Abstruse

Background: Previous analyses derived the relative risk (RR) of dying as a result of low weight-for-age and calculated the proportion of child deaths worldwide owing to underweight.

Objectives: The objectives were to examine whether the take chances of dying because of underweight varies by cause of death and to gauge the fraction of deaths past crusade attributable to underweight.

Design: Data were obtained from investigators of 10 accomplice studies with both weight-for-age category (<−3 SDs, −three to <−2 SDs, −2 to <−i SD, and >−one SD) and cause of decease information. All x studies contributed data on weight-for-age and gamble of diarrhea, pneumonia, and all-cause mortality; however, simply vi studies contributed information on deaths considering of measles, and only 3 studies contributed information on deaths because of malaria or fever. With use of weighted random furnishings models, we related the log bloodshed rate past cause and anthropometric status in each study to derive cause-specific RRs of dying because of undernutrition. Prevalences of each weight-for-age category were obtained from analyses of 310 national nutrition surveys. With use of the RR and prevalence information, we and so calculated the fraction of deaths past cause attributable to undernutrition.

Results: The RR of bloodshed considering of low weight-for-historic period was elevated for each cause of expiry and for all-crusade mortality. Overall, 52.5% of all deaths in young children were attributable to undernutrition, varying from 44.8% for deaths considering of measles to 60.7% for deaths considering of diarrhea.

Decision: A meaning proportion of deaths in immature children worldwide is attributable to low weight-for-age, and efforts to reduce malnutrition should be a policy priority.

INTRODUCTION

Previously information technology was shown that a child′s take chances of dying equally a event of undernutrition, defined as underweight or low weight-for-age, is not express to only those children with the most severe course of malnutrition (one, 2). Rather, there exists a spectrum of risk associated with all degrees of malnutrition. Although the take a chance of dying is highest among the severely malnourished, when i considers the elevated risk of bloodshed associated with moderate malnutrition in combination with their high prevalence worldwide, it becomes articulate that much of the burden of deaths as a result of undernutrition in immature children is owing to moderate, rather than to severe, undernutrition. Pelletier et al (3) estimated that 55% of kid deaths worldwide are owing to undernutrition.

Despite problems in defining cause of death, particularly in developing countries in which one would often need to rely on exact dissection methods, it is of interest to consider whether the relation betwixt underweight and adventure of dying varies by cause of death. For example, are the risks of dying of malaria every bit a result of undernutrition similar or different from the risks of dying of diarrhea? The purpose of this paper is to describe our assay of real information that chronicle weight-for-age and mortality from diarrhea, pneumonia, measles, and malaria—the principal causes of death of children in developing countries. The goals of the analysis were 1) to derive relative risk (RR) of dying as a issue of underweight past cause of death and ii) to combine these estimates with prevalence data to calculate the fraction of deaths owing to undernutrition during babyhood.

SUBJECTS AND METHODS

Ii fundamental inputs are necessary to summate the burden of disease attributable to undernutrition: one) an estimate of the prevalence of low weight-for-age in young children and 2) the estimated RR of dying associated with child weight-for-age. Prevalences of malnutrition defined past depression weight-for-historic period z score in each of 14 World Health Organization (WHO) mortality regions were estimated by analyzing in a standard manner 310 national diet surveys compiled in the WHO Global Database on Kid Growth and Malnutrition (4). For each region, the average weight-for-age z score was calculated, and from this calculation the proportion of children with z scores −1.01 to −2.00 SDs, −two.01 to −3.00 SDs, and <−3.00 SDs were estimated. In a healthy population, we would await only xiii.6%, 2.1%, and 0.1% of children to be classified into each of these categories, respectively. Weight-for-age was chosen because information technology is the most widely used indicator of child nutritional status in developing countries, besides as in the studies bachelor for analysis.

We used a variety of approaches to provide data for the second key input. Originally, a survey of the published literature was conducted to gather data to approximate the relation betwixt anthropometric status and cause-specific mortality (5). Because bachelor data were insufficient to achieve our goal, nosotros contacted investigators with relevant data (published or unpublished) and asked them to contribute specific report results for our analysis. Briefly, on the basis of published reports of studies examining causes of death among children in developing countries, nosotros constructed a list of investigators to exist contacted to collaborate in this endeavor. We so augmented this list with knowledge about the beingness of other unpublished studies that might provide appropriate data for this purpose and contacted investigators to enquire whether they knew of other studies that might provide appropriate data.

Each investigator was asked to provide the post-obit information: one) a description of their study; 2) the number of children or child-years of study with weight-for-age z scores at the showtime of the follow-up period of <−three.00 SDs, −2.01 to −3.00 SDs, −1.01 to −two.00 SDs, 0.0 to −1.0 SD and >0.0 SDs; and 3) deaths in each category attributable to diarrhea, pneumonia, measles, malaria, other, and total (all cause) during the follow-up menses. During analyses, nosotros collapsed the last 2 categories of anthropometric status and then that the reference group would contain those children with weight-for-historic period z scores >−ane.0 SD. This plummet was done because the number of deaths by crusade for children >−1 SD was limiting in some studies and considering the data did not bespeak substantial differences in rates of expiry past cause between those rates with z scores −1 to 0 SDs and >0 SDs.

In all, 12 studies were identified, and data from ten studies were included in the current analysis (6–16). We excluded 2 data sets: one from Republic of peru because the study provided insufficient deaths for the assay, and one from a case-control study in Brazil because it did not fit the overall analytic approach taken with the other studies, which were prospective cohort designs.

