S. A. KAMEN1R
Neonatal Resuscitation and Newborn Outcomes in
Rural Kenya
by Steven A. Kamenir, MD
Department of Cardiology-Bader 202, Children's Hospital, Harvard Medical School, 300 Longwood Avenue,
Boston, MA 02115, U.S.A.
Summary
Neonatal resuscitation methods vary in developing countries. This study describes the delivery
experience at one rural Kenyan mission hospital, retrospectively analysing delivery data and
newborn outcomes for a 12-month period, and prospectively characterizing neonatal resuscitation
practices. Thirty-six of 878 newborns (4 per cent) suffered unfavourable outcomes, significantly
associated with caesarean, breech, and vacuum deliveries (nine infants, P < 0.01) and birthweight of
2000 g or less (10 infants, P < 0.001). Observed neonatal resuscitation practices were inconsistent and
notable for umbilical vein injections given in lieu of bag and mask ventilation. A basic neonatal
resuscitation protocol was developed. It is concluded that at one Kenyan hospital, unfavourable
newborn outcomes were significantly associated with delivery other than normal vaginal and with
birthweights of 2000 g or less. Neonatal resuscitation methods could be modified for use in this setting,
and might be most useful for term infants delivered by caesarean, breech, or vacuum deliveries.
Introduction
It is estimated that about 6 per cent of all newborns
delivered in the United States will require some type
of neonatal resuscitation.1'2 While standardized in the
United States, neonatal resuscitation practices vary in
developing countries.3'4 In Nairobi, Kenya, perinatal
mortality—the time period from 28 weeks of gestation
to 8 days post-partum—was estimated at 95 per 1000
live births, with about 50 per cent of perinatal mortality
preventable.5 In other developing countries, perinatal
mortality rates have been estimated to range from 27
to 94 per IOOO.6"8 One multicentre study in Africa
documented 5-min Apgar scores of 6 or less in 23 per
cent of newborns; of these, 41 per cent received no
neonatal resuscitation.3 Data on perinatal morbidity and
mortality for rural Kenya, as well as for many developing
countries are not known. Little data exist documenting
neonatal resuscitation practices in these settings. Table 1
highlights selected comparisons of vital statistical data
from Kenya and the United States.9
This study describes the delivery experience at one
rural mission hospital in western Kenya. Newborn
outcomes at this hospital were analysed for associations
with type of delivery and birthweight. The neonatal
resuscitation methods employed at the Kenyan hospital
were characterized. Lastly, a standardized basic neonatal
resuscitation protocol was proposed for use in this rural
setting.
Acknowledgements
I wish to thank the patients and staff at Friends Lugulu Hospital
for their kindness and thoughtful assistance.
170
Materials and Methods
Friends Lugulu Hospital is a 75-bed rural private mission
general hospital in western Kenya staffed by physicians,
physician assistants, midwives, nurses, and support staff.
Delivery records for the twelve month period ending
November 1992, were examined retrospectively for
number of deliveries, type of delivery, birthweight, and
number of multiple gestation deliveries. Antepartum
records were unavailable. Comparable records were
obtained from a tertiary care urban US teaching hospital
for the same time period.
Newborn outcomes at the Kenyan hospital were
analysed for associations between unfavourable outcomes (defined loosely in the delivery records, but
usually implying death of the infant immediately or
within days of delivery) and delivery type (e.g. caesarean
section, breech, vacuum assisted, or multiple gestation)
or newborn birth weight. More detailed intrapartum
maternal and fetal monitoring, and post-natal Apgar
scoring were at best inconsistently recorded and were not
used in subsequent data analysis.
The standard neonatal resuscitation practices at
Friends Lugulu Hospital were characterized after a
2-month period of prospective observation. Initial
evaluation of resuscitation practices took part during a
brief orientation period; afterwards, the observer became
responsible for managing the resuscitations and ethical
considerations limited a more quantitative description of
the usual resuscitation practices. Nevertheless, these
observations helped in the development of a basic
neonatal resuscitation protocol using already available
equipment.
© Oxford University Press 1997
Journal of Tropical Pediatrics
Vol. 43
June 1997
S. A KAMENIR
TABLE 1
Selected vital statistics9
Population (millions)
Births per year (thousand)
GNP per captita (US dollars)
Infant mortality rate under 1 year
(per thousand infants)
Under 5 year mortality rate
(per thousand children)
United States
Kenya
252.6
4024
21,790
9
24.4
1085
370
52
II
75
Chi-square analysis was used to compare discrete data
from the two hospitals, while both Chi-square analysis
and Fisher's exact test were utilized in looking for
associations between delivery type or birthweight and
newborn outcomes at the Kenyan hospital. A P value of
less than 0.05 was considered statistically significant.
