Modifiable Risk Factors For Community-Acquired Pneumonia in Children Under 5 Years of Age in Resource-Poor Settings - A Case-Control Study
Modifiable Risk Factors For Community-Acquired Pneumonia in Children Under 5 Years of Age in Resource-Poor Settings - A Case-Control Study
13211
1 Institute of Public Health, Kilimanjaro Christian Medical University College, Moshi, Tanzania
2 Section of Pediatric Infectious Diseases, Radboud University Medical Center, Nijmegen, The Netherlands
3 Kilimanjaro Clinical Research Institute, Moshi, Tanzania
4 Kilimanjaro Regional Health Management Team, Moshi, Tanzania
5 Community Health Department, Kilimanjaro Christian Medical Centre, Moshi, Tanzania
Abstract objective Despite the availability of vaccines and antibiotics, pneumonia remains the leading cause
of mortality among children under 5 years of age. The objective of this study was to identify
modifiable risk factors for community-acquired pneumonia (CAP) in children under 5 years of age in
a vaccinated population.
methods A case–control study was conducted between January and December 2017. The cases
included children aged 2–59 months with X-ray-confirmed pneumonia, whereas the controls were
children from the community with no history of respiratory infection. A multivariable logistic
regression model was used to determine the modifiable risk factors for CAP.
results A total of 113 children with X-ray-confirmed pneumonia and 350 healthy children were
enrolled in this study. The median ages for the cases and controls were 13.7 (IQR = 7.2–25.3) and
13.4 (IQR = 6.0–24.8) months respectively. One (0.9%) child died after the enrolment. The
independent predictors of CAP included a lack of exclusive breastfeeding for 6 months (aOR = 1.7,
95% CI = 1.0–2.9), underweight (aOR = 2.1, 95% CI = 1.0–4.5), unclean cooking fuel (aOR = 1.8,
95% CI = 1.0–3.3) and low income (aOR = 2.9, 95% CI = 1.6–5.4). No association was found
between vaccination status and CAP.
conclusion In addition to a lack of exclusive breastfeeding, children from families of low-
economic status were at risk of contracting CAP. Since the risk factors are complex, the study results
call for more concerted efforts by and collaboration among the health, agriculture and development
sectors to address mortality caused by CAP.
keywords community-acquired pneumonia, risk factors, children under 5 years of age, Tanzania
Screened Screened
n = 217 n = 387
Excluded:
Excluded:
n = 3 denied consent
n = 7 denied consent
n = 13 not healthy
n = 5 critically ill
n = 12 inappropriate
n = 20 wheezing
age
A brief description of the study aims was provided to without an immunisation card were included in the study
the parents, and a further detailed discussion was held but were excluded from the analysis of birth weight and
with interested parents. In addition to general characteris- vaccination status. Finally, households in which children
tics (e.g. age of the child, sex of the child, age of the par- under 5 years of age shared a bedroom with more than
ent and education level of the parent), the following two other people were considered overcrowded.
explanatory variables were also collected. Breastfeeding After the interviews, weight was measured using a Seca
status was considered, and exclusive breastfeeding was Digital Baby Scale Model 354 (Seca GmbH & Co. KG,
defined according to the WHO recommendation of Hamburg, Germany), and the weight was recorded to the
6 months [25]. Families using gas or electricity to cook nearest 0.1 kg. The weight-for-age Z-score was derived
were considered to use clean cooking fuel, while biomass, from the WHO child growth standards, and underweight
firewood, charcoal and kerosene were considered was defined as a Z-score below 2 SDs [26]. Finally, for
unclean. A prior history of medications used was col- participants with an unknown HIV status, initial HIV
lected, particularly antibiotics and zinc tablets. Vaccina- screening was conducted by SD bioline HIV-1/2 3.0, a
tion status and birth weight were extracted from child lateral flow immunochromatographic assay (SD STAN-
growth monitoring and vaccination cards. Children DARD DIAGNOSTICS, INC. 65, Borahagal-ro,
Giheung-gu, Yongin-si, Gyeonggi-do, Korea), followed by age- and sex-matched healthy controls (Table 1). Half
confirmation with Uni-Gold HIV, a lateral flow (51.6%) of the controls were exclusively breastfed com-
immunochromatographic assay (TRINITY BIOTECH pared to 36.7% of the cases. Approximately one-quarter
PLC, Bray, Co. Wicklow, Ireland). Pre- and post-test (20.7%) of the children with CAP had a low weight per
counselling was offered to all participants’ parents prior age compared with 6.6% of the healthy controls.