The x studies were conducted in sub-Saharan Africa and Asia ( Tabular array ane). Each study contributed information about deaths as a outcome of diarrhea and pneumonia; however, exposure to infectious diseases such as malaria and measles depends on the ecology of the study setting and health care utilization (ie, measles vaccination rates). Therefore, half-dozen studies (Republic of guinea-bissau, Senegal, Ghana, Nepal, Indonesia, and The Philippines) contributed data on deaths as a result of measles, and iii studies (Republic of guinea-bissau, Senegal, and Ghana) contributed information on malaria-related deaths. In the report in Guinea-bissau, deaths as a issue of malaria were reported as a upshot of fever; thus, deaths equally a result of delirious illnesses other than malaria were necessarily included.

TABLE 1

Prospective accomplice studies with information on weight-for-age z score and survival or death by cause

Weight-for-age z score at the start of the follow-up catamenia
Country Citation Age range Follow-up flow <−iii >−3 to <−2 >−2 to <−i >−one
mo mo
Sudan Fawzi et al (six) 6–72 6 365 985 1123 592
Senegal Garenne et al (7) 0–59 6 1663 2186 2035 1249
Republic of guinea-bissau Andersen (viii) 0–59 6–12 1155 4366 7341 2731
Ghana WHO/CHD (9) 0–12 12 46 183 506 1410
Nepal West et al (10, xi) 0–72 12 1030 2082 2002 892
Bangladesh Arifeen et al (12) 0–11 three 79 250 475 478
Pakistan Khan et al (13), Jalil et al (14) 0–60 120 253 687 1241 1341
Republic of india WHO/CHD (ix) 0–12 12 172 639 1259 1295
Indonesia Sutrisna et al (15) 0–lx 18 207 240 797 4856
Philippines Ricci and Becker (16) 0–59 three 800 3144 4834 5115
Weight-for-age z score at the start of the follow-up period
Land Commendation Historic period range Follow-up period <−3 >−3 to <−two >−ii to <−i >−i
mo mo
Sudan Fawzi et al (6) vi–72 half-dozen 365 985 1123 592
Senegal Garenne et al (vii) 0–59 6 1663 2186 2035 1249
Guinea-Bissau Andersen (8) 0–59 vi–12 1155 4366 7341 2731
Ghana WHO/CHD (ix) 0–12 12 46 183 506 1410
Nepal West et al (10, eleven) 0–72 12 1030 2082 2002 892
Bangladesh Arifeen et al (12) 0–11 3 79 250 475 478
Pakistan Khan et al (13), Jalil et al (14) 0–lx 120 253 687 1241 1341
India WHO/CHD (nine) 0–12 12 172 639 1259 1295
Republic of indonesia Sutrisna et al (xv) 0–60 18 207 240 797 4856
Philippines Ricci and Becker (16) 0–59 3 800 3144 4834 5115

TABLE one

Prospective cohort studies with data on weight-for-historic period z score and survival or expiry by crusade

Weight-for-age z score at the get-go of the follow-up period
Country Citation Age range Follow-upwardly period <−3 >−3 to <−2 >−2 to <−1 >−1
mo mo
Sudan Fawzi et al (half dozen) 6–72 6 365 985 1123 592
Senegal Garenne et al (7) 0–59 half-dozen 1663 2186 2035 1249
Republic of guinea-bissau Andersen (8) 0–59 six–12 1155 4366 7341 2731
Ghana WHO/CHD (ix) 0–12 12 46 183 506 1410
Nepal West et al (10, 11) 0–72 12 1030 2082 2002 892
Bangladesh Arifeen et al (12) 0–11 3 79 250 475 478
Islamic republic of pakistan Khan et al (13), Jalil et al (xiv) 0–threescore 120 253 687 1241 1341
Republic of india WHO/CHD (9) 0–12 12 172 639 1259 1295
Republic of indonesia Sutrisna et al (15) 0–lx 18 207 240 797 4856
Philippines Ricci and Becker (16) 0–59 3 800 3144 4834 5115
Weight-for-age z score at the outset of the follow-up flow
Country Citation Historic period range Follow-upward period <−three >−3 to <−2 >−two to <−1 >−1
mo mo
Sudan Fawzi et al (6) 6–72 6 365 985 1123 592
Senegal Garenne et al (seven) 0–59 6 1663 2186 2035 1249
Republic of guinea-bissau Andersen (viii) 0–59 6–12 1155 4366 7341 2731
Ghana WHO/CHD (ix) 0–12 12 46 183 506 1410
Nepal Due west et al (10, eleven) 0–72 12 1030 2082 2002 892
Bangladesh Arifeen et al (12) 0–11 iii 79 250 475 478
Islamic republic of pakistan Khan et al (13), Jalil et al (14) 0–60 120 253 687 1241 1341
Bharat WHO/CHD (9) 0–12 12 172 639 1259 1295
Indonesia Sutrisna et al (fifteen) 0–60 eighteen 207 240 797 4856
Philippines Ricci and Becker (16) 0–59 three 800 3144 4834 5115

To guess the relations between underweight and mortality risk, we followed the analytic procedures used by Pelletier (17). The steps are described briefly hither. First, we calculated so graphed the mortality rates by anthropometric status and crusade of death for each written report both in elementary terms (per k) and in proportional terms (as the logarithm of the bloodshed risk). 2d, we compared the results and goodness-of-fit statistics of regression analyses of the natural logarithm of mortality by weight-for-age z score category with the results of the unproblematic bloodshed rates past weight-for-age z score category and verified that the models with log mortality rate provided a better fit of the data. For these analyses, we used weighted regression and the weighting scheme of Pelletier (17): [(i/deaths) + (one/children)]. 3rd, the coefficients from the weighted random furnishings models for log mortality charge per unit were used to provide global estimates of the RR and 95% CI of mortality for each weight-for-age category. The midpoint of the anthropometric category was used for estimation, and −0.5 SD was considered the value of the reference weight-for-age category of >−1.0 SD. Fourth, we calculated the population attributable fraction (PAF) of deaths past cause for undernutrition (weight-for-age <−1 SD) for each of the xiv WHO regions with use of a standard formula (eighteen) in which the minimal chance exposure or counterfactual exposure distribution was the distribution of weight-for-age in the loftier-income countries of North America and Western Europe. To obtain a global estimate of the PAF, we used the method described by Vander Hoorn et al (19). Nosotros used information on the number of deaths by crusade and all-cause mortality for each region (20), calculated the number of deaths attributable to undernutrition, summed across the regions, and divided by the global number of deaths every bit a result of each cause. For all-cause mortality, we used our estimated risks of overall mortality associated with underweight and regional estimates of total deaths in children aged birth to 4 y. SAS (version 8; SAS Constitute Inc, Cary, NC) was used for all analyses.

