Back to contents

The management of childhood pneumonia in settings with a high burden of HIV (part 2)

Published: 16 May 2008

Optimising pneumonia case management

Given the complexity of its causes, how best should pneumonia be managed in a setting with a high burden of HIV? Rather than re-invent the wheel, the following sections borrow liberally from Pneumonia, the Forgotten Killer of Children, IMCI materials and the Pocket Book of Hospital Care for Children to describe the current case management of childhood pneumonia approach. It is beyond the scope of this HATIP to examine the entire evidence base surrounding the pneumonia case management approach (please see the resources section for links where you can download these materials). Nevertheless, HATIP would like to highlight a few key points and critical elements, and make a few comments along the way.

The basic principles of case management are simple enough:

  • teach mothers/caregivers to recognise the signs and symptoms of pneumonia
  • get the child with pneumonia to appropriate care
  • treat with the appropriate antibiotics

According to  Pneumonia, the Forgotten Killer of Children, if these steps, and key preventive interventions such as expanded vaccination programmes and cotrimoxazole prophylaxis were universally put into practice, 1.6 million children’s lives could be saved. But evidence also suggests that to work in high HIV burden settings, there will also have to be much more aggressive HIV prevention work, universal access to provider-initiated HIV testing and post-natal follow-up, universal cotrimoxazole for HIV-infected/exposed infants and all-round strengthening of maternal newborn child health services.

But of course, even the most basic activities aren’t being performed the way they should be; and, perhaps, the “simple” case management strategy is easier said than done. Take, for instance, the first step — teaching the mother/caregiver when to suspect pneumonia.

Recognising pneumonia and seeking appropriate care

“It is critical that caregivers understand the importance of this disease and the risk it poses to their children’s health.” Pneumonia, the Forgotten Killer of Children.

Caregivers need to be taught that the hallmark symptoms - cough with either difficulty breathing or rapid breathing -  should prompt them to seek out trained health personnel or a healthcare centre for diagnosis and treatment. But the report cites findings from a recent survey showing that only one of five caregivers knows these symptoms require immediate action.

But since children frequently get coughs from colds, caregivers need a clear explanation of how to recognise “difficulty breathing or rapid breathing.” (IMCI defines difficult breathing as any unusual breathing pattern; rapid breathing is defined below).

And something else that complicates matters is that a child with pneumonia will have other symptoms as well: possibly fever, chills, headaches, loss of appetite or wheezing. By the time the mother/caregiver seeks a trained health worker, there may be symptoms of severe pneumonia such as struggling to breathe, referred to as ‘lower chest wall indrawing’ because the lower part of the chest moves in or retracts noticeably while the child is trying to inhale; flaring nostrils; stridor (a harsh noise made during inhalation — as opposed to wheezing when breathing out), or grunting in infants.

In infants (under two months of age), stridor or grunting signs are danger signs. Children (especially infants) with very severe pneumonia also may have central cyanosis (grey purple skin), be unable to drink or breastfeed, vomit everything, have convulsions, hypothermia, unconsciousness, or lethargy (be abnormally sleepy or difficult to wake) or present with head nodding (a sign of severe respiratory distress).

Some of the more common early symptoms overlap with other diseases which caregivers have been taught to be more afraid of. For instance, fever and rapid breathing are also symptoms of malaria — and campaigns have had some success teaching parents/caregivers in malaria-endemic regions to give antimalarials to their children with fever.  Thus, the caregiver’s first action might be to give antimalarials and wait for a response. But the delay of appropriate treatment could make the difference between life and death.

Delays and inappropriate treatment in rural Uganda

This is just what Källender et al describe happening in rural Uganda in the recent WHO Bulletin.55 Between November 2005 and August 2007, the researchers interviewed thousands of people and reviewed the individual case histories of children who had died of pneumonia to try to understand why these children did not survive. The cause of death was assigned for 164 children, (44) 27% with pneumonia. The researchers conceded that these were verbal/social autopsies and couldn’t possibly be foolproof, but they made some interesting observations nonetheless.

First, is that most of the children were first treated at home: 52% with antimalarials and 27% with antibiotics (about a quarter got both). This was associated with a median 2-day delay in seeking out a qualified healthcare provider, most likely because they were waiting to see improvement on the medication. (More than half the time the medicines were leftovers from neighbours or family and only sometimes came from the private sector or pharmacist).

