Contents - Previous - Next


Types of IUGR

IUGR infants can be subdivided according to whether they are stunted or wasted (or both). Stunted infants have been chronically undernourished in utero, resulting in a proportionate reduction in both skeletal and soft tissue growth. Wasted infants have normal linear growth but reduced tissue mass resulting in a low ponderal reflecting later onset, primarily in the third trimester (Rosso and Winick, 1974; Campbell, 1974; Villar and Belizan, 1982; Villar et al, 1989). Whether this heterogeneity in body composition and aetiology has a differential impact on the risk of mortality and morbidity will be explored.

Low birth weight and mortality

Large population-based studies of neonatal and postneonatal mortality (all causes).

The relative risks of neonatal and postneonatal death among infants of different birth weights have been reviewed previously (Ashworth and Feachem, 1985). Birth weights of 3500-4500 g were associated with the lowest risk. In these large population-based studies, infants were grouped in 500 g intervals whatever their gestational age. Preterm infants, however, are generally regarded as being at greater risk of death in the first year of life, especially neonatal death, than term infants of comparable birth weight (Behrman et al, 1971; Mata, 1978; Victora et al, 1987). Previously calculated relative risks may therefore overestimate the risks for term IUGR infants in developing countries. Therefore for this review, additional data sets have been sought and median values recalculated. This has been done separately for populations with a LBW prevalence of < 7.5% (assumed to reflect preterm births) and for populations where the prevalence of LBW is > 15%. The latter has been suggested as a proxy for IUGR prevalence (World Health Organization, 1995).

Table 1 presents neonatal mortality data from 9 populations in which the majority of LBW infants can be expected to have been preterm. Despite a 30-year span and large reductions in overall neonatal mortality during this period, for any birth-weight interval the relative risks of neonatal death are remarkably consistent across studies. Table 2 gives comparable data for populations where the majority of LBW infants can be expected to be IUGR. Population sizes are much smaller and no national data sets were located. The results are less consistent among the 4 populations, but all show an increased risk of neonatal mortality with decreasing birth weight.

Comparable birth-weight-specific mortality data for the postneonatal period are given in Tables 3 and 4. For populations where most LBW infants would be expected to be preterm, the relative risks by birth weight are again very consistent (Table 3). This is less so for populations where LBW infants are presumed IUGR (Table 4).

An alternative approach to ascertain birth-weight-specific mortalities for IUGR infants is to identify studies where gestational ages are specified. Five such studies were located and 8 data sets for term infants extracted (Tables 5 and 6). All were from the USA, and all except two data sets were restricted to neonatal mortality. By comparing Tables 2 and 5, the results of the two approaches regarding neonatal risk can be compared. Focusing on birth weights 2000-2499 g, the relative risk of neonatal death compared to birth weights 2500-2999 g, ranges from 2.7-5.6 for term infants of known gestation (Table 5) and from 1.4-4.5 for infants presumed to be term (Table 2). For infants presumed to be preterm (Table 1) the risk ranges from 2.8-4.5.

Similarly for postneonatal mortality in infants 2000-2499 g, the relative risk ranges from 2.0-2.2 for term infants of known gestation (Table 6) and from 0.8-2.3 for infants presumed to be term (Table 4). For infants presumed to be preterm (Table 3), the relative risk ranges from 1.8-2.3.

Table 7 compares the median relative risks for neonatal and postneonatal mortality by birth weight derived from the approaches described above. These population-based comparisons indicate that

i) term LBW infants have increased risks of neonatal and postneonatal death compared with infants of adequate birth weight (ABW),

ii) term LBW infants have risks of neonatal and postneonatal death that are of a broadly similar magnitude to those of preterm infants of the same birth weight,

iii) in the 2000-2499 g birth-weight category which accommodates most LBW infants in developing countries, there is a 4-fold increased risk of neonatal death compared with infants 2500-2999 g, and a 2-fold greater risk of postneonatal death (composite median values for the 10 neonatal 'term' studies and the 5 postneonatal 'term' studies).

