Shripada Rao wrote the protocol, performed literature search and selected relevant studies, assessed the methodological quality of the studies, extracted study data, entered the data into RevMan, corresponded with authors of the studies to get additional information, wrote the review and compiled other references.
Premature babies who are not yet ready to digest and absorb full milk feeds are given lipids through a vein to improve their nutrition. The issue of how early the lipids should be introduced to be advantageous while causing no harm is a matter of debate. The review found that, while no side effects were reported, there was no statistically significant benefit of introducing lipids before five days of age. Long term effects of early introduction of lipids in premature babies have not been reported.
Lipids are essential components of parenteral nutrition for preterm infants. Parenteral lipids can be administered through a peripheral vein, and their early introduction offers the potential advantages of increasing energy intake and providing essential fatty acids and fat soluble vitamins. Concerns have been raised about potential adverse effects including chronic lung disease (CLD), increase in pulmonary vascular resistance, impaired pulmonary gas diffusion, bilirubin toxicity, sepsis and free radical stress.
To determine the safety and efficacy of 'early' (≤ 5 days after birth) introduction of lipids to parenterally fed preterm infants.
Eligible studies were identified by searching MEDLINE (December 2004), EMBASE 1980 - 2004, Oxford Database of Perinatal Trials, Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 4, 2004) and CINAHL (December 1982 - December 2004). Abstracts of the Society for Pediatric Research were hand searched from 1980 to 2004 inclusive. No language restrictions were applied.
All randomised or quasi randomised controlled trials comparing 'early' versus 'no early' introduction of lipids to preterm infants.
Data were sought regarding effects on growth and risk of CLD or death, other respiratory morbidities including duration of respiratory support, duration of supplemental oxygen, the need for home oxygen, pneumothorax (PTX), pulmonary haemorrhage and pulmonary interstitial emphysema (PIE), ≥ stage 2 necrotizing enterocolitis (NEC), retinopathy of prematurity (ROP), patent ductus arteriosus (PDA), sepsis, intraventricular haemorrhage (IVH), clinically significant thrombocytopenia and significant jaundice. Methodological quality of eligible studies was assessed according to allocation concealment, blinding of intervention, blinding of outcome assessment and completeness of follow up. When appropriate, meta-analysis was conducted to provide a pooled estimate of effect. For categorical data the Typical relative risk (RR), Typical risk difference (RD) and number needed to treat (NNT) with 95% confidence intervals (CI) were calculated. Continuous data were analysed using weighted mean difference (WMD).
Five studies (n = 397) were included in the review. All studies compared
the effectiveness and safety of 'early' introduction versus 'no early' introduction
of lipids in preterm infants. The timing of introduction of 'early lipids'
ranged from < 12 hours after birth to day five of life. The timing of
introduction of lipids in the 'no early' lipid group ranged from day six
after birth to day 14 after birth. The initial dose ranged from 0.5 - 1 g/kg/day
with gradual daily increments up to a maximum of 2.5 - 3.5 g/kg/day.
For the primary outcomes (growth, death and CLD), there was no statistically
significant difference between the 'early' lipid and 'no early' lipid groups.
Days to regain birth weight: [WMD 0.59 (95% CI -2.41, 3.58); two trials; N = 71].
Rate of weight gain (g/day) during period of hospital stay: [MD -2.40 (95% CI -5.30, 0.50); one trial; N = 129]
Death (irrespective of time): [Typical RR 1.04 (95% CI 0.69, 1.56); Typical RD 0.01 (95% CI -0.07, 0.08); five trials; N = 397]
Neonatal deaths: [Typical RR 1.35 (95% CI 0.78, 2.34); Typical RD 0.05 (95% CI -0.04, 0.13); four trials; N = 268].
CLD: [Typical RR 1.10 (95% CI 0.81, 1.49); Typical RD 0.04 (95% CI -0.09, 0.17); two trials; N = 193].
For the secondary outcomes of other respiratory morbidities including duration of respiratory support, duration of supplemental oxygen, PTX, pulmonary haemorrhage, PIE, NEC, ROP, PDA, sepsis, IVH and significant jaundice, there were no statistically significant differences between 'early' and 'no early' lipid groups.
No statistically significant effects of 'early introduction' of lipids on short term nutritional or other clinical outcomes, either benefits or adverse effects, were demonstrated in the studies reviewed. Based on the currently available evidence, 'early' initiation of lipids (≤ 5 days after birth) can not be recommended for short term growth or to prevent morbidity and mortality in preterm infants.
Lipids are essential components of parenteral nutrition for preterm infants to provide essential fatty acids (EFA) and to meet their high energy needs. Preterm infants have very limited endogenous lipid stores (Koletzko 2002). Parenteral lipids are an attractive source of nutrition early in postnatal life because of their high energy density (Ziegler 2002), energy efficiency (Flatt 1985), isotonicity with plasma, and suitability for administration through a peripheral vein (Skeie 1988). Parenteral lipid infusion enables the delivery of fat soluble vitamins.
Even a short delay of 3-7 days in supplying lipids to parenterally fed preterm infants leads to biochemical EFA deficiency (Gutcher 1991; Friedman 1976; Foote 1991; Lee 1993). EFA deficiency increases antioxidant susceptibility in preterm infants (Tomsits 2000). Prenatal and postnatal EFA deficiency reduces body and brain weights (Van Aerde 2004).
EFA deficiency can be prevented with introduction of as little as 0.5 to 1 g/kg/day lipid infusion (Cooke 1985; Gutcher 1991). Prevention of EFA deficiency in preterm infants could theoretically decrease the prevalence of complications associated with free radical formation such as CLD and ROP (Tomsits 2000). Experimental animal studies suggest that increasing polyunsaturated fatty acids (PUFA) can confer a protective effect against the toxic effects of hyperoxia on the newborn animal lung (Sosenko 1988; Sosenko 1991).
The provision of parenteral lipids to sick preterm infants early in
postnatal life may have risks. Concerns have been expressed about the potential
adverse effects of 'early' lipid infusions in preterm infants including increased
rates of sepsis due to coagulase negative staphylococci (Avila-Figueroa 1998), CLD and mortality (Cooke 1991).
Reported pulmonary complications include deposition of fat globules in capillaries,
alveolar macrophages and pulmonary arteriolar lining cells (Barson 1978; Levene 1980), interference with pulmonary gas diffusion (Greene 1976) and an increase in pulmonary vascular resistance (Prasertsom 1996).
Lipid infusion in adults with adult respiratory syndrome (ARDS) causes deterioration
in lung functions and haemodynamics through activation of inflammatory mediators
(Lekka 2004). Some studies have shown that lipid infusions convey an increased free radical stress (Pitkanen 1991; Pitkanen 1998; Helbock 1993; Neuzil 1995).