“Given the likelihood that some mothers missed the early pneumonia symptoms and only took action when the child was severely sick, a 2-day delay could be detrimental for sick children,” wrote Källender et al. Also, given the source of the drugs, it seems likely that the "dosing, duration, type, and quality of the antibiotics may have been inadequate or inappropriate, leading to resistance development and treatment failures.

Health system failures also played a part in the children’s death. Once appropriate care was sought out, many were taken to the district hospital, which referred them onto the regional hospital because they were too poorly equipped to provide critical care to children. A couple died in transit. Of the 44 children who died of pneumonia, only one had received oxygen.

One take-home message was that the malaria and pneumonia management strategies should be better integrated (rather than separate vertical interventions). The researchers note some major inequities in how the diseases are managed. For instance, even local shopkeepers and trained community workers can dispense antimalarials for presumptive treatment of fever at the community level but policies prohibit dispensing of antibiotics without a prescription.

“To improve quality of care where it occurs and to reduce inappropriate use of drugs, integrated child-health interventions in the home, community and private sector are needed. The feasibility and effect on mortality of training community health workers and drug vendors on management of pneumonia and malaria with prepacked drugs should be tested, while the quality of health-facility care needs to improve,” the researchers concluded.

Expanding the cadres who can offer appropriate care

It may not always be easy for a caregiver to rush their child to appropriate care, particularly when it is remote and they have no transport. So it is crucial to identify the people nearby within the community (including local shopkeepers and, of course, traditional healers) to whom caregivers are likely to go, and to provide those individuals with adequate training and supervision to recognise and treat non-severe pneumonia, and to know how to respond appropriately to more severe cases.

According to basic IMCI materials, lower level health workers assessing a child must first check for general danger signs (such as convulsions, lethargy, vomiting everything or inability to drink or breastfeed) that indicate that the child urgently needs medical attention. A primary health care nurse should be able to complete an assessment, if necessary provide some emergency care to stabilise a child, and provide the pre-referral treatment, such as the first dose of an appropriate antibiotic, before sending the child onto a better equipped facility for treatment.

Clearly, at the very least lay health workers must be trained to recognise danger signals and make immediate referrals — though local programmes will need to determine what other activities they can be entrusted with before referral. Nevertheless, at the very least, lives could be saved if health programmes would meet with community stakeholders to work out the logistics to make any referrals effective ones.

Assessing and classifying the child with cough and difficult breathing

After checking for danger signs, asking about cough and difficult breathing are the highest priority questions in IMCI in children over 2 months of age (neonates are handled somewhat differently). Again, this approach should be standardised across health cadres who are dispensing antimalarials so that the approach to child health is consistent across sectors.

The next question regards how long the child has been coughing (to try to flag whether the child might have TB or some other serious chronic respiratory condition) — but the IMCI handbook recommends referral for TB assessment only after they have been coughing for 21 days. This is problematic, because, again, TB can present as acute pneumonia in children with HIV. 

“Sputum sampling in under-fives is very difficult. This is why it is not part of IMCI,” Dr Shamin Qazi, of the Department of Child and Adolescent Health and Development at WHO, told HATIP.

Yet algorithms will need to be adjusted somehow to speed the diagnosis of TB in children. Otherwise, there is a chance that child with infectious TB will be sent back into the community with only cotrimoxazole or perhaps malaria treatment. (We hope to cover the topic of tuberculosis in children in a future issue).

At the lower levels of care, there are three key steps for classifying the severity of suspected pneumonia.

First, the health worker assesses rapid breathing by counting the child’s breaths per minute. If the child is between 2 and 12 months old, rapid breathing is 50 breaths or more per minute; and if the child is between 12 months and 5 years, rapid breathing is 40 breaths or more per minute.

Then health workers should look and listen for lower chest wall indrawing or stridor.  

Any child with fast breathing alone is to be classified as having pneumonia and could be empirically treated. Any child who also has either lower chest wall indrawing or stridor is classified as having severe pneumonia or very severe disease, and is supposed to be referred to a properly equipped health facility for further assessment and treatment — at least until new operational guidance (due to the results from Hazir et al’s study of home-based care of severe pneumonia in Pakistan) comes out.

In theory, at the hospital setting, the child would be further classified as having severe pneumonia  if they only have lower chest wall indrawing without any of the other danger signs (being cyanosed, unable to drink, reduced level of consciousness) that indicate severe disease.