to HIV testing. None of the children tested were found to A total of 77 cases (68.1%) were self-referral patients,
be positive; those enrolled in the study who were already with approximately one-quarter (23.9%) of all cases
infected had a known HIV status. Eighty (22.9%) parents given antibiotics at home prior to admission. Most cases
of the controls declined the HIV test, while all cases were (72.6%) had severe CAP, and one (0.9%) patient died in
either tested or had a known HIV status. When compar- the ward.
ing the proportions of HIV status among the cases and According to the bivariate analysis, we found several
controls, those with an unknown HIV status were factors that were significantly associated with CAP
excluded. among children under 5 years of age. These factors
included low birth weight, lack of exclusive breastfeed-
ing, incomplete vaccination and undernutrition (Table 1).
Data analysis
In addition, income-generating activities, type of cooking
Data processing and analysis were performed in STATA fuel, and type of house were significant risk factors for
version 14 (StataCorp. 2016. Stata Statistical Software: CAP (Table 2).
Release 14. College Station, TX, USA: StataCorp LLC). According to the multivariate logistic regression, only a
A logistic regression analysis was carried out to examine lack of exclusive breastfeeding, underweight, unclean
the risk factors associated with CAP. Since the matching cooking fuel and poverty remained significantly associ-
was considered weak, the crude odds ratios (cORs) and ated with CAP (Table 3). Children with CAP had almost
corresponding 95% CIs were calculated using uncondi- twice the odds of not being breastfed exclusively for
tional logistic regression. In the bivariate analysis, factors 6 months compared with healthy control children
with a P-value ≤0.2 were included in the final multivari- (aOR = 1.7, 95% CI = 1.0–2.8). Furthermore, the odds
able model. Additionally, several other variables were of being underweight were 2.5-times higher in children
selected a priori based on previous studies. Collinearity with CAP than in the healthy controls (aOR = 2.5, 95%
was identified between the type of house and income-gen- CI = 1.3–5.1). Additionally, children with CAP were
erating activities; therefore, the model with income-gener- more likely to be from a household that used unclean
ating activities was selected. The final model included cooking fuel compared with the healthy controls
both child characteristics and socio-demographic infor- (aOR = 1.7, 95% CI = 1.0–2.8). Lastly, the odds that
mation. Differences with a P-value <0.05 were considered parents of CAP patients were involved in farming or per-
statistically significant. formed unskilled labour, resulting in a low income, were
three times higher than the odds for parents of healthy
controls (aOR = 3.1, 95% CI = 1.7–5.5).
Ethical consideration
The study was approved by the Kilimanjaro Christian
Discussion
Medical University College Research Ethics Review Com-
mittee under certificate number 948 and the Tanzanian This case–control study examined risk factors for CAP
National Health Research Ethics Committee under certifi- among children under 5 years of age in a resource-limited
cate NIMR/HQ/R.8a/Vol.IX/2422. The participants’ par- setting. The main findings of this study support some of
ents or legal guardians were fully informed of the risks the WHO/UNICEF pneumonia risk factors in children,
and the benefits of participating in the study and under- which include lack of exclusive breastfeeding, poor nutri-
stood that participation was voluntary. Before enrolment, tion, indoor air pollution, and poverty. However, some
the parents or legal guardians of the eligible children of the factors previously reported to influence the risk of
signed a written consent form. developing CAP, such as lack of immunisation and over-
crowding, were not found to play a significant role. The
use of antibiotics and zinc tablets offered significant pro-
Results
tection against CAP.
