Procalcitonin and cellulitis: correlation of procalcitonin blood levels with measurements of severity and outcome in patients with limb cellulitis.

BACKGROUND: Procalcitonin levels may be raised in bacterial infections and have been used to guide antibiotic therapy. There is little data on procalcitonin and limb cellulitis. OBJECTIVES: Within a clinical trial of antibiotic therapy, we examined the correlation between clinical observations, blood tests and local measurements of skin damage, with serum procalcitonin levels. METHODS: The data is from a subset of the patients recruited into a clinical trial of antibiotic therapy for cellulitis (clindamycin for cellulitis, NCT01876628) whose procalcitonin levels were correlated with clinical and laboratory measurements. We selected the variables strongly correlated with procalcitonin and evaluated the predictive value of the baseline procalcitonin on the primary trial outcome. RESULTS: 136 patients provided 307 procalcitonin levels which were correlated with 8 variables. The strongest correlations (correlation coefficient of >0.5) with procalcitonin were the affected skin area (0.537), C-reactive protein (0.574) and neutrophil:lymphocyte ratio (0.567). Receiver operator characteristic curves demonstrated poor sensitivity and specificity of procalcitonin in predicting primary outcome. Procalcitonin baseline levels were low but decreased as patients recovered. CONCLUSIONS: Procalcitonin levels are generally low in limb cellulitis and cannot be used to confirm the diagnosis or the need for antibiotic therapy. Procalcitonin is a poor predictor of early improvement.

Recent transmission of tuberculosis in a cohort of HIV-1-infected female sex workers in Nairobi, Kenya.

OBJECTIVES: To describe the epidemiological and clinical characteristics of HIV-related tuberculosis in a female cohort, and to investigate the relative importance of recently transmitted infection and reactivation in the pathogenesis of adult HIV-related tuberculosis. DESIGN: Members of an established cohort of female sex workers in Nairobi were enrolled in a prospective study. Women were followed up regularly and seen on demand when sick. METHODS: Between October 1989 and September 1992 we followed 587 HIV-infected and 132 HIV-seronegative women. Standard protocols were used to investigate common presentations. Cases of tuberculosis were identified clinically or by culture. All available Mycobacterium tuberculosis strains underwent DNA fingerprint analysis. RESULTS: Forty-nine incident and four recurrent episodes of tuberculosis were seen in HIV-infected women; no disease was seen in seronegative sex workers (P = 0.0003). The overall incidence rate of tuberculosis was 34.5 per 1000 person-years amongst HIV-infected participants. In purified protein derivative (PPD) skin test-positive women the rate was 66.7 per 1000 person-years versus 18.1 per 1000 person-years in PPD-negative women. Twenty incident cases (41%) were clinically compatible with primary disease. DNA fingerprint analysis of strains from 32 incident cases identified two clusters comprising two and nine patients; allowing for index cases, 10 patients (28%) may have had recently transmitted disease. Three out of 10 (30%) patients who were initially PPD skin test-negative became PPD-positive. Taken together, 26 incident cases (53%) may have been recently infected. DNA fingerprint analysis also identified two (50%) of the four recurrent tuberculosis episodes as reinfection. CONCLUSIONS: Substantial recent transmission of tuberculosis appears to be occurring in Nairobi amongst HIV-infected sex workers. It may be incorrect to assume in other regions of high tuberculosis transmission that active HIV-related tuberculosis usually represents reactivation of latent infection.

PIP: A 3-year (1989-92) prospective study of 587 HIV-positive and 132 HIV-negative commercial sex workers in Nairobi, Kenya, revealed substantial recent transmission of tuberculosis in the HIV-infected group. The cohort was enrolled at a community clinic that provides counseling, sexually transmitted disease services, and free condoms. In HIV-positive women, 49 incident and 4 recurrent episodes of tuberculosis were diagnosed during the study period; there were no tuberculosis cases among HIV-negative women. The overall incidence rate of tuberculosis was 34.5/1000 person-years among HIV-positive women. 20 incident cases (41%) met the clinical case definition of primary disease. DNA fingerprint analysis of strains from 32 incident cases suggested 10 women (28%) may have had recently transmitted disease. 3 of 10 women who were initially purified protein derivative (PPD) skin test-negative became PPD-positive. Clinical presentation, tuberculin skin testing, and strain clustering data all independently suggested that substantial Mycobacterium tuberculosis transmission was occurring in HIV-infected prostitutes during the study period. As many as 26 (53%) of the 49 patients with incident disease may have recently acquired tuberculosis and DNA fingerprint analysis identified 2 (50%) of the 4 recurrent tuberculosis episodes as reinfection. These findings challenge the assumption that tuberculosis in HIV-infected individuals represents reactivation of latent endogenous infection.

