Risk factors for human immunodeficiency virus infection in Ethiopian children with tuberculosis.

BACKGROUND: Separate risk factors for HIV infection and for tuberculosis (TB) are well-studied, but it is unclear whether these risk factors still apply in the new epidemiologic situation of dual infection. This study examines risk factors associated with seropositivity for HIV in Ethiopian children with clinical TB. METHODS: A prospective, controlled study of children with TB diagnosed in Addis Ababa from December 11, 1995, to January 28, 1997, in which HIV-positive children were compared with HIV-negative children with regard to sociodemographic background, previous medical history and vaccination. RESULTS: HIV prevalence among children with clinical TB was 11.2%. High educational status of mothers, low age, loss of one or two parents and earlier Calmette-Guérin bacillus (BCG) vaccination of the child were factors independently related to HIV infection. CONCLUSION: Factors associated with HIV infection among children with clinical TB include higher education of parents, higher income and better living conditions. The HIV epidemic might thus modify traditional risk factors for tuberculosis. It might also decrease the overall effect of BCG vaccination given that BCG did not provide protection in children infected with HIV. An expected increase of dually infected children who are younger, more in need of hospitalization and often lacking one or both parents will put an additional burden on the Ethiopian health care system.

Etiology of pneumonia, sepsis and meningitis in infants younger than three months of age in Ethiopia.

METHODS: Within a multicenter study coordinated by WHO, an investigation of the etiologic agents of pneumonia, sepsis and meningitis was performed among infants younger than 3 months of age seen at the Ethio-Swedish Children's Hospital in Addis Ababa for a period of 2 years. Of the 816 infants enrolled 405 had clinical indications for investigation. RESULTS: There were a total of 41 isolates from blood cultures from 40 infants. The study showed that the traditionally known acute respiratory infection pathogen Streptococcus pneumoniae was most common in this extended neonatal age group, found in 10 of 41 blood isolates. Streptococcus pyogenes was a common pathogen in this setting (9 of 41 blood isolates), whereas Salmonella group B was found in 5 of 41 isolates. Streptococcus agalactiae, which is a common pathogen in developed countries, was absent. A study of the susceptibility pattern of these organisms suggests that a combination of ampicillin with an aminoglycoside is adequate for initial treatment of these serious bacterial infections, but the combination is not optimal for the treatment of Salmonella infections. Among 202 infants on whom immunofluorescent antibody studies for viruses were performed based on nasopharyngeal aspirates, respiratory syncytial virus was found in 57 (28%) infants, and Chlamydia trachomatis was isolated in 32 (15.8%) of 203 infants.

Oxygen delivery to children with hypoxaemia in small hospitals in developing countries.

Oxygen administration is one of the most important modalities of therapy for a patient with hypoxaemia to prevent death. This review summarises the methods of oxygen delivery applicable in small hospitals in developing countries, and evaluates information about their safety and efficacy. The following criteria were considered: cost and availability, efficiency/oxygen concentration achieved, tolerability/comfort of the method, requirement of humidification, demand for nursing care, and safety of the method and complications. In summary, it is concluded that all low-flow methods, i.e., nasopharyngeal catheters, nasal catheters and prongs, are effective in the oxygenation of sick children with severe pneumonia or bronchiolitis. Nasal prongs are the safest method of oxygen delivery, but nasopharyngeal catheters and nasal catheters are more easily available and less expensive. However, if they are used, they need close supervision to avoid serious complications. Nasal prongs are the method of choice for oxygen delivery in small hospitals in developing countries.

Pattern of resolution of tachypnoea and fever in childhood pneumonia.

Acute lower respiratory infections (ALRI) account for one fifth of deaths among children below five years of age and pneumonia is responsible for about 70% of all ALRI deaths. Interventions with antibiotics have shown reduction in pneumonia case-fatality rates. However, there is room for further reduction of deaths from pneumonia through improved monitoring and follow up system. We studied the pattern of resolution of tachypnoea and fever among 108 children who presented to our outpatient clinic with non-severe pneumonia and among 102 children who were admitted for severe pneumonia. We found that tachypnoea was present in 18% and 23% after 72 hours of initiation of antibiotics and fever resolved completely after 48 hours and 72 hours of initiation of therapy in non-severe cases of pneumonia among children two to 11 months and 12 to 59 months of age respectively. Conversely, among cases of severe pneumonia on day 5 of initiation of treatment, tachypnoea and fever were present in 65% and 51% respectively in children two to 11 months old and in 53% and 60% respectively in children 12 to 59 months old. Respiratory rate increased with increase in body temperature at an average rate of four breaths per minute for every 1 degree C rise. Our study suggests that body temperature and respiratory rate can be used to monitor the clinical course of non-severe pneumonia. Further research is needed to identify other clinical signs that will help the health worker to decide improvement in attacks of severe pneumonia.