A total of 463 children between 2 and 59 months of age Consistent with previous knowledge from a systematic
were eligible and were enrolled in this study, 113 of review [27], we found that a lack of exclusive breastfeed-
whom had X-ray-confirmed pneumonia, and 350 were ing for at least the first 6 months of life was associated
Table 1 Distribution of the child-related factors and bivariate analysis for the risk of acquiring community-acquired pneumonia (CAP;
n = 463)
CI, confidence interval; cOR, crude odds ratio; Pedzinc, zinc tablet; WAZ score, weight-for-age Z-score.
*Variables expressed as the median (IQR).
with CAP in this population. Available evidence suggests Children with undernutrition have impaired immune
that exclusive breastfeeding offers protection against gas- function, which predisposes them to persistent and recur-
trointestinal infection and pneumonia [28]. This protec- rent colonisation and subsequent pneumonia infection;
tion arises through the transfer of immunoglobulins, however, infections also predispose children to malnutri-
primarily IgA, to the mucosal surfaces of the child [29]; tion [33]. This concept of reverse causality is unlikely in
since IgA is the predominant immunoglobulin in the lung our study because of the short duration of illness (a med-
mucosa, it is part of the first-line response to infection ian of 3 days), suggesting that being underweight pre-
[30]. Children not exclusively breastfed for 6 months are ceded CAP infection. These children are not only more
more likely to be vulnerable to respiratory tract infections vulnerable to the development of pneumonia but also at
due to a lack of sufficient mucosal immunoglobulins [31]. risk for other infectious diseases, such as gastrointestinal
Most women feeding other foods to their infants before tract infection and tuberculosis [33]. More importantly,
the recommended 6 months believe that breast milk is malnutrition impacts growth and development, especially
not sufficient for older infants [32]. This finding reflects in younger children [34]. While several risk factors lead
the need for improved health education and awareness of to undernutrition in children, poverty tends to predomi-
the importance of exclusive breastfeeding and the need to nate [35]. Since undernutrition in children is complex,
address community myths surrounding this subject. holistic approaches are likely to produce good outcomes.
In this study, underweight children had a higher risk of As mentioned previously, the association found
CAP, which is consistent with available evidence demon- between the income of the parents and CAP is not sur-
strating a link between malnutrition and CAP [33]. prising. In this study, the income-generating activities of
Table 2 Bivariate analysis of socio-economic, demographic and environmental risk factors for community-acquired pneumonia (CAP)
among children under 5 years of age (n = 463)
the parents were linked to other environmental risk fac- especially those of a younger age, tend to be more exposed
tors for pneumonia. Parents who were less educated were because they are carried by or on the back of their mothers
more likely to be involved in small-scale farming and during the preparation of the family’s food [38]. This contin-
unskilled labour compared to those with a high level of uous and longer exposure to polluted air results in impaired
education. These parents have limited resources to sustain pulmonary alveolar macrophages and epithelial cells [39].
their families; therefore, they ultimately live in houses These findings are consistent with the current body of
with poor ventilation and use unclean cooking fuels. knowledge supporting the association between unclean
Additionally, with food insecurity in the family, these cooking fuel and childhood pneumonia [6, 8]. The use of
children are likely to suffer poor nutrition, which is a clean cooking fuel has not yet received substantial attention
potential risk factor for pneumonia. Consistent with our in the health sector, especially in Tanzania. However, there
findings, several previous studies have reported an associ- are several efforts advocating the use of clean energy as a
ation between family income and the risk of pneumonia strategy to combat climate change. These efforts have a mul-
in children [36, 37]. tiplier effect, including a reduced risk of pneumonia.
Lastly, the use of unclean cooking fuel was the main con- The above-mentioned risk factors for childhood pneu-
tributor to indoor air pollution in the resource-limited set- monia have been suggested to have decreased [40],
tings of this study and increased the risk of CAP. Children, although the trend is different in this resource-poor
Table 3 Multivariate analysis of risk factors associated with community-acquired pneumonia (CAP) among children under 5 years of
age
setting, and a lack of exclusive breastfeeding, indoor air proportion of children had not completed vaccinations
pollution, and malnutrition are the key pneumonia risk per their age. Additionally, with this high-vaccine cover-
factors. The pooled global estimates sometimes mask dis- age, herd immunity is possible for unvaccinated chil-
parities between regions. Therefore, continued monitoring dren. Evidence from numerous studies shows that the
of improvements in nutrition and living standards in dif- PCV offers protection even to unvaccinated children
ferent regions is warranted. [41–43].