The risks and benefits of childhood bacille Calmette-Guérin immunization among adults with AIDS. International MAC study groups.

OBJECTIVE: To define the risks of disseminated bacille Calmette-Guérin (BCG) or disseminated Mycobacterium tuberculosis in adults with AIDS who were immunized with BCG in childhood. DESIGN: HIV-infected patients with CD4 < 200 x 10(6)/l were enrolled from five study sites (New Hampshire, Boston, Finland, Trinidad and Kenya). Prior BCG immunization was determined and blood cultures for mycobacteria were obtained at study entry and at 6 months. Acid-fast bacilli were identified as Mycobacterium tuberculosis complex (MTBC) using DNA probes. MTBC isolates were then typed by both IS6110 restriction fragment length polymorphism and polymerase chain reaction/restriction enzyme analysis. SETTING: Most patients in New Hampshire and Finland were outpatients; most patients in Trinidad were inpatients with terminal illness; and most patients in Kenya were outpatients, although 44 were inpatients with terminal illness. PARTICIPANTS: A total of 566 patients were enrolled, including 155 with childhood BCG immunization; 318 patients had a single study visit and culture, and 248 patients had two study visits and cultures. MAIN OUTCOME MEASURES: Isolation and identification of mycobacteria from blood cultures. RESULTS: Blood cultures were positive for MTBC in 21 patients; none were positive for M. bovis BCG, and 21 were M. tuberculosis-positive. In Trinidad, seven (87%) out of eight isolates of M. tuberculosis were indistinguishable by IS6110 typing; BCG immunization was associated with a decreased risk of bacteremic infection with M. tuberculosis (P = 0.05). CONCLUSIONS: The risk of disseminated BCG among adult AIDS patients with childhood BCG immunization is very low. Childhood BCG immunization is associated with protection against bacteremia with M. tuberculosis among adults with advanced AIDS in Trinidad.

Microbiology of HIV associated bacteraemia and diarrhoea in adults from Nairobi, Kenya.

We undertook a retrospective descriptive comparison of the spectrum of pathogens responsible for bacteraemia and diarrhoea in HIV antibody positive and negative patients over 4 years (1988-92), in Nairobi, Kenya. The study population was recruited from primary to tertiary centres of clinical care and consisted of 2858 adults (15 years or older). There were 415 significant blood culture isolates, 192 from 1785 HIV negative patients and 223 from 953 HIV positive patients. There were 233 significant faecal isolates, 22 from 115 HIV negative patients and 211 from 531 HIV positive patients. The most common pathogens detected in blood were Streptococcus pneumoniae and Salmonella typhimurium and in faeces Shigella flexneri, S. typhimurium and Cryptosporidium parvum. The agents causing illness in HIV positive patients in Nairobi are similar to those prevalent in the HIV negative community and the investigation of a febrile illness with or without diarrhoea in an HIV positive patient should reflect this.