According to Dr English however, this rarely occurs in practice. “If you didn’t have any of those [signs of very severe pneumonia] then you’re severe pneumonia, and amazingly those two things are supposed to be different. Although when you ask people, nobody’s ever considered them as different and they’ve rarely treated them differently except in clinical trials.”

Reducing the child’s suffering during initial and referral assessments

But it is impossible to assess a child who is crying or frightened. The IMCI Handbook recommends that the mother hold the child in her lap to keep her child calm, and that sleeping children should not be woken.

It is important to make the child as comfortable as possible anyway, so anyone assessing a child with suspected pneumonia should keep in mind how the child might be suffering.

“Symptoms associated with lung involvement can be very disturbing to patients. Severe air hunger or a sensation of suffocation can lead to escalating feelings of fear, anxiety, and panic. Relief of symptoms can make a great deal of difference in the quality of life,” write Dr Liz Gwyther and colleagues in the Clinical Guide to Supportive and Palliative Care for HIV/AIDS in Sub-Saharan Africa. The guide makes the following recommendations to relieve the child’s sensation of shortness of breath:

  • Positioning for comfort (extra pillows to raise the chest)
  • Assistance with walking
  • Humidified air (create steam by heating a pan of water)
  • Fanning the face
  • Fresh air
  • For cough, suggest soothing remedies such as honey and lemon, plain or with eucalyptus leaves or neem tree oil, and to loosen sputum, suggest plenty of water and other liquids.

The child should be given appropriate analgesics to manage fever, and opioids might be considered for those who are very distressed, or to alleviate the sensation of shortness of breath The Pocket Book recommends using a rapid-acting bronchodilator for the child who is wheezing, and making sure that any thick secretions in the throat are removed by gentle suction. Effective antibiotic treatment and oxygen therapy (below) are, of course, also palliative measures.

Presumptive treatment

It is not possible to distinguish between the different causes of pneumonia with any degree of certainty on the basis of clinical presentation alone. Since time is of the essence, prompt empiric treatment must be given, which is why it is so important that research be conducted to determine which infections are causing pneumonia locally, so as to refine treatment strategies over the long run.

As noted earlier in the article, treatment recommendations are in a state of flux. What follows comes from the most recent edition of the Pocket Book.

Empiric Treatment Recommendations

(adapted from the Pocket Book — please see for complete details)56

Very severe pneumonia: Ampicillin (50 mg/kg IM every 6 hours) and gentamicin (7.5 mg/kg IM once a day) for 5 days. If there’s a good response, treat in hospital or as an outpatient with oral amoxicillin (15 mg/kg three times a day) plus IM gentamicin once daily for a further 5 days.  Or use chloramphenicol (25 mg/kg IM or IV every 8 hours) until improvement, then continue orally 4 times a day for a total course of 10 days. Or use ceftriaxone (80 mg/kg IM or IV once daily).

If no improvement in 48 hours, treat as for staphylococcal pneumonia: switch to gentamicin (7.5 mg/kg IM once a day) and cloxacillin (50 mg/kg IM or IV every 6 hours), for. When the child improves, continue cloxacillin (or dicloxacillin) orally 4 times a day for a total course of 3 weeks. 

Give oxygen to all children with very severe pneumonia (see more below)

Children should be frequently monitored for signs of improvement (by nurses at least every 3 hours and twice daily by doctors). (See further diagnostic investigations if there are complications). 

HIV confirmed or suspected: Treat as above, but if no improvement within 48 hours, switch to ceftriaxone (80 mg/ kg IV once daily over 30 minutes) if available. If it is not available, use the gentamicin/cloxacillin regimen described above.

In addition, from start of treatment, give high-dose cotrimoxazole (8 mg/kg of trimethoprim and 40 mg/kg of sulfamethoxazole IV every 8 hours or orally 3 times a day) for 3 weeks. (During his talk in Cape Town, Dr Jeena pointed out that the amount of fluid that intravenous cotrimoxazole is given in has a significant effect on outcomes: “You actually require large volumes of fluid and if you don’t include it in your total fluid volume, you end up with pulmonary œdema.” Most countries in the region do not have intravenous cotrimoxazole and use oral suspension or portions of tablets.

