Keywords
feeding intolerance - growth velocity - very low birth weight infants - feeding methods
Adequate nutrition for very preterm infants is essential for optimal growth and neurodevelopmental
outcomes. Nutrition can be provided by either the parenteral or enteral route. Provision
of enteral nutrition is desirable but fraught with the risk of feeding intolerance.[1]
[2] Feeding intolerance is frequently encountered, causing feeding disruptions and delays
in reaching full feeds. Several feeding methods are practiced to reduce feeding intolerance.
The feeding methods used in preterm infants are continuous infusion (CI) or intermittent
bolus by gravity (IBG).[3] Bolus feeding stimulates cyclic surges of gastrointestinal tract hormones; it may
promote motility, maturation of intestinal lining, and enhance protein anabolism.[4]
[5]
[6]
[7] In a busy neonatal intensive care unit (NICU) setting, bolus feeds may be given
inadvertently in a much shorter time despite recommendations from the World Health
Organization to feed over 10 to 30 minutes.[8] This can be avoided by delivering bolus intermittent feeds by an infusion pump in
which the duration can be controlled. The functional limitations of the premature
infant's gastrointestinal system, such as delayed gastric emptying or intestinal transit,
could hinder its ability to handle bolus feeds.[4]
[5]
[6]
[7]
[9]
[10] Alternate method of feeding these infants is by CI, which has the potential benefit
of enhancing duodenal motor function and improving feeding tolerance.[4]
[11]
[12]
, At the same time, the CI has shown a significant fat loss in simulation studies,
which might impact infant growth parameters.[13]
[14]
[15]
Although there are several feeding methods that have the physiological basis and biological
plausibility to improve feed tolerance, only a few have been subjected to experimental
validations through randomized controlled trials (RCT).[9]
[10]
[16]
[17]
[18]
[19]
[20]
[21]
[22] Some trials have compared intermittent bolus feeding with CI feeds, but the results
are inconsistent.[17]
[18]
[19]
[21]
[22] Cochrane metanalysis comparing bolus versus continuous feeds in very low birth weight
infants concluded that the current evidence is inconclusive for determining an optimum
feeding strategy because of the small sample size and methodological limitations.[3] In addition, no clinical trials have systematically evaluated intermittent bolus
feeding given in a controlled time by infusion. With this background, we compared
three modalities of feeding: CI, intermittent bolus by infusion (IBI), and IBG in
very preterm infants for their effect on time to reach full enteral feeds, growth
parameters, and morbidities.
Materials and Methods
Trials Design, Settings, and Participants
This RCT was done in a neonatal intensive care unit of a tertiary care center in northern
India from July 2015 to May 2017. Infants ≤32 weeks and with birth weight ≤1,250 g,
in whom orogastric feeds were started within 72 hours of birth, were included. Those
born with significant congenital anomalies (gastrointestinal anomalies or lethal malformations)
were excluded.
Randomization and Blinding
Infants were randomly assigned to one of the three groups: group A, CI; group B, IBI;
and group C, IBG. This was done by an independent investigator using a computer-generated
random number table using a variable block size of 3, 6, or 9. The allotment was performed
by using sealed, sequentially numbered, opaque envelopes. The nature of the intervention
prevented us from blinding.
Intervention
In the CI group, feeds were delivered by an infusion pump continuously over 24 hours.
The syringe was loaded with expressed human milk every 6 hours or freshly prepared
formula every 2 hours. In IBI group, milk was given every 2 hours and infused over
15 minutes by an infusion pump. In the IBG group, milk was given every 2 hours over
10 to 30 minutes by gravity. The intervention was continued till the infant reached
total direct feeds. Total direct feeds were defined as feeding directly from the mother's
breast or supplemented with a cup, the volume being ad libitum without the need for
an orogastric feed. This was usually attempted at 33 to 34 weeks' postmenstrual age
(PMA), depending on the infant's clinical status.