PIP: Researchers conducted a retrospective analysis of stool specimens from 646 adult patients and of blood cultures from 2738 adult patients to examine the etiology of opportunistic infection in HIV-positive individuals in Nairobi, Kenya, and to compare this etiology with the range of pathogens causing disease in the HIV-negative population. Adults at least 15 years old contributed the stool and blood samples that were received at the Wellcome Trust-Kenya Medical Research Institute during 1988-92. The 415 significant blood culture isolates comprised 192 from 1785 HIV-negative patients and 223 from 953 HIV-positive patients. The most frequently detected pathogens in blood included Streptococcus pneumoniae (58 in HIV-positive cases and 25 in HIV-negative cases) and Salmonella typhimurium (56 in HIV-positive cases; 5 in HIV-negative cases). There were 233 significant stool isolates, 211 from 531 HIV-positive patients and 22 from 115 HIV-negative patients. 20 blood cultures and 21 stool cultures had more than 1 significant pathogen. The most commonly detected organisms in the stools were Shigella flexneri (49 for HIV-positive cases and 9 in HIV-negative cases), S typhimurium (40 in HIV-positive cases and 3 in HIV-negative cases), and Cryptosporidium parvum (45 in HIV-positive cases and 0 in HIV-negative cases). With two exceptions, the spectrum of pathogens associated with infection in HIV-positive patients was the same as that for HIV-negative patients. Physicians should consider this when they investigate and manage febrile illness with or without diarrhea in an HIV-positive patient.

The international epidemiology of disseminated Mycobacterium avium complex infection in AIDS. International MAC Study Group.

OBJECTIVE: To determine rates of disseminated Mycobacterium avium complex (MAC) infection among AIDS patients in developed and developing countries, and to determine whether different rates reflect differences in exposure or immunity, or both. DESIGN: Prospective cohort study. SETTING: University hospitals and outpatient AIDS programs. METHODS: HIV-infected subjects with CD4 counts < 200 x 10(6)/l were interviewed and had CD4 lymphocyte counts, blood cultures for mycobacteria (baseline and at 6 months), and skin tests with purified protein derivative (PPD) and M. avium sensitin. RESULTS: Among 566 study patients rates of disseminated MAC were 10.5-21.6% in New Hampshire, Boston and Finland compared to 2.4-2.6% in Trinidad and Kenya (P < 0.001). PPD skin test reactions > or = 5 mm were present in 20% of patients from Kenya compared to 1% at other sites (P < 0.001). Among patients from the United States and Finland, multiple logistic regression indicated that occupational exposure to soil and water was associated with a decreased risk of disseminated MAC, whereas the following were associated with an increased risk of disseminated MAC: low CD4 count, swimming in an indoor pool, history of bronchoscopy, regular consumption of raw or partially cooked fish/shellfish and treatment with granulocyte colony-stimulating factor. CONCLUSIONS: Rates of disseminated MAC in AIDS are higher in developed than developing countries and are due to both differences in exposure and differences in immunity. These data provide a rationale for prevention of MAC through both active immunization and reduction in exposure to the organism.

Invasive pneumococcal disease in a cohort of predominantly HIV-1 infected female sex-workers in Nairobi, Kenya.

BACKGROUND: HIV infection is a major risk factor for pneumococcal disease in industrialised countries. Although both are common infections in sub-Saharan Africa, few studies have investigated the importance of this interaction. We have followed up a cohort of female sex-workers in Nairobi and report here on the extent of invasive pneumococcal disease. METHODS: A well-established cohort of low-class female sex-workers, based around a community clinic, was followed up from October, 1989, to September, 1992. 587 participants were HIV positive and 132 remained HIV negative. Set protocols were used to investigate common presentations. Cases were identified clinically and radiographically. Streptococcus pneumoniae and other pathogens were diagnosed by culture. FINDINGS: Seventy-nine episodes of invasive pneumococcal disease were seen in the 587 HIV-positive women compared with one episode in the 132 seronegative women (relative risk 17.8, 95% CI 2.5 to 126.5). In seropositive women the incidence rate was 42.5 per 1000 person-years and the recurrence rate was 264 per 1000 person-years. By serotyping, most recurrent events were re-infection. A wide spectrum of HIV-related pneumococcal disease was seen: only 56% of cases were pneumonia; sinusitis was seen in 30% of cases, and occult bacteraemia, a novel adult presentation, in 11%. Despite forty-two bacteraemic episodes, no deaths were attributable to Strep pneumoniae. At first presentation the mean CD4 cell count was 302/microL(SD 191) and was 171/microL (105) for recurrent episodes. During acute Strep pneumoniae infection the CD4 cell count was reversibly suppressed (mean fall in sixteen episodes, 105/microL [123]). The neutrophil response to acute infection was blunted and was correlated with CD4 count (r=0.50, 95% CI 0.29 to 0.66). Strep pneumoniae caused more disease, at an earlier stage of HIV immunosuppression, than Mycobacterium tuberculosis or non-typhi salmonellae. INTERPRETATION: Our study highlights the importance of the pneumococcus as an early but readily treatable complication of HIV infection in sub-Saharan Africa.

