K. ZAMAN ET Ai_
Acute Respiratory Infections in Children: A Communitybased Longitudinal Study in Rural Bangladesh
by K. Zaman,* A. H. Baqui,***, Md. Yunus,* R. B. Sack,* O. M. Bateman,* H. R. Chowdhury* and R. E. Black**
* International Centre for Diarrhoeal Disease Research, GPO Box - 128, Dhaka - 1000, Bangladesh
** Johns Hopkins University School of Hygiene and Public Health, Baltimore, MD 21218, USA
Summary
A community-based longitudinal study conducted in Matlab, a rural area in Bangladesh, investigated
acute respiratory infections (ARI) among children. A cohort of 696 children under 5 years of age was
followed for 1 year yielding 183,865 child-days of observation. Trained field workers visited the study
children every fourth day. Data on symptoms suggesting ARI, such as fever, cough, and nasal discharge,
were collected for the preceding 3 days by recall. To determine the type and severity of ARI, the field
workers conducted physical examinations (temperature, rate of respiration, and chest indrawing) of
children reporting cough and/or fever. The overall incidence of ARI was 5.5 episodes per child-year
observed; the prevalence was 35.4 per hundred days observed. Most of the episodes (96 per cent) were
upper respiratory infections (URI). The Incidence of acute lower respiratory infections (ALRI) was 0.23
per child per year. The incidence of URI was highest in 18-23-month-old children, followed by Infants 6 11 months old. The highest incidence of ALRI was observed in 0-5-roonth-old infants followed by 12-17month-old children. Among 559 children who were followed for 6 months or longer, about 9 per cent did
not suffer any URI episode and about 16 per cent suffered one or more ALRI episodes. About 46 per cent
of URI and 65 per cent of ALRI episodes lasted 15 days or more. The incidence rates of URI were higher
during the monsoon and pre-winter periods, and that of ALRI at the end of the monsoon and during the
pre-winter periods. Sociodemographic variables were not associated with the incidence of URI or ALRI.
The study documents ARI to be a major cause of morbidity among rural Bangladeshi children.
Introduction
Acute respiratory infections (ARI) are important causes
of morbidity and mortality in young children in the
developing world.1 ARI account for nearly one-third of
Acknowledgements
This research was supported by the CDD programme of the
WHO and the International Centre for Diarrhoeal Disease
Research, Bangladesh (ICDDR,B). The ICDDR,B is supported
by countries and agencies which share its concern for the health
problems of developing countries. Current donors include: the
aid agencies of the governments of Australia, Bangladesh,
Belgium, Canada, China, Germany, Japan, the Netherlands,
Norway, Republic of Korea, Saudi Arabia, Sweden, Switzerland,
the United Kingdom, and the United Stales; international
organizations including the Arab Gulf Fund, Asian Development
Bank, International Atomic Energy Centre, the United Nations
Children's Fund (UN1CEF), the United Nations Development
Programme (UNDP), the United Nations Population Fund
(UNFPA), and the World Health Organization (WHO); private
foundations including Ford Foundation, Population Council,
Rockefeller Foundation, and the Sasakawa Foundation; and
private organizations including American Express Bank, Bayer
AG, CARE, Family Health International, Helen Keller International, the Johns Hopkins University, Procter Gamble,
SANDOZ, Swiss Red Cross, and the University of California
Davis and others. We thank Drs Andres de Francisco and
Kh. Zahid Hasan for reviewing the manuscript and
Mr Mahmudul Haque for helping in analysis of the data.
Journal of Tropical Pediatrics
Vol.43
June 1997
all deaths among children below 5 years of age in many
countries including Bangladesh. The overall incidence
rate of ARI in this age group ranges from 3 to 10 episodes
per child per year,4"6 with the highest incidence among
children less than 2 years of age? The incidence rates in
the developing countries are comparable to those of
developed countries, but cause-specific mortality rates
from ARI are 10-50 times higher in developing than in
developed countries.7'8 A community-based longitudinal
study on ARI conducted in Bangladesh a decade prior to
the present study showed that children spent 60 per cent
of the time with ARI.9 This study was designed to
reassess the problem using more precise current definitions of upper and lower ARI in order to plan future
interventions for the control of ARI in children less than
5 years of age in rural Bangladesh.