RESULTS

For each study, the accented bloodshed rates past weight-for-historic period category and cause of decease were calculated and compared graphically with those relating weight-for-age category and log mortality charge per unit. As shown for mortality as a effect of pneumonia in Figure 1, the apparent heterogeneity across study settings in the relations betwixt weight-for-age category and mortality hazard in accented terms (Figure 1A) largely disappears when the relations are regraphed with mortality adventure considered in proportional terms (Figure 1B).

Figure one.

Underweight and mortality as a result of pneumonia. A: Deaths per 1000; B: logarithm of deaths per 1000.

Underweight and mortality every bit a result of pneumonia. A: Deaths per 1000; B: logarithm of deaths per 1000.

FIGURE one.

Underweight and mortality as a result of pneumonia. A: Deaths per 1000; B: logarithm of deaths per 1000.

Underweight and mortality equally a issue of pneumonia. A: Deaths per 1000; B: logarithm of deaths per 1000.

The relations betwixt weight-for-age category and log mortality for each crusade of death were estimated with utilize of weighted random effects regression. From these models, RR (95% CI) of dying by crusade of decease were calculated with use of weight-for-historic period >−one.0 SD as the reference category ( Table 2). At that place are pregnant increased risks of dying associated with depression weight-for-age for overall bloodshed as well equally for each crusade of decease examined. Shown in Table 3 is the estimated mean weight-for-historic period z scores and prevalence of children in each weight-for-age category in each of the 14 WHO regions. For The Americas A and Europe B regions, the average z score is 0, and the prevalences in each category are those expected in a healthy population. In dissimilarity, the boilerplate z scores in sub-Saharan Africa (Afr D and Afr E) and Southeast Asia (Sear B and Sear D) are betwixt −1.35 and −1.ninety, and prevalences of children with moderate (z scores: −two.01 to −3.00 SDs) or astringent (z scores: <−three.00 SDs) malnutrition are quite high.

TABLE ii

Relative risk (95% CI) of mortality overall and past cause associated with low weight-for-age estimated from weighted random effects regression analysis

Cause of death <3 SDs 1 −2 to −3 SDs −ane to −2 SDs >−1 SDs
Diarrhea 12.50 (7.19, 21.73) 5.39 (3.73, 7.79) ii.32 (1.93, two.79) 1.0
Pneumonia 8.09 (iv.36, 15.01) 4.03 (2.67, 6.08) 2.01 (one.63, 2.47) 1.0
Malaria 9.49 (3.25, 27.66) 4.48 (2.20, 9.15) 2.12 (ane.48, 3.02) 1.0
Measles v.22 (2.29, 11.88) three.01 (ane.74, 5.21) 1.73 (one.32, 2.28) 1.0
All causes 8.72 (5.55, 13.72) 4.24 (iii.13, 5.53) 2.06 (1.77, two.39) 1.0
Cause of death <three SDs 1 −2 to −iii SDs −1 to −2 SDs >−1 SDs
Diarrhea 12.50 (7.19, 21.73) five.39 (3.73, seven.79) 2.32 (1.93, 2.79) 1.0
Pneumonia eight.09 (4.36, 15.01) four.03 (2.67, 6.08) two.01 (i.63, 2.47) one.0
Malaria ix.49 (3.25, 27.66) four.48 (2.20, 9.15) 2.12 (one.48, three.02) ane.0
Measles five.22 (ii.29, 11.88) 3.01 (one.74, v.21) 1.73 (1.32, 2.28) 1.0
All causes 8.72 (5.55, 13.72) 4.24 (three.thirteen, 5.53) 2.06 (1.77, ii.39) 1.0

1

Calculated at −three.5, −ii.five, and −1.5 compared with 0.five SD weight-for-age from weighted random effects models. A significant examination for trend is evidenced past a statistically pregnant (P < 0.05) coefficient for weight-for-age in each model.

TABLE two

Relative risk (95% CI) of mortality overall and past cause associated with low weight-for-historic period estimated from weighted random effects regression assay

Cause of death <three SDs ane −ii to −3 SDs −1 to −2 SDs >−one SDs
Diarrhea 12.l (vii.19, 21.73) five.39 (3.73, 7.79) 2.32 (ane.93, 2.79) ane.0
Pneumonia 8.09 (4.36, 15.01) 4.03 (two.67, half dozen.08) two.01 (one.63, 2.47) one.0
Malaria 9.49 (iii.25, 27.66) 4.48 (2.xx, 9.xv) 2.12 (one.48, 3.02) 1.0
Measles v.22 (2.29, 11.88) iii.01 (1.74, v.21) i.73 (1.32, 2.28) 1.0
All causes eight.72 (v.55, 13.72) 4.24 (iii.13, 5.53) 2.06 (one.77, two.39) 1.0
Cause of death <iii SDs 1 −2 to −3 SDs −ane to −two SDs >−1 SDs
Diarrhea 12.fifty (7.19, 21.73) 5.39 (3.73, seven.79) 2.32 (ane.93, 2.79) 1.0
Pneumonia 8.09 (4.36, 15.01) 4.03 (2.67, 6.08) ii.01 (one.63, 2.47) 1.0
Malaria 9.49 (three.25, 27.66) 4.48 (2.20, 9.15) ii.12 (1.48, iii.02) 1.0
Measles v.22 (two.29, 11.88) 3.01 (ane.74, v.21) ane.73 (one.32, 2.28) 1.0
All causes 8.72 (5.55, xiii.72) 4.24 (3.13, 5.53) 2.06 (i.77, 2.39) i.0

i

Calculated at −3.5, −two.5, and −1.v compared with 0.v SD weight-for-age from weighted random effects models. A significant exam for tendency is evidenced by a statistically significant (P < 0.05) coefficient for weight-for-age in each model.