Disseminated Mycobacterium avium infection among HIV-infected patients in Kenya.

Previous studies from Africa have been unable to identify disseminated Mycobacterium avium complex (MAC) infection in patients with advanced human immunodeficiency virus (HIV) infection. We performed mycobacterial blood cultures and CD4 counts on 48 symptomatic adults with advanced HIV infection admitted to the hospital in Nairobi, Kenya over 4 weeks in 1992. Fourteen patients had mycobacteremia; these patients had significantly lower CD4 counts than the patients with negative cultures (14/mm3 vs. 85/mm3; p < 0.01). Three patients (6%) were bacteremic with M. avium (mean CD4 count, 10/mm3) and 11 (23%) were bacteremic with Mycobacterium tuberculosis complex (MTB) (mean CD4 count, 15/mm3). Thus, M. avium bacteremia was detected significantly less frequently in the study population than MTB bacteremia (p = 0.04). The minimum rate for HIV-associated disseminated M. avium infection in patients admitted to the hospital in Nairobi was estimated to be approximately 1%. Patients with mycobacteremia died or were discharged home sick before the diagnosis was made. Disseminated M. avium does occur in adults with advanced HIV infection in sub-Saharan Africa, but is less common than disseminated MTB.

Inadequate response to intradermal hepatitis A vaccine.

To assess the efficacy of the intradermal route of administration of hepatitis A vaccine we conducted a study in hospital laboratory workers. Forty-three volunteers were given three different combinations of intradermal and intramuscular hepatitis A vaccine and compared with 18 controls given intramuscular vaccine only. The geometric mean titres (GMT) after one, two and three intradermal doses of 0.1 ml each were 4.5, 28 and 143 IU l-1 respectively. The GMT after one intramuscular dose in the controls was 163 IU l-1. The results indicate that the response to intradermal hepatitis A vaccine is poor and its use cannot be recommended.

The effect of human immunodeficiency virus type-1 on the infectiousness of tuberculosis.

SETTING: Developing country tertiary referral hospital plus catchment community. OBJECTIVE: To determine the infectiousness of culture-confirmed pulmonary tuberculosis in patients infected with Human Immunodeficiency Virus type-1 (HIV-1). DESIGN: Comparison of the incidence of tuberculosis and the prevalence of tuberculin skin test positivity among the household contacts of both HIV-1 positive and negative cases with pulmonary tuberculosis. RESULTS: Of 255 contacts of HIV-1 negative index cases, 2 were HIV-1 positive and of 102 contacts of HIV-1 positive index cases, 14 were HIV-1 positive (odds ratio (OR) = 20.0 95% Confidence Interval (CI) 4.4-193). 21 cases of tuberculosis were diagnosed among contacts, of whom 3 were HIV-1 positive. The overall unadjusted OR for tuberculosis among contacts of HIV-1 positive index cases was 1.6 (95% CI 0.6-4.3) compared to contacts of HIV-1 negative index cases. Amongst HIV-1 negative contacts alone the OR was 1.5 (95% CI 0.4-4.4). In this group the best predictors of tuberculosis among contacts were female sex of the index case (OR = 3.4 95% CI 1.1-12), sharing the same bed as the index case (OR = 2.6 95% CI 0.9-7.4), and contact's age less than 5 years (OR = 3.3 95% CI 1.1-9.5). HIV-1 positive contacts were more likely to develop tuberculosis than HIV-1 negative contacts (OR = 4.1 95% CI 0.7-17). Tuberculin skin test positivity rates were the same among the HIV-1 negative contacts of HIV-1 positive and negative index cases (OR = 1.1 CI 0.7-1.6). CONCLUSIONS: HIV-1 associated pulmonary tuberculosis is not more infectious than tuberculosis alone. The presence of HIV-1 in a community does not mandate a change in the management of contacts of patients with pulmonary tuberculosis.

