Dealing with childhood pneumonia in developing countries: how can we make a difference?

Short abstract

Perspective on the paper by Hazir et al (see page 291)

The past few years have seen renewed attention focused on the persistent burden of childhood mortality globally. Of the 10.6 million deaths of children under 5 every year, the vast majority occur in a mere 42 countries of the developing world. It is also apparent that despite advances in understanding the pathophysiology and significance of the major causes of child death, most of the known killers such as diarrhoeal disorders and acute respiratory infections (ARI) still continue to take a heavy toll.¹

Most of the deaths from ARI are due to pneumonia. The annual incidence of pneumonia is estimated at 151 million new cases per year, of which 11–20 million (7–13%) cases are severe enough to require hospitalisation.² Serious neonatal infections account for 30–50% of neonatal mortality in different regions and it is difficult to disentangle sepsis and deaths from pneumonia. With the inclusion of neonatal pneumonia, recent estimates indicate that pneumonia is the single largest contributor to child mortality, accounting for almost 28–34% of all under‐5 deaths globally.³ It is also important to note that in contrast to diarrhoeal deaths where mortality rates have reduced dramatically, despite the introduction of a global programme for the control of ARI almost 15 years ago, there has been little change in overall burden of deaths from pneumonia. Figure 1 shows estimates of deaths of children under 5 from pneumonia, and although recent figures represent improvements in estimates rather than increasing trends, it is evident that the global burden of deaths from pneumonia remains unchanged. These composite figures also hide the enormous differentials that exist in ARI mortality rates between countries and between various socioeconomic groups within countries. The bulk of deaths from childhood pneumonia affect the poor who have higher exposure rates to risk factors for developing ARI such as overcrowding, poor environmental conditions, malnutrition and also limited access to curative health services.

Figure 1 Recent trends in pneumonia mortality. GBD, Global Burden of Disease; CHERG, Childhood Epidemiology Reference Group.

Such issues of access to preventive and curative services are the main driving force behind revised strategies for reducing the burden of deaths due to childhood pneumonia in developing countries. Although classic clinical detection algorithms and management strategies for ARI have largely relied on management in health facilities, there is increasing recognition of the importance of reaching the poor in community settings. These strategies involve recognising and managing pneumonia in community settings through community health workers (CHWs).⁴ Notwithstanding emerging evidence of the feasibility and effect of ambulatory management of pneumonia by CHWs, there are several issues that merit careful consideration as this strategy is scaled up. Two key questions are whether current algorithms for clinical diagnosis of pneumonia are robust, and if so, are current antibiotic regimens adequate for the treatment of pneumonia in primary care settings?

These questions have been answered to some extent by Hazir et al⁵ in the current issue of Archives. Although it is reassuring that the currently recommended dose of co‐trimoxazole was effective in most cases, it is uncertain whether the clinical criteria used by health workers correctly identified most cases of bacterial pneumonia. Recent studies by the same group in a consecutive group of children diagnosed with non‐severe and severe pneumonia in ambulatory settings showed that only 14% had radiologically confirmed pneumonia.⁶ Thus it is possible that current diagnostic criteria for non‐severe pneumonia may pick up a significant proportion of children with viral lower respiratory infections, which may not need antibiotic treatment at all. Given the lack of specificity of clinical features, it is also difficult to envisage that CHWs and other health workers will be able to readily diagnose pneumonia on clinical criteria alone without ancillary diagnostic tools or aids. Given the desire to reduce mortality from pneumonia, public health‐policy makers have accepted some overdiagnosis as inevitable in this regard. However, as such treatment regimens and antibiotic use are scaled up in population settings by CHWs, there are implications for public health policy that must be considered.

Few studies on antimicrobial treatment of pneumonia include concomitant microbiological information, and although there is substantial evidence from recent in vitro studies to indicate that resistance to commonly used antibiotics is increasing,⁷,⁸ most studies on treatment have relied on clinical failure rates to assess this aspect. Table 1 indicates treatment failure rates among children with non‐severe pneumonia treated with various regimens over the past 15 years with treatment failure rates ranging from 10% to 23%.⁹,¹⁰,¹¹,¹²,¹³,¹⁴ Although the differences in clinical response rates in some of these studies may relate to different patient populations and variable application of clinical diagnostic criteria, they may also represent genuine differences in antimicrobial resistance patterns. Notwithstanding the findings by Hazir et al,⁵ further studies on the efficacy of amoxicillin are needed in children with radiologically confirmed pneumonia.

