Results

The case notes of 114 vertically HIV infected children who had received their main clinical care from one of the three hospitals were reviewed, covering the period 1984 to 1994. Nineteen children had presented with Pneumocystis carinii pneumonia and were excluded from the analysis. Of the remaining 95 children, 43 were male, and 52 female. Seventy four per cent were black African, 13% white, 10% mixed, and 3% other. Fifty five (58%) were born in the UK, but only 19 were followed up prospectively from birth. Of the children born abroad the median age of arrival in the UK was 2 years (range 0 to 8 years). Thirteen of the 95 children had died. The median age of the 82 surviving children was 4.7 years (range 1.3 to 15.7 years).

LYMPHOCYTIC INTERSTITIAL PNEUMONITIS

A diagnosis of LIP had been made in 31 children (33%), of whom 19 were born in the UK. Only five children had this diagnosis confirmed on open lung biopsy. Age at diagnosis and follow up time since first presentation were similar among children with and without LIP (table1). Among the 19 children born in the United Kingdom, the median age of LIP diagnosis was 1.8 years (range 0.4 to 5.1 years). LIP was diagnosed in 15 of the 39 black African children born in the UK, compared with four of the 16 white or ‘other’ children (χ² = 0.4, p > 0.5). Parotid enlargement (defined as bilateral and persistent for more than three months) was significantly more common in children in the LIP group (68%) than in the non-LIP group (3%) (χ²= 45.5, p < 0.001). A chronic cough which was present in 32% of children overall and lasted a median of three years was also noted significantly more often in the LIP group (χ² = 13.2, p < 0.001) (table 1).

ACUTE LOWER RESPIRATORY TRACT INFECTIONS

Over the study period of 276 patient years of follow up, 40 children (42%) required 68 admissions for acute lower respiratory tract infections comprising a total of 700 hospital bed days. The admission rate for first and recurrent admission per child year of follow up was 0.38 in children with LIP compared with 0.17 among children without LIP (z score 3.73, p = 0.0002: table 1). Although the median length of admission was similar for the LIP and non-LIP groups, multiple admissions were more frequent in the LIP group. A positive culture was documented from blood culture, sputum, nasopharyngeal aspirate, or on bronchoalveolar lavage during only 24 (35%) of the 68 admissions with acute lower respiratory tract infection. A single organism was isolated in 15 episodes, Streptococcus pneumoniae being the commonest isolate (table 2). Dual pathology was found in nine cases, cytomegalovirus being the most common second isolate. No significant differences were seen between the LIP and non-LIP groups. Of the 95 children, all but one had been recorded as taking regular co-trimoxazole as prophylaxis against pneumocystis pneumonia. Intravenous immunoglobulin had been given to 11/95 children (12%) at some stage of the illness, while 32/95 (34%) had taken zidovudine. Fifteen children (16%) had received bronchodilators, and prolonged courses of steroids had been used in three children, all with severe LIP. During admissions for acute lower respiratory tract infection, intravenous antibiotics were given during 63/68 (93%) of episodes, but ventilation was only required in only 4/68 episodes (6%).

Discussion

Lymphocytic interstitial pneumonitis as a manifestation of paediatric HIV disease occurred in nearly one third of children in our study, a similar rate to that reported from North America.2 LIP was originally an AIDS defining diagnosis in the 1987 CDC classification of AIDS in children. In the revised classification of paediatric HIV disease,8 it was recognised that children with LIP as a first AIDS indicator disease had an improved survival compared with other AIDS indicator diseases,6 and LIP has been redefined as a stage B disease indicator. However, this study confirms that LIP is associated with considerable respiratory morbidity. Although the overall rate of acute lower respiratory tract infection in this study (0.25 episodes/child/year of follow up) is the same as that previously reported in HIV infected children,5 the incidence of such infections in children with LIP was twice that observed in the children with no LIP. The number of children with LIP who had recurrent admission for acute lower respiratory tract infection was particularly striking. The incidence of acute lower respiratory tract infections in children with LIP is approximately 10-fold higher than that reported in an American community based study of non-HIV infected children.9 This study provides evidence of the considerable chronic respiratory morbidity associated with LIP. Some children develop chronic lung damage and bronchiectasis, with a clinical picture similar to cystic fibrosis. While bronchodilators may be of some benefit, other specific treatments have not been systematically studied in LIP. Rubinstein et al have reported improved oxygenation in a small number of children with severe LIP after starting steroids,10 but no controlled trials of steroid treatment have been carried out.

The aetiology of LIP remains uncertain. The principal hypotheses suggest that LIP represents an abnormal lymphoproliferative response, either to HIV alone, or due to superinfection with another virus, the most likely candidate being the Epstein-Barr virus.11 12In the majority of children, a presumptive diagnosis of LIP is now made on radiological and clinical grounds. Lung biopsy had been performed in only five children in this study, at a time when there was little experience of LIP in HIV infected children. Parotid enlargement is confirmed as a useful marker for the presence of LIP in paediatric HIV infection.

Both paediatricians and parents need to be aware of the increased risk of pneumonia in children with LIP, and acute infections should be promptly treated with systemic antibiotics and physiotherapy. An improved understanding of the viral aetiology and immunopathology of LIP is required. Development of a scoring system would facilitate multicentre trials. The most urgent management strategies that need to be evaluated are treatment trials of inhaled or oral steroids, combined with the use of long term continuous antibiotic prophylaxis against encapsulated bacteria.