Free fatty acids released after lipolysis of the parenteral lipids may displace
bilirubin from albumin binding sites, resulting in increased levels of unbound
bilirubin and an increased risk of kernicterus (Spear 1985). Thrombocytopenia was reported as one of side effects of lipid infusion (Lipson 1974). Most of the current literature suggests that there is no effect of parenteral lipid infusion on platelet number or function (Spear 1990; Herson 1989).
The American Academy of Pediatrics (AAP 2003)
gives recommendations about the rate of lipid infusion, but not guidelines
about how early lipids can be started. Reviews on the subject recommend early
initiation of lipids on the first or second day of life at a low dose of
0.5-1 g/kg/day, gradually increasing to 3-3.5 g/kg/day, and with daily infusion
over 20-24 hour period (Putet 2000; Ziegler 2002; Koletzko 2002; Innis 2002; Van Aerde 2004).
A meta-analysis on the subject published in abstract form (Fox 1998 and Wilson 1998) reported that 'early' lipid infusion when compared to 'late' lipid infusion is not associated with an increased risk CLD or death. It did not address the issue of potential advantages of early lipid infusion like physical growth and reduction in incidence of free-radical-injury associated diseases of prematurity like ROP, NEC and IVH. Hence we thought it is justified to do a systematic review and meta-analysis.
Primary objectives:
To assess the effect of 'early' (≤5 days) introduction of lipids on growth and risk of CLD or death in parenterally fed preterm infants.
Secondary objectives:
To assess the effect of 'early' introduction of lipids on other respiratory morbidities including duration of respiratory support, duration of supplemental oxygen, the need for home oxygen, PTX , pulmonary haemorrhage and PIE, ≥ stage 2 NEC, ROP, PDA, sepsis, IVH, clinically significant thrombocytopenia and significant jaundice.
Subgroup analyses were planned on the basis of gestational age (≤ 30 weeks, more than 30 weeks).
Randomised or quasi randomised controlled trials reporting at least one of the clinical outcomes.
Preterm infants (less than 37 weeks) who needed parenteral nutrition. Postnatal age at study entry could be ≤ 5 days.
Primary
1. Physical growth: days to regain birth weight, rate of weight gain (g/kg/day) during period of hospital stay.
2. Death: before discharge and neonatal death (≤ 28 days).
3. CLD: Oxygen therapy or any form of respiratory support at ≥ 28 days of life.
Secondary
1. Duration of respiratory support (days).
2. Duration of supplemental oxygen (days).
3. Need for home oxygen therapy.
4. Pneumothorax (PTX) diagnosed by transillumination or by X ray chest.
5. Pulmonary haemorrhage needing alterations in respiratory care or causing haemodynamic instability.
6. Pulmonary interstitial emphysema (PIE) diagnosed radiologically.
7. Necrotizing enterocolitis (NEC): Stage 2 or more on Bell's staging system (Bell 1978).
8. Retinopathy of prematurity (ROP): any stage of ROP during the weeks
after birth observed by direct or indirect ophthalmoscope as defined by ICROP
(ICROP 1984) classification.
9. Significant patent ductus arteriosus (PDA) diagnosed clinically or
by echocardiograph needing treatment either conservatively by fluid restriction,
diuretics, indomethacin /ibuprofen or surgery.
10. Sepsis: positive blood culture.
11. Intraventricular haemorrhage (IVH): all grades and severe (grade 3 or 4) of Papile classification (Papile 1978).
12. Clinically significant thrombocytopenia needing platelet transfusion.
13. Significant jaundice, necessitating phototherapy or exchange transfusion.
The standard search strategy of the Cochrane Neonatal Review Group was used. Trials were identified by MEDLINE (1966 to December 2004), EMBASE 1980 - 2004, Oxford Database of Perinatal Trials, Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 4, 2004) and CINAHL (December 1982 - December 2004). Abstracts of the Society for Pediatric Research from 1980 to 2004 inclusive were hand searched.
MEDLINE was searched using the following MeSH terms or text words: infant, newborn/OR infant, low birth weight/OR infant, very low birth weight/OR infant, premature AND intravenous fat emulsion, intralipid, intravenous lipid, intravenous fat, lipid infusion, parenteral lipid and lipid emulsion in different combinations. Articles so obtained were hand searched to identify randomised and quasi randomised controlled trials which addressed the issue of 'early' introduction of parenteral lipids to preterm infants.
Reference lists of published narrative and systematic reviews were also reviewed. No language restriction was applied. The authors of all published studies were contacted to clarify reported data or provide additional information including mean and standard deviations. Sosenko 1993 provided clarification of existing data, but no additional information was available. Brownlee 1993, Alwaidh 1996 (contact author Ryan) and Gilbertson 1991 (contact author Kovar) responded, but no additional information was available from these three authors. There was no response from Hammerman 1988.
Standard methods of the Cochrane Neonatal Review Group were used.
Selection process: Randomised and quasi randomised controlled trials fulfilling the selection criteria were included. Selection was done separately by two investigators (KS and SR).
Criteria for assessing the methodological quality of the studies: Two authors (KS and SR) assessed each study separately. Quality of included trials was evaluated independently by the authors, using the following criteria: Allocation concealment, blinding of intervention, blinding of outcome assessment and completeness of follow up. There are three potential answers to these questions: yes, no, cannot tell.
Data extraction and entry: Both authors extracted, assessed and coded separately all data for each study, using a form that was designed specifically for this review. Any standard error of the mean was replaced by the corresponding standard deviation. For each study, final data was entered into RevMan by one author (SR) and then checked by a second author (KS). At each stage any disagreement was resolved by discussion.
Statistical analysis: Standard method of the Cochrane Neonatal Review Group was used, using a fixed effect model for meta-analysis. In assessing the treatment effects, we used the Typical relative risk and Typical risk difference, with 95% confidence intervals, for categorical outcomes. If risk difference (RD) was statistically significant, number needed to treat (1/RD) was calculated as well. For outcomes measured on a continuous scale we used the weighted mean difference (WMD), with 95% confidence intervals. Heterogeneity tests were performed to assess the appropriateness of pooling the data. Results of the I2 statistic are reported.
Seven randomised control trials assessing the effect of 'early' introduction of lipids on various outcomes were identified.
Excluded studies
Two studies (Wilson 1997 and Ibrahim 2004)
were excluded from the review because, in each study, both 'early' and 'no
early' lipid groups received lipids ≤ 5 days of life.
Included studies
Five studies were included in this review.
