisfaction with hospital course.

20
Design/Methods

• Population
• Strong and Highland Hospital
• Postpartum women with babies born at 35 weeks
• Written survey of 300 mothers
• Total of 60 medication dependant diabetics
  (30/30)
• Total of 120 healthy mothers whose infants stayed
  on birth center (60/60)
• Total of 120 mothers whose infants went to
  NICU/SCN for asymptomatic chorio (60/60)

21
Design/Methods

• Primary Outcome
• Mothers with diabetes
• Increased maternal satisfaction with their
  hospital course during the first day of their
  infants life.
• Increase on average by 0.5 points (1-10).
• Additional Anticipated Results
• Earlier initiation of breastfeeding and improved
  maternal-infant bonding in the first day of life.

22
Design/Methods

• Statistics
• Stata 10 Effect Size Table (? 0.05, power 0.8)
• Mean satisfaction score of 2 groups 7 vs. 7.5
  with SD 1 for both groups
• 63 surveys per group

23
Data Analysis

• Comparison of mean satisfaction scores
• Students t test will be used for pair-wise
  comparisons of the group means
• Analysis of variance (ANOVA) will be used for
  comparison of means of all three groups

24

•  
• Dear New Mother,
• The following questions are being asked about
  your post-partum hospital experience and your
  satisfaction with your hospital stay. This
  information is being collected as part of a
  research study and all information will remain
  confidential. Thank you.
• 1. What did you plan to feed your baby before
  your baby was born? Breastmilk ?, Formula ?, or
  Both ?
• 2. Did you try to breastfeed your baby during
  your hospital stay?  Yes ? or No ?
• If No, please skip to question number 6.
•  
• 3. When did you first put your baby to breast? 
  Within the first hour after birth ?
  Between 1 and 4 hours
  of age ?
• After 4 hours of age ?
• 4. Did your babys first attempt at
  breastfeeding occur as soon after birth as you
  hoped? Yes ? or No ?
• If no, why not? _____________________________
  _________
  
  
• If no, was this a source of stress
  for you?  Yes ? or No ?
•  

25
Survey Collection
105/150 70
26
Mean Satisfaction ScoreAll Mothers
8.26 CI7.6-8.8
8.34 CI 7.6-9
Two-sample t-test, p value 0.8
27
Statistically Significant Results

• Well infant mother satisfaction vs. IMDDM mother
  satisfaction.
• Time to first breastfeed Well vs. IMDDM

28
Mean Satisfaction ScoreWell vs. IMDDM
9.1 CI 8.7-9.6
7.1 CI 6.2-8.0
Two sample t-test, p value 0.00
29
Satisfaction ScoreIMDDM
Satisfaction Score (IMDDM) Frequency Percent
1 1 3
4 4 13
5 2 6
6 5 16
7 4 13
8 3 10
9 6 20
10 5 16
Total 30 100
38
30
Satisfaction ScoreWell
Satisfaction Score (Well) Frequency Percent
1 1 1.6
2 1 1.6
6 2 3.3
7 3 5
8 4 6.6
9 10 16.6
10 39 65
Total 60 100
6.5
31
Time to First FeedWell vs. IMDDM
First Feed (SMHHH) Well IMDDM Total
lt1hour 29 6 35
1-4 hours 12 4 16
gt4 hours 3 13 16
44 23 67
p0.00, Pearson chi2
32
Almost Statistically Significant

• Trend towards statistical significance
• Time to first breastfeed HH vs. SMH

33
Time to First Breast FeedSMH vs. HH
First Feed (All infants) SMH HH Total
lt1hour 15 22 37
1-4 hours 13 10 23
gt4 hours 12 7 19
40 39 79
p0.22, Pearson chi2
34
Time to First BreastfeedSMH vs. HH
First Feed (All infants) SMH HH Total
lt1hour 15 22 37
1hour 25 17 42
40 39 79
p0.09, Pearson chi2
35
Not Statistically Significant

• Mean satisfaction score in chorio vs. IMDDM.
• Hospital differences in feeding plans.
• Hospital differences in home feeding plans.
• Hospital differences in viewing breastfeeding as
  a challenge if infant admitted to NICU/SCN.

36
Mean Satisfaction ScoreIMDDM vs. Chorio
7.4 CI 6.1-8.7
7.1 CI 6.2-8.0
P value 0.6 Two Sample t-test
37
Whats to Come?

• Continued tracking of score use
• Incorporate HH data
• September 2009 (1 year)
• Currently collecting surveys
• Anticipate completion by end of April 2009
• Final data analysis
• Comparing surveys before and after risk score use
• PAS
• Retrospective study component 2009

38
Future Possibilities

• Apply score to LGA infants?
• Apply score for GDMA1 or Type II (no meds)?