Feeding Protocol
Infants weighing 1,000 to 1,250 g were started on feeds at a volume of 80 mL/kg/d
(total enteral nutrition), and the increment was done at a rate of 20 mL/kg/d till
infants reached full enteral feeds of 180 mL/kg/d. Infants with absent/reversed end
diastolic flow in antenatal doppler, perinatal asphyxia (Apgar's score ≤3 at 5 min),
or those who were initially hemodynamically unstable (requiring inotropes) were kept
nil per oral for 24 to 48 hours as per clinical status. Incremental feeding regimes,
as mentioned below, were subsequently started in these infants. Infants weighing <1,000 g
or severe small for gestational age (SGA, <3rd centile) were started on parenteral
nutrition and minimal enteral feeds at a volume of 10 to 20 mL/kg/d on day 1. Parenteral
nutrition was started at a fluid rate of 80 to 100 mL/kg/d with total calories of
60 to 70 kcal/kg/d, amino acids of 2 to 2.5 g/kg/d, and lipids of 1 to 2 g/kg/d. Fluids/feeds
were increased at a rate of 20 mL/kg/d as per clinical discretion. Parenteral nutrition
was continued until a 100 mL/kg/d feed volume was achieved. Every attempt was made
to give mother's own milk. In case of its nonavailability, preterm formula containing
1.8 g of protein and 79 calories in 100 mL was used. Expressed human milk was collected
and stored as per standard protocol. The milk was fortified with the bovine milk fortifier
(FM 85; Nestle, Vevey, Switzerland) once the infant reached a feed volume of 100 mL/kg/d.
The fortification was continued till the infant was feeding completely from the mother's
breast or till discharge. Feed volume was advanced till the infant reached a volume
of 180 mL/kg/d. The intervention was continued till the infants reached total direct
feeds as defined earlier. Feeding intolerance was defined as the presence of any of
the following: increase in abdomen girth >2 cm in between feeds, presence of abdominal
signs such as abdominal wall discoloration, erythema, or tenderness, and presence
of hemorrhagic or bilious residuals. Gastric residuals were not checked routinely
unless abdominal girth increased by >2 cm in between feeds. Nil per oral hour was
calculated.
Outcome Assessment
The primary outcome was time taken to reach full feeds, i.e., feed volume of 180 mL/kg/d,
and tolerated for at least 48 hours. It was calculated from the day feeding was initiated.
Secondary outcomes were episodes of feeding intolerance, nil per oral hours, time
to regain birth weight, the proportion of infants who developed necrotizing enterocolitis
(NEC) stage 2 or beyond,[23] bronchopulmonary dysplasia (BPD; oxygen requirement at 36 wk of PMA),[24] patent ductus arteriosus (PDA) requiring either medical or surgical treatment, retinopathy
of prematurity (ROP) requiring treatment, and intraventricular hemorrhage (IVH) ≥2
on cranial ultrasound,[25] culture-proven late-onset sepsis, growth velocity (GV) at the stoppage of intervention
and discharge, duration of hospital stay, and all-cause mortality.
Naked weights of infants at approximately the same time of the day were measured using
electronic balance scales that were accurate to 5 g (Sunrise Digital Baby Scale).
The weighing scale was calibrated every 6 months and maintained in a logbook. Occipitofrontal
circumference (OFC) and length were recorded weekly till the time of discharge. OFC
was measured using a paper tape placed across the frontal bones above the eyebrows
and over the occipital prominence on the back of the head. Infantometer was used to
measure length to the nearest of 1 mm. The average of the two independent measurements
for OFC and length was taken. Weight GV was noted by the 2-Point Average Weight model
from time to regain birth weight till discharge. It was calculated by dividing the
total weight difference at 2 points by the number of days and average weight. Length
difference from birth to discharge weekly was noted to calculate length increment
per week. OFC gain per week was calculated from birth till discharge.
Sample Size Calculation
The sample size was calculated from a previous study done in our institute.[26] The mean duration to reach full feeds in infants ≤1,250 g was 10.6 days with a standard
deviation (SD) of 5.7 days. A total of 171 infants with 57 infants in each group were
required to detect a difference of 3 days to reach full feeds with a power of 80%
and a two-sided significance of 5%.