Methods
The study was conducted in Matlab, a rural area of
Bangladesh located 45 km southeast of Dhaka, the
capital of Bangladesh. Matlab is a low lying floodprone area intersected by numerous canals and rivers.
There are four distinct seasons, the monsoon period,
July-September with an average of 152cm rainfall
(temperature range 23-38°C), the pre-winter period,
October-November (temperature range 16-33°C), the
dry winter season, December-February (temperature
© Oxford University Press 1997
133
K. ZAMANETAL.
range 13-29°C), and the dry hot period (March-June
(temperature range 26-38T). The population density
of this area is about 700/Km2 and the infant mortality
rate is about 90 per 1000 live births.10 The principal
occupations in the Matlab area are fanning and fishing.
Since 1966, the International Centre for Diarrhoeal
Disease Research, Bangladesh (ICDDR.B) has maintained a Demographic Surveillance System (DSS) in
this area comprised of periodic censuses and intervening longitudinal registration of vital events." A central
diarrhoeal hospital, staffed by physicians and paraprofessionals provides free therapy for all admitted
patients.
The study was carried out in three villages of the
Matlab DSS area. A cohort of 696 children aged 0-59
months was followed prospectively from May 1988 to
April 1989, yielding 183,865 child-days of observation.
A total of 575 children was recruited at the beginning of
the study and 10-12 newboms were recruited monthly.
Of these, 512 children were followed for a full year and
559 were followed for 6 months or longer duration.
Details of recruitment procedures and surveillance
methods have been described previously.12 Briefly, the
field workers for the study were recruited from the same
village as the study participants. They had at least 10th
grade of education, and given 2 months theoretical and
practical training in Matlab hospital and in the field.
They were trained how to conduct interviews, to record
accurately a child's respiratory rate and temperature, and
to recognize cases of chest indrawing. The field workers
were provided with paracetamol syrup and packets of
oral rehydration salts (ORS). They visited all the study
children every fourth day. Data on symptoms suggesting
ARI (fever, cough, nasal discharge) were collected for
the preceding 3 days by recall. Children with reported
cough and/or fever on the day of visit were examined by
the field workers. These children were examined for the
temperature, the rate of respiration, and the presence of
chest indrawing. Upper respiratory infection (URI) was
defined as the presence of fever (>37.8°C) with cough
and/or nasal discharge.13 Acute lower respiratory infection (ALRI) was defined as the presence of cough and
respiratory rate >50/min with or without chest
indrawing.'3 An episode was considered to be an ALRI
when the child developed ALRI at the beginning of the
episode or started with URI followed by ALRI. For some
analyses, cases of URI and ALRI were grouped together
as cases of ARI. A new episode was considered possible
when the child was free of symptoms of ARI for at least 1
week. The number of days for which information was
available for each child were used to calculate the
number of child-days observed as the denominator.
Incidence and prevalence rates were calculated using
person time as the denominator.14 Data from all 696
children were analysed to calculate the age-specific
incidence/prevalence rates, and seasonality of URI and
ALRI. Since some children were followed for a short
period of time, we restricted other analyses to 559
children who were followed for at least 6 months or
longer duration. Children with ALRI needing hospitalized care were referred by the field workers to Matlab
hospital.
Data on socio-economic conditions (sex of the child,
family size, average monthly income, maternal age,
maternal education) were collected from study households at the beginning of the study. About three quarters
of the study children had a Z score below —2 weight-forage and height-for-age, and about one-quarter had a Z
score below —2 for weight-for-height. The immunization status of the children at the beginning of the study
was very low. Only 2 - 3 % of children had received
diphtheria-pertussis-tetanus and BCG vaccine. After 3
months of follow-up all children aged 9-23 months were
immunized against measles.
Data quality was ensured through proper training of
workers, fortnightly meetings, standardization of
TABLE 1
Incidence and prevalence rates of upper and lower respiratory infections by age of study children (n = 696),
Matlab, Bangladesh, May 1988 to April 1989
URI
Age group
(months)
0-5
6-11
12-17
18-23
24-29
30-35
36-41
42-47
48-59
Total
Child days
observed
10,651
20,456
20,292
23,649
27.236 •
24.018
21,335
17,839
18,389
183,865
No. of episodes Incidence*
ALRI
Prevalencet
No. of episodes
Incidence
Prevalence
162
352
262
422
379
311
284
261
239
5.55
6.28
4.71
6.51
5.08
4.73
4.86
5.34
4.74
4155
38.44
30.81
34.82
34.29
27.69
30.02
26.24
32.67
15
11
19
16
18
10
15
9
4
051
0.20
0.34
0.25
0.24
0.15
0.26
0.18
0.08
8.87
0.76
5.20
2.53
1.84
1.97
4.38
2.49
0.86
2672
5.30
32.59
117
0.23
2.86
* Per child-year observed, incidence rate = no. of new episodes (URI or ALRI)/total days observed • 365.
t Per 100 days observed, prevalence rate = no. of days with URI or ALRI/total days observed • 100.