Results
Delivery statistics from the US and Kenyan hospitals are
summarized in Table 2. Significantly fewer caesarean
sections were performed at Friends Lugulu Hospital (6 v.
19 per cent, P < 0.001). For vaginal deliveries, significantly more breech deliveries occurred at the Kenyan
hospital (3 v. 0.5 per cent, P< 0.001), while vacuum
assisted deliveries were significantly greater at the US
hospital (1 v. 0.6 per cent, P < 0.05). No significant
differences were found in the number of low birthweight
infants (2500 g or less) or in the number of multiple
gestation deliveries.
Table 3 details newborn outcomes at Friends Lugulu
Hospital. Thirty-six of 878 infants (4 per cent) had unfavourable outcomes. Significant associations were found
between unfavourable outcomes and delivery other than
normal vaginal (e.g. combined caesarean, vacuum, and
breech deliveries; P < 0.01) and with birth weight of
2000g or less (P < 0.001). Multiple gestation deliveries
were not significantly associated with unfavourable
newborn outcomes.
Prospective observation of neonatal resuscitation
practices revealed several variations from the standard
US protocol. Resuscitation equipment was infrequently
checked, incompletely stocked, and often unavailable
when needed. For example, of the three available bag
and mask ventilators, only one functioned well; one was
broken, and one was inappropriately large. Problem
deliveries were infrequently anticipated resulting in
delayed calls for additional caregivers. For many
deliveries, one attendant cared for both the mother and
the infant, often leaving the infant unattended in the
immediate minutes after delivery.
Seven deliveries were observed during the orientation
period before the observer was required to intervene.
Three of seven infants were dried partially or completely, six of seven were bulb suctioned for nasal and
TABLE
2
1 ABLE Z.
Delivery statistics: comparison between one Kenyan and US hospital
Urban US hospital
Rural Kenyan hospital
4098
4225
3326(81.2%)
22 (0.5%)
54(1.3%)
772(18.8%)
697(16.1%)
116(2.8%)
849
878
802 (94.5%)
26 (3.1%) P< 0.001
5 (0 6%) P < 0.05
47 (5.5%) P < 0.001
143 (16.3%) NS
28 (3.3%) NS
Toial deliveries
Total infants
All vagina] deliveries
breech vagina]
vacuum vaginal
Caesarean section
Birth weight < 2500 g
Multiple gestation
TABLE 3
Newborn outcomes: Friends Lugulu Hospital, Kenya
(total infants delivered: 878)
Outcomes
Total
Normal vagina]
Delivery other than normal vagina]
caesarean section
vacuum vaginal
breech delivery
Multiple gestation
Birth weight < 2000 g
Journal of Tropical Pediatrics
Vol. 43
June 1997
Favourable
Unfavourable
842
773(91.8%)
69 (8.2%)
42 (5.0%)
4 (0.5%)
23 (2.7%)
52 (6.2%)
18(2.1%)
36
27 (75.0%)
9 (25.0%) P< 0.01
5(13.9%)
1 (2.8%)
3 (8.3%)
5 (8.8%) NS
10 (27.8%) P< 0.001
171
S A. KAMENIR
• Appropriate anticipation of high risk deliveries and
fetal distress, allowing extra attendants to be called for
help prior to delivery.
• Minimum basic resuscitation equipment to be prepared in advance. This equipment would include the
following: dry towel, bulb suction, and bag and mask
ventilator.
• Routine resuscitation for all neonates to include: nasal
and oropharyngeal suctioning, drying, and proper
stimulation.
• If newborn remains apnoeic, begin bag and mask
ventilation (with oxygen, if available).
• Abandon routine use of umbilical vein injections.
FIG. 1. Proposed basic neonatal resuscitation protocol.
oral secretions, and two of seven were stimulated
appropriately. For the subset of three apnoeic infants
(1-min Apgar scores of 1, 3, and 3), two of three were
dried, suctioned, and stimulated. None of the apnoeic
infants was ventilated, but one infant was given
bicarbonate via umbilical vein injection. Five-minute
Apgar scores for each apnoeic infant were 6. These
trends of inconsistent drying and stimulation of newborns and, for apnoeic infants, the choosing of umbilical
vein injections of glucose or bicarbonate over bag and
mask ventilation were seen repeatedly at later deliveries
prior to arrival of additional help.
