Increasing incidence of Kaposi's sarcoma in black South Africans in KwaZulu-Natal, South Africa (1983-2006).

The aim of the study was to describe the temporal trends in the incidence of Kaposi's sarcoma (KS) in black South Africans in KwaZulu-Natal (KZN). The study was designed as a retrospective record review. The incidence of Kaposi's sarcoma was estimated using administrative records for patients receiving care for KS through public sector oncology clinics in KZN, 1983-2006. Annual age-standardized incidence rates were calculated using provincial census data for the denominator. Age-specific rates were calculated for the pre-AIDS (1983-1989) and for the generalized AIDS epidemic eras (2006). Age-standardized incidence of KS increased in KZN from <1:100,000 in 1990 to at least 15:100,000 in 2006; this increase was observed in both men and women. There was a shift in the peak age-specific incidence rates from the sixth decade of life in the pre-AIDS era to the fourth and fifth decades in the AIDS era. In conclusion, KS is a growing public health problem in KZN, South Africa. These data reinforce the need for comprehensive national access to and roll-out of antiretroviral drugs, given their success in prevention and treatment of KS in first-world settings.

Characteristics of HIV-1-associated Kaposi's sarcoma among women and men in South Africa.

Despite the increase of HIV-1-associated Kaposi's sarcoma (KS), little is known about HIV-associated KS in the African setting, particularly among women. A descriptive study of the demographic, clinical, immunological and virological features of AIDS-associated KS from KwaZulu-Natal, South Africa was undertaken. Consecutively, recruited patients were clinically staged; CD4/CD8 cell counts, HIV-1 viral loads and clinical parameters were evaluated. Of the 152 patients (77 male and 75 female) 99% were black. Females were significantly younger (P = 0.02) and had poorer disease prognosis (odds ratio [OR] = 2.7, 95% confidence interval [CI] = 1.4-5.4, P = 0.003) and were more likely to have extensive cutaneous KS when compared with males (OR = 3.1, 95% CI = 1.4-6.7, P = 0.003). One-third of patients had coexisting HIV-related disease, most commonly tuberculosis, and these were more frequent in females (56.7 vs. 43.3%). In conclusion, HIV-associated KS in South Africans has an equal female-to-male ratio. Females are younger and have more severe disease than males.

Breast health problems are rare in both HIV-infected and HIV-uninfected women who receive counseling and support for breast-feeding in South Africa.

BACKGROUND: Breast problems, including mastitis, can interfere with the duration and exclusivity of breast-feeding. However, there are no large prospective studies documenting the prevalence, duration, and timing of such problems in breast-feeding women, particularly those who are infected with human immunodeficiency virus (HIV). METHODS: Women enrolled prenatally underwent a breast-feeding counseling intervention until 6 months after delivery. Breast health problems were documented per breast for 180 days after delivery, with 14-day recall histories. RESULTS: Breast health problems were rare, and there were no significant differences between HIV-infected and HIV-uninfected women for any of the following conditions: engorgement, 39 HIV-infected women (3.5%) versus 33 HIV-uninfected women (2.7%; P=.30); breast thrush, 17 (1.5%) versus 12 (1.0%; P=.25); bleeding nipple, 6 (0.5%) versus 4 (0.3%; P=.45); and mastitis/abscess, 11 (1.0%) versus 6 (0.5%; P=.17). Most problems occurred during the first month after birth, with few additional mothers experiencing problems after this point: at 1 and 6 months, 13% and 17% of all mothers, respectively, had experienced a minor or major breast health problem, including sore nipples. Women who had not exclusively breast-fed their infants were more likely to experience any of the breast health problems than were women who had exclusively breast-fed their infants (time-dependent variable; adjusted odds ratio, 1.46; 95% confidence interval, 1.13-1.87; P=.003). HIV-infected women who experienced any serious breast health problem (i.e., bleeding nipple, pus oozing from a nipple or breast, or mastitis/abscess) were 3.55 times (95% confidence interval, 0.86-14.78 times; P=.08) more likely to transmit HIV postnatally to their infant. CONCLUSIONS: With encouragement to exclusively breast-feed, women experienced few breast health problems. When those problems did occur, HIV-infected women with bleeding nipple, pus oozing from a nipple or breast, or mastitis/abscess were more likely to transmit HIV to their infants.