134
Journal of Tropical Pediatrics
Vol.43
June 1997
K. ZAMANCTAL.
methods, continuous supervision and duplicate morbidity surveillance on a sample of study children.
Results
A total of 2823 ARI episodes was identified of which
2789 were new (incident) cases and the remaining 34
cases were present at the time of enrolment. The overall
incidence of ARI in the study children was 5.5 episodes
per child-year observed. Most of the episodes (96 per
cent) were URI (incidence 5.3 per child per year). Only 4
per cent of the total episodes were ALRI; the incidence
of ALRI was 0.23 episodes per child per year (Table 1).
The incidence rate of URI was highest in children 18-23
months old followed by infants 6-11 months. The
highest incidence of ALRI was observed in infants below
6 months of age.
The overall prevalence of ARI was 35.4 per hundred
days observed; the prevalence of both URI and ALRI
were highest in infants under 6 months of age (Table 1).
About 46 per cent of URI and 65 per cent of ALRI
episodes lasted 15 days or more. The median duration of
URI episodes was 13 days and that of ALRI episodes
was 28 days. The distribution of children by incidence
rates of URI is shown in Fig. 1; 9 per cent of the children
did not experience a URI episode during the period of
study, 7 per cent experienced one, 7 per cent experienced
two, 9 per cent experienced three and another 10 per cent
experienced four episodes. About 11 per cent suffered 10
or more URI episodes.
The distribution of the study children by incidence
rates of ALRI was highly skewed; 84 per cent of the
children did not experience any ALRI episode, 11 per
cent experienced one episode per child year and about 5
per cent experienced two or more episodes per child year
(data not shown).
The distribution of children by ARI prevalence
rates showed that 8 per cent children suffered from
ARI for 1-4 per cent of the days, 5 per cent for 5-9
per cent of the days. About 31 per cent children
io6
*
4
lirm
2
0
0
1
2
10 11 12 13
Prevalence of ARV100 days
FIG. 2. Frequency distribution of study children (n =
559) by prevalence rates of acute respiratory infections,
Matlab, Bangladesh, May 1988 to April 1989.
May Jun Jul Aug Sep Oct Nov Dec Jm
Ftto Mar Apr
Month
FIG. 3. Seasonal distribution of incidence rates of upper
and lower respiratory infections among study children
(n = 696), Matlab, Bangladesh, May 1988 to April 1989.
suffered from ARI for > 50 per cent of the total days
observed (Fig. 2).
The incidence rates of URI were high during JulyNovember (monsoon and pre-winter period) and those of
ALRI in September-November (end of monsoon and
pre-winter period). The highest peak for both URI and
ALRI was in September (end of monsoon period)
(Fig. 3). There was no significant association between
the sociodemographic variables studied (sex of the child,
family size, average monthly income, maternal age,
maternal education) and the incidence of URI or ALRI
(data not shown).
A total of 27 referred children with severe ALRI was
admitted in Matlab hospital. There were three deaths
among study children during the study period. Two of
these were related to ALRI.
Eplsodes/chlld/year
FIG. I. Frequency distribution of study children (n =
559) by incidence rates of upper respiratory infections,
Matlab, Bangladesh, May 1988 to April 1989.
Journal of Tropical Pediatrics
Vol.43
June 1997
Discussion
This study clearly demonstrates that ARI is an important
cause of morbidity among rural Bangladeshi children.
135
K. ZAMANETAL
The overall incidence of ARI was in the same range as
obtained from studies in both developing and developed
countries.4""* However, the prevalence was much lower
than reported by Black el al.9 from the same community. This might be due to differences in surveillance
methods and/or the definitions used for ARI episodes. The
study by Black et al. was based on alternate day surveillance and used a less restrictive definitions.9 Furthermore, the incidence rate of ALRI was much lower than the
rates reported from Costa Rica15 and Gambia.16
Some methodologic differences need to be considered
when comparing rates of ARI from different studies.