Based on these observations, a basic neonatal
resuscitation protocol was developed (Fig. 1). The minimum equipment needed included a dry towel, a bulb
suction device, and a bag and mask ventilator. Routine
care for all infants entailed nasal and oropharyngeal
suctioning, drying, and properly stimulating each infant.
For apnoeic newborns, bag and mask ventilation was
begun (with oxygen, if available). Routine umbilical
vein injections were discouraged.
This protocol was successfully used by the author to
resuscitate a term infant born by emergency caesarean
section for bradycardia and thick meconium stained
fluid. Apgar scores were 1 and 5 at I and 5 min,
respectively. After approximately 25 min of bag and
mask ventilation with room air, the infant's heart rate
increased above 100 beats per minute and respirations
became spontaneous. The child nursed well and went
home uneventfully after the mother's recovery.
These methods were taught to the staff at Friends
Lugulu Hospital and were readily accepted as an
alternative to the current standard of care. Time
constraints prevented prospective evaluation of this
resuscitation protocol.
Discussion
Neonatal resuscitation can prevent morbidity and
mortality from birth asphyxia. Even the most severely
compromised infants can survive with few sequelae with
aggressive resuscitation. 1 ^ 12 Only rarely (0.12 percent
of deliveries) do neonates require chest compressions
172
and inotropic medications at delivery; most respond
readily to airway stabilization and temporary positive
pressure ventilation.13 There is a greater need (approaching 50 per cent) for neonatal resuscitation at deliveries
with antenatal fetal distress and those requiring instrumental and operative intervention.
The results of this study describing the delivery
experience of one hospital in rural Kenya show a lower
incidence of caesarean section deliveries and a greater
incidence of higher risk breech deliveries compared to an
urban US teaching hospital. These data are comparable
to others obtained in Kenya and east Africa showing
caesarean section rates of 6-12 per cent and breech
delivery rates of about 2 per cent. 315 The findings of
a significant association between unfavourable outcomes
and both delivery other than normal vaginal and birth
weight 2000g, or less are also consistent with data
from this region.15
Most striking were the observations of neonatal
resuscitation, where umbilical vein injections of
sodium bicarbonate and glucose took precedence over
manuevers to assist the ventilation of distressed infants.
The practice of using intravenous bicarbonate and
glucose in the delivery room dates back more than
thirty years.16 More recently, the use of sodium
bicarbonate has been questioned1 because of possible
hypernatraemia, intraventricular haemorrhage, hypercarbia if the infant is inadequately ventilating, and transient
lowering of intramyocardial pH.2 Current recommendations limit its use only for documented or presumed
metabolic acidosis once adequate ventilation has been
established and other measures have proven unsuccessful.18 Clearly, the emphasis of all resuscitation
practices should emphasize bag and mask ventilation
as the primary intervention for apnoeic newborns.
Aggarwal estimated that 50 per cent of the perinatal
deaths in Nairobi were preventable for a variety of
reasons, including perinatal monitoring.5 The only data
available describing neonatal resuscitation in this region
documented that in infants with five minute Apgars of 6
or less, 31 per cent were resuscitated with suction and
stimulation, 18 per cent with bag and mask ventilation,
and 41 per cent had no interventions.3 Observations of
deliveries at Friends Lugulu Hospital suggest similar
patterns of neonatal resuscitation. Given the increased
rate of higher risk vaginal deliveries in this region,
improved neonatal resuscitation efforts could have a
significant impact on newborn outcomes.
The equipment necessary for the basic neonatal
resuscitation protocol proposed here likely would be
available to facilities even with limited resources. Of
note, intubation and use of medications are not pan of
the protocol, making it more accessable to a variety of
caregivers with less training required. Equally important to proper care of newborns is identifying perinatal
risk factors, monitoring labour using such tools as the
partogram,19 and responding quickly to signs of fetal
distress.3 At least at this one hospital, the teaching of this
neonatal resuscitation protocol was well received. It is
Journal of Tropical Pediatrics
Vol. 43
June 1997
S. A. KAMENIR
speculated that this protocol might be generally applicable in other developing countries where resources are
limited.
