Abstract
Aims
We compared the radiological features and outcome of WHO defined severe pneumonia among HIV infected and exposed uninfected children randomised to receive penicillin or oral amoxicillin in Durban, South Africa.
Methods
Of 425 children aged between 3 and 59 months with WHO defined severe pneumonia, 366 had anonymous HIV testing performed. Outcome was assessed by failure to improve at 48 h after enrolment or deterioration within 14 days. Chest radiographs were evaluated according to WHO defined radiological criteria for pneumonia and internationally standardised radiological criteria. Findings were stratified for HIV status.
Results
82 (22.4%) children were HIV infected, 40 (10.9%) were HIV exposed and 244 (66.7%) were HIV uninfected. The day 14 outcome in children <12 months of age was significantly worse in HIV‐1 infected than HIV uninfected children (OR 2.8 (95% CI 1.35 to 3.5), p = 0.002), while HIV‐1 infected and uninfected children aged ⩾12 months had equivalent outcomes. Parental penicillin and oral amoxicillin had equivalent response rates in all HIV groups. According to the WHO radiological classification, children who failed WHO standard antimicrobial treatment had significantly higher “other consolidates/infiltrates” than “endpoints for consolidation” in the HIV infected group (OR 5.45 (95% CI 1.58 to 21.38), p<0.002), while the reverse was true for HIV exposed uninfected children (OR 4.13 (95% CI 0.88 to 20.57), p<0.036).
Conclusions
The WHO standard treatment guideline for severe pneumonia is inadequate for HIV‐1 infected infants. The increased prevalence of “other consolidates/infiltrates” among HIV‐1 infected children who failed standard treatment supports the addition of co‐trimoxazole to WHO standard treatment.
The global incidence of acute lower respiratory infections (ALRI) is 154 million new episodes per annum with 7–13% of patients requiring hospitalisation.1 ALRI causes 1.9 million deaths among children annually, accounting for over a fifth of all deaths in Africa.2 The World Health Organization (WHO) response to this burden of disease has been the introduction of standard case management guidelines. The benefit of this intervention in HIV non‐endemic areas was recently shown in a meta‐analysis where pneumonia related mortality among neonates, infants and children between 0 and 4 years of age was reduced by 27%, 27% and 20%, respectively.3 The HIV‐1 epidemic has altered the prevalence, presentation and outcome of ALRI.
The role of chest radiographs in the management of children with ALRI has had mixed reviews. Some studies suggest that radiographs help confirm the diagnosis of pneumonia in only 36% of episodes, help change the diagnosis in approximately 20% of cases and result in a change in treatment plan in 34%.4,5 The chest radiographic features of HIV associated pneumonia in childhood have scarcely been reported. Sivit et al described the radiological features of Pneumocystis jiroveci (carinii) pneumonia in infancy (PCP) and found that a ground glass interstitial appearance with pulmonary air cysts and thoracic air leak syndromes were common.6 The WHO has defined radiological criteria to assist in the determination of bacterial pneumonia in their vaccine trials on streptococcal pneumonia.7 While several studies have described the aetiology of HIV related pneumonia in children, none have evaluated the differences in radiology and response rates to standard WHO therapy.
We therefore performed a prospective nested substudy of a larger international pneumonia study8 to define the radiological features on admission and clinical response at day 14 of HIV‐1 infected and exposed children with WHO defined severe pneumonia treated with either oral amoxicillin or injectable penicillin.
Methods
Children between 3 and 59 months of age with WHO defined severe pneumonia, enrolled at one of the nine study sites, King Edward VIII Hospital, Durban, South Africa, were entered into this study. The criteria for WHO defined severe pneumonia included presence of cough, tachypnea and chest indrawing and standardised treatment includes parental penicillin for 48 h followed by oral amoxicillin or oral amoxicillin alone. Treatment failure was defined as failure to improve on prescribed therapy by 48 h or deterioration in respiratory status as evidenced by increasing respiratory rates and chest indrawing, increasing oxygen requirements and onset of danger signs at any point during the 14‐day study period.