Table 1 Treatment outcomes for non‐severe pneumonia.

Authors (location)	Years	Target population, age range	Treatment regimen	Treatment failure rate
Keeley et al9 (Zimbabwe)	1987–1988	614 children, age 3 months–12 years	Co‐trimoxazole 20/4 mg/kg/day twice daily for 5 days	2%
			Procaine penicillin 300 000 IU/kg/day I/M once daily for 5 days	1%
Straus et al10 (two sites in Pakistan)	1991–1992	595 children, age 2–59 months	Amoxicillin 45 mg/kg/day for 5 days	13%
			Co‐trimoxazole 20/4 mg/kg/day for 5 days	23%
COMET Study11 (nine sites in Pakistan)	1995–1996	1143 children, age 2–59 months	Co‐trimoxazole 20/4 mg/kg/day for 5 days	19%
			Co‐trimoxazole 40/8 mg/kg/day for 5 days	21%
CATCHUP Study Group12 (eight sites in Pakistan)	1998–1999	1471 children, age 2–59 months	Amoxicillin 50 mg/kg/day for 5 days	16%
			Co‐trimoxazole 20/4 mg/kg/day for 5 days	19%
MASCOT Pneumonia Study group13 (seven sites in Pakistan)	1999–2001	2000 children, age 2–59 months	Amoxicillin 45 mg/kg/day for 3 days	21%
			Amoxicillin 45 mg/kg/day for 5 days	20%
ISCAP Study14 (seven referral hospitals in India)	2000–2002	2188 children, age 2–59 months	Amoxicillin 31–54 mg/kg/day for 3 days	11%
			Amoxicillin 31–54 mg/kg/day for 5 days	10%

CATCHUP, Co‐triamoxazole Amoxicillin Trial in Children Under 5 years for Pneumonia; COMET, Cotrimoxazole Multicentre Efficacy; MASCOT, Multicentre Amoxycillin Short Course Therapy.

These findings underscore the continued need for preventive strategies, as well as alternative improved diagnostic and treatment regimens. Clinical approaches for the management of childhood pneumonia are considerably hampered by the lack of a gold standard, as classic microbiological methods have poor sensitivity and current algorithms lack sufficient specificity. It is therefore likely that community strategies for the recognition and management of pneumonia by ancillary health workers that rely on simple clinical criteria, other than auscultation, will over diagnose bacterial pneumonia. There are legitimate concerns that widespread use of first‐line antibiotics for all ARIs will lead to loss of effectiveness. Table 1 summarises the outcomes from recent therapeutic trials for the treatment of non‐severe pneumonia, indicating treatment failure rates exceeding 15% in many cases. It is therefore imperative that antibiotic regimens for both dosage and duration be evidence based and their use restricted as much as possible.

It is therefore important that developing countries look at a combination of strategies for reducing the burden and mortality from pneumonia. These include the important role of preventive strategies such as control of environmental factors (eg, indoor air pollution)¹⁵ dealing with prevalent micronutrient deficiencies such as zinc and vitamin A deficiencies and promotion of household behaviours such as exclusive breast feeding¹⁶,¹⁷ and hand washing.¹⁸ Many of these preventive strategies have health benefits that far exceed mere reduction in respiratory infections, such as reduction in diarrhoea burden and improvement in nutrition indices.

There has been considerable progress in vaccination strategies for the prevention of childhood infections and pneumonia. The importance of measles and pertussis vaccine in reducing child mortality is well established. Recent public health successes with large scale introduction of Haemophilus influenzae type b vaccine in developing countries¹⁹,²⁰ also indicate that this vaccine must now become part of the universal expanded programme on immunisation vaccination package. Similar preliminary success with pneumococcal conjugate vaccine in Gambia²¹ and South Africa²² also raises the hope that we may soon have vaccines against the common bacterial pathogens which cause pneumonia.