PIP: Using data on tuberculosis (TB) index cases over age 15 years seen at the Infectious Diseases Hospital in Nairobi and the Ngaira Avenue Chest Clinic over September 1, 1989 and October 10, 1990, and their contacts, the authors determined the infectiousness of culture-confirmed pulmonary TB in patients infected with HIV-1. Comparing the incidence of TB and the prevalence of tuberculin skin test positivity among the household contacts of HIV-1 positive and negative cases with pulmonary TB found HIV-1-associated pulmonary TB to be no more infectious than TB alone. The presence of HIV-1 in a community therefore does not require a change in the management of contacts of patients with pulmonary TB.

Evidence of previous infection with Mycobacterium avium-Mycobacterium intracellulare complex among healthy subjects: an international study of dominant mycobacterial skin test reactions.

Skin tests with 0.1 mL of intermediate-strength Mycobacterium tuberculosis purified protein derivative (PPD) and 0.1 mL of Mycobacterium avium sensitin were conducted on 484 healthy subjects from diverse geographic sites. Reactions of > or = 5 mm to one antigen that exceeded the reaction to the other by > or = 3 mm were considered M. avium- or PPD-dominant. PPD-dominant reactions were more frequent at sites where routine Bacille Calmette-Guérin immunization is done or where there are high rates of tuberculosis: New Hampshire, 2%; Boston, 7%; Finland, 14%; Trinidad, 26%; and Kenya, 28%. However, rates of M. avium-dominant reactions ranged from 7% to 12% at all sites. Analysis of dominant reactions based on a more stringent 10-mm minimum reaction size showed similar trends. These data suggest that exposure to MAC is similar in developed and developing countries but that broad mycobacterial immunity is greater in developing countries and may contribute to the lower rates of disseminated MAC infections in AIDS in these areas.

Isolation of Mycobacterium avium complex from water in the United States, Finland, Zaire, and Kenya.

Disseminated infection with organisms of the Mycobacterium avium complex (MAC) is a common complication of AIDS in the United States and other developing countries, but it is rare or absent in sub-Saharan Africa. To assess the comparative likelihood of exposure to MAC in these geographic areas, we used a standard protocol to culture 91 water samples from environmental sites and piped water supply systems in the United States, Finland, Zaire, and Kenya. MAC was isolated from all geographic areas and from 22 of 91 (24%) samples. Isolation rates were 13 of 47 (28%) for environmental samples and 9 of 44 (20%) for water supply samples. Overall isolation rates were 18 of 52 (35%) samples in the United States and Finland, whereas they were 4 of 39 (10%) samples in Zaire and Kenya (P = 0.015). MAC isolation rates from water supply systems were 8 of 25 (32%) samples in the United States and Finland and 1 of 19 (5%) samples in Zaire and Kenya (P = 0.056). MAC was isolated from hospital water in the United States and Finland but not in hospital water in Zaire and Kenya. Serovar determinations showed that six of eight isolates from the United States were serovar 4 or 8. One MAC isolate from Zaire was identified as an "X" mycobacterium. These data suggest that exposure to MAC in water is likely in diverse areas of the world, but that the likelihood of human exposure to the organism in water may be slightly less in sub-Saharan Africa than in developed countries in the Northern Hemisphere.

PIP: Between March 1990 and February 1992, microbiologists collected 91 water samples from various environmental sites (lakes, ponds, rivers, streams, harbors, marshes, and standing water) and from piped municipal and private water supply systems to determine the likelihood of human exposure to Mycobacterium avian complex (MAC) in New Hampshire and Boston in the US, Finland, Kenya, and Zaire. They wanted to examine the international distribution of MAC to determine whether the observation of AIDS patents in Africa not having MAC infection is association with differences in the environmental distribution of MAC. Overall isolation rates for environmental samples and for water supply samples stood at 28% and 20%, respectively. MAC isolation rates for all samples in the 2 developed countries were significantly higher than they were in the 2 Sub-Saharan African countries (35% vs. 10%; p = .015). The rates for water supply systems were higher in the US and Finland than they were in Kenya and Zaire (32% vs. 5%; p = .056). None of the water supply samples from hospitals in Kenya and Zaire tested positive for MAC, while about 20% in the US and 50% in Finland did. Serovars 4 and 8 of M. avian, which have been linked to infection in AIDS patients, accounted for 75% of the environmental M. avium isolates in the US. An X mycobacterium was found in an MAC isolate from Zaire. These findings indicate that the probability of human exposure to MAC in water is less than Sub-Saharan Africa than it is in developed countries in the northern hemisphere.