Table 1. Relative risks of neonatal mortality by birthweight compared to birthweight of 2500-2999g. Selected countries with LBW prevalence < 7.5%

Birth weight (g)

UK (1993)

Belgium (1986-87)

USA California (1977)

USA (1986-87)

USA (1980)

USA (5 regions) (1974-75)

Norway (1967-68)

USA Oregon (1959-66)

USA (1960)

1000-1499

99.6a

49.6

52.7

34.0

46.7

99.6a

41.0

47.5

55.0

1500-1999

10.4

15.9

13.5

12.9

13.5

17.5

16.6

20.6

19.6

2000-2499

2.8

3.3

3.0

3.7

4.0

3.8

4.5

4.4

4.4

2500-2999

1.0

1.0

1.0

1.0

1.0

1.0

1.0

1.0

1.0

3000-3499

0.5

0.4

0.3

0.4

0.5

0.5

0.4

0.4

0.4

3500-3999

0.4

0.4

0.3

0.3

0.3

0.3

0.2

0.3

0.3

4000-4499

0.4

0.4

0.3

0.3

0.4

0.4

0.2

0.2

0.4

4500+

0.4

0.4

0.4

0.9

0.9

0.4

0.4

0.3

0.8

Neonatal mortality rateb

4.1

4.8

5.1

5.6

7.3

8.2

8.5

13.8

16.9

% LBW

6.5c

4.9

7.0d

5.9

6.0

6.1

5.2

5.2

6.8

No. of births

673,000
all

229,000
single

290,000
single

7.4
million single

3.5
million single/all races

234,000
single

700,000
all

40,000
white/single

3.6 million
white

Reference

Platt and Pharoah, 1995

Buekens et al, 1995

Williams and Chen, 1982

Buekens et al, 1995

Buehler et al, 1987

Starfield et al, 1982

Erickson and Bjerkedal, 1982

Behrman et al, 1971

Chase, 1969

a includes all births <1500g
b neonatal mortality rate: Deaths per 1000 live births except for Williams and Chen (1982) (per 1000 vaginal births) and Erickson and Bjerkedal (1982) (per 1000 total births)
c1992 figure
d estimated from Erickson and Bjerkedal (1982).

Table 2. Relative risks of neonatal mortality by birthweight compared to birthweight of 2500-2999 g. Selected countries with LBW prevalence >15.0%

Birth weight (g)

USA (Denver) (1974-80)

India (Delhi) (1969-72)

India (N. Arcot) (1969-75)

Guatemala (Santa Maria Cauque) (1964-72)

1000-1499

32.8

94.6

16.7a

-

1500-1999

9.4

30.9

4.4

27.3

2000-2499

2.1

4.5

1.4

3.4

2500-2999

1.0

1.0

1.0

1.0

3000-3499

0.3

1.1

0.8

0

3500-3999

0.3

1.1

0.6

-

4000-4499

0

1.1

1.4

-

4500+

0

1.1

1.4

-

Neonatal mortality rateb

17.5

21.2

34.8

39.0

% LBW

15.2

23.2

31.9

41.7

No. of births

14,400 all

4,590 single

4,220 single

416 single

Reference

Koops et al, 1982

Ghosh et al, 1979

Rao and Inbaraj, 1978

Mata, 1978

a includes all births <1500g
b neonatal mortality rate: Deaths per 1000 live births

Table 3. Relative risks of postneonatal mortality by birthweight compared to birthweight of 2500-2999g. Selected countries with LBW prevalence < 7.5%

Birth weight (g)

UK (1993)

Norway (1967-1968)

USA (1980)

USA (5 regions) (1974-75)

USA (1960)

1000-1499

14.7a

2.6

9.3

12.2a

6.3

1500-1999

3.8

3.0

4.2

3.7

3.5

2000-2499

2.3

1.8

2.1

1.9

1.9

2500-2999

1.0

1.0

1.0

1.0

1.0

3000-3499

0.6

0.6

0.6

0.6

0.6

3500-3999

0.4

0.5

0.4

0.5

0.5

4000-4499

0.4

0.4

0.4

0.5

0.5

4500+

0.4

0.5

0.4

0.5

0.5

Postneonatal mortality rateb

2.1

3.2

3.7

4.3

5.4

% LBW

6.5c

5.2

6.0

6.1

6.8

No. of births

673,000
all

700,000
all

3.5 million
single/all races

234,000
single

3.6 million
white

Reference

Platt and Pharoah, 1995

Erickson and Bjerkedal, 1982

Buehler et al, 1987

Starfield et al, 1982

Chase, 1969

a includes all births <1500g
b postneonatal mortality rate: Deaths per 1000 neonatal survivors, except for Erickson and Bjerkedal (1982) (per 1000 total births) and Platt and Pharoah (1995) and Starfield et al (1982) (per 1000 live births)
c1992 figure