Profile of cardiac diseases in Ethiopian children.

A review of cardiac admissions to the Ethio-Swedish Children's Hospital from January 1981 to December 1988 revealed 365 cases. Patients were included in the study if they had clinical, laboratory and echocardiographic (M, 2D-Modes, doppler) proofs of the specific disease entity. One hundred and ten patients were eligible. This accounted for 0.65% of all admissions (N = 16,905). Rheumatic heart disease accounted for 54.5% (N = 60), congenital heart disease for 35.5% (N = 39), and acquired heart disease of nonrheumatic origin for 10% (N = 11) or cardiac admissions. Of patients with rheumatic heart disease, the mitral valve was involved in 41.7% (N = 25). Pure mitral stenosis was seen in 8.3% (N = 5). Ventricular septal defect was seen in 33.3% (N = 13) of patients with congenital heart disease. The mean hospital stay for patients with rheumatic heart disease was 35.4 days, and 24.7 days for patients with congenital heart. As rheumatic heart disease predominates, methods for decreasing its incidence are discussed.

Oxygen therapy for children with acute respiratory infections in small hospitals.

Oxygen administration is one of the most important modalities of therapy for a patient with hypoxaemia to prevent death and disability from common conditions such as acute lower respiratory tract infections. Oxygen needs to be available at all times in hospitals, however, it is too expensive for many developing countries. There is little information for health professionals regarding indications for initiating oxygen therapy, selecting appropriate method of oxygen administration and deciding on the source for oxygen. A review of the literature using medline citations and cross references from published articles and other manuscripts was made. The review described the two main sources of oxygen for small hospital-cylinders and oxygen concentrators and their advantages and disadvantages. It also looked at the evidences for clinical indications to initiate and discontinue oxygen therapy. Studies on efficient and safe methods of administration of oxygen were reviewed as well. The review concluded that oxygen may be administered in children with cyanosis, chest indrawing, inability to drink or breastfeed, tachypnea with respiratory rate above 70/minute or in a child who develops restlessness and improves on oxygen. The initial capital cost of concentrators is high but the running cost is low and it does not require transport while oxygen cylinders are expensive to transport and need continuous refilling. The safest method of oxygen administration are the prongs followed by the nasal catheters.

A four-year cohort study of HIV seropositive Ethiopian infants and children: clinical course and disease patterns.

Sixty-two HIV positive infants who were admitted to the Missionaries of Charity Orphanage in Addis Ababa were followed From July 25, 1991 to July 30, 1995 for a total period of 4 years. Regular clinical examination and treatment by a paediatrician was being offered to these infants in addition to monitoring of their HIV serostatus every three months until the age of 18 months and every year after that. Among those aged above 18 months, 14 children were HIV sero-positive and alive and 4 children were HIV sero-positive but died. Thirty-three children had sero-reverted to negative. The mother-to-child transmission was crudely estimated at 29% to 47%. Among the clinical signs, generalized lymphadenopathy, hepatomegaly, splenomegaly, wasting, stunting and delayed motor development were found more commonly in the definitely HIV positive children. Upper respiratory tract infections, acute diarrhoea, pneumonia, pyogenic skin infections, sepsis and candidal infections were the commonest causes of illness. Comparison of the HIV positive with the HIV negative cases indicates that the risk is higher for the indefinitely HIV positive group for episodes of acute diarrhoea, pneumonia, sepsis and candidal infections.

PIP: Mother-to-child rates of HIV transmission are high in Africa. Findings are presented on 62 HIV-positive infants admitted to the Missionaries of Charity Orphanage, Addis Ababa, who were followed from July 25, 1991, to July 30, 1995. The infants were provided with regular clinical examination and treatment by a physician, as well as the monitoring of their HIV serostatus every 3 months until age 18 months and every year thereafter. Among infants over age 18 months, 14 were HIV seropositive and alive, and 4 were HIV positive, but died. 11 children were HIV positive and died before age 18 months and 33 seroreverted to HIV seronegative status. The level of mother-to-child HIV transmission was 29-47%. Among the clinical signs presented, generalized lymphadenopathy, hepatomegaly, splenomegaly, wasting, stunting, and delayed motor development were more often found in the definitely HIV-positive children. Upper respiratory tract infections, acute diarrhea, pneumonia, pyogenic skin infections, sepsis, and candidal infections were the most commonly seen illnesses.

