The impact of home-based HIV counseling and testing on care-seeking and incidence of common infectious disease syndromes in rural western Kenya.

BACKGROUND: In much of Africa, most individuals living with HIV do not know their status. Home-based counseling and testing (HBCT) leads to more HIV-infected people learning their HIV status. However, there is little data on whether knowing one's HIV-positive status necessarily leads to uptake of HIV care, which could in turn, lead to a reduction in the prevalence of common infectious disease syndromes. METHODS: In 2008, Kenya Medical Research Institute (KEMRI) in collaboration with the Centers for Disease Control and Prevention (CDC) offered HBCT to individuals (aged ≥13 years) under active surveillance for infectious disease syndromes in Lwak in rural western Kenya. HIV test results were linked to morbidity and healthcare-seeking data collected by field workers through bi-weekly home visits. We analyzed changes in healthcare seeking behaviors using proportions, and incidence (expressed as episodes per person-year) of acute respiratory illness (ARI), severe acute respiratory illness (SARI), acute febrile illness (AFI) and diarrhea among first-time HIV testers in the year before and after HBCT, stratified by their test result and if HIV-positive, whether they sought care at HIV Patient Support Centers (PSCs). RESULTS: Of 9,613 individuals offered HBCT, 6,366 (66%) were first-time testers, 698 (11%) of whom were HIV-infected. One year after HBCT, 50% of HIV-infected persons had enrolled at PSCs - 92% of whom had started cotrimoxazole and 37% of those eligible for antiretroviral treatment had initiated therapy. Among HIV-infected persons enrolled in PSCs, AFI and diarrhea incidence decreased in the year after HBCT (rate ratio [RR] 0.84; 95% confidence interval [CI] 0.77 - 0.91 and RR 0.84, 95% CI 0.73 - 0.98, respectively). Among HIV-infected persons not attending PSCs and among HIV-uninfected persons, decreases in incidence were significantly lower. While decreases also occurred in rates of respiratory illnesses among HIV-positive persons in care, there were similar decreases in the other two groups. CONCLUSIONS: Large scale HBCT enabled a large number of newly diagnosed HIV-infected persons to know their HIV status, leading to a change in care seeking behavior and ultimately a decrease in incidence of common infectious disease syndromes through appropriate treatment and care.

Epidemiology of respiratory syncytial virus infection in rural and urban Kenya.

BACKGROUND: Information on the epidemiology of respiratory syncytial virus (RSV) infection in Africa is limited for crowded urban areas and for rural areas where the prevalence of malaria is high. METHODS: At referral facilities in rural western Kenya and a Nairobi slum, we collected nasopharyngeal/oropharyngeal (NP/OP) swab specimens from patients with influenza-like illness (ILI) or severe acute respiratory illness (SARI) and from asymptomatic controls. Polymerase chain reaction assays were used for detection of viral pathogens. We calculated age-specific ratios of the odds of RSV detection among patients versus the odds among controls. Incidence was expressed as the number of episodes per 1000 person-years of observation. RESULTS: Between March 2007 and February 2011, RSV was detected in 501 of 4012 NP/OP swab specimens (12.5%) from children and adults in the rural site and in 321 of 2744 NP/OP swab specimens (11.7%) from those in the urban site. Among children aged <5 years, RSV was detected more commonly among rural children with SARI (odds ratio [OR], 2.0; 95% confidence interval [CI], 1.2-3.3), urban children with SARI (OR, 8.5; 95% CI, 3.1-23.6), and urban children with ILI (OR, 3.4; 95% CI, 1.2-9.6), compared with controls. The incidence of RSV disease was highest among infants with SARI aged <1 year (86.9 and 62.8 episodes per 1000 person-years of observation in rural and urban sites, respectively). CONCLUSIONS: An effective RSV vaccine would likely substantially reduce the burden of respiratory illness among children in rural and urban areas in Africa.