39
Sources

• Cornblath M, et al. Controversies Regarding
  Definition of Neonatal Hypoglycemia Suggested
  Operational Thresholds. PEDIATRICS 2000 105
  1141-45.
• Brand PLP, et al. Neurodevelopmental outcome of
  hypoglycemia in healthy, large for gestational
  age, term newborns. Arch Dis Child 2005 90
  78-81.
• Plagemann A, et al. Impact of Early Neonatal
  Breast-feeding on Psychomotor and
  Neuropsychological Development in Children of
  Diabetic Mothers. Diabetes Care 2005 March Vol
  28.
• Johnson TS, et al. Fetal Growth Curves Does
  Classification of Weight-for-Gestational Age
  Predict Risk of Hypoglycemia in the Term Newborn?
  J of Midwifery Womens Health 2006 51 39-44.
• Riskin A, et al. Infant of a diabetic mother.
  UpToDate Jan 2008.
• Avery, et al. Diseases of the Newborn.
• McGowan, Jane. Neonatal Hypoglycemia.
  Neoreviews July 1999.
• Cordero L, et al. Management of Infants of
  Diabetic Mothers. Arch Pediatr Adolesc Med 1998
  152 249-254.
• Van Howe RS, et al. Hypoglycemia in Infants of
  Diabetic Mothers Experience in a Rural
  Hospital. Am J Perin 2006 23 105-110.
• Yang J, et al. Fetal and Neonatal Outcomes of
  Diabetic Pregnancies. Obstetrics Gynecology
  2006 108 no 3, part 1 644-650.
• Nold JL, et al. Infants of diabetic mothers.
  Pediatr Clinic N Am 2004 51 619-637.
• Bertini AM, et al. Perinatal Outcomes and the
  use of oral hypoglycemia agents. J. Perinat.
  Med. 2005 33 519-523.
• Rozance PJ, et al. Hypoglycemia in newborn
  infants Features associated with adverse
  outcomes. Biol Neonate 2006. 90 (2) 74-86.
• Cowett, Richard. Neonatal Care of the Infant of
  the Diabetic Mother. NeoReviews 2002. E190
• Chan, SW. Neonatal Hypoglycemia. UpToDate Jan
  2008.
• Stewart DR, et al. Neonatal Small Left Colon
  Syndrome. Ann. Surgery December 1997.

40
Infants of Diabetic MothersPrevious Study at
Strong

• Haidar-Ahmad et al
• January 2003-June 2005
• Retrospective study
• Asymptomatic infants of medication dependant
  diabetic mothers (IMDDM)
• 35 weeks
• Primary Outcome
• Hypoglycemia (BGlt 40mg/dL) requiring intravenous
  dextrose

41
Infants of Diabetic MothersPrevious Study at
Strong

• Haidar-Ahmad et al
• Data
• Four statistically significant risk factors
• Risk score for needing IV dextrose
• ? risk score ? ?risk IV dextrose

42
Hypoglycemia Risk Score
Risk factor Score0 Score1
Maternal BG pre-delivery lt120 mg/dl 120 mg/dl
Maternal age (AMA) lt35 years 35 years
Neonatal size AGA SGA or LGA
Neonatal DR BG 40 lt120 mg/dl lt40 or 120 mg/dl
43
Risk Score and Primary Outcome
Score Treated Total 12 n () Not treated Total 115 n () Total
0 0 54 (100) 54
1 4 (8) 49 (92) 53
2 5 (31) 11 (69) 16
3 3 (75) 1 (25) 4
Fishers exact Plt0.001
44
NBN/BIRTH CENTER vs.
NICU/SCN
45
During this study period 9.5 of infants born at
SMH required transfer to NICU.
NBN/BIRTH CENTER vs.
NICU/SCN
46
IMDDM StudyResults of Haidar-Ahmad

• If we use score 0 to triage to the NBN
• Reduce admissions to NICU by 43
• PPV 0.16 Sensitivity 1
• NPV 1 Specificity 0.47
• If we use score 0 or 1 to triage to the NBN
• Reduce admissions to NICU from DR by 83
• PPV 0.4 Sensitivity 0.67
• NPV 0.96 Specificity 0.9
• Risk of transfer to NICU/SCN from NBN/Birth
  Center is 15

47
Infants of Diabetic MothersOngoing Research

• Hypothesis 1
• A hypoglycemia risk score can be used in the
  delivery room to predict need for intravenous
  glucose in asymptomatic infants of medication
  dependant diabetic mothers.

48
Infants of Diabetic MothersOngoing Research

• Hypothesis 2
• Avoiding unnecessary separation of mother and
  infant after birth will change maternal
  satisfaction and aid in earlier establishment of
  breastfeeding.