Childhood meningitis in a tertiary hospital in Addis Ababa: clinical and epidemiological features.

Hospital based prospective study of 136 children aged 3 months to 12 years admitted as cases of meningitis between June 1996 and May 1997 was undertaken to assess the epidemiological features, evaluate the outcome and measure the duration of clinical improvement after initiation of treatment. Infants accounted for 79 (58%) of cases and 23 (68%) of deaths. Next to fever and vomiting, neurologic signs were the commonest presentation. H. influenzae, S. pneumoniae and N. meningitidis accounted for 90% of culture isolates. Delayed presentation, partial treatment, altered sensorium at admission and pneumococcal meningitis were risk factors for mortality. The morbidity and mortality rates were 25% of total cases each. The mean duration of fever was 3.7 (+/- 2.27) days after treatment. The morbidity and mortality rates are still high in this hospital.

Epidemic meningococcal meningitis in children. A retrospective analysis of cases admitted to ESCH (1988).

One hundred twenty four cases of meningococcal meningitis were seen at the Ethio-Swedish Children's Hospital (ESCH) during the epidemic period December 1, 1987 to January 31, 1989. Data on demographic and clinical profile of patients were collected and analyzed. Two thirds of patients were above 5 years of age. Fifty percent of the patients came from the "mercato area" of the city of Addis Abeba, Higher 3, 4, 5 & 6. Thirteen cases were from outside Addis Abeba. The main clinical presentations were fever and vomiting in all age groups and headache in those above 5 years of age. The classical meningeal signs were rare in those below 5 years of age. The most common neurological deficit was loss of hearing. Mortality was very low (only 3 deaths). Continuous surveillance of demographic and clinical indicators is recommended as they could be useful early warning signs of an epidemic.

Prevalence of bacterial pathogens in children with acute respiratory infection in Addis Ababa.

A study was conducted in the Ethio-Swedish Children's Hospital and different schools and kindergartens in Addis Ababa to determine the prevalence of bacterial agents that are associated with acute respiratory infection in children from 1998-1999. A total of 883 subjects were studied, out of which 77% were cases from the Ethio-Swedish Children's Hospital and 23% were controls from different schools and kindergartens. From each case and control throat and nasopharyngeal specimens were collected. Culture and different biochemical tests were used to isolate the potential bacterial pathogens. Clinical findings like cough, difficult breathing and fever were correlated with laboratory findings. S. pneumoniae and H. influenzae type b were the most commonly isolated bacteria in both throat and nasopharyngeal specimens; 74% and 70% in the cases and 2% and 5% in the control groups, respectively. This paper discusses the association between throat and nasopharyngeal carriership of bacteria and acute respiratory infection in children in Addis Ababa.

Mothers' perceptions of signs and symptoms of acute respiratory infections in their children and their assessment of severity in an urban community of Ethiopia.

Standard case management strategy has been recommended to reduce the high mortality rate in children with acute respiratory infections (ARI). Appropriate case management has been shown to prevent such deaths, but only if families recognize signs of possible pneumonia and seek care promptly from a trained health worker. The purpose of the present study was to assess mothers' perception and interpretation of ARI signs and symptoms in relation to that of a physician in an urban community in Addis Ababa. Two hundred and twenty-two mothers who brought their children to hospital with cough or difficulty in breathing and an equal number of control mothers were studied. Most mothers did not recognize these signs, including the key signs of pneumonia-rapid breathing and chest indrawing. While it was shown that between two physicians there was good agreement on kappa values above 70% for most ARI signs, there was little agreement between physicians and mothers or between mothers whose children came for ARI problems and mothers of hospital controls. The few mothers who recognized these signs did not interpret them as serious. The study concludes by recommending intensive health education and further ethnographic studies on community beliefs about ARI in children, with particular emphasis on documentation of the terms, signs and symptoms by which families recognize the illness.