Table 4. Relative risks of postneonatal mortality by birthweight compared to birthweight of 2500-2999 g. Selected countries with LBW prevalence >15.0%

Birth weight (g)

India (Delhi) (1969-72)

India (N. Arcot) (1969-75)

Guatemala (Santa Maria Cauque) (1964-72)

1000-1499

9.3

4.5a

-

1500-1999

5.3

2.3

7.0

2000-2499

2.3

1.1

0.8

2500-2999

1.0

1.0

1.0

3000-3499

0.7

0.6

0.5

3500-3999

07

0.5

-

4000-4499

0.7

0

-

4500+

0.7

0


Postneonatal mortality rateb

25.3

51.6

60.0

% LBW

23.2

31.9

41.7

No. of births

4,590 single

4,220 single

416 single

Reference

Ghosh et al, 1979

Rao and Inbaraj, 1978

Mata, 1978

a includes all births <1500g
b postneonatal mortality rate: Deaths per 1000 neonatal survivors.

Table 5. Risks of neonatal mortality by birthweight for term infants

Birth weight (g)

USA (1980)

USA (California) (1980-81)

USA (1980)

USAa (Oregon) (1959-66)

USA (Denver) (1958-68)

USAa (California) (1970-76)

1000-1499

-

17.5

-

-

-

59.2

1500-1999

-

13.1

-

17.9b

18.1

18.7

2000-2499

5.6

4.7

3.0

4.4

2.7

4.4

2500-2999

1.0

1.0

1.0

1.0

1.0

1.0

3000-3499

0.4

0.4

0.6

0.4

0.5

0.4

3500-3999

0.3

0.3

0.5

0.5

0.6

0.3

4000-4499

0.3

0.2

0.5

0.1

1.9

0.5

4500+

0.5

0.3

1.4

0

1.9

-

Neonatal mortalityc

1.7

1.7

3.0

3.5

4.8

10.5

Gestation (weeks)

40-41

38-41

40-41

40

38-42

38-40

No. of births

861,000 white single

213,500 white male

122,000 black single

15,500 white

9,500 all

1 million single

Reference

Sappenfield et al, 1987

Binkin et al, 1985

Sappenfield et al, 1987

Behrman et al, 1971

Lubchenco et al, 1972

Williams et al, 1982

a estimated from 250g birthweight intervals
b1750-1999g
c neonatal mortality rate: Deaths per 1000 live births for the gestations specified.

Table 6. Relative risks of postneonatal mortality by birthweight for term infants

Birthweight (g)

USA (1980)

USA (1980)

1000-1499

-

-

1500-1999

-

-

2000-2499

2.2

2.0

2500-2999

1.0

1.0

3000-3499

0.5

0.6

3500-3999

0.4

0.5

4000-4499

0.3

0.3

4500+

0.3

0.8

Postneonatal mortalitya

2.4

5.2

Gestation (weeks)

40-41

40-41

No. births

861,000 white single

122,000 black single

Reference

Sappenfield et al, 1987

Sappenfield et al, 1987

a Postneonatal mortality rate: deaths per 1000 neonatal survivors

Table 7. Median relative risks of neonatal and postneonatal mortality by birthweight

Postneonatal

preterm & terma

mostly pretermb

mostly termc

only termd

Birthweight (g)

1000-1499

50.0

49.6

32.8

38.3

1500-1999

18.0

15.9

18.3

18.0

2000-2499

4.0

3.8

2.7

4.4

2500-2999

1.0

1.0

1.0

1.0

3000-3499

0.5

0.4

0.5

0.4

3500-3999

0.4

0.3

0.4

0.4

>4000

0.4

0.4

0.7

0.4

a median values (n= 13 studies) from Ashworth and Feachem (1985)

b median values from Table 1 (n = 9 studies)

c median values from Table 2 (n = 4 studies)

d median values from Table 5 (n = 6 studies)

Postneonatal

preterm & terme

mostly pretermf

mostly termg

only termh

Birthweight (g)