Severe pneumonia: benzylpenicillin alone (50,000 units/kg IM or IV every 6 hours) for at least 3 days. Once the child oral amoxicillin (25 mg/kg twice a day for two more days. If no improvement, treatment to be switched to chloramphenicol (25 mg/kg q8 hrs IM or IV) until improvement, after which point the drug should be continued orally for 10 days. If there was still no improvement, treat as very severe pneumonia.

For severe pneumonia in HIV-infected or exposed children, treatment is the same as described above HIV in for very severe pneumonia. See attached document.

Non-severe pneumonia: on an outpatient basis: cotrimoxazole (4 mg/kg trimethoprim/20 mg/kg sulfamethoxazole twice a day) for 5 days or amoxicillin (25 mg/kg twice a day) for 5 days if in an HIV setting. However, amoxicillin is preferred if the child is already on cotrimoxazole prophylaxis or lives in a setting where cotrimoxazole is commonly used (in other words, where bacteria may be less sensitive to the drug).

Treatment duration in children with HIV: There are few  data to indicate what the optimal duration of antibiotics should be in children with HIV. Also, should children who are known to be HIV-infected or exposed perhaps be offered broad spectrum antibiotics from the start (including treatment for staphylococcal pneumonia) rather than waiting for poor outcomes?

In the DPPS study, the odds ratio of a poor outcome in a child with HIV was 10.3. In this high HIV incidence study (65% of the participants), 42% below the age of 12 months failed standard therapy. In the APPIS study, at 14 days, 41% of the HIV infected infants with severe pneumonia failed treatment or had passed away.57

Further diagnostic investigations

Diagnostic evaluations take the backseat in pneumonia case management (and will continue to do so until new real time diagnostic tests become available that are practical in resource limited settings).

“It's often difficult to pin-point the cause of pneumonia in an operational setting,” Dr Doug Wilson, Head of Medicine at Edendale Hospital in KwaZulu Natal told HATIP. “Both sputum – which is hard to get out of kids anyway - and blood cultures have low yield — a nightmare if you're worried about drug resistance.”

But there are situations where further diagnostic investigations could prove useful, such as in children who are not responding to treatment, to diagnose uncommon HIV-related causes of pneumonia, or PCP in settings where there is not such a high prevalence of HIV. 

Although the treatment algorithm included recommends treatment for staphylococcal pneumonia if the standard regimen fails, suggestive investigations include chest x-rays (with pneumatocoele or pneumothorax with effusions), Gram stained sputum smears (revealing numerous Gram-positive cocci), or heavy growth of S. aureus in cultured sputum or empyema fluid. The child may also have septic skin pustules.58

As already noted, bacteraemia is common, so blood cultures may be useful. In practice, though, health workers rarely send for cultures of bacterial infections, which is part of the reason why there isn’t a good sense for which infections are actually causing pneumonia.59

If a pyogenic (pus-laden) effusion forms in the pleural cavity, it should be drained and then analysed for infectious organisms (for protein and glucose content, cell count and differential count, Gram and Ziehl-Neelsen staining, and bacterial and Mycobacterium tuberculosis culture).60

The need to diagnose PCP in infant may sound questionable in a setting with a high prevalence of HIV. But having a high index of suspicion or a confirmed diagnosis could help guide choices regarding alternative treatment regimens and supportive care, such as corticosteroids.

The Pocket Book includes WHO guidelines for the diagnosis of PCP, which state that: “PCP should be presumptively diagnosed in any child who has severe or very severe pneumonia and bilateral interstitial infiltrates on chest X-ray. Consider the possibility of pneumocystis pneumonia in children, known or suspected to have HIV, whose ordinary pneumonia does not respond to treatment. Pneumocystis pneumonia occurs most frequently in infants and is often associated with hypoxia. Fast breathing is the most common presenting sign, respiratory distress is out of proportion with chest findings; fever is often mild.  Peak age is 4–6 months.”

Wijesingh and Graham reviewed the evidence base for the guidance in detail in The Journal of Tropical Pediatrics in 2007.61

Dr Jeena also presented some recent analysis from the DPPS (see table), stressing “these children don’t have chest signs - upon auscultation [listening with a stethoscope], there’s nothing to hear. Multivariate analysis basically showed the higher the HIV viral load, the greater the risk for PCP; they were all under 6 months, had clinical cyanosis; and again no chest signs.”