First, we used a more restrictive definition of URI, taking
the presence of fever as a necessary sign although fever
had not been shown to be a necessary indicator for
ALRI.' 7 Secondly, we used a cut off point of 50
respirations/minute irrespective of the age of the child to
define a case of ALRI. This may have resulted in
underestimation of ALRI cases in children 12-59
months age since a new criterion of 40 respirations/
min has been suggested by WHO in that age group, but a
partially offsetting over-estimation of ALRI in children
under 2 months old, a group in which WHO now
suggests a cut off of 60 respirations/min.17 Finally, we
have calculated the ARI incidence rates for each child
using total days observed as the denominator. The rate
may be less than in other studies13 where child-days at
risk (days without ARI) were used as the denominator.
The frequent morbidity surveillance by a group of
trained field workers resident in the respective village
should have resulted in identification of most of the
episodes.
The highest incidence of ALRI in our study was in the
first 6 months of life, similar to previous reports.18 The
median duration of an ALRI episode in our study was
similar with Philippine study13 but much longer than in
American community study.19 This longer duration of
ALRI episode might be due to inclusion of days with
URI in an ALRI episode and the high prevalence of
malnutrition among study children." We observed a
higher incidence of URI during the monsoon and prewinter season, whereas ALRI peaked at the end of
the monsoon and in the pre-winter seasons. In some
countries the highest incidence was observed during
winter and in others in the warm season, however,
the seasonal peaks of ALRI were often different from
those of ARI.' 3 2 0
No community-based aetiological study on ARI has
been conducted in Bangladesh. A hospital-based study in
Bangladesh revealed that respiratory syncytial virus
(RSV) was the most common ALRI pathogen in children
under 5 years of age.21 In a community-based study in
Colombia RSV was the most common viral agent for
children below 18 months of age.18
The mortality was substantially low among our study
children. We observed only three deaths during the
follow-up periods. Using the age specific mortality rate of
the community during the same period of time22 the
expected number of deaths would be 10. Factors
136
associated with this low mortality may be the early
recognition of symptoms by frequent home visits of the
field workers, timely referral of severely sick children to
Matlab hospital, provision of ORS, measles immunization to all eligible children, and treatment facilities for
minor illnesses at the household level.
We did not find any association between low socioeconomic status or family size with the occurrence of
URI or ALRI. Such a result agrees with findings in
Burkina-Faso,23 but differs from those in India.24 In the
Indian study higher incidence of ARI was observed
among children from low socio-economic group than
children from the high socio-economic group.
We have demonstrated that children in this setting
have high rates of both URI and ALRI. Our ability to
control URI is limited since these are mostly viral
infections for which no vaccine exist, however, control
strategies can be directed at ALRI which have a high
case fatality rate. Other studies have shown that ALRI,
predominantly pneumonia, is due to bacterial infections26 for the majority of the cases and that case
management with effective antibiotics can reduce ALRI
mortality by half.27 This study indicates that approximately 15 per cent of all under five children need to be
treated each year using current WHO ALRI case
management guidelines. Although this is a sizable
commitment of health resources, it has been shown to
be highly cost effective health intervention in developing
countries.
Additional technologies e.g. vaccines
against RSV and bacterial pneumonic pathogens, are
needed to sustain reductions in ALRI mortality in
developing country children.
References
1. Denny FW, Loda FA. Acute respiratory infections are the
leading cause of death in children in developing countries.
Am J Trop Med Hyg 1986; 35: 1-2.
2. Stansfield SK. Acute respiratory infections in the developing world: strategies for prevention, treatment and control.
Pediat Infect Dis J 1987; 6: 622-9.
3. Spika JS, Munshi MH, Wojtyniak B, el al. Acute lower
respiratory infections- a major cause of death in children in
Bangladesh. Annl Trop Paediat 1989; 9: 33-9.
4 Tupasi TE, Velmonte MA, Sanvictores MEG, et al Determinants of morbidity and mortality due to acute respiratory
infections: implications for intervention. J Infect Dis 1988;
157:615-23.
5. Wafula EM, Onyango FE, Mirza WM, et al. Epidemiology
of acute respiratory tract infections among young children
in Kenya. Rev Infect Dis 1990; 12 (suppl 8): S1035-8.