Table three

Mean z score and percentage of children in each weight-for-age category in each Earth Health Organization (WHO) region i

Prevalence
WHO region 2 Hateful z score <−3 SDs >−3 to <−two SDs >−two to <−ane SDs >−1 SDs
%
Africa D −1.54 7.1 25.1 38.3 29.5
Africa E −one.5 vi.8 24.2 38.3 30.7
The Americas A 0 0.ane ii.1 13.6 84.2
The Americas B −0.35 0.5 four.5 20.8 74.2
The Americas D −0.84 1.6 10.8 31.3 56.three
Eastern Mediterranean B −0.half dozen 0.eight 7.3 26.3 65.vi
Eastern Mediterranean D −1.33 4.7 20.iv 37.8 37.1
Europe A 0 0.one 2.1 xiii.6 84.2
Europe B −0.57 0.7 6.9 25.7 66.vii
Europe C −0.05 0.ii 2.iv 14.5 82.9
Southeast Asia B −one.35 five.0 20.8 37.9 36.three
Southeast Asia D −i.9 xiii.4 32.v 35.8 xviii.iii
Western Pacific Region A −0.22 0.3 iii.5 xviii.0 78.two
Western Pacific Region B −1.0 2.3 13.half-dozen 34.1 l.0
Prevalence
WHO region two Hateful z score <−3 SDs >−3 to <−2 SDs >−2 to <−1 SDs >−1 SDs
%
Africa D −ane.54 7.ane 25.1 38.3 29.5
Africa E −1.5 6.8 24.two 38.3 thirty.7
The Americas A 0 0.one 2.1 13.half dozen 84.ii
The Americas B −0.35 0.five iv.v 20.8 74.2
The Americas D −0.84 i.6 ten.8 31.3 56.3
Eastern Mediterranean B −0.half-dozen 0.viii vii.iii 26.iii 65.half dozen
Eastern Mediterranean D −1.33 4.7 xx.4 37.8 37.1
Europe A 0 0.one two.one 13.half dozen 84.ii
Europe B −0.57 0.7 half dozen.9 25.7 66.7
Europe C −0.05 0.2 2.four 14.5 82.ix
Southeast Asia B −i.35 five.0 20.8 37.9 36.3
Southeast Asia D −i.9 xiii.4 32.five 35.8 18.three
Western Pacific Region A −0.22 0.3 three.5 xviii.0 78.2
Western Pacific Region B −i.0 2.3 13.half dozen 34.1 50.0

ane

Expected percentages of children in each weight-for-age category in a healthy population are those reported here for The Americas A and Europe A.

ii

Regions for the Global Burden of Illness project. A indicates very low child mortality and very low developed mortality; B, depression child mortality and low developed mortality; C, low child bloodshed and high adult bloodshed; D, high kid mortality and loftier adult mortality; E, high child mortality and very high adult mortality.

Table three

Mean z score and percent of children in each weight-for-age category in each World Wellness Arrangement (WHO) region 1

Prevalence
WHO region 2 Hateful z score <−3 SDs >−iii to <−2 SDs >−2 to <−1 SDs >−one SDs
%
Africa D −1.54 7.1 25.1 38.3 29.5
Africa Eastward −i.five 6.eight 24.two 38.3 thirty.7
The Americas A 0 0.1 2.ane 13.half-dozen 84.two
The Americas B −0.35 0.five 4.five twenty.8 74.2
The Americas D −0.84 1.6 10.8 31.three 56.3
Eastern Mediterranean B −0.half dozen 0.viii 7.3 26.three 65.6
Eastern Mediterranean D −ane.33 4.seven xx.4 37.8 37.1
Europe A 0 0.one ii.1 xiii.half dozen 84.2
Europe B −0.57 0.7 6.9 25.7 66.7
Europe C −0.05 0.2 2.4 14.v 82.9
Southeast Asia B −1.35 5.0 20.8 37.ix 36.three
Southeast Asia D −ane.nine 13.4 32.5 35.8 eighteen.3
Western Pacific Region A −0.22 0.3 3.5 18.0 78.ii
Western Pacific Region B −1.0 2.3 13.6 34.i 50.0
Prevalence
WHO region 2 Mean z score <−iii SDs >−three to <−2 SDs >−2 to <−1 SDs >−ane SDs
%
Africa D −i.54 7.1 25.1 38.3 29.5
Africa E −one.v vi.8 24.two 38.3 30.7
The Americas A 0 0.1 2.ane xiii.half dozen 84.2
The Americas B −0.35 0.5 four.5 twenty.8 74.two
The Americas D −0.84 ane.6 10.eight 31.3 56.3
Eastern Mediterranean B −0.vi 0.8 7.3 26.3 65.6
Eastern Mediterranean D −ane.33 4.7 20.4 37.viii 37.1
Europe A 0 0.1 two.1 13.six 84.two
Europe B −0.57 0.7 6.nine 25.7 66.7
Europe C −0.05 0.2 ii.4 14.five 82.ix
Southeast Asia B −i.35 5.0 twenty.8 37.9 36.iii
Southeast Asia D −1.9 13.iv 32.five 35.viii eighteen.3
Western Pacific Region A −0.22 0.3 3.v 18.0 78.two
Western Pacific Region B −1.0 ii.iii 13.6 34.i 50.0

1

Expected percentages of children in each weight-for-age category in a healthy population are those reported here for The Americas A and Europe A.

two

Regions for the Global Brunt of Disease project. A indicates very depression child mortality and very depression adult mortality; B, depression child bloodshed and low developed mortality; C, depression child mortality and high adult bloodshed; D, high child mortality and high developed mortality; E, loftier child mortality and very high adult bloodshed.