Children were included in the study if they had fast breathing, cough, fever and chest indrawing (WHO defined severe pneumonia), they did not respond to two doses of salbutamol if wheeze was present, and parents had provided written informed consented. Children with danger signs of very severe pneumonia, severe malnutrition (weight for age z scores less than −3), chronic cardiopulmonary conditions, asthma or measles, or penicillin allergies were excluded at screening. All children had a posterior/anterior chest radiograph performed on admission.
The primary endpoint of the study was to determine the impact of HIV disease on the outcome of WHO defined severe pneumonia and the usefulness of chest radiography in determining the risk of treatment failure.
A second consent for anonymous linked HIV‐1 testing was obtained with HIV pretest counselling and a clear explanation of the following protocol. Samples were taken, labelled with the study number and stored for later analysis. These samples were tested at the end of patient recruitment and just before data analysis. An HIV‐1 ELISA antibody test (Determine, Abbott Laboratories, Johannesburg, South Africa) was performed. Samples from children less than 15 months old with a positive HIV‐1 ELISA test were re‐tested with an HIV‐1 DNA PCR test (Amplicor, Roche Diagnostics, Basel, Switzerland), while children over 15 months old had a repeat HIV‐1 ELISA test. Children with a positive HIV DNA PCR result or those with two positive HIV‐1 ELISA tests (if >15 months old) were regarded as HIV‐1 infected, while children under 15 months of age who were HIV‐1 ELISA positive but PCR negative were regarded as HIV exposed. Children who were HIV ELISA negative were regarded as HIV uninfected. Patients who refused this test were still included in the study but were regarded as HIV untested.
The study was approved by the ethics committee of the University of KwaZulu‐Natal and monitored by an independent data and safety monitoring board (DSMB). Initially all children with WHO defined severe pneumonia regardless of their HIV status were enrolled with consent. During the second interim analysis, the DSMB recommended expansion of the exclusion criteria to omit children with known or clinically suspected HIV disease based on the evidence that PCP had become increasingly recognised as an important pathogen in HIV seropositive children and was not covered by the study treatments. This change was accepted and implemented. Anonymous HIV‐1 screening continued until the end of the study.
The radiographic findings were interpreted according to predefined criteria by independent radiologists. Radiological features and the site of involvement were evaluated according to the proposed WHO radiological classification for pneumonia.9,10
A WHO defined endpoint for consolidation was a dense opacity that may be fluffy consolidation (dense, often homogenous (confluent) opacity) of a portion or whole of a lobe, often containing air bronchograms and sometimes associated with a pleural effusion (dense opacification within the pleural space). Atelectasis, a dense opacity and a positive silhouette sign with the mediastinal border, was considered to be an endpoint for consolidation.
Other consolidation/infiltrates were defined as linear and patchy in‐homogenous airspace densities in a lacy pattern involving both lung fields and featuring peribronchial thickening, perivascular cuffing, nodular or recticular nodular changes and multiple small areas of atelectasis and hyperinflation. These included interstitial infiltrates and bronchopneumonia with perihilar/paratracheal lymphadenopathy.
A chest radiograph was considered to be normal when no abnormal opacities were seen.
The admission radiographic findings of HIV infected exposed and uninfected children were compared. The radiological findings of children who failed to clinically respond were evaluated according to HIV status.
Statistical analysis
All data were analysed with SPSS 11.5 for Windows (SPSS, Chicago, IL). Analyses were performed using the Student t test for continuous variables and the χ2 and Fisher exact tests for categorical variables. Any two‐tailed p value of p<0.05 was considered statistically significant.
Results
A total of 425 children with WHO defined severe pneumonia were entered into the study. Of these, HIV‐1 status was determined in 366 (86.1%). Fifty nine children (13.9%) did not have their status determined and were classified as HIV untested. The outcome and radiological features in these children did not bias the data and consequently these children were omitted from further analysis. Eighty two of the 366 children tested (22.4%) were HIV infected, 40 (10.9%) were HIV exposed uninfected and 244 (66.7%) were HIV uninfected. The mean age of these subjects was 17 months (range 3–58 months) and the male to female ratio was 1.24:1 (table 1).