He added that they do not have much fever, although they show all the classical signs of respiratory distress, tend to be more hypoxaemic than children with bacterial infection, and often show signs of HIV/AIDS.

Significant findings in PCP cases (n=33) DPPS


Odds Ratio (95% Confidence Interval

P value

No chest signs on auscultation

10.06 (4.64-21.7)


Danger Signs

6.09 (1.5-53.7)


Under 6 months



Clinically cyanosed

5.94 (2-17)


Z score <3 sd weight age

0.233 (0.54-1)


He said that chest x-rays are actually variable. “It depends on where you pick up the disease, whether you pick up the disease earlier on — and that’s when you get your interstitial patchy changes with hyperinflation [ground glass], or whether you pick the disease up later, when you’ve got AIDS where they’ve got a full spectrum of alveolar consolidation,” he said.

A good quality sputum specimen can be used to diagnose PCP, however, certain techniques such as lung aspiration and bronchoalveolar lavage may be too invasive and could exacerbate respiratory problems in infants.

CMV also presents challenges for diagnosis in most resource-limited settings. Clinical presentation is very similar to PCP, with radiological evidence of interstitial pneumonitis.

“CMV serology is not sensitive,” said Dr Jeena. “Often in our situation, we rely on the CMV DNA PCR to actually identify — and in a study we published, we showed a sensitivity of 90 to 100% and a positive predictive value for cytopathic disease of 80%.”

Cotrimoxazole resistance

Another matter is how to manage children failing standard PCP treatment.

“PCP has poor outcomes even in developed countries and even with appropriate therapy including ICU,” said Dr. Zar, with a mortality of 47% (PCP) vs 18% (non-PCP pneumonia).62  

Dr Jeena stressed: “Even in Great Ormond Street [a centre of excellence in the UK], once upon a time, they were doing poorly with PCP (with a 2 year survival of only 12.5%). Now [due to ART] they don’t see PCP, that’s the good news.” 

The poor outcomes despite cotrimoxazole raise several issues. For instance, how much is due to the emergence of resistant P jirovecii strains? A study by Dr Zar and colleagues found resistance in 13% of strains in South Africa,63 though she noted that many have questioned the impact of resistant strains on outcomes. “Poor outcomes are more likely to be due to coinfections than to resistance,” Dr Zar told HATIP.

However Dr Jeena cited several studies where the presence of a mutation conferring resistance to cotrimoxazole was associated with poorer outcomes.64, 65, 66

“Now in many parts of Africa you won’t be able to diagnose resistance,” he said. “So the current recommendation is that we probably should be looking at some clinical marker: if by seven days a child does not respond to cotrimoxazole, we should be thinking maybe this is resistance and then considering a change of therapy. There aren’t any studies to actually prove this yet but we really need to do some work around this thing because there’s no doubt — even with appropriate therapy for PCP — these children still don’t do very well.”

He listed a number of possible alternatives including clindamycin and primaquine (the most effective) or trimethoprim with dapsone.


However, the fact that the highest mortality in children with PCP is during the first few days on therapy, suggests that some other mechanism may be to blame for many of the failures. Specifically, failing to manage the inflammation associated with the condition.

But there are mixed opinions about the use of corticosteroids which may reduce inflammation, but at the same time increase susceptibility to many common pathogens.

Corticosteroids have long been given to people with PCP and are associated with a reduction in mortality (in patients with A/a gradient > 35 mmHg or room air PaO2<70 mmHg); and a recent Cochrane Review concluded that even though the number and size of trials was small, the evidence suggested a beneficial effect in patients who were hypoxaemic.67

Dr Jeena points out that most of those data come from industrialised countries however. “The developing countries are very different from the developed countries in that the microbiological load of opportunistic infections is much higher in developing countries. We need to have a guarded approach in this regard. If you’ve got a very high degree of suspicion, or you’ve got criteria which indicate that you require this type of therapy, only then should we be using corticosteroids otherwise you may put your patient at risk.”

At the same time, clinicians must decide quickly. “In adult data, corticosteroids were only effective if used in the first 48 hours,” said Dr Zar. “So if there is a high suspicion of PCP and moderate or severe illness, corticosteroids should be used early.”

Prevention: strengthening MNCH services the best policy?