Alwaidh 1996 was a single centre study performed at Liverpool Childrens's Hospital, England.
The quality of the trials was assessed using the criteria of the Neonatal Review Group. Assessment was predominantly based on allocation concealment, blinding of intervention, blinding of outcome assessment and completeness of follow up.
Allocation concealment:
Allocation concealment was accomplished in Alwaidh 1996, Hammerman 1988 and Sosenko 1993 using sealed envelopes. It was not clear if allocation was concealed in Brownlee 1993. Gilbertson 1991 was quasi random (alternatively, not randomly assigned) and allocation was not concealed.
Blinding of intervention:
Sosenko 1993 and Hammerman 1988
reported that blinding of intervention was not possible because of the obvious
recognizability of lipid emulsion. It is not clear whether intervention was
blinded in Alwaidh 1996, Brownlee 1993. Intervention was not blinded in Gilbertson 1991.
Blinding of outcome assessment:
Only Sosenko 1993 was credited with blinding of outcome assessment. It is not clear whether outcome assessment was blinded in Hammerman 1988; Alwaidh 1996 and Brownlee 1993. It was not blinded in Gilbertson 1991.
Completeness of follow up:
Two trials (Alwaidh 1996 and Sosenko 1993) reported complete follow up. Brownlee 1993 excluded infants who died from further analysis. Gilbertson 1991 excluded data from three infants from analysis because they required TPN for less than a week. Hammerman 1988
did not have complete follow up for ROP, 11/20 in 'early' lipid group versus
17/22 in 'no early' lipid group having their eyes examined.
Five studies fulfilled our selection criteria and were included in this review (Alwaidh 1996; Brownlee 1993; Gilbertson 1991; Hammerman 1988 and Sosenko 1993).
They included a total of 397 preterm infants. There was no disagreement regarding
inclusion/exclusion of studies, quality assessment or data extraction. Available
data were pooled and analysed as listed below. A planned subgroup analysis
of preterm infants ≤ 30 weeks gestation was not possible because of lack
of published data. The authors of all the studies were contacted but additional
data were unavailable.
Primary outcome measures
Physical growth
a. Days to regain birth weight (01.01)
Four studies (Alwaidh 1996; Gilbertson 1991; Hammerman 1988; Sosenko 1993) involving a total of 268 preterm infants reported this outcome. No individual study found a statistically significant difference between 'early' lipid and 'no early' lipid groups. Sosenko 1993 did not report the actual data. Alwaidh 1996 reported the data as median (range) 15 (0-21) days in 'early' versus 16 (8-30) days in 'no early' lipid group. Only two studies (Gilbertson 1991; Hammerman 1988) reported the data in a form that could be pooled for meta analysis. The meta-analysis of these two studies did not support a significant effect [WMD 0.59 days (95% CI -2.41, 3.58); two trials; N = 71].
b. Rate of weight gain (g/day) during period of hospital stay (01.02)
Only Brownlee 1993 reported this outcome. There was no significant difference in daily weight gain (mean ± SD) to discharge (18.6 ± 7.7 versus 21 ± 9.1 g/day, 'early' versus 'no early' lipid respectively); [MD -2.40 (95% CI -5.30, 0.50); one trial; N = 129].
Death
a. Death before discharge (01.03)
No study reported this specific outcome.
b. Death (irrespective of time) (01.04)
All five studies (Alwaidh 1996; Brownlee 1993; Gilbertson 1991; Hammerman 1988; Sosenko 1993) involving a total of 397 preterm infants reported this outcome. No individual study found a statistically significant effect of 'early' versus 'no early' lipid on this outcome. Meta analysis of all the studies did not show a significant effect of introduction of 'early' lipids on overall mortality [Typical RR 1.04 (95% CI 0.69, 1.56); Typical RD 0.01 (95% CI -0.07, 0.08); five trials; N = 397]. These are presented as post hoc findings as we did not include this outcome in our protocol.
c. Neonatal death (01.05)
Four studies involving a total of 268 preterm infants reported this outcome (Alwaidh 1996; Gilbertson 1991; Hammerman 1988; Sosenko 1993). No individual study found a statistically significant effect of 'early' versus 'no early' lipid on this outcome. Meta analysis of these four studies did not show a significant difference in the incidence of neonatal deaths between 'early' and 'no early' lipid group [Typical RR 1.35 (95% CI 0.78, 2.34); Typical RD 0.05 (95% CI -0.04, 0.13); four trials; N = 268].
Chronic lung disease (01.06)
The definition of CLD varied widely in different studies. Hammerman 1988 did not define bronchopulmonary dysplasia (BPD). They found a higher incidence of BPD in 'early' lipid versus 'no early' lipid group (14/20 versus 10/22; P < 0.10). Sosenko 1993 defined CLD as the need for supplemental oxygen for at least 28 of the first 60 days of age associated with abnormal chest X ray findings. They found no significant difference in the incidence of CLD between 'early' and 'no early' lipid group (20/70 versus 24/63 respectively). In the Gilbertson 1991 trial, diagnosis of BPD was based on history plus radiological appearance in infants requiring supplemental oxygen after 28 days of life. Only two studies (Alwaidh 1996; Brownlee 1993) involving a total of 193 preterm infants reported this outcome as per our definition (need for oxygen/respiratory support at ≥ 28 days of postnatal life). Meta-analysis of these two studies showed no significant difference in incidence of CLD between 'early' versus 'no early' lipid group [Typical RR 1.10 (95% CI 0.81, 1.49); Typical RD 0.04 (95% CI -0.09, 0.17); 2 trials; N = 193].
Secondary outcome measures
Duration of respiratory support (days) (01.07)
All the five studies (Alwaidh 1996; Brownlee 1993; Gilbertson 1991; Hammerman 1988; Sosenko 1993) involving a total of 397 preterm infants reported this outcome. Hammerman 1988
reported a statistically significant increase in the number of days of mechanical
ventilation in the 'early' lipid group compared to the 'no early' lipid group
(37 ± 35 versus 21 ± 18 days respectively). No other study
showed a statistically significant effect of 'early' versus 'no early' lipids
on this outcome. Brownlee 1993 reported the data as median (range): 8.5 (1-45) days in 'early' versus 8 (1-95) days in 'no early' lipid group. Alwaidh 1996 also reported the data as median (range): 15 (2-91) days in 'early' versus 13 (2-60) days in 'no early' lipid group. Sosenko 1993
reported the data separately for two weight categories. In the 600- 800 g
category, the median duration of mechanical ventilation was 40 days in the
'early' versus 46 days in the 'no early' lipid group. In the 801-1000 g category,
it was 37 days in the 'early' versus 25 days in 'no early' lipid group. Only
two studies (Gilbertson 1991; Hammerman 1988)
reported the data in a form that could be pooled for meta-analysis. The meta-analysis
of these two studies showed no statistically significant difference between
'early' versus 'no early' lipid group [WMD 0.89 days (95% CI -8.61,10.40;
two trials; N = 71]. However caution should be exercised in interpreting
this result because the analysis showed significant heterogeneity between
the two studies.