Randomized trial testing the effect of vitamin A supplementation on pregnancy outcomes and early mother-to-child HIV-1 transmission in Durban, South Africa. South African Vitamin A Study Group.

OBJECTIVE: Poor vitamin A status has been associated with a higher risk for mother-to-child transmission of HIV-1 and there is contradictory evidence on the impact of vitamin A on perinatal outcome. We therefore assessed the effect of vitamin A supplementation to mothers on birth outcome and mother-to-child transmission of HIV-1. DESIGN AND METHODS: In Durban, South Africa 728 pregnant HIV infected women received either vitamin A (368) or placebo (360) in a randomized, double-blind trial. The vitamin A treatment consisted of a daily dose of 5000 IU retinyl palmitate and 30 mg beta-carotene during the third trimester of pregnancy and 200000 IU retinyl palmitate at delivery. HIV infection results were available on 632 children who were included in the Kaplan-Meier transmission analysis. Results are reported on mother-to-child transmission rates up to 3 months of age. RESULTS: There was no difference in the risk of HIV infection by 3 months of age between the vitamin A [20.3%; 95% confidence interval (CI), 15.7-24.9] and placebo groups (22.3%; 95% CI, 17.5-27.1), nor were there differences in foetal or infant mortality rates between the two groups. Women receiving vitamin A supplement were, however, less likely to have a preterm delivery (11.4% in the vitamin A and 17.4% in the placebo group; P = 0.03) and among the 80 preterm deliveries, those assigned to the vitamin A group were less likely to be infected (17.9%; 95% CI, 3.5-32.2) than those assigned to the placebo group (33.8%; 95% CI, 19.8-47.8). CONCLUSION: Vitamin A supplementation, a low-cost intervention, does not appear to be effective in reducing overall mother-to-child transmission of HIV; however, its potential for reducing the incidence of preterm births, and the risk of mother-to-child transmission of HIV in these infants needs further investigation.

Persistent and chronic lung disease in HIV-1 infected and uninfected African children.

OBJECTIVE: The causes of persistent lung disease (PLD) and chronic lung disease (CLD) are unknown in HIV-infected children in developing countries. We describe the causes and course of PLD and CLD in HIV-infected and uninfected children. METHOD: Of 194 children with lung disease persisting for at least 1 month who were seen at the paediatric respiratory clinic over a 2-year period, 42 underwent invasive investigations after failed initial management over 3 months. PLD was defined as the presence of clinical and radiological features of lung disease for more than 1 month, and CLD as these features for more than 3 months. RESULTS: One hundred and thirty-eight (71%) of the 194 children with PLD were HIV-infected, 52 (27%) were not infected and four (2%) were of undetermined HIV status. Forty-eight per cent of the HIV-infected children and 52% of the HIV-uninfected children responded to initial treatment over 3 months; the presumptive diagnoses in these were tuberculosis, interstitial pneumonitis, bronchiectasis and post-ventilation lung syndrome. Of the 28 HIV-infected children with CLD who underwent invasive investigations 16 (57%) had lymphoid interstitial pneumonitis, eight (29%) had tuberculosis and four (14%) had non-specific interstitial pneumonitis. Of the 14 HIV-uninfected children with CLD who had invasive testing there were four cases (29%) each of tuberculosis and interstitial pneumonitis, three (22%) cases of bronchiectasis and one case of each of extrinsic allergic alveolitis, crytogenic fibrosing alveolitis and non-Hodgkin's lymphoma. CONCLUSION: This is the first set of data on the causes of CLD in HIV-infected children in a developing country. Every effort should be made to identify the infectious agent, whether M. tuberculosis or a secondary bacterial infection in LIP, in order to treat most appropriately these children with lung disease.