1000-1499

6.1

9.3

6.9

-

1500-1999

3.4

3.7

5.3

-

2000-2499

1.8

1.9

1.1

2.1

2500-2999

1.0

1.0

1.0

1.0

3000-3499

0.6

0.6

0.6

0.5

3500-3999

0.5

0.5

0.6

0.4

>4000

0.5

0.4

-

0.3

e median values (n = 8 studies) taken from Ashworth and Feachem (1985)

f median values from Table 3 (n = 5 studies)

g median values from Table 4 (n = 3 studies)

h median values from Table 6 (n = 2 studies)

Community-based studies of mortality (all causes and cause-specific)

Smaller, community-based studies may offer certain advantages compared with the large population-based studies described above, which rely on routinely collected data. These include better estimates of gestational age and opportunities to control for potential confounding variables. If families with LBW (or IUGR) infants are more disadvantaged in terms of income, housing, parental education etc. than families of ABW infants, then they may be more exposed to infection postnatally and have inferior medical care when ill. Consequently higher mortality in LBW or IUGR infants could be due, at least in part, to their environment, rather than birth weight per se. On the other hand, smaller studies have insufficient size to allow birthweight-specific mortalities to be calculated with confidence, which limits their usefulness in estimating the potential impact on mortality of interventions to reduce IUGR.

Table 8 summarises the relative risks of mortality in LBW infants from all causes (7 studies), and from respiratory and diarrhoeal infections (4 and 3 studies respectively). Risks are calculated for birth weight as a dichotomous variable. For most, the comparison is with infants 2500 g, but in one study the modal birth weight (3000-3499 g) has been used. Only 4 of the 12 studies listed were specifically of term infants. Only in 2 studies were confounding variables controlled. Where cause-specific relative risks are available within the same study, the levels of risk for diarrhoea and acute respiratory infections (ARI), appear lower than for all causes combined.

Table 8 also gives data for early childhood mortality (4 studies). The findings suggest that low birth weight continues to have an adverse effect on mortality beyond infancy.

Table 8. Summary of community-based studies of birthweight mortality

Country (ref)

Design

Gestation

Age (months)

Sample size (deaths)

Birthweight (g)

Risk ratio (95% confidence interval)

Outcome

Brazil

cohort

term

0-6

393

3000-3499

1.0


Lira et al, 1996




(12)

1500-2499

10.2 (2.2-46.7)

all causes







6.6a (1.4-31.2)


India

cohort

term

0-11

4590

2500

1.0


Ghosh et al, 1979




(213)

2000-2499

2.6

all causes

India

cohort

term

0-11

4220

2500

1.0


Rao & Inbaraj, 1978




(362)

< 2500

1.7

all causes

Guatemala

cohort

term

0-11

385

2500

1.0


Mata, 1978




(24)

< 2500

1.7

all causes

Mata, 1975



12-47

(39)


1.8


Indonesia

cohort

term + preterm

0-11

687

2500

1.0


Kusin et al, 1989


(9% LBW)


(83)

< 2500

3.4

all causes

Nigeria

cohort

term + preterm

0-11

4334

> 2500

1.0


Ayeni & Oduntan, 1978


(11 % LBW)


(133)

£ 2500

5.8

all causes

Brazil

cohort

term + preterm

0-11

5914

2500

1.0


Barros et al, 1987


(9% LBW)


(215)

< 2500

11.0 (8.7-14.4)

all causes







6.7 (3.0-14.9)

ARI







2.5 (0.9-6.7)

diarrhoea

Victora et al, 1992



12-59

(29)


3.3

all causes

Brazil

case-control

term + preterm

0.25-11

1070

2500

1.0


Victora et al, 1988


(9% LBW)


(357)

1500-2499

1.9a (1.1-3.6)b

ARI







2.0a (1.1-3.6)b

diarrhoea

India

cohort

term + preterm

0-11

659

2500

1.0


Datta et al, 1987


(32% LBW)


(19)

< 2500

8.0

ARI

UK

cohort

term + preterm

1-50

5522

2500

1.0

bronchitis

Douglas & Mogford, 1953




(40)

< 2500

3.6

+pneumonia

USA

cohort

term + preterm


51,931

2500

1.0


Read et al, 1994


(9% LBW)

1-11

(371)

1500-2499

2.4 (1.4-4.0)

infectious




12-84

(258)


2.5 (1.3-4.5)

disease

USA

cohort

term + preterm

1-11

193,733

2500

1.0


Gibson & Alexander, 1985




(93)

< 2500

7.1

diarrhoea

a adjusted for confounders
b 90% confidence intervals.

 


Contents - Previous - Next