These 2 pieces of information were combined to estimate the PAF for undernutrition overall and for each WHO region ( Table 4). Equally shown, the PAF for The Americas A and Europe A groups are 0%, whereas the PAF for all-cause mortality for the poorest regions of the world are more fifty%. Overall, it is estimated that 52.5% of deaths in immature children worldwide are attributable to undernutrition, with the proportion of deaths varying from 44.8% for measles to 60.seven% for diarrhea.

TABLE iv

Population attributable fraction (PAF) of deaths by cause and World Wellness Organization (WHO) region

PAF by cause of decease
WHO region 1 All causes Diarrhea Pneumonia Measles Malaria
%
Africa D 56.1 64.vii 54.6 45.2 57.7
Africa E 55.ane 61.7 53.half dozen 44.2 56.7
The Americas A 0.00 0.00 0.00 0.00 0.00
The Americas B 13.2 sixteen.1 12.six 9.3 13.eight
The Americas D 32.iv 38.1 31.2 24.iii 33.vii
Eastern Mediterranean B 22.6 27.1 21.seven xvi.5 23.7
Eastern Mediterranean D 49.7 56.3 48.2 39.ii 51.three
Europe A 0.00 0.00 0.00 0.00 0.00
Europe B 21.iii 25.5 20.4 15.5 22.3
Europe C 2.2 2.eight two.ane 1.v 2.iii
Southeast Asia B 50.four 57.1 49.0 39.9 52.0
Southeast Asia D 64.viii 71.i 63.iv 53.viii 66.iv
Western Pacific Region A 8.i 10.0 7.7 5.7 8.5
Western Pacific Region B 38.four 44.5 37.1 29.3 39.8
Globe 52.5 60.7 52.iii 44.8 57.3
PAF by cause of death
WHO region 1 All causes Diarrhea Pneumonia Measles Malaria
%
Africa D 56.one 64.7 54.6 45.two 57.7
Africa E 55.1 61.7 53.six 44.2 56.7
The Americas A 0.00 0.00 0.00 0.00 0.00
The Americas B thirteen.2 16.1 12.6 9.3 13.eight
The Americas D 32.4 38.1 31.2 24.3 33.7
Eastern Mediterranean B 22.6 27.1 21.7 16.v 23.7
Eastern Mediterranean D 49.seven 56.three 48.2 39.2 51.3
Europe A 0.00 0.00 0.00 0.00 0.00
Europe B 21.3 25.5 xx.iv 15.5 22.iii
Europe C 2.two ii.8 2.ane 1.v 2.iii
Southeast Asia B 50.4 57.1 49.0 39.9 52.0
Southeast Asia D 64.8 71.1 63.iv 53.8 66.four
Western Pacific Region A 8.i 10.0 7.7 5.7 8.5
Western Pacific Region B 38.4 44.5 37.1 29.three 39.8
Earth 52.5 60.seven 52.3 44.viii 57.3

1

Regions for the Global Burden of Disease project. A indicates very depression child mortality and very low adult mortality; B, low child mortality and low adult mortality; C, low kid mortality and high adult bloodshed; D, high child mortality and loftier adult mortality; Due east, high child mortality and very loftier developed bloodshed.

Table iv

Population owing fraction (PAF) of deaths by crusade and World Health Organization (WHO) region

PAF by cause of death
WHO region 1 All causes Diarrhea Pneumonia Measles Malaria
%
Africa D 56.1 64.7 54.six 45.2 57.7
Africa E 55.1 61.7 53.half-dozen 44.2 56.7
The Americas A 0.00 0.00 0.00 0.00 0.00
The Americas B xiii.2 16.1 12.6 nine.3 13.eight
The Americas D 32.4 38.one 31.2 24.3 33.7
Eastern Mediterranean B 22.half-dozen 27.1 21.7 xvi.5 23.7
Eastern Mediterranean D 49.7 56.iii 48.2 39.2 51.iii
Europe A 0.00 0.00 0.00 0.00 0.00
Europe B 21.three 25.5 20.iv 15.v 22.3
Europe C 2.two ii.viii 2.1 1.v 2.iii
Southeast Asia B 50.iv 57.1 49.0 39.nine 52.0
Southeast Asia D 64.8 71.1 63.4 53.eight 66.4
Western Pacific Region A eight.ane 10.0 7.7 5.vii viii.5
Western Pacific Region B 38.4 44.5 37.1 29.3 39.8
World 52.5 60.7 52.3 44.eight 57.3
PAF by cause of death
WHO region 1 All causes Diarrhea Pneumonia Measles Malaria
%
Africa D 56.1 64.7 54.half dozen 45.ii 57.7
Africa East 55.1 61.7 53.vi 44.2 56.7
The Americas A 0.00 0.00 0.00 0.00 0.00
The Americas B 13.2 xvi.i 12.vi nine.3 13.viii
The Americas D 32.4 38.1 31.2 24.three 33.7
Eastern Mediterranean B 22.6 27.1 21.seven 16.5 23.7
Eastern Mediterranean D 49.vii 56.three 48.two 39.2 51.3
Europe A 0.00 0.00 0.00 0.00 0.00
Europe B 21.3 25.5 xx.4 15.5 22.3
Europe C 2.2 2.8 ii.1 1.five two.3
Southeast Asia B fifty.four 57.1 49.0 39.9 52.0
Southeast Asia D 64.8 71.1 63.4 53.8 66.4
Western Pacific Region A 8.i 10.0 seven.vii 5.7 viii.5
Western Pacific Region B 38.4 44.5 37.1 29.iii 39.viii
World 52.5 60.vii 52.3 44.eight 57.3