Notably, some of the known risk factors for CAP, such Lastly, no significant association was found between
as immunisation and parental cigarette smoking, were parental smoking and the risk of CAP. This finding dif-
not found to be significantly associated with pneumonia fers from that of a recent review that found that parental
in this population. Several factors might have contributed smoking was a strong predictor of lower respiratory tract
to these conflicting findings, such as the high vaccination infection [44]; however, in this population, only a small
coverage in the population. For example, only a small proportion of parents were smokers.
The most important strength of this study is that inci- Exchange Service (DAAD). The funding bodies had no
dent cases were used, and all cases were confirmed by X- role in the design of the study, collection, analysis, and
ray, thus preventing the inclusion of non-pneumonia interpretation of the data or in writing the manuscript.
cases with infection-related respiratory tract diseases. Sec-
ond, the enrolled cases were derived from different types
of health facilities, which provided us with a broad view References
of the risk factors, as different types of patients are 1. WHO. Fulfilling the Health Agenda for Women and Chil-
admitted to each of these hospitals. dren. The 2014 Report ER. WHO: Geneva, 2014.
The findings should be interpreted with some caution 2. UNICEF. Pneumonia: Current Status + Progress [Internet].
because of the risk of biases. In terms of selection bias, UNICEF DATA: Monitoring the Situation of Children and
some of the children with severe CAP needing urgent medi- Women. 2018. (Available from: https://data.unicef.org/topic/
cal care in the intensive care unit were not enrolled; the child-health/pneumonia/) [8 Oct 2018]
parents of these children were not emotionally capable of 3. Rudan I, Boschi-Pinto C, Biloglav Z, Mulholland K, Camp-
bell H. Epidemiology and etiology of childhood pneumonia.
participating in the interview. Such exclusion may have
Bull World Health Organ 2008: 86: 408–416.
resulted in under-reporting of some of the risk factors and
4. WHO. Ending Preventable Child Deaths from Pneumonia
may have therefore biased the results in the direction of no and Diarrhoea by 2025: The integrated Global Action Plan
association. Similarly, a priori selection of the three wards for Pneumonia and Diarrhoea (GAPPD). WHO/UNICEF:
may have introduced a bias related to the controls; Geneva, Switzerland, 2013; 1–61.
although comparable, some minor differences exist 5. (UNICEF) UNCF. One is too many Ending child deaths
between the cases and controls. Though not significant, an from pneumonia and diarrhoea. 2016.
inverse association was observed between parental smok- 6. Kilabuko JH, Nakai S. Effects of cooking fuels on acute res-
ing and CAP. However, the risk of a social desirability bias piratory infections in children in Tanzania. Int J Environ
is likely in this study, as the parents of children with CAP Res Public Health 2007: 4: 283–288.
may have underreported their smoking habits because they 7. Karki S, Fitzpatrick AL, Shrestha S. Risk factors for pneu-
monia in children under 5 years in a teaching hospital in
knew the health consequences of smoking.
Nepal. Kathmandu Univ Med J (KUMJ) 2014: 12: 247–252.
8. Mahalanabis D, Gupta S, Paul D, Gupta A, Lahiri M,
Conclusion Khaled MA. Risk factors for pneumonia in infants and
young children and the role of solid fuel for cooking: a
More efforts are needed to address the lack of exclusive case-control study. Epidemiol Infect 2002: 129: 65–71.
breastfeeding. In addition, more concerted efforts should 9. Ujunwa F, Ezeonu C. Risk factors for acute respiratory tract
be invested to combat malnutrition, which is also linked infections in under-five children in Enugu Southeast Nigeria.
with family poverty, and thus minimise preventable Ann Med Health Sci Res 2014: 4: 95–99.