“Children are dying of diseases that are preventable and treatable with tools that are simple i.e. IMCI; prevention of HIV infection of young women; prevention of transmission of HIV from mother-to-child; and cotrimoxazole prophylaxis.” Saving Children: 200568 

Ultimately, the best way to reduce HIV’s impact on pneumonia is to control it through preventive interventions starting as far upstream as possible, by scaling up family-based prevention and provider-initiated testing and counselling in general and ANC HIV testing and prevention counselling in particular. But other evidence-based preventive activities must be scaled up as well. General MNCH strengthening is essential: stakeholders, policy makers, funders and researchers should consider the possibility that, with these preventive interventions in place, providing the best care possible for all children may lead to the best possible outcomes in children with HIV and pneumonia.

Antenatal HIV testing, and HIV prevention or treatment for the mother

Having a child die of HIV-related pneumonia is the last way that one should have to find out about one’s own HIV status. Provider-initiated testing and counselling and other HIV testing outreach services must be universally available to reach all pregnant women.

Women should receive targeted care and support with ART, when indicated for their own health, if positive, and if HIV-negative, targeting counselling services to prevent HIV transmission to women during the ante and post natal periods (when women are at an especially high risk for HIV acquisition). For instance, data from last year’s International AIDS Society Meeting in Sydney showed an 8% incidence of recent seroconversion (within the previous six months) among pregnant women attending clinics in Botswana.69 Follow-up testing of mothers at vaccination visits in South Africa also suggest this.70 Engaging her male partner(s) in testing and counselling may be important for the success of prevention efforts.


HIV-positive women should receive the best available regimen to prevent mother to child transmission, preferably ART, and preferably for as long as the mother breastfeeds. The continuing low uptake of PMTCT simply isn’t acceptable and ultimately impacts on the quality of care that every child receives.

Accurate child health cards

Documentation of maternal HIV status (with indicators such as HIV-positive, last test date negative, refused test) or the child’s HIV status on the caregiver-held child health cards may be essential for the well being of all children in HIV-endemic settings. If included in the initial assessment, having a child health card available documenting current HIV exposure status (treating untested as HIV-exposed) could allow the implementation of more aggressive community case management of pneumonia  to be attempted in high HIV burden settings. However, in order to not disclose a child or mother’s HIV status, community care workers would have to be trained to respect HIV confidentiality.

Another alternative would be training HIV community care workers/home based care teams in the implementation of the spectrum of community-based MNCH interventions (malaria, TB screening, etc) to offer families with HIV an alternate entry point to care.

Universal access to cotrimoxazole prophylaxis

WHO and UNICEF recommend cotrimoxazole prophylaxis for all HIV-positive children, as well as for infants born to HIV-infected mothers, in order to prevent pneumonia (and other infections) and UNICEF has made it a target to provide cotrimoxazole (and/or antiretroviral treatment) by 2010 to 80 per cent of children in need. And yet, in the most recent update on the campaign, Children and AIDS Second Stocktaking Report, which came out last month, only a small number of countries are reporting on rolling out cotrimoxazole.71

“Out of an estimated 4 million children in need of cotrimoxazole prophylaxis (HIV-exposed and HIV-infected), only 4% are currently receiving this intervention,” wrote Zachariah et al in a review of the major barriers preventing the scale-up of cotrimoxazole prophylaxis.72 Of course, one of the chief reasons for this is the lack of coverage of PMTCT programmes and their appalling lack of follow-up.

Among the numerous specific actions proposed to tackle these challenges, the authors suggest that cotrimoxazole be made as an essential component of routine maternal and child-health services at all levels of the health system. One option, “the universal option” is cotrimoxazole prophylaxis for all infants and children born to mothers confirmed or suspected of living with HIV. This strategy may only be considered in settings with a high prevalence of HIV, high infant mortality caused by infectious diseases, or limited health infrastructure.

If suspected includes a mother who has refused the offer of HIV testing, this option may indeed be universal enough.

Given that there is usually good participation in at least the first vaccination visit around 6 weeks, MNCH programmes in high HIV burden countries could offer mothers another chance to test (including those who were negative at or prior to delivery), and provide cotrimoxazole for the child to all of those without an HIV-negative test result. The timing is crucial, because although cotrimoxazole reduces the risk of malaria and a variety of bacterial infection in children, and the risk of PCP is in infants,  the peak incidence of PCP is between 2 and 6 months, so the window of opportunity is narrow.