Duration of supplemental oxygen (days) (01.08)
Four studies (Brownlee 1993; Gilbertson 1991; Hammerman 1988; Sosenko 1993) involving 333 preterm infants reported this outcome. Only Hammerman 1988
reported a statistically significant increase in the number of days on supplemental
oxygen in the 'early' lipid group compared to 'no early' lipid group (51
± 39 versus 28 ± 23 days). Other three studies (Brownlee 1993; Gilbertson 1991; Sosenko 1993) did not find any statistically significant differences. Brownlee 1993
reported the data as median (range): 19.5 (2 - 75 days) versus 20.5 (2 -
127) days; 'early' versus 'no early' lipid respectively. Sosenko 1993
reported the data as median. No significant differences in this outcome were
noted in both birth weight categories. In the 600-800 g category it was 31.5
versus 40 days ('early' versus 'no early' lipid); In the 801-1000 g category,
it was 18 versus 17 days ('early' versus 'no early' lipid). Only two studies
(Gilbertson 1991; Hammerman 1988)
reported the data in a form that could be pooled for meta-analysis. The meta-analysis
of these two studies showed no statistically significant difference between
'early' and 'no early' lipid groups [WMD 5.50 days (95% CI -8.22, 19.22);
two trials; N = 71]. However caution should be exercised in interpreting
this result because the analysis showed significant heterogeneity between
the two studies.
Home oxygen (01.09)
Only one study (Hammerman 1988) reported this outcome. The need for home oxygen therapy was higher in 'early' lipid compared to 'no early' lipid group (7/20 versus 0 /22 ). [RR 16.43 (95% CI 1.00, 270.41); RD 0.35 (95% CI 0.14, 0.56); one trial; N = 42].
Pneumothorax (01.10)
Only one study (Sosenko 1993) reported
this outcome. There was no significant difference in the incidence of PTX
between 'early' and 'no early' lipid groups [RR 0.54 (95% CI 0.21, 1.40);
RD -0.07 (95% CI -0.18, 0.04); one trial; N = 133].
Pulmonary haemorrhage (01.11)
Only one study (Sosenko 1993) reported
this outcome. In the 600-800 g category, there was a statistically significant
increase in the incidence of pulmonary haemorrhage in the 'early' lipid versus
'no early lipid' group (11/42 versus 3/37). In the 801-1000 g category, there
was no significant difference between 'early' and 'no early' lipid groups
(6/28 versus 4/26). When the data were combined, there was a trend towards
increase in pulmonary haemorrhage in 'early' lipid group compared to 'no
early' lipid group which was of borderline statistical significance [RR 2.19
(95% CI 0.97, 4.92); RD 0.13 (95% CI 0.00, 0.26); one trial; N = 133].
Pulmonary interstitial emphysema (01.12)
Only one study (Sosenko 1993) reported
this outcome. There was no significant difference in the incidence of PIE
between 'early' versus 'no early' lipids [RR 0.99 (95% CI 0.45, 2.17); RD
0.00 (95% CI -0.13, 0.12); one trial; N = 133].
Necrotizing enterocolitis (≥ stage 2) (01.13)
None of the studies reported this outcome specifically.
Necrotizing enterocolitis (any stage) (01.14)
Three studies reported this outcome (Gilbertson 1991; Hammerman 1988; Sosenko 1993) and none of them showed a statistically significant difference. Meta analysis did not reveal a significant difference in the incidence of NEC between 'early' versus 'no early' lipid groups [Typical RR 0.82 (95% CI 0.34, 1.98); Typical RD -0.02 (95% CI -0.10, 0.06); three trials; N = 204)]. These are presented as post hoc findings as we did not include this outcome in our protocol.
Retinopathy of prematurity (any stage) (01.15)
Hammerman 1988 reported an increase in the incidence of ROP (any stage) in the 'early' lipid group compared to 'no early' lipid group ( 8/11 v 4/17). Gilbertson 1991 and Sosenko 1993 did not show any significant difference. Meta-analysis of these three studies (Gilbertson 1991; Hammerman 1988 and Sosenko 1993) did not show a significant difference in the incidence of ROP between 'early' v 'no early' lipid groups [Typical RR 1.02 (95% CI 0.74, 1.41); Typical RD 0.01 (95% CI -0.12, 0.14); three trials; N = 204].
Patent ductus arteriosus (clinically significant) (01.16)
None of the studies reported on this outcome as per our definition. But four studies (Brownlee 1993; Gilbertson 1991; Hammerman 1988 and Sosenko 1993) reported on the incidence of PDA and none of them showed a statistically significant difference between the two groups. Brownlee 1993 reported that there was no significant difference between the groups regarding the incidence of PDA, but data were not given. Meta-analysis of the other three studies (Gilbertson 1991; Hammerman 1988 and Sosenko 1993) showed a trend towards decrease in the incidence of PDA in 'early' versus 'no early' lipid groups which did not reach statistical significance [Typical RR 0.84 (95% CI 0.66, 1.06); Typical RD -0.10; (95% CI -0.23, 0.03); three trials; N = 204)]. These are presented as post hoc findings as we did not include this outcome in our protocol.
Sepsis (01.17)
None of the studies reported on the incidence of sepsis as defined in our protocol (positive blood culture). Gilbertson 1991 defined it as either blood culture positive or suggestive clinical picture with hematologic evidence indicating infection. There was no statistically significant difference in the incidence of sepsis between 'early' and 'no early' lipid groups (2/16 and 5/13 respectively). Sosenko 1993 did not define sepsis. They reported sepsis survivors (%). There were no differences between the two groups. Hammerman 1988 reported in the discussion that there was no clinical evidence of increased susceptibility to bacterial infection between the groups, but data were not given. Meta-analysis could not be performed.
Intraventricular haemorrhage (any grade) (01.18)
Four studies (Brownlee 1993; Gilbertson 1991; Hammerman 1988 and Sosenko 1993) involving 333 preterm infants reported this outcome. None of them showed a statistically significant difference in the incidence of IVH (any grade) between the 'early' and 'no early lipid' groups. Brownlee 1993 reported that there was no significant difference between the groups regarding the incidence of IVH, but data were not given. Meta analysis of three studies (Gilbertson 1991; Hammerman 1988 and Sosenko 1993) showed a trend towards decrease in the incidence of IVH in the 'early' lipid group which did not reach statistical significance [Typical RR 0.74 (95% CI 0.53, 1.04); Typical RD -0.12 (95% CI -0.26, 0.01); three trials; N = 204)].