49
Goals of my IMDDM study

• Retrospectively apply and validate the score
• Change in practice protocol for admission
  location of asymptomatic IMDDMs 35wks
• Compare maternal satisfaction and breastfeeding
  success during hospital stay

50
Test Validation in a Retrospective IMDDM Cohort

• Population
• IMDDM 35 weeks and their mothers at Highland
• Design
• Two sided continuity corrected chi-square test of
  equal proportions (? 0.05, power 0.8)
• 10 of infants with a risk score 0-1 will
  require IV dextrose
• 40 infants with a risk score gt1 will require IV
  dextrose
• 71 charts to review
• Primary Outcome
• Hypoglycemia requiring intravenous dextrose

51
Test Validation in a Retrospective IMDDM Cohort

• Analysis
• Score component bivariate analysis
• Revision dependent on statistically significant
  results
• Compare results to those of Haidar-Ahmad
• Exact logistic regression
• 4 score components
• Overall score performance
• Success at predicting hypoglycemia requiring IV
  dextrose
• Anticipated Results
• Verification that risk score performs similarly
  in separate population of IMDDMs
• Risk score of 0-1 should have NPV 0.9 in order
  to proceed with change in clinical practice

52
Retrospective Data AnalysisRisk Score Components
Bivariate analysis of need for IV Dextrose
Risk Factor Total Not treated with D10 bolus/IVF Treated with D10 bolus/IVF Risk Ratio p-value
AMA n () 25 23 2 RR 1.2 p 0.69
Maternal BG pre-delivery gt120 n () 8 7 1 RR 2 p 0.45
SGA or LGA n () 9 7 2 RR 5 p 0.04
DR BG lt40 or 120 6 3 3 RR 18 p 0.002
Stata 10, Fishers exact for p-values
53
Additional Retrospective Data Analysis
Patient Factor Test Result
Parity Kruskal-Wallis 0.04
BMI T-test 0.0005
GA weeks T-test 0.007
Vag vs. C/S Fishers exact 0.05
Class of DM Fishers exact 0.09
Infant sex Fishers exact 0.36
Med used Fishers exact 0.5
Infant size Fishers exact 0.54
54
Risk Score ComponentsHaidar Bivariate Analysis
Risk Score Components Scheurer Bivariate Analysis
Risk Factor Total Not treated with D10 bolus/ IVF Treated with D10 bolus/ IVF Risk Ratio p-value
AMA n () 28 21 7 5 p 0.001
Maternal BG pre-delivery gt120 n () 30 23 7 4.5 p 0.003
SGA or LGA n () 30 24 6 3.2 p 0.02
DR BG lt40 or 120 9 6 3 4.4 p 0.01
Risk Factor Total Not treated with D10 bolus/ IVF Treated with D10 bolus/ IVF Risk Ratio p-value
AMA n () 25 23 2 1.2 p 0.69
Maternal BG pre-delivery gt120 n () 8 7 1 2 p 0.46
SGA or LGA n () 9 7 2 5 p 0.04
DR BG lt40 or 120 6 3 3 18 p 0.00
55
Exact logistic regression             Number of
obs         78                                   
               Model score    
20.88848                                          
        Pr gt score       0.0039         glu
Odds Ratio    Suff.  2Pr(Suff.)   95
Conf. Interval          aga  
2.127882           2      0.8775      .0993077   
24.56013       matbg    2.350372          1     
0.9083      .0374158    50.60609       drbg2  
 3.803723           6      0.0252     
1.162711    14.71454         ama  
1.928495           2      0.9136     
.1159265    31.11316
Area under ROC curve 0.82
56
Score PerformanceApplication of Score to 78
Infants
Score No IV Dextrose n () IV Dextrose n () Total
0 40 (98) 1 (2) 41
1 27 (100) 0 27
2 2 (50) 2 (50) 4
3 4 (80) 1 (20) 5
4 0 1 (100) 1
Total 73 5 78
DR blood glucose weighted as 2 points P value
0.000, Fishers exact. IV Dextrose D10 bolus,
Continuous D10IVF, or Both
57
Score Performance
Total Risk Score No IV Dextrose IV Dextrose Total
Score 0 or 1 (to NBN/BC) 67 1(1.5) 68
Score 2 (to NICU/SCN) 6 4 (40) 10
Total P 0.001 Fishers exact 73 5 78
If we use score 0 or 1 to triage to the NBN/BC
use DR blood glucose ? 40 as 2 points PPV
0.40 Sensitivity 0.80 NPV 0.98 Specificity
0.92
58
Risk Score ValidationSummary

• Total risk score of 0 or 1 to triage to birth
  center/newborn nursery
• With delivery room BG 2 points
• 1.5 of babies that go to BC/NBN end up needing
  IV dextrose
• 40 of babies that go to NICU end up needing IV
  dextrose
• p-value 0.001 (Fishers exact)
• With each factor 1 point
• 3 of babies that go to BC/NBN end up needing IV
  dextrose
• 30 of babies that go to NICU end up needing IV
  dextrose
• p-value 0.01 (Fishers exact)

59
Quality Assurance

• Time to IV dextrose
• Review all IMDDMs requiring transfer to NICU/SCN
  
• Admit log daily review
• Myself
• Research staff
• No longer than 45min-1hr after 1st BG ? 24mg/dL
• If gt 1hour delay, case reviewed by primary CTG
  reviewer(s)
• Serum glucose sent upon admission to NICU
• If delayed dextrose therapy occurs in gt 5 cases
• Practice protocol will be referred to medical
  director and division chiefs for review
• Zahis study 4/107 (4) with score 0 or 1 would
  have required transfer to NICU for hypoglycemia
• 4 infants in 2.5 years 1.6 infants/year

60
Evaluation of the IMDDM Clinical Practice Change

• Score use
• Admit location
• Transfers

61
Thank you!