Statistical Analysis
Data were analyzed using SPSS software version 19. Quantitative data with normal distribution
were compared using the Student's t-test and one-way analysis of variance test as applicable. Quantitative data with
skewed distribution were analyzed using the Mann–Whitney U test. Nonquantitative data
were compared using the chi-square or the Fischer's exact test. Time to event was
analyzed by using the Kaplan–Meier survival analysis. A two-sided p-value of less than 0.05 was considered significant.
Ethics Approval
The trial was approved by the institutional ethics committee and registered with the
Clinical Trial Registry of India. Written informed consent was obtained from parents
of eligible infants at the time of initiation of feeds.
Results
A total of 185 eligible infants were screened for enrollment. Among these infants,
39 were excluded due to various reasons ([Fig. 1]). A total of 146 infants were randomized, 49 infants each in CI and IBI group and
48 infants in IBG group. Of these, 113 infants could reach the primary outcome and
were analyzed. The baseline characteristics in the three groups were comparable. The
mean (SD) gestation in CI, IBI, and IBG group were 28.4 (2.2), 28.5 (1.9), and 28.6
(1.8) weeks, respectively. The mean (SD) birth weight was 962 (196.3), 974 (184.0),
and 961 (171.2) g in CI, IBI, and IBG group, respectively. The proportion of SGA infants
and those with antenatal doppler showing absent/reversed end diastolic flow were comparable
([Table 1]). Time to reach full feeds in the CI, IBI, and IBG group was similar (median [interquartile
range]: 13 [10–16], 11.5 [9–17], and 13 [9.5–14.2] d, respectively, p = 0.71; [Table 2]). The proportion of infants who developed feeding intolerance in the CI, IBI, and
IBG group were similar. There was no difference in time to regain birth weight and
other growth parameters at discharge ([Tables 2] and [3]). Other morbidities like the length of hospital stay, late-onset sepsis, NEC ≥2,
BPD, IVH ≥2, PDA requiring treatment, ROP requiring treatment, and mortality were
similar ([Table 4]). In subgroup analysis in infants weighing ≤1,000 g, there was no difference in
the feeding, growth-related outcomes, and duration of hospital stay ([Table 5]).
Table 1
Baseline characteristics
|
Baseline characteristics
|
Group A
CI (n = 49)
|
Group B
IBI (n = 49)
|
Group C
IBG (n = 48)
|
p-Value
|
|
Gestation (wk)[a]
|
28.4 (2.2)
|
28.5 (1.9)
|
28.6 (1.8)
|
0.91
|
|
Birth weight (g)[a]
|
962 (196.3)
|
974 (184.0)
|
961 (171.2)
|
0.92
|
|
Male
|
25 (51.0%)
|
27 (55.1%)
|
23 (47.9%)
|
0.77
|
|
SGA
|
9 (18.4%)
|
13 (26.5%)
|
11 (22.9%)
|
0.62
|
|
Vaginal delivery
|
13 (26.5%)
|
10 (20.4%)
|
11 (22.9%)
|
0.82
|
|
Complete ANS
|
40 (81.61%)
|
33 (67.3%)
|
39 (81.2%)
|
0.13
|
|
Gestational hypertension
|
19 (38.8%)
|
10 (20.4%)
|
18 (37.5%)
|
0.09
|
|
A/REDF
|
9 (18.4%)
|
10 (20.4%)
|
11 (22.9%)
|
0.85
|
|
Multiple gestation
|
19 (38.8%)
|
21 (42.9%)
|
19 (39.5%)
|
0.68
|
|
PPV at birth
|
10 (20.4%)
|
7 (14.3%)
|
8 (16.7%)
|
0.40
|
Abbreviations: A/REDF, absent/reversed end diastolic flow; ANS, antenatal steroids;
CI, continuous infusion; IBG, intermittent bolus by gravity; IBI, intermittent bolus
by infusion; PPV, positive pressure ventilation; SGA, small for gestational age.