Method of feeding and transmission of HIV-1 from mothers to children by 15 months of age: prospective cohort study from Durban, South Africa.

OBJECTIVE: To determine the risk of HIV transmission by infant feeding modality. DESIGN AND SETTING: A prospective study in two hospitals in Durban, South Africa. PARTICIPANTS: A total of 551 HIV-infected pregnant women enrolled in a randomized trial of vitamin A. INTERVENTIONS: Women self-selected to breastfeed or formula feed after being counselled. Breastfeeders were encouraged to practice exclusive breastfeeding for 3-6 months. MAIN OUTCOME MEASURES: Cumulative probabilities of detecting HIV over time were estimated using Kaplan-Meier methods and were compared in three groups: 157 formula-fed (never breastfed); 118 exclusively breastfed for 3 months or more; and 276 mixed breastfed. RESULTS: The three feeding groups did not differ in any risk factors for transmission, and the probability of detecting HIV at birth was similar. Cumulative probabilities of HIV detection remained similar among never and exclusive breastfeeders up to 6 months: 0.194 (95% CI 0.136-0.260) and 0.194 (95% CI 0.125-0.274), respectively, whereas the probabilities among mixed breastfeeders soon surpassed both groups reaching 0.261 (95% CI 0.205-0.319) by 6 months. By 15 months, the cumulative probability of HIV infection remained lower among those who exclusively breastfed for 3 months or more than among other breastfeeders (0.247 versus 0.359). CONCLUSION: Infants exclusively breastfed for 3 months or more had no excess risk of HIV infection over 6 months than those never breastfed. These findings, if confirmed elsewhere, can influence public health policies on feeding choices available to HIV-infected mothers in developing countries.

T-helper cell responses to HIV envelope peptides in cord blood: protection against intrapartum and breast-feeding transmission.

BACKGROUND: Acquired HIV-specific cell-mediated immune responses have been observed in exposed-uninfected individuals, and it has been inferred, but not demonstrated, that these responses constitute a part of natural protective immunity to HIV. This inference was tested prospectively in the natural exposure setting of maternal-infant HIV transmission in a predominantly breast-fed population. METHODS: Cord blood from infants of HIV-seropositive women in Durban, South Africa, were tested for in vitro reactivity to a cocktail of HIV envelope peptides (Env) using a bioassay measuring interleukin-2 production in a murine cell line. Infants were followed with repeat HIV RNA tests up to 18 months of age to establish which ones acquired HIV-infection. RESULTS: T-helper cell responses to Env were detected in 33 out of 86 (38%) cord blood samples from infants of HIV-seropositive women and in none of nine samples from seronegative women (P = 0.02). Among infants of HIV-seropositive mothers, three out of 33 with T-helper responses to Env were already infected before delivery (HIV RNA positive on the day of birth), two were lost to follow-up, and none of the others (out of 28) were found to be HIV infected on subsequent tests. In comparison, six out of 53 infants unresponsive to Env were infected before delivery, and eight out of 47 (17%) of the others were found to have acquired HIV infection intrapartum or post-partum through breast-feeding (P = 0.02). CONCLUSIONS: T-helper cell responses to HIV envelope peptides were detected in more than one-third of newborns of HIV-infected women; no new infections were acquired by these infants at the time of delivery or post-natally through breast-feeding if these T-helper cell responses were detected in cord blood.

Variation in breastmilk HIV-1 viral load in left and right breasts during the first 3 months of lactation.