Association of the CT values of real-time PCR of viral upper respiratory tract infection with clinical severity, Kenya.

Quantitative real-time polymerase chain reaction (qRT-PCR) assay of the upper respiratory tract is used increasingly to diagnose lower respiratory tract infections. The cycle threshold (CT ) values of qRT-PCR are continuous, semi-quantitative measurements of viral load, although interpretation of diagnostic qRT-PCR results are often categorized as positive, indeterminate, or negative, obscuring potentially useful clinical interpretation of CT values. From 2008 to 2010, naso/oropharyngeal swabs were collected from outpatients with influenza-like illness, inpatients with severe respiratory illness, and asymptomatic controls in rural Kenya. CT values of positive specimens (i.e., CT values < 40.0) were compared by clinical severity category for five viruses using Mann-Whitney U-test and logistic regression. Among children <5 years old we tested with respiratory syncytial virus (RSV), inpatients had lower median CT values (27.2) than controls (35.8, P = 0.008) and outpatients (34.7, P < 0.001). Among children and older patients infected with influenza virus, outpatients had the lowest median CT values (29.8 and 24.1, respectively) compared with controls (P = 0.193 for children, P < 0.001 for older participants) and inpatients (P = 0.009 for children, P < 0.001 for older participants). All differences remained significant in logistic regression when controlling for age, days since onset, and coinfection. CT values were similar for adenovirus, human metapneumovirus, and parainfluenza virus in all severity groups. In conclusion, the CT values from the qRT-PCR of upper respiratory tract specimens were associated with clinical severity for some respiratory viruses.

Viral and bacterial causes of severe acute respiratory illness among children aged less than 5 years in a high malaria prevalence area of western Kenya, 2007-2010.

BACKGROUND: Few comprehensive data exist on the etiology of severe acute respiratory illness (SARI) among African children. METHODS: From March 1, 2007 to February 28, 2010, we collected blood for culture and nasopharyngeal and oropharyngeal swabs for real-time quantitative polymerase chain reaction for 10 viruses and 3 atypical bacteria among children aged <5 years with SARI, defined as World Health Organization-classified severe or very severe pneumonia or oxygen saturation <90%, who visited a clinic in rural western Kenya. We collected swabs from controls without febrile or respiratory symptoms. We calculated odds ratios for infection among cases, adjusting for age and season in logistic regression. We calculated SARI incidence, adjusting for healthcare seeking for SARI in the community. RESULTS: Two thousand nine hundred seventy-three SARI cases were identified (54% inpatient, 46% outpatient), yielding an adjusted incidence of 56 cases per 100 person-years. A pathogen was detected in 3.3% of noncontaminated blood cultures; non-typhi Salmonella (1.9%) and Streptococcus pneumoniae (0.7%) predominated. A pathogen was detected in 84% of nasopharyngeal/oropharyngeal specimens, the most common being rhino/enterovirus (50%), respiratory syncytial virus (RSV, 22%), adenovirus (16%) and influenza viruses (8%). Only RSV and influenza viruses were found more commonly among cases than controls (odds ratio 2.9, 95% confidence interval: 1.3-6.7 and odds ratio 4.8, 95% confidence interval: 1.1-21, respectively). Incidence of RSV, influenza viruses and S. pneumoniae were 7.1, 5.8 and 0.04 cases per 100 person-years, respectively. CONCLUSIONS: Among Kenyan children with SARI, RSV and influenza virus are the most likely viral causes and pneumococcus the most likely bacterial cause. Contemporaneous controls are important for interpreting upper respiratory tract specimens.

Additional diagnostic yield of adding serology to PCR in diagnosing viral acute respiratory infections in Kenyan patients 5 years of age and older.

The role of serology in the setting of PCR-based diagnosis of acute respiratory infections (ARIs) is unclear. We found that acute- and convalescent-phase paired-sample serologic testing increased the diagnostic yield of naso/oropharyngeal swabs for influenza virus, respiratory syncytial virus (RSV), human metapneumovirus, adenovirus, and parainfluenza viruses beyond PCR by 0.4% to 10.7%. Although still limited for clinical use, serology, along with PCR, can maximize etiologic diagnosis in epidemiologic studies.