Note: All values are expressed as n (%) unless specified otherwise.
a Mean (standard deviation).
Table 2
Outcomes in infants who reached primary outcome
|
Outcomes
|
Group A
CI (n = 41)
|
Group B
IBI (n = 38)
|
Group C
IBG (n = 34)
|
p-Value
|
|
Full feeds (d)[a]
|
13 (10–16)
|
11.5 (9–17)
|
13 (9.5–14.2)
|
0.71
|
|
100 mL/kg/d (d)[a]
|
7 (5–11)
|
6 (4–12)
|
9 (5–10.2)
|
0.76
|
|
Duration of PN[a]
|
8 (6–13)
|
10 (7–14)
|
9 (6–12)
|
0.81
|
|
TBW (d)[a]
|
9 (6–12)
|
11 (6.5–13)
|
10 (7–12)
|
0.38
|
|
DOI day[a]
|
37 (28–55)
|
43 (27.5–57.5)
|
35 (25–57)
|
0.70
|
|
Feeding intolerance, n (%)
|
21 (51.2%)
|
20 (52.6%)
|
22 (64.7%)
|
0.45
|
|
NPO hours[a]
|
60 (22–102)
|
56 (10–102)
|
48 (18–120)
|
0.69
|
|
EHM (%)
|
63.3
|
62.1
|
66.1
|
0.53
|
Abbreviations: CI, continuous infusion; DOI, duration of Intervention; EHM, expressed
human milk percent of total milk consumed during neonatal intensive care unit stay;
IBG, intermittent bolus by gravity; IBI, intermittent bolus by infusion; NPO, nil
per oral; PN, parenteral nutrition; TBW, time to regain birth weight.
a Median (interquartile range).
Table 3
Growth characteristics
|
At discharge
|
Group A
CI (n = 40)
|
Group B
IBI (n = 37)
|
Group C
IBG (n = 33)
|
p-Value
|
|
Weight growth velocity (g/kg/d)
|
13.6 (3.3)
|
13.2 (3.6)
|
14.1 (2.9)
|
0.57
|
|
Length increment (cm/wk)
|
0.73 (0.17)
|
0.75 (0.17)
|
0.77 (0.17)
|
0.70
|
|
OFC gain (cm/wk)
|
0.66 (0.12)
|
0.61 (0.14)
|
0.62 (0.14)
|
0.34
|
Abbreviations: CI, continuous infusion; IBG, intermittent bolus by gravity; IBI, intermittent
bolus by infusion; OFC, occipitofrontal circumference.
Note: Variables expressed as mean (standard deviation).
Table 4
Other secondary outcomes
|
Outcome
|
Group A
CI (n = 49)
|
Group B
IBI (n = 49)
|
Group C
IBG (n = 48)
|
p-Value
|
|
Length of hospital stay[a]
|
44 (29.5–62.5)
|
50 (32.5–70)
|
43 (26–64.7)
|
0.58
|
|
Late-onset sepsis
|
25 (51%)
|
23 (46.9%)
|
20 (41.7%)
|
0.65
|
|
NEC ≥ stage 2
|
1 (2%)
|
2 (4.1%)
|
2 (4.2%)
|
0.80
|
|
BPD
|
8 (16.3%)
|
10 (20.4%)
|
8 (17.4%)
|
0.86
|
|
IVH ≥ grade 2
|
0 (0%)
|
1 (2.0%)
|
2 (4.2%)
|
0.35
|
|
PDA requiring treatment
|
1 (2.0%)
|
1 (2.0%)
|
1 (2.2%)
|
0.44
|
|
ROP requiring treatment
|
3 (6.1%)
|
4 (8.2%)
|
1 (2.2%)
|
0.51
|
|
Mortality
|
4 (8.2%)
|
7 (14.3%)
|
5 (10.4%)
|
0.77
|
Abbreviations: BPD, bronchopulmonary dysplasia; CI, continuous infusion; IBG, intermittent
bolus by gravity; IBI, intermittent bolus by infusion; IVH, intraventricular hemorrhage;
PDA, patent ductus arteriosus; ROP, retinopathy of prematurity.