Intraventricular haemorrhage (Grade 3 and 4) (01.19)
Only Hammerman 1988 involving 42 preterm infants reported this outcome and gave the actual data. There was no statistically significant difference between 'early' lipid and 'no early' lipid groups (2/20 versus 6/22 respectively) [RR 0.37 (95% CI 0.08, 1.61); RD -0.17 (95% CI -0.40, 0.06); one trial; N = 42]. Sosenko 1993 reported that there was no significant difference in the incidence of grade 3 and 4 IVH between two groups but data were not given.
Clinically significant thrombocytopenia needing platelet transfusion (01.20)
None of the studies reported this specific outcome. Hammerman 1988 reported no significant difference in the incidence of thrombocytopenia between 'early' lipid (5/20) versus 'no early' lipid group (6/22). The definition of thrombocytopenia was not clear. Gilbertson 1991 defined thrombocytopenia as platelet count < 150 ×103 . There was no significant difference between 'early' ( 5/13) and 'no early' lipid (7/16) groups.
Significant jaundice (01.21)
Only Gilbertson 1991 involving a total of 29 preterm infants reported this outcome and gave the data. Significant jaundice was defined as serum bilirubin > 200 mmol/litre (11.7 mg/dl) and requirement for phototherapy. They found no significant difference between 'early' v 'no early' lipid groups [RR 1.14 (95% CI 0.47, 2.75); RD 0.05 (95% CI -0.31, 0.41); one trial; N = 29]. Brownlee 1993 reported that there was no significant difference in the incidence of significant jaundice between the two groups, but definition and data were not given. Hammerman 1988 reported no significant difference in the bilirubin levels between 'early' versus 'no early' lipid groups on day one, three and five.
The outcome of 'early' administration of lipids on benefits and adverse effects in 397 preterm infants from five RCTs have been reported in this review. For the main outcomes of growth, death and CLD, there was no statistically significant difference between 'early' v 'no early' lipid administration. There were no significant differences in the secondary outcomes of respiratory morbidity, significant jaundice, NEC, ROP, PDA, sepsis, IVH and thrombocytopenia. An increased incidence of mortality and pulmonary haemorrhage was reported by Sosenko 1993 in the subgroup of babies with birth weight 600-800 g. However significantly fewer babies in this subgroup had received antenatal corticosteroids which may have contributed to the morbidity and mortality. Hammerman 1988 reported that 'early' administration of lipids to preterm infants is associated with increased incidence of CLD and ROP which was postulated as mediated by increased levels of the vasoconstrictor thromboxane B2. Similar results were not replicated by the subsequent RCTs. The meta-analysis revealed that there was a trend towards beneficial effect of 'early' lipid administration on the incidence of PDA and CNS haemorrhage but both did not reach statistically significant levels.
A meta-analysis (published as abstract: Fox 1998; Wilson 1998)) of six prospective controlled trials in preterm infants also reported that 'early' lipid infusion is not associated with an increased risk of death or CLD when compared to 'late' lipid infusion. In addition to Alwaidh 1996; Brownlee 1993; Gilbertson 1991; Hammerman 1988 and Sosenko 1993 which are included in this review, they also included Wilson 1997 in the analysis. As mentioned previously, we excluded Wilson 1997 because in both 'early' and 'no early' lipid groups, lipids were started ≤ 5 days of life. In addition to effects on death and CLD that were reported in that meta-analysis, our review compared the effect of 'early' versus 'no early' introduction of lipids on physical growth and other morbidities of prematurity.
Essential fatty acid deficiency increases antioxidant susceptibility in preterm infants (Tomsits 2000). Prevention of EFA deficiency by 'early' introduction of lipid emulsions could theoretically decrease the prevalence of complications associated with free radical formation such as BPD and ROP. Experimental animal studies suggest that increasing lung lipid polyunsaturated fatty acids (PUFA) can confer a protective effect against the toxic effects of hyperoxia on the newborn animal lung (Sosenko 1988; Sosenko 1991). These theoretical advantages of 'early' lipid administration were generally not found in clinical trials. The possible explanation may be because the morbidities of the preterm infant are multifactorial in nature rather than just related to free radical injury. Secondly the lipid emulsion itself carries an increased free radical stress (Pitkanen 1998; Helbock 1993). Elevated levels of oxidized lipids can be formed during its clinical use, especially when combined with phototherapy. Because lipid hydroperoxides are cytotoxic, inadvertent infusion of rancid lipid may add to the numerous problems encountered by preterm infants (Neuzil 1995). The apparent lack of benefit on short term physical growth may be due to the very small time difference (five days) between 'early' and 'no early' lipids as defined in this review.
Essential fatty acids are necessary for brain development (Jensen 2002). It is not clear whether a short duration of deficiency of essential fatty acids will affect long term neurodevelopmental outcomes in preterm infants. The studies in our review did not address the issue of long term neurodevelopmental outcomes.
No statistically significant effects of 'early' introduction of lipids on short term nutritional or other clinical outcomes, either benefits or adverse effects, were demonstrated in the studies reviewed. Based on the currently available data, 'early' introduction of lipids (≤ 5 days after birth) cannot be recommended for short term growth or to prevent morbidity and mortality in preterm infants.
An RCT with sample size adequate to estimate effects on long term growth, respiratory and neurodevelopmental outcomes is required to determine efficacy and safety of 'early' introduction of lipids in preterm infants. Future research needs to concentrate on infants less than 28 weeks gestation as these infants have high risk of short term morbidity, mortality and long term neurodevelopmental sequelae.
We are thankful to Sosenko IRS, Brownlee KG, Ryan SW (Alwaidh 1996) and Kovar IZ (Gilbertson 1991) for responding to our queries. We are also very thankful to Tricia Scolaro, chief librarian at Princess Margaret Hospital for Children, Perth, Western Australia for her help in literature search. We are very much thankful to Jane Bell, Research Officer, Australasian Coordinating Network for the Cochrane Neonatal Review Group for her valuable suggestions in editing the review.