Table 1 Baseline characteristics of the children.
| Characteristics | HIV infected, n = 82 (%) | HIV exposed uninfected, n = 40 (%) | HIV uninfected, n = 244 (%) | OR (95% CI) HIV infected vs uninfected |
|---|---|---|---|---|
| Male | 48 (58.5) | 24 (60) | 133 (54.5) | 1.07 (0.87 to 1.33) |
| <12 months of age | 27 (32.9) | 37 (92.5) | 88 (36.1) | 0.91 (0.64 to 1.3) |
| Malnourished | 17 (20.7) | 4 (10) | 30 (12.3) | 1.69 (0.98 to 2.89) |
| Respiratory rate >60 bpm | 36 (43.9) | 9 (22.5)* | 84 (34.4)* | 1.28 (0.94 to 1.72) |
| Saturation <90% | 5 (6.1) | 5 (12.5)† | 17 (7)† | 0.88 (0.33 to 2.3) |
| Amoxicillin treatment group | 32 (39) | 20 (50) | 130 (53.3) | 0.73 (0.55 to 0.98) |
*OR 0.65 (95% CI 0.36 to 1.19); †OR 1.79 (95% CI 0.7 to 4.59).
Statistical analysis was not performed on the age of HIV exposed versus HIV infected or uninfected children because of skewed sample by definition.
Treatment failure rates on day 14
The overall treatment failure rate for the 425 children enrolled on an intent‐to‐treat analysis was 16.7% (n = 72) and this was equally distributed in the oral amoxycillin and parental penicillin arms (16.7 vs 17.1%, respectively). Among the children where HIV status was determined (n = 366), the treatment failure rate was 17.7% (n = 65), being equally distributed in both treatment arms for all HIV groups. Treatment failure rates among the HIV infected, HIV exposed and HIV uninfected children were 24.4%, 22.5% and 14.8%, respectively (table 2). A similar trend was seen for deaths. Age less than 1 year was a predictor of worse outcome in the HIV infected and uninfected groups (p = 0.027 and p = 0.04, respectively). Higher treatment failure rates were recorded for the HIV‐1 infected than the HIV uninfected infants. In children older than 12 months of age, similar rates of treatment failure were recorded for HIV‐1 infected and uninfected children.
Table 2 Treatment outcome at 14 days by age and HIV status.
| Characteristics | HIV‐1 infected, n = 82 (%) | HIV‐1 exposed, n = 40 (%) | HIV uninfected, n = 244 (%) | OR (95% CI) and p values for the comparison of outcomes HIV infected vs HIV uninfected |
|---|---|---|---|---|
| Overall treatment failures (n = 65) | 20 (24.4) | 9 (22.5) | 36 (14.8) | 1.53 (0.92 to 2.55), p = 0.02 |
| Treatment failures | ||||
| Age 2–11 months | 11/27 (40.7) | 9/37 (24.3)* | 17/88 (19.3) | 2.8 (1.35 to 3.5), p = 0.002 |
| Age 12–59 months | 9/55 (16.4) | 0/3 (0) | 19/156 (12.2) | 1.4 (0.88 to 2.8), p = 0.59 |
| Deaths* (n = 5) | 3 (3.6) | 1 (2.5) | 1 (0.4) | – |
*Deaths also reflect non‐responders and no statistical analyses were performed because of small sample size.
Radiographic characteristics on admission
Overall, 39.6% of children with severe pneumonia had a WHO defined radiographic endpoint equivalent of consolidation and 59.6% had other infiltrates. Three children had normal chest radiographs (table 3). “Endpoint for consolidation” was recognised in a significantly higher proportion of HIV infected than HIV uninfected children, while “other infiltrates” (patchy consolidation) was more common in HIV uninfected than HIV‐1 infected children. Among the HIV infected and HIV exposed group, segmental parenchymal lung involvement was more commonly seen in the left lower lobe (51.1% and 50.0%, respectively), while HIV uninfected children frequently had bilateral and paracardiac/perihilar patchy changes (29.2% and 26.5%, respectively). Lobar changes were rarely seen in any of three groups. A higher proportion of HIV infected children with “other infiltrates” than “endpoint for consolidation” failed to respond to either oral amoxicillin or injectable penicillin, while a higher proportion of HIV exposed uninfected children with “endpoint for consolidation” failed therapy. Fifty eight (15.8%) children with WHO clinically defined severe pneumonia had chest radiographic features suggestive of bronchiolitis (hyperinflation with air trapping).