Note: Variables expressed as n (%) unless specified otherwise.
a Median (interquartile range).
Table 5
Outcomes related to feeds in infants ≤1,000 g
|
Outcomes
|
Group A
CI (n = 19)
|
Group B
IBI (n = 20)
|
Group C
IBG (n = 18)
|
p-Value
|
|
Full feeds (d)[a]
|
15 (15–20)
|
14.5 (9.2–18.5)
|
14 (11.7–18)
|
0.24
|
|
100 mL/kg/d (d)[a]
|
10 (6–13)
|
8.5(5–12)
|
10 (6–12.5)
|
0.44
|
|
Duration of PN[a]
|
11 (8–15)
|
10 (7–14)
|
11 (6–15.7)
|
0.57
|
|
TBW (d)[a]
|
10 (6–12)
|
13 (9–18)
|
11.5 (7–14)
|
0.25
|
|
Feeding intolerance, n (%)
|
12 (63.2%)
|
13 (65.0%)
|
15 (83.3%)
|
0.33
|
|
NPO hours[a]
|
96 (49–150)
|
46 (11–102)
|
72 (30–147)
|
0.15
|
|
NEC ≥ 2, n (%)
|
0 (0%)
|
2 (10%)
|
1 (5.6%)
|
0.37
|
|
DOI day[a]
|
51 (39–86)
|
56 (50–76)
|
55 (38.2–64.7)
|
0.80
|
|
Length of hospital stay[a]
|
72 (50–103)
|
68 (59–84)
|
60 (46–86)
|
0.60
|
Abbreviations: CI, continuous infusion; DOI, duration of intervention; IBG, intermittent
bolus by gravity; IBI, intermittent bolus by infusion; NEC, necrotizing enterocolitis;
NPO, nil per oral; PN, parenteral nutrition; TBW, time to regain birth weight.
a Median (interquartile range).
Fig. 1 Flowchart of infant enrollment.
Discussion
The primary objective of our study was to evaluate the three modalities of feeding,
i.e., by CI, IBI over a period of 15 minutes and IBG in very preterm infants ≤32 weeks'
gestation or ≤1,250 g birth weight for their impact on time to reach full enteral
feeds. We included intermittent feeding by infusion pump as an innovative feeding
method with the premise that in a busy intensive care unit, delivering intermittent
feeding by gravity within the recommended duration may be a challenge. No randomized
controlled studies have systematically evaluated these three methods to assess feed
tolerance in preterm infants. Although multiple trials have been conducted on continuous
feeding versus intermittent bolus feeding by gravity, the results are conflicting
and inconclusive.
In our study, there was no difference in the time (days) to reach full feeds in the
three groups. Both bolus and continuous feeding strategies have some physiological
advantages, but these did not translate into a clinically significant outcome of earlier
achievement of full enteral feeds. Akintorin et al[17] randomized a similar cohort of preterm infants with birth weight 700 to 1,250 g
in two groups: continuous versus intermittent bolus feeding by gravity. There was
no difference in the days to reach 100 kcal/kg/d for at least 48 hours. A similar
intervention done by Silvestre et al[18] did not find a difference in days to reach full feeds in the two groups. Our results
are consistent with the findings of these two RCTs, which included similar cohorts
and comparable interventions. Similar findings were observed by Rövekamp-Abels et
al,[21]
[22] who enrolled infants with higher birth weight (<1,750 g) in comparison to our study.