1

Regions for the Global Burden of Affliction projection. A indicates very depression child mortality and very depression adult bloodshed; B, depression child mortality and low adult bloodshed; C, depression kid bloodshed and high adult mortality; D, high child mortality and high adult mortality; E, high child mortality and very high adult mortality.

Word

These analyses bespeak that undernutrition in young children contributes significantly toward the global burden of disease. Indeed, childhood underweight is the leading cause of global burden of disease (21, 22). Deaths attributable to undernutrition embrace 53% of all babyhood deaths, echoing the previous approximate of 55% of all deaths to young children (1, iii). Amid the principal causes of death in young children, threescore.7% of deaths equally a result of diarrhea, 52.iii% of deaths equally a issue of pneumonia, 44.8% of deaths every bit a upshot of measles, and 57.3% of deaths equally a result of malaria are owing to undernutrition. These attributable fractions are large considering undernutrition is the underlying cause for most deaths associated with severe infections and because undernutrition is notwithstanding highly prevalent in many regions of the world. These numbers place prevention of undernutrition among children every bit one of the acme priorities for action in efforts to reduce child mortality.

It should be emphasized that the calculations made by Pelletier et al (3) related to childhood deaths, ie, deaths among children aged 1–four y. This specificity was related to the express availability of information sets with subjects aged younger than 1 y. Equally shown in Tabular array i, many of the studies available for our analyses were prospective cohort designs that monitored the vital status of children aged birth to 4 y. Heuristically, one could apply our results to all deaths during infancy, and nosotros have washed so. However, information technology is also understood that a sizable number of deaths during the neonatal period relate more than to events during pregnancy, birth, and the transition to extrauterine life and less to the infectious causes of death studied here.

The studies included here used specific protocols to appraise both anthropometric status and cause of death. With minimal training, weight measures are highly accurate and precise; therefore, errors in exposure cess are unlikely to have affected our results. However, it is too true that the estimates of the weight-for-age distribution for each WHO region were imputed on the footing of the all-time available information to date; thus, discrepancies with present extent of underweight are possible as new surveys become bachelor. In almost cases, cause of death was determined through exact autopsy that uses standard protocols (23). Several reasons exist why the results for measles and malaria deaths should be interpreted with caution. First, a reduced number of studies contributed to these analyses, half-dozen for measles and iii for malaria. Second, past studies on the influence of low weight-for-age on measles mortality yielded equivocal results (24–26). Third, attributing deaths to malaria is problematic, and, equally in the study in Republic of guinea-bissau, deaths as a result of malaria were not distinguished from deaths as a effect of other causes of fever unless these illnesses had symptoms or signs consequent with other causes such equally diarrhea or pneumonia. The results for Republic of guinea-bissau are consistent with those from Senegal, and, although the results from Ghana suggest a different blueprint of relation betwixt underweight and bloodshed take chances, the number of deaths as a result of malaria is quite small (n = 8). Further studies would need to be included to refine these estimates.

It is of import to consider that low weight-for-age is commonly associated with deficiencies of micronutrients (27). For example, zinc deficiency contributes to poor growth in immature children (28). The association of low weight-for-age with these deficiencies could hateful that some of the risk attributed to underweight should exist attributable to specific micronutrient deficiencies; however, these other deficiencies also occur in children who are not underweight. Moreover, zinc supplementation in populations that have a moderately loftier prevalence of zinc deficiency has had like furnishings on reducing infectious disease morbidity in children who were classified to be undernourished or non (29), and vitamin A supplementation reduces mortality in vitamin A-scarce populations regardless of anthropometric condition (11, thirty). Although such evidence is not conclusive, it does advise that the relations observed hither may only in role be attributed to the effects of concurrent micronutrient deficiencies.

Undernutrition contributes to the morbidity burden amongst children as well, merely this contribution is not consequent across illnesses. Although evidence suggests that low weight-for-age increases hazard of having pneumonia, diarrhea, or a clinical malarial attack, it may not touch the risk of measles, which is ubiquitous even in well-nourished unvaccinated children (22). The relatively greater hazard of dying of infectious disease than for the gamble of having the communicable diseases is understandable, given the synergy between illness and malnutrition (31, 32). This synergy was most clearly demonstrated for diarrhea illnesses (33), and our analyses propose an extension to pneumonia, malaria, and measles every bit well.

Information technology is well understood that young children with moderate-to-severe undernutrition are at increased risk of dying, but note that in these analyses children with weight for age z scores betwixt −i.01 and −2.00 SDs were about twice as likely to dice as children with z scores >−1 SD (the reference grouping). We have labeled this grouping every bit likely suffering from mild undernutrition, although some have argued against using this term because one would await a sizable proportion (thirteen.vi%) of a salubrious reference population to autumn into this category and this expectation needs to be considered during analysis (34). Information technology is of import to recognize that in these analyses we have used the expected prevalences from the reference population equally the counterfactual, which substantially removes them from the burden estimates, and, thus avoids this trouble. We are not arguing for a redefinition of malnutrition to include all children <−1.00 SD, a topic not dealt with in this paper, only our results practice emphasize the point that in populations with loftier prevalences of undernutrition, 30–40% of children have z scores between −1.00 and −ii.00 SDs and have small however significantly increased gamble of dying.