PIP: In Ethiopia, acute respiratory infections (ARI) account for 33% of infant mortality and 20% of mortality among children younger than age 5 years. The author interviewed 222 mothers who brought their children to Ethio-Swedish Children's Hospital in Addis Ababa with cough or difficulty breathing to assess their perception and interpretation of ARI signs and symptoms in relation to those of a physician. 230 control mothers were also studied. Most mothers did not recognize rapid breathing and chest indrawing as key signs of pneumonia. Good agreement was observed between two physicians on kappa values above 70% for most ARI signs, but there was little agreement between physicians and mothers or between mothers whose children came for ARI problems and mothers of hospital controls. The few mothers who recognized the signs did not interpret them as serious. The author recommends intensive health education and further ethnographic studies on community beliefs about ARI in children, with particular emphasis upon documentation of the terms, signs, and symptoms by which families recognize the illness.

Managing pneumonia.

PIP: Health workers should assess all children with a cough or difficult breathing for pneumonia. They should refer any child with severe pneumonia to a hospital for admission. At the hospital, a distinction is usually made between severe or very severe pneumonia among children 2 months to 5 years old. Signs or symptoms unique to very severe pneumonia are cyanosis and inability to drink. If a child has these signs and has convulsions, hospital personnel should consider a lumbar puncture to check for meningitis. Chest indrawing may also be present in very severe pneumonia cases. Chest indrawing in children with no cyanosis who are able to drink constitutes severe pneumonia. Health workers need to look for a variety of nonspecific signs of severe or very severe pneumonia in babies younger than 2 months: not feeding well, convulsions, abnormally sleepy, fever (38 degrees Celsius), fast breathing (=or+ 60 breast/minute), cyanosis, grunting, or apnea. These signs are also signs for meningitis or sepsis in young infants. Treatment for all 3 conditions is benzylpenicillin plus gentamicin for at least 14 days. Oxygen treatment is also indicated for these young infants. Treatment for both severe and very severe pneumonia cases includes oxygen and an antibiotic (benzylpenicillin and chloramphenicol, respectively). Hospitals should have in stock at all times essential antibiotics (benzylpenicillin, cloxacillin, chloramphenicol, and gentamicin) and an oxygen supply (oxygen cylinder or oxygen concentrator). When the oxygen supply is limited, children with very severe pneumonia should be the priority. Oxygen needs to be delivered at a flow rate of 1-2 liters/minute via nasal prongs or a nasal catheter. Admitted pneumonia cases with fever (39 degrees Celsius) should receive paracetamol to treat the fever. Hospital workers need to keep the airway of pneumonia cases clear and to encourage them to drink and/or breast feed.^ieng

Pneumococcal and Haemophilus influenzae meningitis in a children's hospital in Ethiopia: serotypes and susceptibility patterns.

Streptococcus pneumoniae and Haemophilus influenzae are responsible for most pyogenic meningitis cases in children in Ethiopia. Resistance of S. pneumoniae and H. influenzae to penicillin and chloramphenicol respectively has been reported globally. Resistance has been related to specific serotypes of S. pneumoniae or to beta-lactamase-producing H. influenzae strains. This study describes the serotypes/ serogroups and susceptibility pattern of the two organisms causing meningitis in Ethiopian children. There were 120 cases of meningitis caused by S. pneumoniae (46) and H. influenzae (74) over a period of 3 years (1993-95). Nineteen children died from pneumococcal and 28 from haemophilus meningitis. Penicillin-resistant pneumococcal meningitis (4/8 = 50%) caused a greater mortality rate than penicillin-susceptible pneumococcal meningitis (15/38 = 39%). Common serotypes accounting for 76% of S. pneumoniae were type 14, 19F, 20, 1, 18 and 5; and serotypes 14, 19F and 7 (accounting for 17% of strains) showed intermediate resistance to penicillin G. 97% of the H. influenzae isolates were type b, and in only two cases beta-lactamase-producing. 72% of isolates of the S. pneumoniae we identified belong to serotypes preventable by a 9-valent vaccine. Our study highlights the possibility of resistant pyogenic meningitis in children in Ethiopia due to emerging resistant strains of S. pneumoniae and H. influenzae isolates.

The Butajira rural health project in Ethiopia: mortality pattern of the under fives.