None
| Study | Methods | Participants | Interventions | Outcomes | Notes | Allocation concealment |
| Alwaidh 1996 | Concealment of allocation- Yes (using sealed envelopes). Blinding of intervention- Can't tell Blinding of outcome assessment- Can't tell Completeness of follow up- Yes. | N
= 64. Infants with BW less than 1500 g who were admitted to the Intensive
Care Unit and required parenteral nutrition. Exclusion criteria were not
mentioned. The median (range) GA was 28 (23-31) weeks in 'early' lipid v
28 (23-31) weeks in 'no early' lipid group. Median (range) BW 997 g (536-1353)
in 'early' v 1006 g (542-1486) in 'no early' lipid group. | 'Early' lipid: N =32 'No early' lipids: N =32. 'Early lipid' group were introduced lipids on day 5. 'No early' lipid group were introduced lipids on day 14. Parenteral lipids as intralipid (Kabivitrium) 20% was administered in an initial dose of 0.5 g/kg/day, increasing over 5 days to 3 g/kg/day . Lipid was discontinued when >/= 50% of fluid requirements were met by milk feeds. | Total days of ventilation, age at regaining BW, age at starting milk feeds, inspired ambient oxygen concentrations on day 28 and at 36 weeks post conception. | A | |
| Brownlee 1993 | Concealment of allocation- Can't tell Blinding of intervention- Can't tell Blinding of outcome assessment- Can't tell Completeness of follow up- No | N
= 129. Preterm infants with BW less than or equal to 1750 g still requiring
IPPV at 12 hours of age with radiologic features of RDS. Infants with severe
congenital anomalies or pulmonary hypoplasia were excluded. 'Early' lipid group - median GA 29 (23 - 33), median BW 1144 g (539 -1748) . 'No early' lipid group - median GA -29 (24 - 36), median BW 1147 g (415 -1647). | 'Early' lipid group: N = 63. 'No early' lipid group: N = 66. Infants randomized to the 'early' lipid group received parenteral nutrition within the first 36 hours; those in the 'no early lipid' group received parenteral nutrition on the sixth completed day. Intralipid 20% was started at a dose of 0.5 g/kg/day and increased daily by this amount to a maximum of 3.5 g/kg/day. Lipid infusions were continuous over 24 hours. Lipid intake was reduced to 1.5 g/kg/day if the serum bilirubin concentration increased to more than 200 micromols/litre, if the infant was thought to have sepsis, or if the C reactive protein was greater than 20 mg/l. | IPPV days, IPPV + CPAP days, oxygen therapy days, age at discharge days, weight gain at 2 weeks of age, 4 weeks of age, daily weight gain to discharge, plasma triglyceride concentrations. | 25 infants who died after entry into the trial were excluded from further analysis. | B |
| Gilbertson 1991 | Concealment of allocation- No. Quasi randomised by alternatively assigning. Blinding of intervention- No Blinding of outcome assessment- No Completeness of follow up- No | N = 29 Infants BW < 1500 g, less than 6 hours of age on admission to the NICU, ventilator dependence, requirement for intensive medical and nursing care as defined by the criteria of the British Paediatric Association, and estimated need for total parenteral nutrition for at least one week. Infants with major congenital abnormalities and infants of diabetic mothers were excluded. The mean gestational age of infants in the 'early' lipid group was 28.60 ± 2.12 weeks and in the 'no early' lipid group was 28.80 ± 2.09 weeks. The mean birth weight of infants in the 'early' lipid group was 1150 ± 240 g and in the 'no early' lipid group was 1090 ± 324 g. | 'Early' lipid: N = 16 No early lipid: N = 13. 'Early' lipid group received TPN with lipids increasing from 1g/kg/day on the first day of life to 3 g/kg/day by day 4; 'no early' lipid group received an isocaloric, isovolumetric regimen that differed only in that it contained no lipid until eighth day and had a higher glucose concentration. Lipid was administered as intralipid 20% and infused at a constant rate over 20 hours. | Days
of ventilation, days in oxygen, maximum Fio2, maximum PIP, PaO2, PaCO2, BPD,
jaundice, septicemia, thrombocytopenia, periventricular haemorrhage, NEC,
PDA and ROP. | Data were given as mean and SEM. SD was calculated by reviewers. The authors excluded data from 3 infants in the 'no early' lipid group. | C |
| Hammerman 1988 | Allocation concealment- Yes (using sealed envelopes). Blinding of Intervention- No. Blinding of outcome assessment- Can't tell Completeness of follow up- Can't tell. | N
=42. Preterm infants with BW < 1750 g with RDS, who had not received any
nutrition by day 3 of life and who were expected to receive parenteral nutrition
for at least five subsequent days. Infants who were expected to receive enteral nutrition within the first week of life, severe jaundice or who were receiving Indomethacin were excluded. 'Early lipid' group: BW 1166 ± 431g, GA 30 ± 3 weeks. 'No early' lipid group: BW 1086 ± 384 g, GA 29 ± 2 weeks. | 'Early lipid' group: N =20. 'No early' lipid group: N =22. 'Early' lipid group infants were begun on a regimen of TPN and 0.5 g/kg/day of Vitrium (lipid) concurrently. lipid infusion was increased at a rate of 0.5 g/kg/day to a maximum of 2.5 g/kg/day. 'No early' lipid group did not receive lipids until day 8. | Days to regain BW, PDA, IVH, NEC, bilirubin, thrombocytopenia, ROP, death, thromboxane B2 levels, AaDo2 on day 1, 3 and 5, BPD, ventilatory support (days), supplemental oxygen (days), home on oxygen (number of infants), plasma 6-keto-PGF1 alpha (day 1, 3 and 5). | It appears as though more patients had eye examinations in 'early lipid' group (11) compared to 'no early' lipid group (17). Thromboxane B2 levels were significantly elevated on day 3 and day 5 of study in the 'early' lipid group. This was hypothesised to be the mediator of increased respiratory morbidity and ROP in the 'early' lipid group. | A |
| Sosenko 1993 | Allocation concealment- Yes (using sealed envelopes). Blinding of intervention- No Blinding of outcome assessment - Yes Completeness of follow up- Yes | N = 133. Entry criteria - inborn infants, 600-1000 g BW, requirement of mechanical ventilation at 6 postnatal hours for the 600-800 g and requirement of mechanical ventilation plus supplemental oxygen at 6 post natal hours for the 801-1000 g infants Exclusion criteria - major congenital anomaly, clinical evidence of congenital infection, previability or terminal condition. Gestational age of participants was not given. | 'Early' lipid group: N = 70. 'No early' lipid group: N = 63. Infants assigned to 'early' lipid group received 20% intralipid starting at less than 12 hours, at a dosage of 0.5 g/kg/day for the first 24 hours and then increasing 0.5 g/kg/day every 24 hours, until a dose of 1.5 g/kg was reached and was maintained through the seventh postnatal day. Lipid infusions were maintained for a 24 hour period. Infants in the 'no early' lipid group received no intralipid for until after the seventh day. The initiation of intravenous aminoacid was begun at 2 or 3 days of age in both groups. | Pulmonary
end points - incidence of CLD, duration and quantity of oxygen exposure and
mechanical ventilation for the first 2 postnatal months, mortality rate,
time of death and duration of hospital stay. Other outcome measures - PIE,
PTX, pulmonary haemorrhage, PDA, NEC, ROP, sepsis. Biochemical endpoints - fatty acids from tracheal fluid and plasma. | Most
of the data were given separately for 600-800 g and 801-1000 g groups and
whenever possible, data were combined. Most of the data were given as percentages.