These findings underscore the need to make the comeback of the nutritional status of children a priority. In addition to reducing growth faltering, investments in child diet programs would support and complement disease-specific prevention and command programs in developing countries. In round numbers, this ways that ane 000 000 pneumonia deaths, 800 000 diarrhea deaths, 500 000 malaria deaths, and 250 000 measles deaths could be prevented by eradication of child undernutrition. Although the potential for diet programs to reduce diarrhea morbidity and bloodshed was previously recognized (35), the result that a large proportion of kid deaths equally a result of malaria could be prevented by child nutrition interventions is noteworthy. With electric current programmatic efforts, rates of undernutrition among children are declining in near countries at one% per year or less (36). We and others contend that this amount of progress is unacceptable (37, 38). Strategies to more finer reduce child undernutrition past using experiences gained from successful nutrition programmers (39) are urgently needed.

We thank the researchers who provided us with original data anthropometric status and cause-specific mortality.

LC, MdO, MB, and RB contributed to the conceptualization, data assay, and writing of the manuscript. None of the authors take any financial or personal involvement in any company or organization sponsoring the research.

REFERENCES

ane

Pelletier

DL

,

Frongillo

EA

Jr,

Habicht

J-P

.

Epidemiologic prove for a potentiating result of malnutrition on child mortality

Am J Pub Health

1993

;

83

:

1130

–

3

.

2

Pelletier

DL

,

Frongillo

EA

Jr,

Schroeder

DG

,

Habicht

J-P

.

A methodology for estimating the contribution of malnutrition to child mortality in developing countries

.

J Nutr

1994

;

124

:

2106S

–

22S

.

3

Pelletier

DL

,

Frongillo

EA

Jr,

Schroeder

DG

,

Habicht

J-P

.

The effects of malnutrition on child mortality in developing countries

.

Bull WHO

1995

;

73

:

443

–

8

.

iv

de Onis

M

,

Blössner

G

.

The WHO Global Database on Child Growth and Malnutrition: methodology and applications

.

Int J Epidemiol

2003

;

32

:

518

–

26

.

five

Rice

AL

,

Sacco

L

,

Hyder

A

,

Black

RE

.

Malnutrition equally an underlying cause of babyhood deaths associated with infectious diseases in developing countries

.

Bull WHO

2000

;

78

:

1207

–

21

.

vi

Fawzi

WW

,

Spiegelman

DL

,

El Amin

A

,

Nestel

,

P

,

Mohamed

KA

.

A prospective study of malnutrition in relation to child mortality in the Sudan

.

Am J Clin Nutr

1997

;

65

:

1062

–

9

.

7

Garenne

M

,

Maire

B

,

Fontaine

O

,

Dieng

K

,

Briend

A

.

Risques de décès associés à différents états nutritionnels chez l′enfant d′Âge préscolaire [Risks of dying associated with different nutritional condition in pre-school aged children]

.

Rapport final.

Dakar, Senegal

:

ORSTOM

,

September

1987

:

246

.

Reprinted past CEPED, Paris

,

2000

.

8

Andersen

Yard

.

Anthropometric measurements in health programmes: epidemiological and statistical aspects.

PhD thesis

.

Academy of Copenhagen

,

1997

.

9

WHO/CHD Immunisation-linked vitamin A supplementation study group

.

Randomised trial to assess benefits and rubber of vitamin A supplementation linked to immunization in early infancy. WHO/CHD immunization-linked vitamin A supplementation study group

.

Lancet

1998

;

352

:

1257

–

63

.

10

West

KP

Jr,

LeClerq

SC

,

Shrestha

SR

, et al.

Effects of vitamin A on growth of vitamin-A scarce children: field studies in Nepal

.

J Nutr

1997

;

127

:

1957

–

65

.

11

West

KP

Jr,

Pokhrel

RP

,

Katz

J

, et al.

Efficacy of vitamin A in reducing preschool child bloodshed in Nepal

.

Lancet

1991

;

338

:

67

–

71

.

12

Arifeen

S

,

Black

RE

,

Antelman

G

,

Baqui

A

,

Caulfield

L

,

Becker

Due south

.

Exclusive breastfeeding reduces acute respiratory infections and diarrhea deaths among infants in Dhaka slums

.

Pediatrics

2001

;

108

:

E67

.

thirteen

Khan

SR

,

Jalil

F

,

Zaman

S

,

Lindblad

BS

,

Karlberg

J

.

Early child health in Lahore Islamic republic of pakistan: 10

.

Bloodshed Acta Paediatr Scand

1993

;

82

(

suppl

):

109

–

17

.

14

Jalil

F

,

Lindblad

BS

,

Hanson

LA

, et al.

Early child health in Lahore Pakistan: I. Study Design

.

Acta Paediatr Scand

1993

;

82

(

suppl

):

iii

–

xvi

.

15

Sutrisna

B

,

Reingold

A

,

Kresno

S

,

Harrison

GG

,

Utomo

B

.

Care-seeking for fatal illness in young children in Indramayu, west Java, Republic of indonesia

.

Lancet

1993

;

342

:

787

–

nine

.

16

Ricci

JA

,

Becker

Due south

.

Risk factors for wasting and stunting among children in Metro Cebu, Philippines

.

Am J Clin Nutr

1996

;

63

:

966

–

75

.

17

Pelletier

DL

.

The relationship between child anthropometry and mortality in developing countries: implications for policy, programs and hereafter research

.

J Nutr

1994

;

124

:

2047S

–

81S

.

18

Eide

GE

,

Heuch

I

.

Attributable fractions: fundamental concepts and their visualization

.