The mechanism and risk factors associated with mother-to-child transmission of HIV-1 through breastfeeding remain unclear; breastmilk viral load may be an important determinant of transmission. Analysis of breastmilk cell-free viral load in samples taken from each breast at 1, 6 and 14 weeks postpartum showed that HIV-1 is shed intermittently and load may differ considerably between breasts of an individual woman at any given time. Breastmilk HIV-1 load was undetectable in approximately one-third of samples.

Vitamin A supplementation reduces measles morbidity in young African children: a randomized, placebo-controlled, double-blind trial.

The effects of vitamin A supplementation on measles morbidity are unclear. Sixty hospitalized children aged 4-24 mo with complicated measles received a World Health Organization--(WHO) recommended dose of vitamin A or placebo. The two groups were comparable in known covariants of measles severity: weight-for-age percentiles, overcrowding, rash, total lymphocytes, and serum concentrations of zinc, albumin, prealbumin, retinol-binding protein, and vitamins A and E. Ninety percent of the patients had hyporetinemia. Integrated morbidity scores, determined by severity of condition (eg, diarrhoea, herpes, and respiratory-tract infection) were assigned on day 8 and 6 wk and 6 mo; these were reduced by 82%, 61%, and 85%, respectively, in the supplemented group, which was mainly due to reduced respiratory-tract infection. There was one death in the placebo group. At 6 wk weight gain was significant in the supplemented group. Despite the selected sample, attention to multiple covariates enhances the validity of the data obtained and supports the current WHO recommendations for vitamin A supplementation during measles.

Vitamin A deficiency among children in a periurban South African settlement.

Preschool children (aged 3-6 y) who were living in an informal settlement within metropolitan Durban, South Africa, were assessed for vitamin A status. The serum retinol concentration of 169 children tested was 0.73 +/- 0.26 mumol/L (mean +/- SD). Nine children (5%) had vitamin A deficiency (< 0.35 mumol/L) and 75 children (44%) had low vitamin A concentrations (< 0.70 mumol/L). Conjunctival impression cytology (CIC) in 185 children revealed that 18% had poor vitamin A status as defined by two abnormal conjunctival specimens. The CIC test was a feasible and reproducible method; however, it correlated poorly with the traditionally accepted serum retinol threshold of deficiency in this population where overt vitamin A deficiency is not prevalent. This survey demonstrated that regardless of the measurement tool, there is a prevalence of subclinical vitamin A deficiency in this typical periurban informal settlement and accordingly we suggest that these children should be targeted for vitamin A-intervention strategies.

Prevention and treatment of perinatal HIV-1 infection in the developing world.

Measures to reduce mother-to-child transmission of HIV-1 are discussed according to the presumed timing of transmission. Long course antiretrovirals can substantially decrease intrauterine transmission. Antiretrovirals and caesarean section reduce vertical transmission by preventing intrapartum transmission; but caesarean section is unsuitable for most developing countries. Short and very short course antiretrovirals have an efficacy rate of about 50% in non-breastfeeding populations compared with just under 40% at 6 months and between 23 and 30% at 15-24 months in breastfeeding women. The latter is due to the fact that postnatal transmission occurs as long as breastfeeding continues. Breastfeeding accounts for a third to a half of overall transmission. Exclusive rather than mixed breastfeeding appears to diminish the risk of transmission, and HIV-infected women who choose to breastfeed should be advised to give breastmilk exclusively for a maximum of 5-6 months.

Characterization of proteinuria in asymptomatic family members and household contacts of children with hepatitis B virus-associated membranous nephropathy.