Table 3 Description and comparison of radiological endpoints with HIV status among children with WHO defined severe pneumonia that responded or failed to respond to standard WHO recommended therapy.
| HIV status | WHO defined radiological features of pneumonia | Statistics | |||||
|---|---|---|---|---|---|---|---|
| Endpoint for consolidation | Other infiltrates | Normal chest radiograph | |||||
| Overall prevalence, n (%) | Prevalence in non‐ responders, n (%) | Overall prevalence, n (%) | Prevalence in non‐ responders, n (%) | Overall prevalence, n (%) | Prevalence in non‐ responders, n (%) | Comparison of “endpoint for consolidation” with“other infiltrates” in non‐responders, OR (95% CI) and p values | |
| HIV infected (n = 82) | 45 (54.9) | 5 (11.1) | 37 (45.1) | 15 (40.5) | 0 (–) | 0 | 5.45 (1.58 to 21.38), p = 0.002 |
| HIV exposed (n = 40) | 16 (40) | 5 (31.3) | 23 (57.5) | 3 (13.0) | 1 (2.5) | 1 (100) | 4.13 (0.88 to 20.57), p = 0.036 |
| HIV uninfected (n = 244) | 84 (34.6)* | 20 (13.7) | 158 (64.8)* | 25 (15.8) | 2 (0.8) | 1 (66.7) | 0.82 (0.52 to 12.1), p = 0.72 |
| Overall radiological | 145 (39.6) | 20 (13.7) | 218 (59.6) | 43 (19.7) | 3 (0.008) | 2 (66.7) | 0.65 (0.52 to 21.38), p = 0.62 |
| features | |||||||
Discussion
The main finding of this study was the three‐fold poorer response of HIV infected infants with WHO defined severe pneumonia to recommended standard antibiotic therapy than HIV exposed and uninfected infants. Overall, infants with WHO defined severe pneumonia had a significantly poorer outcome than children aged 12 months and over, regardless of HIV status. These findings have practical relevance in resource‐poor countries and suggest the need to revise WHO standard acute respiratory infection (ARI) case management guidelines in HIV endemic areas. We did not investigate the aetiological agents in children with treatment failure, but recent ante‐mortem and post‐mortem studies of HIV infected infants have confirmed an increased burden of P jiroveci and other Gram‐negative bacterial pathogens at this age.11 In HIV endemic areas, the routine administration of high dose co‐trimoxazole and broad spectrum antibiotics for the treatment of severe pneumonia in infancy should be considered. Further studies to confirm the benefits of these interventions prior to revising WHO ARI case management guidelines are needed. On the other hand, children over the age of 12 months respond well to standard antimicrobial therapy regardless of their HIV status.
The main radiological finding of this study was the higher treatment failure rates in HIV infected children with “other consolidation/infiltrates” as compared to “endpoint for consolidation”, even though the latter was more common in the HIV infected group. “Other consolidation/infiltrates” may represent viral pathogens or PCP, which are unlikely to respond to standard WHO recommended antimicrobials, supporting the need for co‐trimoxazole. Although 60% of children had WHO defined “other consolidation/infiltrates” overall, HIV infected children had more “endpoints for consolidation” than “other consolidation/infiltrates”. Children with “endpoints for consolidation” would be expected to respond to WHO standardised antimicrobial therapy as the WHO radiographic classification was based on an attempt to identify bacterial pneumonia as “endpoint consolidation” in the vaccine trials.7 The predictive value of radiological appearance in determining the aetiology of pneumonia in HIV infected children is further confounded by the frequent occurrence of mixed infections with bacteria and viruses. Studies have confirmed the role of purulent Gram‐positive and Gram‐negative bacteria, cytomegalovirus and tuberculosis in these children.11,12 Madhi et al in a recent publication on the role of influenza in lower respiratory tract infection showed a 78.1% bacterial co‐infection rate in HIV‐1 infected children as compared to a 35.1% bacterial co‐infection rate in uninfected children.13 The distribution of lung disease in determining the likelihood for HIV disease was of limited value.
What is already known on this topic
Children with WHO defined severe pneumonia are treated worldwide with the standardised antimicrobial therapy of either parental penicillin or oral amoxicillin.