A total of 492 deaths of children below 5 years of age were registered during a 2-year period of demographic surveillance in a rural population of Ethiopia, where an epidemiologic study base population of 28,780 individuals was established in 1987. Data were collected by lay-reporters using a verbal autopsy method. The under-five cumulative mortality rate was 209 per 1000 children. When sub-divided into infants and children 1-4 years, the respective yearly mortality rates were 101 and 32.3 per 1000. There were considerable variations within the district by Peasants' Associations. Rural Lowlands experienced the highest mortality rates, especially for children 1-4 years. Mortality trends over a 2-year period indicate a significant increase for the child population, but not for infants. Similar trends were observed for boys and girls although the rates for boys were generally higher especially during infancy. More deaths occurred in the months of April, June, and July, and October and November indicating two peak seasons in both years. More deaths occurred in Peasant's Associations that were furthest from the health centre. Major probable causes of death were acute respiratory infections, measles, and diarrhoea. It is concluded that even in rural areas of a developing country it is possible to collect from mothers the much needed and valid fertility and mortality data through epidemiological surveillance by using lay-reporters.

Rapid etiological diagnosis of pyogenic meningitis by coagglutination, latex agglutination and immunoosmophoresis of cerebrospinal fluid, serum and urine.

Three rapid diagnostic methods for the detection of bacterial antigen-COA (Coagglutination), LA (latex agglutination) and IEOP (immunoelectroosmophoresis); DM (direct microscopy after Gram staining) and culture of cerebrospinal fluid (CSF) have been compared in 53 bacterial meningitis patients and in 13 controls. Urine samples were assayed for antigens in connection with CSF in 32 for COA, LA, and IEOP. The cause of meningitis was S. pneumoniae in 31, H. influenzae type b in 16 and N. meningitidis in 6 using the methods for diagnosis by LA in 100%, COA in 94%, culture in 89%, IEOP in 81% and DM in 78% of the CSF samples. None of the CSF from controls were positive by any of the methods. LA and COA detected antigen in CSF in all 4 patients on 4th and 5th day of treatment but none of the tests were helpful 7-10 days later. LA was effective in detecting antigen in urine in 10/10 (100%) H. influenzae type b, 17/20 (85%) S. pneumoniae prior to antibiotic treatment. IEOP was helpful in antigen detection better in serum than in urine in 5/6 H. influenzae type b and 9/11 S. pneumoniae meningitis. In general we have found LA and COA as simple, rapid specific and sensitive tests and can be applied even in field situations and that LA can be used for antigen detection in urine and IEOP in serum. A large scale study is needed in order to standardize methods for antigen detection in serum and urine.

The Butajira project in Ethiopia: a nested case-referent study of under-five mortality and its public health determinants.

During one year of follow-up, 306 deaths of children under the age of 5 years were included in a concurrent case-referent study that was based on a population estimated at 28,780 in 1987. A total of 612 live referents, matched for age, sex and study area, were also selected from the study population through density sampling. Data were collected by lay reporters by verbal autopsy. For the study period the estimated cumulative under-five mortality rate was 293 and the infant (0-11 months old) mortality rate was 136 per 1000. Major probable causes of death were diarrhoeal disease or acute respiratory infections (ARI). The relative importance of parental and environmental characteristics was assessed using conditional multiple logistic regression analysis. Under-five mortality was associated with paternal illiteracy, maternal ethnicity, and not being in the committee of people's organizations. Parental factors affected the infants relatively more than they did the children, especially with regard to ARI mortality. This was also noted with "absence of window", a proxy measure for evaluating the type of housing. In terms of etiological fractions a greater number of under-five deaths could be ascribed to parental than environmental conditions, with relatively more infants being affected than children.

PIP: In Butajira district, Ethiopia (130 km south of Addis Ababa), 2 trained field workers went to each household of 306 5 children who died between October 1988 and September 1989 and of 612 age, sex, and study area matched referents to interview caretakers as part of a study analyzing determinants of 5 mortality, especially parental and environmental determinants. Leading causes of infant and child deaths were acute respiratory infections (ARI) (33 and 20%, respectively) and diarrheal disease (23 and 32%, respectively). The multivariate analysis showed that Silti ethnicity of the mother (odds ration = 1.74), paternal illiteracy (OR = 1.45), and nonmembership in people's organizations (OR = 1.95) were parental risk factors. All children with all 3 parental factors were at a 4.6 times higher risk of 5 mortality than those with no parental risk factors. Infants with 3 parental risk factors were at a considerably greater mortality risk than were 1-4 year old children with 3 parental risk factors (OR, 10.7 vs. 2.5). Children with 3 parental risk factors living in the highlands faced a higher risk of death than those in the lowlands (15.5 vs. 2.8). 3 parental risk factors carried a higher risk of mortality pursuant to ARI than to diarrhea (14.8 vs. 2.4). The only environmental risk factor was no window in the house (OR = 1.54). Absence of windows was especially risky for infants (OR = 2.7) and for children with ARI (OR = 1.8 vs. 1.3 with diarrhea). Lack of windows was linked to crowding and poor housing. Public health interventions to eliminate parental or environmental factors would result in 55% or 31% fewer deaths, respectively. These findings suggest that parental factors have a greater impact on 5 mortality than environmental factors and that both parental and environmental factors and that both parental and environmental factors have a greater impact on infant than on child mortality.