The biosocial background in which the hepatitis B virus (HBV) carrier state with membranous nephropathy (MN) develops was studied by evaluating HBV carriage and proteinuria among 195 family members and household contacts of 31 index HBV carrier children with MN. Unrelated individuals from the communities of these index cases who were negative for HBV served as controls (n = 123). HBV was determined by using third-generation enzyme-linked immunosorbent assay, slot-blot hybridization, and nested polymerase chain reaction. Patterns of proteinuria were determined by using sodium dodecyl sulfate-polyacrylamide gel electrophoresis; immunoglobulin G and haptoglobulin were suggestive of MN. Seventy-two members (36.9%) of the study group (n = 195) were HBV carriers; 21 of these carriers (29.2%) had proteinuria. Twenty-eight members (41.2%) of the study group who were HBV negative (n = 68) and 26.8% of the controls showed proteinuria. This lack of association between HBV carriage and proteinuria remained when controlled for sex and family relationship. HBV was not protective against the development of proteinuria. Proteinuria suggestive of MN was strongly associated with an abnormal protein-creatinine ratio (P: = 0.001), but was not significantly different between subjects and controls (8.7% versus 6.5%; P: = 0.5). Genetic influences or environmental exposures in these subjects may be responsible for the proteinuria, suggesting underlying glomerular basement membrane damage. Discordance between the HBV carrier state and patterns of proteinuria in the study group suggest that HBV and MN may not be causally related or may reflect exceptional interaction between specifically vulnerable individuals and HBV.

Vitamin A supplementation enhances specific IgG antibody levels and total lymphocyte numbers while improving morbidity in measles.

The effect of vitamin A supplementation on selected factors of immunity was tested in African children (ages 4 to 24 months with complicated measles) during a randomized double-blind intervention trial. Placebo (n = 31) and treated groups (n = 29) had similar baseline characteristics. The supplemented group had significant reductions in morbidity (expressed as integrated morbidity scores) during the acute (Day 8, P = 0.006) and chronic (Day 42, P = 0.02; 6 months; P = 0.002) phases. In the treated group there was an increase in total number of lymphocytes (Day 42, P = 0.05) and measles IgG antibody concentrations (Day 8, P = 0.02), both of which have consistently been previously shown to correlate more closely with outcome in measles than other immunologic, clinical and radiologic factors. Interleukin 2 and plasma complement values were unaffected by vitamin A supplementation. These findings reinforce results from animal studies that show that the pathways of vitamin A activity in decreasing morbidity and mortality are partly founded on selective immunopotentiation.

Predicting perinatal human immunodeficiency virus infection by antibody patterns.

The evolution of human immunodeficiency virus type 1 (HIV-1) antibody titers determined by enzyme-linked immunosorbent assay between birth and 18 months of age was investigated in 118 babies born to HIV-1-seropositive South African mothers. By 18 months 41 (34.7%) children were diagnosed as HIV-1-infected by standard criteria. All 77 uninfected babies cleared maternal antibodies by 15 months; 94.5% of these babies seroreverted by 12 months. By 9 months of age a significant difference (P < 0.05) was noted between antibody decay rates in infected and uninfected children. Of the children subsequently shown to be uninfected, 95.8% demonstrated > or = 50% decay in antibody titers between 6 and 9 months; only 1 in the infected group showed a similar pattern (sensitivity, 97.8%; specificity, 93.8%). The approach of assessing the progression of antibody decay in infected and uninfected babies makes it a feasible and useful tool for estimating vertical transmission rates and diagnosis of perinatal HIV-1 infection earlier than standard practice.

Severe, rapidly progressive human immunodeficiency virus type 1 disease in newborns with coinfections.