Data were extrapolated from figures and percentages when possible. Authors
clarified the existing data. | A |
| Study | Reason for exclusion |
| Ibrahim 2004 | Both 'early' and 'no early lipid' groups received lipids within first five days after birth. |
| Wilson 1997 | Both 'early' and 'no early' lipid groups received lipids within first five days after birth. |
Alwaidh MH, Bowden L, Shaw B, Ryan SW. Randomised trial of effect of delayed intravenous lipid administration on chronic lung disease in preterm neonates. Journal of Pediatric Gastroenterology and Nutrition 1996;22:303-6.
Brownlee 1993 {published data only}
Brownlee KG, Kelly EJ, Ng PC, Kendall-Smith SC, Dear PR. Early or late parenteral nutrition for the sick preterm infant? Archives of Disease in Childhood 1993;69:281-3.
Gilbertson 1991 {published data only}
Gilbertson N, Kovar IZ, Cox DJ, Crowe L, Palmer NT. Introduction of intravenous lipid administration on the first day of life in the very low birth weight neonate. Journal of Pediatrics 1991;119:615-23.
Hammerman 1988 {published data only}
Hammerman C, Aramburo MJ. Decreased lipid intake reduces morbidity in sick premature neonates. Journal of Pediatrics 1988;113:1083-8.
Sosenko 1993 {published data only}
Sosenko IR, Rodriguez-Pierce M, Bancalari E. Effect of early initiation of intravenous lipid administration on the incidence and severity of chronic lung disease in premature infants. Journal of Pediatrics 1993;123:975-82.
Ibrahim HM, Jeroudi MA, Baier RJ, Dhanireddy R, Krouskop RW. Aggressive early total parental nutrition in low-birth-weight infants. Journal of Perinatology 2004;24:482-6.
Wilson 1997 {published data only}
Wilson DC, Cairns P, Halliday HL, Reid M, McClure G, Dodge JA. Randomised controlled trial of an aggressive nutritional regimen in very low birth weight infants. Archives of Diseases in childhood 1997;77:F4-11.
* indicates the primary reference for the study
Commmittee on Nutrition. Nutritional needs of preterm infants. In: Ronald E. Kleinman, editor(s). Pediatric Nutrition Handbook. 5 edition. American Academy of Pediatrics, 2003.
Avila-Figueroa C, Goldman DA, Richardson DK, Gray JE, Ferrari A, Freeman J. Intravenous lipid emulsions are the major determinant of coagulase - negative staphylococcal bacteremia in very low birth weight newborns. Pediatric Infectious Disease Journal 1998;17:10-7.
Barson AJ, Chistwick ML, Doig CM. Fat embolism in infancy after intravenous fat infusions. Archives of Diseases in Childhood 1978;53:218-23.
Bell MJ, Ternberg JL, Feigin RD, Keating JP, Marshall R, Barton L, Brotherton T. Neonatal necrotizing enterocolitis. Therapeutic decisions based upon clinical staging. Annals of surgery 1978;187:1-7.
Cooke RJ, Zee P, Yeh YY. Safflower oil emulsion administration during parenteral nutrition in preterm infant.1.Effect on essential fatty acid status. Journal of Pediatric Gastoenterology and Nutrition 1985;4:799-803.
Cooke RWI. Factors associated with chronic lung disease in preterm infants. Archives of Diseases in Childhood 1991;66:776-779.
Flatt JP, Ravussin E, Acheson KJ, Jequier E. Effects of dietary fat on postprandial substrate oxidation and on carbohydrate and fat balances. Journal of Clinical Investigation 1985;76:1019-24.
Foote KD, MacKinnon MJ, Innis S. Effect of early introduction of formula versus fat free parenteral nutrition on essential fatty acid status of preterm infants. Americal Journal of Clinical Nutrition 1991;54:93-97.
Fox GF, Wilson DC, Ohlsson A. Effect of early V/S late introduction of intravenous lipid to preterm infants on death and chronic lung disease (CLD) - results of meta-analysis. Pediatric Research 1998;21:abstract No.250.
Friedman Z., Danon A, Stahilman MT, Oates JA. Rapid onset of essential fatty acid deficiency in the newborn. Pediatrics 1976;58:640-9.
Greene HL, Hazlett D, Demaree R. Relationship between intralipid induced hyperlipemia and pulmonary function. American Journal Of Clinical Nutrition 1976;29:127-35.
Gutcher GR, Farrell PM. Intravenous infusion of lipids for the prevention of essential fatty acid deficiency in the preterm infant. American Journal Of Clinical Nutrition 1991;54:1024-28.
Helbock HJ, Motchnik PA, Ames BN. Toxic hydroperoxides in intravenous lipid emulsions used in preterm infants. Pediatrics 1993;91:83-7.
Herson VC, Block C, Eisenfeld L, Maderazo EG, Krause PJ. Effects of intravenous fat infusion on neonatal neutrophil and platelet function. Journal of Parenteral and Enteral Nutrition 1989;13:620-2.
The committee for the classification of Retinopathy of prematurity. An international classification of retinopathy of prematurity. Archives of Ophthalmology 1984;102:1130-4.
Innis SM. Lipid metabolism in the preterm infant. NeoReviews 2002;3:e50-e51.
Jensen CL, Heird WC. Lipids with an emphasis on long-chain polyunsaturated fatty acids. Clinics in Perinatology 2002;29:261-81.
Koletzko B. Parenteral lipid infusion in infancy: Physiological basis and clinical relevance. Clinical Nutrition 2002;21, supplement 2:53-65.
Lee EJ, Simmer K, Gibson RA. Essential fatty acid deficiency in parenterally fed preterm infants. Journal of Pediatrics and Child Health 1993;29:51-5.
Lekka ME, Liokatis S, Nathanail C, Galani V, Nakos G. The impact of intravenous fat emulsion administration in acute lung injury. American Journal of Respiratory and Critical Care Medicine 2004;169:638-44.