Stat Meth Med Res

2001

;

10

:

159

–

93

.

19

Vander Hoorn

S

,

Ezzati

M

,

Rodgers

A

,

Lopez

AD

,

Murray

CJL

.

Estimating owing burden of disease from exposure and hazard information

. In:

Comparative quantification of health risks: the global and regional burden of disease due to 25 selected major run a risk factors.

Cambridge, MA

:

World Health Organization/Harvard Academy Printing

(in press).

20

World Health Organization (WHO)

.

The World Health Report 2002: reducing risks, promoting healthy life.

Geneva

:

Globe Health System

,

2002

.

21

Ezzati

Thousand

,

Lopez

Advert

,

Rodgers

A

,

Vander Hoorn

South

,

Murray

CJL

, and the

Comparative Chance Cess Collaborating Group

.

Selected major risk factors and global and regional burden of disease

.

Lancet

2002

;

360

:

1347

–

lx

.

22

Fishman

Southward

,

Caulfield

LE

,

de Onis

M

,

Blossner

Thou

,

Mullany

L

,

Blackness

RE

.

Malnutrition and the global burden of disease: underweight

. In:

Comparative quantification of health risks: the global and regional burden of illness due to 25 selected major risk factors.

Cambridge, MA

:

Globe Wellness Organisation/Harvard Academy Press

(in press)

.

23

Blindside

AT

,

Bang

RA

, and

SEARCH squad

.

Diagnosis of causes of childhood deaths in developing countries by exact autopsy; suggested criteria

.

Bull WHO

1992

;

70

:

499

–

507

.

24

Chen

LC

,

Chowdhury

A

,

Huffman

SL

.

Anthropometric assessment of energy-poly peptide malnutrition and subsequent risk of bloodshed amongst preschool aged children

.

Am J Clin Nutr

1980

;

33

:

1836

–

45

.

25

Aaby

P

,

Bukh

J

,

Lisse

IM

,

Smits

AJ

.

Measles mortality, state of nutrition, and family construction: a community written report from Guinea-Bissau

.

J Infect Dis

1983

;

147

:

693

–

701

.

26

Aaby

P

.

Malnutrition and overcrowding/intensive exposure in severe measles infection: review of community studies

.

Rev Infect Dis

1988

;

x

:

478

–

91

.

27

Bhan

MK

,

Sommerfelt

H

,

Strand

T

.

Micronutrient deficiency in children

.

Br J Nutr

2001

;

85

(

S2

):

S199

–

203

.

28

Brown

KH

,

Peerson

JM

,

Allen

LH

.

Result of zinc supplementation on children′s growth: a meta-analysis of intervention trials

.

Bibl Nutri Dieta

1998

;

54

:

76

–

83

.

29

Bhutta

ZA

,

Blackness

RD

,

Brown

KH

, et al.

Prevention of diarrhea and pneumonia past zinc supplementation in children in developing countries: pooled analysis of randomized controlled trials. Zinc Investigators′ Collaborative Group

.

J Pediatr

1999

;

135

:

689

–

97

.

30

Sommer

A

,

Tarwotjo

I

,

Djunaedi

E

, et al.

Impact of vitamin A supplementation on childhood mortality: a randomized controlled community trial

.

Lancet

1986

;

1

:

1169

–

73

.

31

Scrimshaw

NS

,

Taylor

CE

,

Gordon

JE

.

Interaction of nutrition and infection.

Geneva

:

World Health Organization

,

1968

.

(Monograph Series 57.)

32

Shankar

AH

.

Nutritional modulation of malaria morbidity and mortality

.

J Infect Dis

2000

;

182

(

suppl

):

S37

–

53

.

33

Black

RE

,

Brown

KH

,

Becker

S

.

Malnutrition is a determining gene in diarrheal duration, just not incidence, amid young children in a longitudinal study in rural Bangladesh

.

Am J Clin Nutr

1991

;

39

:

87

–

94

.

34

Globe Wellness Organization (WHO)

.

Concrete status: the employ and interpretation of anthropometry.

Report of a WHO Good Commission

.

Geneva

:

Earth Health Organization

,

1995

.

35

Becker

Due south

,

Black

RE

,

Brownish

KH

.

Relative furnishings of diarrhea, fever, and dietary energy food intake on weight gain in rural Bangladeshi children

.

Am J Clin Nutr

1991

;

53

:

1499

–

1503

.

36

de Onis

M

,

Frongillo

EA

,

Blössner

M

.

Is malnutrition failing? An analysis of changes in levels of kid malnutrition since 1980

.

Balderdash WHO

2000

;

78

:

1222

–

33

.

37

Black

RE

,

Morris

SS

,

Bryce

J

.

Where and why are 10 million children dying each year?

Lancet

2003

;

361

:

2226

–

34

.

38

Jones

G

,

Steketee

RW

,

Black

RE

,

Bhutta

ZA

,

Morris

SS

;

Bellagio Child Survival Report Grouping

.

How many kid deaths can nosotros prevent this yr?

Lancet

2003

;

362

:

65

–

71

.

39

Allen

LH

,

Gillespie

SR

.

What works? A review of the efficacy and effectiveness of diet interventions.

Un Administrative Committee on Coordination Subcommittee on Diet (ACC/SCN)

,

Geneva

, and

Asian Development Bank

,

Manila, Philippines

,

2001

.

FOOTNOTES

2

Supported in role past the Globe Health Organization Department of Child and Adolescent Health and Development and by the Family Health and Child Survival (FHACS) Cooperative Understanding between the United states of america Bureau for International Development (USAID), Part of Health and Nutrition, and The Johns Hopkins Bloomberg Schoolhouse of Public Health, Department of International Health.

© 2004 American Order for Clinical Nutrition