Infection and morbidity in patients with tuberculosis in Nairobi, Kenya.

OBJECTIVE: To examine the role of acute infection as a cause of morbidity in patients with tuberculosis. DESIGN: Cross-sectional documentation of predefined acute morbid events. SETTING: Infectious Diseases Hospital, Nairobi, Kenya. PATIENTS: Adults (> or = 15 years), inpatients and outpatients with a diagnosis of tuberculosis presenting with one or more of a series of clinical features. A new event was defined as one occurring at least 1 week after the initial event. INTERVENTIONS: Patients' treatment was modified depending on the results of laboratory investigations. MAIN OUTCOME MEASURES: There were 642 events from 398 patients, 235 HIV-positive patients had 438 events and 163 HIV-negative patients had 204 events (P < 0.0001). Forty-two out of the 235 (18%) HIV-positive patients were bacteraemic compared with nine out of the 163 (6%) HIV-negative patients (P = 0.0003). The most common isolates from blood were Salmonella typhimurium and Streptococcus pneumoniae. RESULTS: Faecal specimens were obtained more commonly from HIV-positive patients (P < 0.001), and often contained bacterial pathogens. CONCLUSIONS: Many of the causes of morbidity in patients with tuberculosis and HIV are not due to tuberculosis or antituberculous therapy, and will not be identified without microbiological investigation.

PIP: Tuberculosis (TB) is a common complication of HIV in Africa. A 1988-89 study further confirmed that considerable morbidity and mortality from acute bacterial infection occurred in HIV patients. It has also been found that anti-TB therapy seems to be as effective in HIV-positive as in HIV-negative TB patients. This paper reports on the level and nature of infectious morbidity suffered by HIV-positive patients receiving treatment for TB. The assessment is based upon a sample of inpatients and outpatients at the Infectious Diseases Hospital in Nairobi. Patients were aged 15 years and older, with a TB diagnosis presenting with 1 or more of a series of clinical features. 642 morbid events were seen in 398 patients: 235 HIV-positive patients had 438 event and 163 HIV-negative patients had 204 events. 18% of the HIV-positive patients versus 6% of the HIV-negative patients were bacteremic. Salmonella typhimurium and Streptococcus pneumoniae were most commonly isolated from sera, while fecal specimens were obtained more commonly from HIV-positive patients and often contained bacterial pathogens. The authors conclude that many causes of morbidity in patients with TB and HIV are not due to TB or anti-TB therapy and will not be identified without microbiological investigation. These results suggest that even with effective anti-TB chemotherapy HIV-positive patients will remain or become unwell.

Quantitative bacillary response to treatment in HIV-associated pulmonary tuberculosis.

A group of 122 patients with culture-proven pulmonary tuberculosis were recruited to examine the concentrations of Mycobacterium tuberculosis in sputum and the relationship to HIV-1 antibody status. They were followed for up to 28 days from the start of antituberculous chemotherapy to assess the early bacillary response to two chemotherapeutic regimens. Of 67 treated with streptomycin, thiacetazone, and isoniazid 17 were HIV positive, and subsequently 55, of whom 20 were HIV positive, were treated with streptomycin, rifampin, isoniazid, and pyrazinamide. The mean initial concentration of M. tuberculosis in the sputum of the HIV-negative patients was significantly higher than in HIV-positive patients (6.95 and 6.34 log colony-forming units respectively; p = 0.019). The HIV-positive patients had less radiologic evidence of disease and significantly fewer zones of lung affected with cavities. The response to treatment was similar, but with HIV-positive patients more likely to become culture negative by 28 days. The differences that exist between HIV-positive and HIV-negative patients are minor, and standard regimens are at least as effective in HIV-positive patients in the first month of treatment.

Biochemical mis-identification of Brucella melitensis and subsequent laboratory-acquired infections.