Levene MI, Wigglesworth JS, Desai R. Pulmonary fat accumulation after intralipid infusion in the preterm infant.. Lancet 1980;2:815-8.
Lipson AH, Pritchard J, Thomas G. Thrombocytopenia after intralipid infusion in a neonate. Lancet 1974;2:1462-3.
Neuzil J, Darlow BA, Inder TE, Sluis KB, Winterbourn CC, Stocker R. Oxidation of parenteral lipid emulsion by ambient and phototherapy lights: potential toxicity of routine parenteral feeding. Journal of Pediatrics 1995;126:785-90.
Papile L, Burstein R, Koffler H. Incidence and evolution of subependymal and intraventricular hemorrhage: A study of infants with birth weights less than 1500 gm. Journal of Pediatrics 1978;92:529-34.
Pitkanen O, Hallman M, Andersson S. Generation of free radicals in lipid emulsion used in parenteral nutrition. Pediatric Research 1991;29:56-9.
Pitkanen OM. Evidence for increased oxidative stress in preterm infants eventually developing chronic lung disease. Seminars in Neonatology 1998;2:199-205.
Prasertsom W, Phillipos EZ, Van Aerde JE, Robertson M. Pulmonary vascular resistance during lipid infusion in neonates. Archives of Diseases in Childhood 1996;74:F95-8.
Putet G. Lipid metabolism of the micropremie. Clinics in Perinatology 2000;27:57-69.
Skeie B, Askanazi J, Rothkopf MM, Rosenbaum SH, Kvetan V, Thomashow B. Intravenous fat emulsions and lung function: a review. Critical Care Medicine 1988;16:183-94.
Sosenko IR, Innis SM, Frank L. Polyunsaturated fatty acids and protection of newborn rats from oxygen toxicity. Journal of Pediatrics 1988;112:630-7.
Sosenko IR, Innis SM, Frank L. Intralipid increases lung polyunsaturated fatty acids and protects newborn rats from oxygen toxicity. Pediatric Research 1991;30:413-7.
Spear ML, Stahl GE, Paul MH, Egler JM, Pereira GR, Polin RA. The effect of 15-hour fat infusions of varying dosage on bilirubin binding to albumin. Journal of Parenteral and Enteral Nutrition 1985;9:144-7.
Spear ML, Spear M, Cohen AR, Pereira GR. Effect of fat infusions on platelet concentration in premature infants. Journal of Parenteral and Enteral Nutrition 1990;14:165-8.
Tomsits E, Rischak K, Szollar L. Effects of early nutrition on free radical formation in VLBW infants with respiratory distress. Journal of the American College of Nutrition 2000;19:237-41.
Van Aerde JE , Michaelann S W, Feldman M and Clandinin MT. Accretion of lipid in the fetus and newborn. In: Polin, Fox, Abman, editor(s). Fetal and Neonatal physiology. 3 edition. Vol. 1. W.B Saunders, 2004:400.
Wilson DC, Fox GF, Ohlsson A. Meta-analysis of effects of early or late introduction of intravenous lipids to preterm infants on mortality and chronic lung disease (abstract). Journal of Pediatric Gastroenterology and Nutrition 1998;26:599.
Ziegler EE, Thureen PJ, Carlson SJ. Aggressive nutrition of the very low birth weight infant. Clinics in Perinatology 2002;29:225-244.
| Comparison or outcome | Studies | Participants | Statistical method | Effect size |
|---|---|---|---|---|
| 01 Early lipid versus no early lipids | ||||
| 01 Days to regain birth weight | 2 | 71 | WMD (fixed), 95% CI | 0.59 [-2.41, 3.58] |
| 02 Rate of weight gain during period of hospital stay (g/day) | 1 | 129 | WMD (fixed), 95% CI | -2.40 [-5.30, 0.50] |
| 03 Death before discharge | 0 | 0 | RR (fixed), 95% CI | No numeric data |
| 04 Death (irrespective of time) | 5 | 397 | RR (fixed), 95% CI | 1.04 [0.69, 1.56] |
| 05 Neonatal death | 4 | 268 | RR (fixed), 95% CI | 1.35 [0.78, 2.34] |
| 06 Chronic lung disease | 2 | 193 | RR (fixed), 95% CI | 1.10 [0.81, 1.49] |
| 07 Duration of respiratory support (days) | 2 | 71 | WMD (fixed), 95% CI | 0.89 [-8.61, 10.40] |
| 08 Duration of supplemental oxygen (days) | 2 | 71 | WMD (fixed), 95% CI | 5.50 [-8.22, 19.22] |
| 09 Home oxygen | 1 | 42 | RR (fixed), 95% CI | 16.43 [1.00, 270.41] |
| 10 Pneumothorax | 1 | 133 | RR (fixed), 95% CI | 0.54 [0.21, 1.40] |
| 11 Pulmonary haemorrhage | 1 | 133 | RR (fixed), 95% CI | 2.19 [0.97, 4.92] |
| 12 Pulmonary interstitial emphysema | 1 | 133 | RR (fixed), 95% CI | 0.99 [0.45, 2.17] |
| 13 Necrotizing enterocolitis (>/= stage 2) | 0 | 0 | RR (fixed), 95% CI | No numeric data |
| 14 Necrotizing enterocolitis (any stage) | 3 | 204 | RR (fixed), 95% CI | 0.82 [0.34, 1.98] |
| 15 Retinopathy of prematurity (any stage) | 3 | 204 | RR (fixed), 95% CI | 1.02 [0.74, 1.41] |
| 16 Patent ductus arteriosus (clinically significant) | 3 | 204 | RR (fixed), 95% CI | 0.84 [0.66, 1.06] |
| 17 Sepsis | 0 | 0 | RR (fixed), 95% CI | No numeric data |
| 18 Intraventricular haemorrhage (any grade) | 3 | 204 | RR (fixed), 95% CI | 0.74 [0.53, 1.04] |
| 19 Intraventricular haemorrhage (grade 3 and 4) | 1 | 42 | RR (fixed), 95% CI | 0.37 [0.08, 1.61] |
| 20 Clinically significant thrombocytopenia needing platelet transfusion | 0 | 0 | RR (fixed), 95% CI | No numeric data |
| 21 Significant jaundice | 1 | 29 | RR (fixed), 95% CI | 1.14 [0.47, 2.75] |
| The review is published as a Cochrane review in The
Cochrane Library, Issue 2, 2005 (see http://www.thecochranelibrary.com for
information). Cochrane reviews are regularly undated as new evidence emerges
and in response to comments and criticisms, and The Cochrane Library should
be consulted for the most recent version of the Review. |