deaths among children, such as those caused by CAP. To 10. Jackson S, Mathews KH, Pulanic D et al. Risk factors for sev-
attain sustainable development goal number three, ‘ensur- ere acute lower respiratory infections in children: a systematic
review and meta-analysis. Croat Med J 2013: 54: 110–121.
ing healthy lives and promoting well-being for all at all
11. Madhi SA, Levine OS, Hajjeh R, Mansoor OD, Cherian T.
ages’, the other risk factors described in this study should
Vaccines to prevent pneumonia and improve child survival.
also be addressed, as most of them are linked. Bull World Health Organ 2008: 86: 365–372.
12. WHO. Pneumococcal conjugate vaccine for childhood
Acknowledgements immunization – WHO position paper. Wkly Epidemiol Rec
2007: 82: 93–104.
First, the authors would like to acknowledge the children 13. WHO. WHO position paper on Haemophilus influenzae
who participated in this study and the parents who type b conjugate vaccines. Wkly Epidemiol Rec 2006: 81:
allowed their children to take part. Second, the authors 445–452.
thank the hospital administrators for their permission to 14. GAVI Alliance. Tanzania immunization Annual Progress
enrol patients. Third, the authors are grateful to the Report 2013 [Internet]. (Available from: https://www.gavi.
org/country/tanzania/documents/) [17 Sep 2018].
research assistant who helped with data collection.
15. Schmidt W-P, Cairncross S, Barreto ML, Clasen T, Genser
Lastly, the authors extend their gratitude to the members
B. Recent diarrhoeal illness and risk of lower respiratory
of the Institute of Public Health, Kilimanjaro Christian infections in children under the age of 5 years. Int J Epi-
Medical University College, for their input. This research demiol 2009: 38: 766–772.
was supported by the Fogarty International Center of the 16. Mahende C, Ngasala B, Lusingu J et al. Aetiology of acute
National Institutes of Health under Award Number febrile episodes in children attending Korogwe District Hospi-
D43TW010138 and partly by the German Academic tal in north-eastern Tanzania. PLoS One 2014: 9: e104197.
17. D’Acremont V, Kilowoko M, Kyungu E et al. Beyond 33. Schlaudecker EP, Steinhoff MC, Moore SR. Interactions of
malaria–causes of fever in outpatient Tanzanian children. N diarrhea, pneumonia, and malnutrition in childhood: recent
Engl J Med 2014: 370: 809–817. evidence from developing countries. Curr Opin Infect Dis
18. Elfving K, Shakely D, Andersson M et al. Acute uncompli- 2011: 24: 496–502.
cated febrile illness in children aged 2–59 months in zanz- 34. Perkins JM, Kim R, Krishna A, McGovern M, Aguayo VM,
ibar – aetiologies, antibiotic treatment and outcome. PLoS Subramanian SV. Understanding the association between
One 2016: 11: e0146054. stunting and child development in low- and middle-income
19. Kahabuka C, Kv ale G, Hinderaker S. Factors associated countries: next steps for research and intervention. Soc Sci
with severe disease from malaria, pneumonia and diarrhea Med 2017: 193: 101–109.
among children in rural Tanzania – a hospital-based cross- 35. Akombi BJ, Agho KE, Hall JJ, Wali N, Renzaho AMN,
sectional study. BMC Infect Dis 2012: 12: 219. Merom D. Stunting, wasting and underweight in Sub-
€
20. Ostberg W, Howland O, Mduma J, Brockington D. Tracing Saharan Africa: a systematic review. Int J Environ Res Pub-
improving livelihoods in rural africa using local measures of lic Health 2017: 14: 863.
wealth: a case study from central Tanzania, 1991–2016. 36. Fonseca Lima EJ, Mello MJ, Albuquerque MF et al. Risk
Land 2018: 7: 44. factors for community-acquired pneumonia in children
21. World Health Organization, Department of Maternal N, under five years of age in the post-pneumococcal conjugate
World Health Organization. Revised WHO classification vaccine era in Brazil: a case control study. BMC Pediatr
and treatment of pneumonia in children at health facilities: 2016: 16: 157.