The radiological features of these children are non‐specific and do not help in the choice of antimicrobial therapy.
What this study adds
HIV‐1 infected infants with WHO defined severe pneumonia had a three‐fold increased rate of treatment failure following standardised antimicrobial therapy with either parental penicillin or oral amoxicillin.
HIV‐1 infected children with WHO defined “other consolidation/infiltrates” on chest radiography are more likely to fail on the standardised antimicrobial therapy of parental penicillin or oral amoxicillin than children with “endpoint for consolidation”.
Some limitations of this study must be taken into consideration. The lack of lateral or follow‐up chest radiographs and the omission of microbiology testing to determine the aetiology of treatment failures in this cohort prevent confirmation of the reasons for the poorer outcomes.
In conclusion, this study clearly demonstrates the negative effect of HIV‐1 infection on the outcome of severe pneumonia in infancy. According to WHO defined radiological criteria, a preponderance of “other infiltrates” was demonstrated in these children. The addition of co‐trimoxazole to standard antibiotics should be urgently considered.
Abbreviations
ALRI - acute lower respiratory infections
DSMB - data and safety monitoring board
PCP - Pneumocystis jiroveci (carinii) pneumonia
Footnotes
Competing interests: None.
References
- 1.UNICEF State of the world's children 2001. New York: UNICEF, 2001
- 2.Williams B G, Gouws E, Boschini‐Pinto C.et al Estimate of worldwide distribution of child deaths from acute respiratory infections. Lancet Infect Dis 2002225–32. [DOI] [PubMed] [Google Scholar]
- 3.Sawazal S, Black R E, for the Pneumonia Case Management Trials Group Effect of pneumonia case management on mortality in neonates, infants, and preschool children: a meta‐analysis of community‐based trials. Lancet 20033547–555. [DOI] [PubMed] [Google Scholar]
- 4.Donelly L F. Maximising the usefulness of imaging children with community‐acquired pneumonia. AJR Am J Roentgenol 1999172505–512. [DOI] [PubMed] [Google Scholar]
- 5.Swingler G H, Hussey G D, Zwarenstein M. Randomized controlled trial of clinical outcome after chest radiograph in ambulatory lower respiratory infection in children. Lancet 1998351404–408. [DOI] [PubMed] [Google Scholar]
- 6.Sivit C J, Millar C R, Rakusan T A.et al Spectrum of chest radiographic abnormalities in children with AIDS and Pneumocystis carinii pneumonia. Pediatr Radiol 199525389–392. [DOI] [PubMed] [Google Scholar]
- 7.Klugman K P, Madhi S A, Huebner R E.et al A trial of a 9‐valent pneumococcal conjugate vaccine in children with and those without HIV infection. N Engl J Med 2003349(14)1341–1348. [DOI] [PubMed] [Google Scholar]
- 8.Addo‐Yobo E, Chisaka N, Hassan M.et al Oral amoxicillin versus injectable penicillin for severe pneumonia in children aged 3 to 59 months: a randomized multi centre equivalency study. Lancet 20043641141–1148. [DOI] [PubMed] [Google Scholar]
- 9.Markowitz R I, Ruchelli E. Pneumonia in infants and children: radiological‐pathological correlation. Semin Roentgenol 199833(2)151–162. [DOI] [PubMed] [Google Scholar]
- 10.Cherian T. WHO standard radiological definitions. Proceedings from a WHO consultative forum on ARI. Geneva: WHO, October 2003
- 11.Chintu C, Mudenda V, Lucas S.et al Lung disease at necropsy in African children dying from respiratory illnesses: a descriptive necropsy study. Lancet 2002360985–990. [DOI] [PubMed] [Google Scholar]
- 12.Zar H J, Tannenbaum E, Hanslo D.et al Sputum induction as a diagnostic tool for community‐acquired pneumonia in infants and young children from a high prevalence area. Pediatr Pulmonol 200336(1)58–62. [DOI] [PubMed] [Google Scholar]
- 13.Madhi S A, Ramsamy N, Bessellar T G.et al Lower respiratory tract infections associated with influenza A and B viruses in an area with high prevalence of pediatric human immunodeficiency type 1 infection. Pediatr Infect Dis J 200221(4)291–297. [DOI] [PubMed] [Google Scholar]