Two randomized trials have shown conflicting results and favor either continuous or
intermittent feeding methods. Dsilna et al[10] randomized preterm infants 24 to 29 weeks' gestation and birth weight <1,200 g in
three groups: continuous nasogastric, intermittent nasogastric, and intermittent orogastric
feeds. Infants who received continuous feeds reached full feeds earlier as compared
with both intermittent groups taken together. Further, the authors reported that infants
with birth weight <850 g randomized to the continuous feeding group had greater benefits
in terms of reaching full feeds than the whole cohort. On the other hand, another
RCT by Dollberg et al[9] showed results in favor of intermittent bolus feeds over continuous feeds. They
enrolled infants <1,250 g and observed that infants in the intermittent bolus group
reached full feeds 8 days earlier than continuous feeds. However, this was a small
pilot study in which only 28 infants were randomized. Cochrane meta-analysis[3] included these studies and found no difference in time to achieve full enteral feeds.
On the contrary, in a recent metanalysis, Wang et al included one additional study[21]
[22] and reported that the time to reach full feeds was longer in continuously fed infants
than in intermittent bolus fed infants.[27] The discrepancy in findings of these two metanalysis calls for well-designed randomized
controlled studies to guide an optimum feeding strategy in very preterm infants.
Tolerance to achieve full feeds and successful weaning from parenteral fluids are
important determinants of clinical outcomes in preterm infants. In our study, there
was no difference in the proportion of infants who developed feeding intolerance or
duration of nil per oral hours in the three groups. Various authors have defined feeding
intolerance differently but have not observed differences in continuous or bolus feeding
methods.[16]
[17]
[19]
[20] In contrast, Rövekamp-Abels et al[21]
[22] observed less gastric residual volume and feed interruptions in the bolus feeding
method. We also did not find significant differences in other secondary outcomes like
days to reach feed volume of 100 mL/kg/d, duration of parenteral nutrition, time to
regain birth weight, length of hospital stays, and proportions of infants who developed
NEC.
Growth in preterm infants is an important clinical outcome and has been evaluated
in most studies that have compared various feeding strategies. In our study, there
was no difference in weight GV (g/kg/d), length increment (cm/wk), and OFC gain (cm/wk)
in the three groups. Three other studies have reported similar results.[16]
[18]
[20] On the contrary, Schanler et al[19] observed that infants fed by the continuous feeding method gained weight slower
than infants fed by the intermittent bolus feeding method. Preclinical studies have
shown that there is a greater loss of fat in the CI method, postulating poor weight
gain.[13]
[14]
[15] Cochrane systematic review[3] on continuous and intermittent feeding methods in preterm infants observed a trend
toward early discharge for infants less than 1,000 g birth weight with results favoring
the continuous feeding method. With this background, we performed a subgroup analysis
to evaluate the impact of three feeding strategies in infants <1,000 g birth weight.
Contrary to the meta-analysis, we did not find a difference in the duration of hospital
stay in the three groups.
The strength of our study is that we compared a novel method of intermittent feeding
by infusion given over a stipulated period with other conventional feeding methods.
In addition, we had an evidence-based aggressive feeding strategy derived from our
unit's experience[28] and contemporary publications.[29]
[30] Feeding protocol and feeding intolerance were well-defined to maintain uniformity
in the study. However, our study had certain limitations. Intervention could not be
blinded, and the study sample size could not be completed. The calculated sample size
was 171 infants, but we could randomize only 146 infants because this study was time-bound,
being part of a postgraduate dissertation. In addition, due to the high transfer out
and mortality rates, only 113 infants could reach the primary outcome. We also found
that the growth parameters of enrolled infants were below recommended norms.[31] This could be due to fortification commencement after infants reached 100 mL/kg/d
of feeds, slow increments of feeding at 15 to 20 mL/kg/d, and a significant proportion
of severe SGA infants. A more aggressive fortification and a larger incremental regime
would probably have resulted in better growth.
Conclusion
In preterm infants ≤32 weeks of gestation and birth weight ≤ 1,250 g, there was no
difference in time to reach full enteral feeds in the three feeding methods: continuous
feeding or intermittent feeding by infusion pump or intermittent feeding by gravity.
Other feeding-related outcomes, growth parameters, and clinical morbidities were also
similar.