Epidemiology, seasonality, and burden of influenza and influenza-like illness in urban and rural Kenya, 2007-2010.

BACKGROUND: The epidemiology and burden of influenza remain poorly defined in sub-Saharan Africa. Since 2005, the Kenya Medical Research Institute and Centers for Disease Control and Prevention-Kenya have conducted population-based infectious disease surveillance in Kibera, an urban informal settlement in Nairobi, and in Lwak, a rural community in western Kenya. METHODS: Nasopharyngeal and oropharyngeal swab specimens were obtained from patients who attended the study clinic and had acute lower respiratory tract (LRT) illness. Specimens were tested for influenza virus by real-time reverse-transcription polymerase chain reaction. We adjusted the incidence of influenza-associated acute LRT illness to account for patients with acute LRT illness who attended the clinic but were not sampled. RESULTS: From March 2007 through February 2010, 4140 cases of acute LRT illness were evaluated in Kibera, and specimens were collected from 1197 (27%); 319 (27%) were positive for influenza virus. In Lwak, there were 6733 cases of acute LRT illness, and specimens were collected from 1641 (24%); 359 (22%) were positive for influenza virus. The crude and adjusted rates of medically attended influenza-associated acute LRT illness were 6.9 and 13.6 cases per 1000 person-years, respectively, in Kibera, and 5.6 and 23.0 cases per 1000 person-years, respectively, in Lwak. In both sites, rates of influenza-associated acute LRT illness were highest among children <2 years old and lowest among adults ≥50 years old. CONCLUSION: In Kenya, the incidence of influenza-associated acute LRT illness was high in both rural and urban settings, particularly among the most vulnerable age groups.

Etiology and Incidence of viral and bacterial acute respiratory illness among older children and adults in rural western Kenya, 2007-2010.

BACKGROUND: Few comprehensive data exist on disease incidence for specific etiologies of acute respiratory illness (ARI) in older children and adults in Africa. METHODOLOGY/PRINCIPAL FINDINGS: From March 1, 2007, to February 28, 2010, among a surveillance population of 21,420 persons >5 years old in rural western Kenya, we collected blood for culture and malaria smears, nasopharyngeal and oropharyngeal swabs for quantitative real-time PCR for ten viruses and three atypical bacteria, and urine for pneumococcal antigen testing on outpatients and inpatients meeting a ARI case definition (cough or difficulty breathing or chest pain and temperature >38.0 °C or oxygen saturation <90% or hospitalization). We also collected swabs from asymptomatic controls, from which we calculated pathogen-attributable fractions, adjusting for age, season, and HIV-status, in logistic regression. We calculated incidence by pathogen, adjusting for health-seeking for ARI and pathogen-attributable fractions. Among 3,406 ARI patients >5 years old (adjusted annual incidence 12.0 per 100 person-years), influenza A virus was the most common virus (22% overall; 11% inpatients, 27% outpatients) and Streptococcus pneumoniae was the most common bacteria (16% overall; 23% inpatients, 14% outpatients), yielding annual incidences of 2.6 and 1.7 episodes per 100 person-years, respectively. Influenza A virus, influenza B virus, respiratory syncytial virus (RSV) and human metapneumovirus were more prevalent in swabs among cases (22%, 6%, 8% and 5%, respectively) than controls. Adenovirus, parainfluenza viruses, rhinovirus/enterovirus, parechovirus, and Mycoplasma pneumoniae were not more prevalent among cases than controls. Pneumococcus and non-typhi Salmonella were more prevalent among HIV-infected adults, but prevalence of viruses was similar among HIV-infected and HIV-negative individuals. ARI incidence was highest during peak malaria season. CONCLUSIONS/SIGNIFICANCE: Vaccination against influenza and pneumococcus (by potential herd immunity from childhood vaccination or of HIV-infected adults) might prevent much of the substantial ARI incidence among persons >5 years old in similar rural African settings.