AIM: To describe a severe form of rapidly progressive HIV-1 infection manifesting in the neonatal period. METHOD: Prospective cohort study, King Edward VIII Hospital, Durban, South Africa. HIV-1-exposed neonates with hepatosplenomegaly, lymphadenopathy or persistent pneumonia within the first 28 days of life were investigated for perinatal infections. Confirmation of neonatal HIV-1 infection, HIV-1 subtype and clinical outcomes were studied. RESULTS: Twenty-three (72%) of 32 symptomatic HIV-1-exposed neonates recruited at a mean of 15.2 days were HIV-1-infected. HIV-1 infection was detected in 5 patients who were tested within 48 h of birth, confirming congenital infection. Congenital infection was not excluded in any case. Median neonatal viral load at recruitment was 471,932 copies/ml and median CD4 was 777 cells/mm3. The predominant clinical presentation was growth retardation and prematurity. Perinatal infections detected included: tuberculosis (8), syphilis (6) and cytomegalovirus (10). All of the neonates with perinatal tuberculosis were HIV-1-coinfected. Maternal and neonatal viral load and CD4 at recruitment were not statistically different between the groups with tuberculosis vs. other coinfections. Gag gene sequence analysis confirmed closely aligned HIV-1 subtype C in mothers and neonates. Nineteen (83%) died by 9 months, with a mean age at death of 3.5 months. CONCLUSIONS: A distinct group of HIV-1-infected babies may clinically manifest in the neonatal period with perinatal coinfections, subsequent rapidly progressive HIV-1 and early death.

Household clustering and intra-household transmission patterns of hepatitis B virus infection in South Africa.

The objective of this study was to investigate household clustering of hepatitis B virus (HBV) infection in South Africa in order to understand intra-household patterns of virus transmission that would provide information on potential risk factors of HBV infection. Subjects were the household contacts of 28 hepatitis B surface antigen (HBsAg) positive children (index-carrier), 22 hepatitis B surface antibody positive children (index-past-infection) and 35 children with no serological evidence of HBV infection (index-negative). Evidence of HBV infection (at least one positive HBV marker) was present in 73.7%, 48.7% and 38.2% and HBsAg was present in 19.9%, 8.7% and 2.9% of household contacts of index-carrier (N = 186), index-past-infection (N = 150) and index-negative (N = 207) children respectively. The clustering of HBV infection and HBsAg was present in all subgroups of household contacts regardless of the degree of relatedness to the index child. As age increased, the cumulative prevalence of HBV infection increased while the likelihood of being HBsAg positive decreased. Regardless of age, males were more likely (p less than 0.01) than females to be HBsAg positive. We conclude that HBV infection clusters in households; that there is an increased susceptibility, which is probably not genetically mediated, of becoming an HBV carrier in certain households and that males are at greater risk of being HBV carriers. Intra-household horizontal person-to-person transmission, the precise mechanisms of which are not known, is important in South African blacks.

The prevalence and transmission of hepatitis B virus infection in urban, rural and institutionalized black children of Natal/KwaZulu, South Africa.

The sera of statistically selected urban (805), rural (238) and institutionalized (127) black children were tested for markers of hepatitis B virus (HBV) infection. The age-standardized (6-14 years) prevalence rates of HBs antigenaemia for comparison between urban, rural and institutionalized children were 10%, 18.5% and 25.1% and the HBV exposure rates were 31.4%, 62.1% and 72.0% respectively. In the newborn to six years age group the prevalence rates of HBsAg and HBV exposure were 2.5% and 7.1% for urban children and 53.1% and 70.3% for institutionalized children. Peak prevalences of HBsAg occurred in the 6-8 year age group and were 14.4% and 22.6% in urban and rural children respectively. Hepatitis Be Antigen (HBeAg) was detected in 46.5% and antibodies to hepatitis Be antigen (HBeAb) in 10.0% of all HBsAg positive children. Multiple mechanisms involving horizontal rather than vertical transmission appeared to be important in urban children, with HBV exposure in females being significantly associated with ear-piercing (p less than 0.001) and scarification (p less than 0.05). In addition, HBsAg was detected in 25 of 29 pools of bloodfed mosquitoes caught at the children's institution and was negative in all four pools of unfed mosquitoes, suggesting that these arthropods may also be one factor in the horizontal spread of HBV infection. Familial clustering of HBV infection was suggested by a significantly higher (p less than 0.01) prevalence of HBsAg amongst family contacts of HBsAg positive urban children (17.7%) than in the control groups of family contacts of HBsAb positive children (8%) and children who were negative for all HBV markers (2.4%). The significance and implications of these findings are discussed.