Universal access to ART for HIV-positive children

Ultimately the burden of pneumonia in infants with HIV could be relieved by treating all children with antiretrovirals, but this will require a scale-up of infant HIV testing. Infant HIV testing was discussed in detail in edition 100 of HATIP.

A table presented by Dr Zar illustrates how ART can dramatically reduce the incidence of respiratory infections in children (although it doesn’t account for other changes in care such as prophylaxis and the conjugate pneumococcal vaccine). 

Incidence of respiratory infections in children on HAART vs. pre-HAART

OI category




IR per 100 Child yrs


IR per 100

Child yrs


Bacterial pneumonia










Dissem. Myco av./ MOTT











Universal access to immunisation

“Vaccines against the two leading bacterial causes of child pneumonia deaths, Haemophilus influenzae type b (Hib) and Streptococcus pneumoniae (pneumococcus), can further improve child survival by preventing about 1,075,000 child deaths per year,” Madhi et al wrote in a review of the vaccines in the recent WHO Bulletin.75

Although the vaccines are less effective in children with HIV not on ART (65% and 54%, respectively), compared with children without HIV (83% and 90%, respectively), “because of a 20–40 times increased risk of illness from these bacteria, HIV-infected children still derive a significant protective effect and the absolute burden of invasive disease and pneumonia prevented by the vaccines exceeds that of HIV non-infected children”

Newer conjugate vaccines may offer even broader protection — though they will be more expensive.

HIV funding partners might want to consider helping out, because making certain that all children are vaccinated is in the best interest of children (and adults) with HIV since the reduction in infectious bacterial illness could improve herd immunity, as discussed previously.

Addressing other risk factors of pneumonia

A number of other factors also increase the risk of pneumonia in children.

“Environmental factors, such as living in crowded homes and exposure to parental smoking or indoor air pollution, may also have a role to play in increasing children’s susceptibility to pneumonia and its severe consequences,” according to the Pneumonia, Forgotten Killer of Children report. In fact, a recent randomised controlled trial in Guatemala reported that decreasing indoor air pollution by installing stoves with a chimney achieved a marked reduction in cases of severe pneumonia.76

Lack of exclusive breastfeeding and poor nutrition are risk factors for pneumonia, and there is evidence that providing zinc in settings where zinc deficiency is common may reduce the risk of pneumonia.77

Handwashing and hygiene in the home may be particularly important where people are living with HIV.78

Infection control continues to be neglected in healthcare facility settings, leading to the spread of pneumonia and potentially polymicrobial mixes in hospitalised children with HIV. (An upcoming HATIP will investigate how hospitals in resource-limited settings are unsafe.)

Advocating for MNCH strengthening

In high burden settings, HIV impacts on all children with pneumonia. HIV programmes and donor partners have to step up to the plate, advocate and work to strengthen MNCH services, to make certain that effective broad-spectrum community based support mechanisms are in place to respond quickly to the various causes of child illness, and that there is improved surveillance of the causes of pneumonia/drug susceptibility.

For instance, children with severe pneumonia should not be turned away from a district hospital because it is poorly equipped. And in many settings, few facilities have adequate supplies of oxygen or pulse oximeters — a device that should help healthcare workers recognise hypoxaemia, improve survival and make oxygen delivery more efficient. According to the GAPP report, WHO has been promoting the availability of oxygen and pulse oximeters to smaller hospitals in developing countries. 

In the recent WHO Bulletin, Enarson et al report on The Child Lung Health Programme (CLHP), an initiative scaling up oxygen delivery throughout paediatric wards in Malawi. The Government of Malawi, the International Union Against Tuberculosis and Lung Disease and the Bill and Melinda Gates Foundation deserve credit for demonstrating that this is feasible in a low income setting.

There should be closer collaboration between HIV/PMTCT programmes and MNCH programmes to identify common needs. For instance, in South Africa at least, there is no such thing as post-natal care between the birth and vaccination visit.

Perhaps HIV-related community-based mechanisms for follow-up visits and maternal support (groups like Mothers2Mothers) or home-based care groups could be utilised and expanded to bridge this gap by making a least one follow-up visit at 3 or 4 weeks after birth for families.