The Butajira Rural Health Project in Ethiopia: a nested case-referent (control) study of under-5 mortality and its health and behavioural determinants.

The associations between under-5 mortality and its health and behavioural determinants were investigated in a rural district of Ethiopia, by means of a concurrent case-referent technique nested within a study-base population established in 1987. Three-hundred and six infant and child deaths registered over a period of 1 year, were contrasted with 612 controls, matched for age, sex and study area. Data were collected by trained non-medical workers using a structured questionnaire. Breastfeeding and supplementary feeding came out as strongly protective against under-5 mortality, even when controlling for parental and environmental determinants. Early termination of breastfeeding was demonstrated to have a substantial impact on mortality, particularly on that caused by diarrhoea. Late introduction of supplementary feeding, particularly of protein origin, was also associated with increased under-5 mortality. When the relative impacts of parental, environmental and behavioural determinants are compared, the greater impact of parental factors can be demonstrated, especially among infants.

Evaluation of clinical pallor in the identification and treatment of children with moderate and severe anaemia.

BACKGROUND: Anaemia from malaria is a common problem in developing countries. Blood transfusion in developing countries is available in few hospitals. Children who are severely anaemic and may require urgent blood transfusion usually present to peripheral first-level health facilities from where they must be referred to hospitals. Since most peripheral facilities do not determine haemoglobin levels, the decision on referral has to be made on clinical grounds. OBJECTIVES: To evaluate the sensitivity and specificity of clinical pallor of the palms, nailbeds, conjunctivae, buccal mucosa or tongue against haemoglobin values and their reproducibility among health workers. METHODS: A total of 2540 children 2 months to 5 years of age presenting to a rural health centre in Ethiopia were enrolled. Clinically detected pallor was compared with measured blood haemoglobin concentrations. RESULTS: Any anaemia (haemoglobin < 11 g/dl) was found in 61% of the children. Severe anaemia (haemoglobin < 5 g/dl) was found in 4%. The presence of any pallor clinically correlated with moderate anaemia (haemoglobin level < 8 g/dl) could be detected with a sensitivity of 95% and a specificity of 64-68% when the palm and nailbeds were used and a sensitivity of 84% and a specificity of 81% when the conjunctivae were used. Severe anaemia was detected clinically as severe pallor in 50-56% of cases (with a specificity of 95-96%). Agreement between physicians was highest for conjunctivae and nailbed pallor (87%) and lowest for palm pallor (73%). Using multivariate analysis, identification of a systolic ejection murmur or altered sensorium, the presence of splenomegaly or malarial parasitaemia were independently predictive of severe and moderately severe anaemia. CONCLUSIONS: Moderate and severe anaemia can be identified clinically in most cases for treatment and referral purposes. A systolic ejection murmur, altered sensorium, the presence of splenomegaly or malarial parasitaemia may be used as additional tools in considering urgent referral for blood transfusion.

A controlled estimate of the risk of HIV infection in Ethiopian children with tuberculosis.

To quantify the risk of human immunodeficiency virus (HIV) infection in children with tuberculosis (TB) a hospital-based, 1-year prospective, case-control study was performed in Addis Ababa, Ethiopia. Children with TB were compared to a control group of children admitted for elective surgery. The control group was also compared to a recent census of the background population. The crude odds ratio for HIV infection was 8.6 (95 % CI 2.2-73). After adjustment for possible confounders in a multivariate regression model, the odds ratio for HIV infection in children with TB was found to be 12.7 (95% CI 2.9-55). Of several independent determinants of TB assessed in the study, this association was the strongest. Until HIV transmission has reached its peak, an increasing burden of dual infection among Ethiopian children is to be expected.