Access to voluntary counseling and testing for HIV in developing countries.

The counseling that precedes and follows testing of subjects for HIV has become, quite unexpectedly, a focal point for assessment of the ethical propriety, availability, and appropriateness of health services during the AIDS epidemic. It can be anticipated that in the worst affected regions, Voluntary Confidential Counseling and Testing (VCCT) will be an integral component of "...access to comprehensive, essential, quality health care" which is WHO's goal of "Health for All" in the next century. The role, purpose, location, and methods of VCCT, which were reviewed at the previous Global Strategies Conference in 1997, are summarized. Currently understood objectives of VCCT include acceptance of the test, provision of care for HIV-infected individuals (particularly pregnant women), prevention of HIV transmission, and psychosocial support. Many countries in Africa are gradually instituting VCCT as part of their Primary Health Care package. For example "...access to care, counselling and support" for HIV/AIDS and STDs is one of the top 10 national priorities in South Africa. However, closer examination in the country reveals personnel and skill shortages, inability of half the primary health care (PHC) clinics to provide antenatal services, and HIV testing being offered in only 56%. Condom availability is generally good, but termination of pregnancy is undertaken in a bare 27% of hospitals. In other regions of Africa, VCCT is also deficient in many respects: medical services are often unavailable, support is absent, availability is restricted and there are few trained counselors. Consequently, workloads are heavy. Requirements for effective counseling will be listed. The global determinants of inequities in accessing VCCT, such as the GNP and the crushing debt burden borne by poor countries, are discussed. A third of women worldwide receive no antenatal care, and just 60% of the roughly 133 million annual births throughout the world are attended by trained health personnel. Even when VCCT services are available, they are often not acceptable. The overwhelming majority of African women appear to accept HIV testing, but only a proportion (59-61% in recent intervention trials) return for the results. Obstacles to be overcome for provision of VCCT services are identified. Evidence for a positive impact of VCCT services includes facilitated decision-making, acceptance and coping with HIV, improved family and community acceptance, increased condom use, and reduced gonorrhea rates and HIV transmission.

Admission patterns and outcomes in a paediatric intensive care unit in South Africa over a 25-year period (1971-1995).

OBJECTIVE: To describe admission and outcome patterns of diseases managed at a paediatric intensive care unit (PICU) in a developing country between 1971 and 1995, in order to provide data which will assist in improving the management of diseases and the rational allocation of health resources. METHOD: This study is based on data collated from annual audits conducted by the head of the PICU, King Edward VIII Hospital, Durban, South Africa. This unit serves a childhood population of about 3-4 million aged from 0-12 years. RESULTS: The proportion of children admitted to the ICU as a percentage of the general paediatric admission has increased from 1.5% in 1971 to 7% in 1995. During this period, 7580 children were admitted to the unit, an average of 303 cases per annum. The overall mortality rate was 35.44% (range 29.9 to 45.4%); over 90% of the children admitted were intubated and 80% required intermittent positive pressure ventilation. Common childhood vaccine preventable infections have declined substantially and have been replaced by HIV related syndromes, lower respiratory tract infections, metabolic disorders, septicaemia and hyaline membrane disease. The mean duration of ICU stay per survivor over the study period 1991-1995 was 13.891 days. Tetanus, septicaemia and HIV related diseases required the longest ICU stay per survivor (>20 days), while accidental injuries, neonatal apnoea and asthma required the shortest duration of ICU stay per survivor (<6 days); 23.9% of all deaths occurred in the first 24 h. CONCLUSION: The profile of diseases in children admitted to this PICU has changed considerably over 25 years. Some of these changes can be attributed to the shift in emphasis to primary health care, especially higher vaccination coverage rates. Profitable utilisation of limited ICU facilities would probably be enhanced by the use of outcome measures such as mortality and mean number of ICU days of stay per survivor.