6. Denny FW, Collier AM, Henderson FW. Acute respiratory
infections in day care. Rev Infect Dis 1986; 8: 527-32.
7. World Health Organization. A programme for controlling
acute respiratory infections in children. Memorandum from
a WHO meeting. Bull Wld Hlth Org 1984; 62: 47-58.
8. Mohs E. Acute respiratory infections in children: possible
control measures. Bull Pan Am Hllh Org 1985; 19: 82-7.
9. Black RE, Brown KH, Becker, S, Yunus M. Longitudinal
studies of infectious diseases and physical growth of
children in rural Bangladesh. I. Patterns of morbidity. Am J
Epidemiol 1982; 115: 305-14.
Journal of Tropical Pediatrics
Vol. 43
June 1997
K. ZAMAN ET AL.
10. International Centre for Diarrhoeal Disease Research,
Bangladesh. Annual Report, 1993.
11. International Centre for Diarrhoeal Disease Research,
Bangladesh. Demographic Surveillance Syslem-Matlab,
Volume 1, methods and procedures, ICDDR, B Sci Rep
1978:9.
12. Baqui AH, Black RE, Yunus M, Hoque ARA, Chowdhury
HR, Sack RB. Methodological issues in diarrhoeal diseases
epidemiology: definition of diarrhoeal episodes. Inl J
Epidemiol 1991; 20: 1057-63.
13. Selwyn BJ. The epidemiology of acute respiratory tract
infection in young children: comparison of findings from
several developing countries. Coordinated Data Group of
BOST1D researchers. Rev Infect Dis 1990; 12: S870-88.
14. Rothman KJ. Modern Epidemiology. Hypothesis testing
with crude person-time data. Boston: Little, Brown and
Company, 1986: 153-6.
15. James JW. Longitudinal study of the morbidity of diarrheal
and respiratory infections in malnourished children. Am J
ClinNutr 1972; 25: 690-4.
16. de Francisco A, Morris J, Hall AJ, Schellenberg JRMA,
Greenwood BM Risk factors for mortality from acute
lower respiratory tract infections in young Gambian
Children. Int J Epidemiol 1993; 22: 1174-82.
17. World Health Organization. Acute respiratory infections in
children: case management in small hospitals in developing
countries. A manual for doctors and other Senior Health
Workers. WHO/ARI/90.5 Geneva: WHO, 1990.
18. Borrero IH, Fajardo LP, Bedoya AM, Zea A, Carmona F,
de Borrero MF. Acute respiratory tract infections among a
birth cohort of children from Cali, Colombia, who were
Journal of Tropical Pediatrics
Vol. 43
June 1997
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
studied through 17 months of age. Rev Infect Dis 1990; 12
(suppl 8): S950-6.
Monto AS, Ullman BM. Acute respiratory illness in an
American community; the Tecumseh study. J Am Med Ass
1974; 227: 164-9.
Denny FW. Clyde WA. Acute lower respiratory tract
infections in non-hospitaJized children. J Pediat 1986; 108:
635-46.
Rahman M, Huq F, Sack DA, el al. Acute lower respiratory
tract infections in hospitalized patients with diarrhoea in
Dhaka, Bangladesh. Rev Infect Dis 1990; 12 (suppl 8):
S899-906.
International Centre for Diarrhoeal Disease Research,
Bangladesh: Demographic Surveillance System-Matlab;
registration of demographic events—1988. ICDDR.B Sci
Rep 1993: 71.
Lang T, Lafaix C, Fassin D, et al. Acute respiratory
infections: a longitudinal study of 151 children in BurkinaFaso. Int J Epidemiol 1986; 15: 553-61.
Verma IC, Menon PSN. Epidemiology of acute respiratory
disease in north India. Ind J Pedial 1981; 48: 37-40.
Denny FW, Clyde WA. Acute respiratory tract infections:
An overview. Paediatr Res 1983; 17: 1026-9.
Shann F. Etiology of severe pneumonia in children in
developing countries. Paediat Infect Dis J 1986; 5: 247-52.
Sazawal S, Black RE. Meta-analysis of intervention trials
on case-management of pneumonia in community settings.
Lancet 1992; 340: 528-33.
World Development Report 1993: Investing in health. New
York: World Bank Oxford University Press, 1993.
137