Brucella species are mis-identified in the API 20NE system as Moraxella phenylpyruvica (profile number 1200004). Since some Brucella spp. grow readily in routine blood culture medium and may be isolated from patients without clinically obvious brucellosis, the risk of laboratory-acquired brucellosis exists. We describe two such cases.

The presentation and outcome of HIV-related disease in Nairobi.

The range of clinical presentations of HIV-related disease in Africa has not been adequately described, despite the fact that many hospitals have to rely heavily on clinical diagnosis. Six hundred adult medical patients seen in the Casualty Department of the main Government hospital in Nairobi were enrolled in a study of the presentation and outcome of HIV-related disease: 506 of these patients were admitted, of whom 19 per cent (95) were HIV seropositive. The remaining 94 were dealt with as outpatients: 11 percent (10) of these were seropositive. A history of prior treatment for sexually transmitted disease and, if male, being uncircumcised, were associated with being seropositive. Three presentations were strongly associated with HIV infection: acute fever with no focus except the gastrointestinal tract (enteric fever-like illness), acute cough with fever (community-acquired pneumonia) and chronic diarrhoea with wasting. The WHO clinical case definition (CCD) for AIDS missed a substantial amount of HIV-related morbidity (sensitivity 39 per cent) and misidentified many seronegative patients (positive predictive value 59 per cent). In comparison with the Centers for Disease Control surveillance definition for AIDS, the CCD was specific (91 per cent) and sensitive (79 per cent) but only had a positive predictive values of 30 per cent: the CCD may therefore be a poor surveillance tool for AIDS. Seropositive patients were much more likely to die than were seronegative patients (39 per cent vs. 15 per cent mortality). Enteric fever-like illness was the presentation which most commonly proved fatal. A wider spectrum of disease is associated with underlying HIV immunosuppression than has previously been described in Africa.

Extrapulmonary and disseminated tuberculosis in HIV-1-seropositive patients presenting to the acute medical services in Nairobi.

We studied 506 consecutive adult acute medical admissions to hospital in Nairobi; 95 (18.8%) were seropositive for HIV-1, and 43 new cases of active tuberculosis (TB) were identified. TB was clearly associated with HIV infection, occurring in 17.9% of seropositive patients compared with 6.3% of seronegatives [odds ratio (OR) 3.2; 95% confidence limits (CL) 1.6-6.5]. Extrapulmonary disease was more common in seropositive than seronegative TB patients (nine out of 17 versus five out of 26; OR 4.7; 95% CL 1.01-23.6); this accounted for most of the excess cases of TB seen in seropositive patients. Mycobacteraemia was demonstrated in two of eight seropositive TB patients but in none of 11 seronegative TB patients. No atypical mycobacteria were isolated. The World Health Organization (WHO) clinical case definition for African AIDS did not discriminate well between seropositive and seronegative TB cases. Five out of seven seropositive women with active tuberculosis had delivered children in the preceding 6 months and were lactating, compared with only one out of eight seronegative tuberculous women. An association between recent childbirth, HIV immunosuppression and the development of TB is suggested.

Life-threatening bacteraemia in HIV-1 seropositive adults admitted to hospital in Nairobi, Kenya.

During 6 months, 506 consecutive adult emergency admissions to hospital in Nairobi were enrolled in a study of bacteraemia and HIV infection. 19% were HIV-1 antibody positive. Significantly more HIV-seropositive than seronegative patients had bacteraemia (26% vs 6%). The predominant organisms isolated from the seropositive patients were Salmonella typhimurium and Streptococcus pneumoniae. Mortality was higher in the seropositive than in the seronegative bacteraemic patients. The findings suggest that non-opportunistic bacteria are important causes of morbidity and mortality in HIV-infected individuals in Africa.

Taxonomic investigation of Legionella pneumophila using monoclonal antibodies.

A panel of 19 monoclonal antibodies was used to produce patterns of immunofluorescent staining of 468 isolates of Legionella pneumophila. Twelve monoclonal antibodies were selected that divided L. pneumophila into 17 phenons which, in the majority of cases, conform to serogroup divisions. These phenons are more easily defined than the present serogroups, and isolates can be placed in them with little ambiguity. The standardized set of monoclonal antibodies was also used to define the subgroups of serogroup 1.