evidence summaries. 2014. 26. 37. Johnson WBR, Abdulkarim AA. Childhood pneumonia in
22. Mackenzie G. The definition and classification of pneumo- developing countries. Afr J Respir Med 2013: 8: 4–9.
nia. Pneumonia (Nathan Qld) 2016: 8: 14. 38. Smith K, Mehta S, Feuz F. Indoor smoke from household
23. Bartlett JG, Dowell SF, Mandell LA, File TM, Musher DM, Fine solid fuels. In: Ezzati MR, Lopez A Murray C (eds). Com-
MJ. Practice guidelines for the management of community- parative Quantification of Health Risks: Global and Regio-
acquired pneumonia in adults. Clin Infect Dis 2000: 31: 347– nal Burden of Disease Due to Selected Major Risk Factors.
382. World Health Organization: Geneva, 2004.
24. Kirkwood BR, Sterne JAC. Essential Medical Statistics (2nd 39. Chauhan AJ, Johnston SL. Air pollution and infection in res-
edn). Blackwell Science: Hoboken, NJ, 2003, 501. piratory illness. Br Med Bull 2003: 68: 95–112.
25. WHO. WHO | Breastfeeding [Internet]. World Health Orga- 40. McAllister DA, Liu L, Shi T et al. Global, regional, and
nization; 2018. (Available from: http://www.who.int/topics/ national estimates of pneumonia morbidity and mortality in
breastfeeding/en/) [4 Aug 2018]. children younger than 5 years between 2000 and 2015: a
26. WHO Multicentre Growth Reference Study Group. WHO systematic analysis. Lancet Glob Health 2019: 7: e47–e57.
Child Growth Standards based on length/height, weight and 41. Centers for Disease Control and Prevention (CDC). Direct
age. Acta Paediatr Suppl 2006: 450: 76–85. and indirect effects of routine vaccination of children with 7-
27. Kramer MS, Kakuma R. Optimal duration of exclusive breast- alent pneumococcal conjugate vaccine on incidence of inva-
feeding. Cochrane Database Syst Rev 2012: (8): CD003517. sive pneumococcal disease–United States, 1998–2003.
28. Lamberti LM, Zakarija-Grkovic I, Fischer Walker CL et al. MMWR Morb Mortal Wkly Rep 2005: 54: 893–897.
Breastfeeding for reducing the risk of pneumonia morbidity 42. Haber M, Barskey A, Baughman W et al. Herd immunity
and mortality in children under two: a systematic literature and pneumococcal conjugate vaccine: a quantitative model.
review and meta-analysis. BMC Public Health 2013: 13 Vaccine 2007: 25: 5390–5398.
(Suppl 3): S18. 43. Miller E, Andrews NJ, Waight PA, Slack MP, George RC.
29. Hurley WL, Theil PK. Perspectives on immunoglobulins in Herd immunity and serotype replacement 4 years after
colostrum and milk. Nutrients 2011: 3: 442–474. seven-valent pneumococcal conjugate vaccination in England
30. Boyton RJ, Openshaw PJ. Pulmonary defences to acute res- and Wales: an observational cohort study. Lancet Infect Dis
piratory infection. Br Med Bull 2002: 61: 1–12. 2011: 11: 760–768.
31. Hanson LA, Korotkova M, Telemo E. Breast-feeding, infant 44. Jones LL, Hashim A, McKeever T, Cook DG, Britton J,
formulas, and the immune system. Ann Allergy Asthma Leonardi-Bee J. Parental and household smoking and the
Immunol 2003: 90(6 Suppl 3): 59–63. increased risk of bronchitis, bronchiolitis and other lower
32. Kazaura M. Exclusive breastfeeding practices in the Coast respiratory infections in infancy: systematic review and
region, Tanzania. Afr Health Sci 2016: 16: 44–50. meta-analysis. Respir Res 2011: 12: 5.
Corresponding Author James Samwel Ngocho, Institute of Public Health, Kilimanjaro Christian Medical University College,
Box 2240 Moshi, Tanzania. E-mail: [email protected]