This study had several potential limitations. Delivery
records from only one Kenyan hospital were analysed;
it is likely that delivery and newborn resuscitation
practices vary widely amongst the many Kenyan health
care facilities. Furthermore, the hospital records at
Friends Lugulu Hospital were limited in detail, standardization of terms, and completeness. Prenatal records
were unavailable. Prospective observation of resuscitation practices were limited by the ethical need to
intervene for apnoeic infants. Lastly, in this region of
Kenya, many mothers deliver their infants at home. The
outcomes of these newborns are unknown.
It is concluded that at one rural Kenyan hospital,
unfavourable newborn outcomes were significantly
associated with delivery other than normal spontaneous
vaginal delivery and with birth weights of 2000g or less.
Neonatal resuscitation methods could be modified for
use in this developing country setting using simple,
inexpensive, and acceptable equipment and methods.
Newborn resuscitation might be most useful for term
infants delivered by caesarean section, breech, or
vacuum assisted deliveries, and for low birth weight
neonates (if appropriate continued support is available).
Prospective studies are necessary to document changes
in morbidity and mortality using the proposed basic
neonatal resuscitation protocol.
References
1. Emergency Cardiac Care Committee and Subcomittees,
American Heart Association. Guidelines forcardiopulmonary resuscitation and emergency cardiac care, VII. Neonatal
Resuscitation. J Am Med Ass 1992; 268: 2276-81.
2. Leuthner SR, Jansen RD, Hageman JR. Cardiopulmonary
resuscitation of the newborn. Pediat Clin N Am 1994; 41:
893-907.
3. Kinoti SN. Asphyxia of the newborn in east, central and
southern Africa. E Afr Med J 1993; 70: 422-33.
Journal of Tropical Pediatrics
Vol.43
June 1997
4. Mounanga M, Mboussou M, Essemi ML, Zinsou RD.
Devenir des cnfanls reanimes. Experience d'une matemite
universilaire africaine. Rev Fr Gynecol Obstet 1992; 87:
30-4.
5. Aggarwal VP, Mati JKG. Review of perinatal mortality at
Kenyatta national hospital Nairobi. J Obstet Gynec E Cent
Afr 1982; 1.
6. Adewunmi OA, Dawodu AH, Marinho AO. Perinatal
mortality surveys in an African teaching hospital: II The
influence of clinico-pathologic and other factors on
perinatal deaths. E Afr Med J 1984; 61: 778-86.
7. Kishan J, Soni AL, Elzouki AY, Mir NA. Perinatal
mortality and neonatal survival in Libya J Trop Pediat
1988; 34: 32-3.
8. Fauveau V, Wojtyniak B, Mostafa G. Sarder AM,
Chakraborty J. Perinatal mortality in Matlab, Bangladesh:
a community-based study. Int J Epidemiol 1990; 19:
606-12.
9. UNICEF. State of the world's children. Geneva: UNICEF,
1993.
10. Scott H Outcome of very severe birth asphyxia. Arch Dis
Childh 1976; 51: 712-16.
11. Steiner H, Neligan G. Perinatal cardiac arrest: quality of the
survivors. Arch Dis Childh 1975; 50: 696-702.
12. Thomson AJ, Searle M, Russell G. Quality of survival after
severe birth asphyxia. Arch Dis Childh 1977; 52: 620-6.
13. Perlman JM, Risser R. Cardiopulmonary resuscitation in
the delivery room: associated clinical events. Arch Pediat
Adolesc Med 1995; 149: 20-5.
14. Primhak RA, Herber SM, Whincup G, Milner RDG. Which
deliveries require paediatricians in attendance? Br Med J
1984; 289: 16-18.
15. Mati JKG, Aggarwal VP, Sanghvi HCG, Lucas S, Corkhill
R. The Nairobi birth survey III. Labour and delivery.
J Obstet Gyn E Centr Afr 1983; 2: 47-56.
16. Behrman R. Alkali therapy in the delivery room. J Pediat
1966; 69: 173-4.
17. Hein H. The use of sodium bicarbonate in neonatal
resuscitation: help or harm? Pediatrics 1993; 91: 496-7.
18. American Heart Association/American Academy of Pediatrics. Textbook of Neonatal Resuscitation. Elk Grove
Village, IL: AAP, 1990.
19. Dujardin B, De Schampheleire I, Sene H, Ndiaye F. Value
of the alert and action lines on the partogram. Lancet 1992;
339: 1336-9.
173