Services could be offered such as home-based family rapid HIV testing, infant feeding counselling and support, cotrimoxazole if appropriate, scheduling vaccination visits, counselling about a safe home environment (indoor pollution, hand washing, basic hygiene), as well as reinforcing messages about what illnesses should prompt caregivers to seek out trained healthcare personnel and other essential MNCH services. It’s a tall order perhaps, but it could be a part of what HIV programmes should contribute to, and a way to become better integrated with, improved MNCH services.

Advocates of children with HIV need to be advocates for MNCH services, because without a continuum of care, many children with HIV and pneumonia won’t have much of a chance.

Resource list

Childhood pneumonia resources

The Child Healthcare Problem Identification Programme in South Africa published the following reports:

Palliative care resources



[55] Källander K et al. Delayed care seeking for fatal pneumonia in children aged under five years in Uganda: a case-series study. Bulletin of the World Health Organization 86:332–338, 2008.

[56] Pocket book of hospital care for children: guidelines for the management of common illnesses with limited resources. Geneva: WHO; 2005.

[57] Jeena P, Thea DM, Macleod MB, the APPIS Group. Failure of standard antimicrobial therapy in children aged 3-59 months with mild or asymptomatic HIV infection and severe pneumonia. Bull World Health Organ 2006;84:269-75.

[58] Pocket Book, op. cit.

[59] Qazi S, op. cit.

[60] Pocket Book, op. cit.

[61] Wijesingh S and Graham S. Clinical Review Evidence behind the WHO Guidelines: Hospital Care for Children What are the Clinical Indicators of PCP? Journal of Tropical Pediatrics 53(1): 1-7, 2007.

[62] Zar H. PIDJ, op cit.

[63] Zar H et al. Prevalence of dihydropteroate synthase mutants in HIV-infected South African children with Pneumocystis jiroveci pneumonia. Clin Infect Dis. 2004 Oct 1;39(7):1047-51.

[64] Helweg-Larsen J et al. Effects of mutations in Pneumocystis carinii dihydropteroate synthase gene on outcome of AIDS-associated P. carinii pneumonia. Lancet. 354(9187):1347-51, 1999.

[65] Kazanjian P et al. Pneumocystis carinii mutations associated with sulfa and sulfone prophylaxis failures in AIDS patients. AIDS. 12(8):873-8, 1998.

[66] Ma L et al. Mutations in the dihydropteroate synthase gene of human-derived Pneumocystis carinii isolates from Italy are infrequent but correlate with prior sulfa prophylaxis. J Infect Dis. 185(10):1530-2, 2002.

[67] Briel M et al. Adjunctive corticosteroids for Pneumocystis jiroveci pneumonia in patients with HIV-infection. Cochrane Database Syst Rev. Issue 3:CD006150, 2006.

[68] Patrick ME and Stephen CR, eds. Saving Children : 2005 A survey of child healthcare in South Africa.

[69] Moyo S et al. Estimation of HIV incidence in among pregnant women attending antenatal clinics In Botswana in 2005 using serological test for recent seroconversion.

[70] Rollins H et al. HIV prevalence rates amongst 6 week old infants in South Africa: the case for universal screening at immunisation clinics. Sixteenth International AIDS Conference, Toronto, abstract THAC0104, 2006. See aidsmap article:

[71] UNICEF. Children and AIDS. Second Stocktaking Report, 2008.

[72] Zachariah R et al. Scaling-up co-trimoxazole prophylaxis in HIV-exposed and HIV-infected children in high HIV-prevalence countries. Lancet Infectious Diseases 7: 686–93, 2007.

[73] Gona P et al. Incidence of opportunistic and other infections in HIV-Infected children in the HAART era. JAMA Vol. 296 No. 3, 2006.

[74] Dankner WM et al. CD4 correlates of opportunistic infections in children infected with the human immunodeficiency virus managed before highly active antiretroviral therapy. Pediatric Infectious Disease Journal. 20(1):40-48, 2001.

[75] Madhi SA, Bulletin of the World Health Organization, op. cit.

[76] Qazi S, op. cit.

[77] Black, R.E. Zinc deficiency, infectious disease and mortality in the developing world. Journal of Nutrition, vol. 133, 2003.

[78] Luby, S.P., et al. Effect of handwashing on child health: a randomised controlled trial. The Lancet, vol. 366, pp. 225-233, 2005.

This content was checked for accuracy at the time it was written. It may have been superseded by more recent developments. NAM recommends checking whether this is the most current information when making decisions that may affect your health.