Secondary household transmission of 2009 pandemic influenza A (H1N1) virus among an urban and rural population in Kenya, 2009-2010.

BACKGROUND: In Kenya, >1,200 laboratory-confirmed 2009 pandemic influenza A (H1N1) (pH1N1) cases occurred since June 2009. We used population-based infectious disease surveillance (PBIDS) data to assess household transmission of pH1N1 in urban Nairobi (Kibera) and rural Lwak. METHODS: We defined a pH1N1 patient as laboratory-confirmed pH1N1 infection among PBIDS participants during August 1, 2009-February 5, 2010, in Kibera, or August 1, 2009-January 20, 2010, in Lwak, and a case household as a household with a laboratory-confirmed pH1N1 patient. Community interviewers visited PBIDS-participating households to inquire about illnesses among household members. We randomly selected 4 comparison households per case household matched by number of children aged <5. Comparison households had a household visit 10 days before or after the matched patient symptom onset date. We defined influenza-like illnesses (ILI) as self-reported cough or sore throat, and a self-reported fever ≤8 days after the pH1N1 patient's symptom onset in case households and ≤8 days before selected household visit in comparison households. We used the Cochran-Mantel-Haenszel test to compare proportions of ILIs among case and comparison households, and log binomial-model to compare that of Kibera and Lwak. RESULTS: Among household contacts of patients with confirmed pH1N1 in Kibera, 4.6% had ILI compared with 8.2% in Lwak (risk ratio [RR], 0.5; 95% confidence interval [CI], 0.3-0.9). Household contacts of patients were more likely to have ILIs than comparison-household members in both Kibera (RR, 1.8; 95% CI, 1.1-2.8) and Lwak (RR, 2.6; 95% CI, 1.6-4.3). Overall, ILI was not associated with patient age. However, ILI rates among household contacts were higher among children aged <5 years than persons aged ≥5 years in Lwak, but not Kibera. CONCLUSIONS: Substantial pH1N1 household transmission occurred in urban and rural Kenya. Household transmission rates were higher in the rural area.

The population-based burden of influenza-associated hospitalization in rural western Kenya, 2007-2009.

OBJECTIVE: To estimate the burden and age-specific rates of influenza-associated hospitalization in rural western Kenya. METHODS: All 3924 patients with respiratory illness (defined as acute cough, difficulty in breathing or pleuritic chest pain) who were hospitalized between June 2007 and May 2009 in any inpatient health facility in the Kenyan district of Bondo were enrolled. Nasopharyngeal and oropharyngeal swabs were collected and tested for influenza viruses using real-time reverse transcriptase polymerase chain reaction (RT-PCR). In the calculation of annual rates, adjustments were made for enrolled patients who did not have swabs tested for influenza virus. FINDINGS: Of the 2079 patients with tested swabs, infection with influenza virus was confirmed in 204 (10%); 176, 27 and 1 were found to be RT-PCR-positive for influenza A virus only, influenza B virus only, and both influenza A and B viruses, respectively. Among those tested for influenza virus, 6.8% of the children aged < 5 years and 14.0% of the patients aged ≥ 5 years were found positive. The case-fatality rate among admitted patients with PCR-confirmed infection with influenza virus was 2.0%. The annual rate of hospitalization (per 100,000 population) was 699.8 among patients with respiratory illness and 56.2 among patients with influenza (with 143.7, 18.8, 55.2, 65.1 and 57.3 hospitalized patients with influenza virus per 100,000 people aged < 5, 5-19, 20-34, 35-49 and ≥ 50 years, respectively). CONCLUSION: In a rural district of western Kenya, the rate of influenza-associated hospitalization was highest among children aged less than 5 years.

Differing burden and epidemiology of non-Typhi Salmonella bacteremia in rural and urban Kenya, 2006-2009.

BACKGROUND: The epidemiology of non-Typhi Salmonella (NTS) bacteremia in Africa will likely evolve as potential co-factors, such as HIV, malaria, and urbanization, also change. METHODS: As part of population-based surveillance among 55,000 persons in malaria-endemic, rural and malaria-nonendemic, urban Kenya from 2006-2009, blood cultures were obtained from patients presenting to referral clinics with fever ≥38.0°C or severe acute respiratory infection. Incidence rates were adjusted based on persons with compatible illnesses, but whose blood was not cultured. RESULTS: NTS accounted for 60/155 (39%) of blood culture isolates in the rural and 7/230 (3%) in the urban sites. The adjusted incidence in the rural site was 568/100,000 person-years, and the urban site was 51/100,000 person-years. In both sites, the incidence was highest in children <5 years old. The NTS-to-typhoid bacteremia ratio in the rural site was 4.6 and in the urban site was 0.05. S. Typhimurium represented >85% of blood NTS isolates in both sites, but only 21% (urban) and 64% (rural) of stool NTS isolates. Overall, 76% of S. Typhimurium blood isolates were multi-drug resistant, most of which had an identical profile in Pulse Field Gel Electrophoresis. In the rural site, the incidence of NTS bacteremia increased during the study period, concomitant with rising malaria prevalence (monthly correlation of malaria positive blood smears and NTS bacteremia cases, Spearman's correlation, p = 0.018 for children, p = 0.16 adults). In the rural site, 80% of adults with NTS bacteremia were HIV-infected. Six of 7 deaths within 90 days of NTS bacteremia had HIV/AIDS as the primary cause of death assigned on verbal autopsy. CONCLUSIONS: NTS caused the majority of bacteremias in rural Kenya, but typhoid predominated in urban Kenya, which most likely reflects differences in malaria endemicity. Control measures for malaria, as well as HIV, will likely decrease the burden of NTS bacteremia in Africa.

Risk factors for hospitalized seasonal influenza in rural western Kenya.

BACKGROUND: Risk factors for influenza hospitalization in Africa are unknown, including the role of HIV. METHODS: We conducted a case-control study of risk factors for hospitalized seasonal influenza among persons in rural western Kenya, a high HIV prevalence area, from March 2006-August 2008. Eligible cases were ≥five years old, admitted to health facilities with respiratory symptoms, and had nasopharyngeal/oropharyngeal swab specimens that tested positive for influenza A or B by real-time reverse transcription-PCR. Three randomly selected age-, sex- and neighborhood-matched controls were enrolled per case. A structured questionnaire was administered and home-based HIV testing was performed. Risk factors were evaluated using conditional logistic regression. RESULTS: A total of 64 cases (38 with influenza A and 26 with influenza B) and 190 controls were enrolled. The median age was 16 years (range 5-69 years). Among cases, 24.5% were HIV-infected versus 12.5% of controls (p = 0.004). Among persons ≥18 years old, 13 (59%) of 22 tested cases were HIV-positive compared with 15 (24%) of 62 tested controls (p = 0.005). In multivariable analysis, HIV-infection was associated with hospitalization due to influenza [adjusted Odds Ratio (aOR) 3.56, 95% CI 1.25-10.1]. The mean CD4 count among HIV-infected cases and controls was similar (399 vs. 387, respectively, p = 0.89). Chronic lung disease (aOR 6.83, 95% CI 1.37-34.0) was also associated with influenza hospitalization in multivariable analysis. Active pulmonary tuberculosis was associated with influenza hospitalization in bivariate, but not multivariable, analysis. CONCLUSIONS: People with HIV infection and chronic lung disease were at increased risk of hospitalized influenza in rural Kenya. HIV infection is common in many parts of sub-Saharan Africa. Influenza vaccine might prevent severe influenza in these risk groups.

The burden of common infectious disease syndromes at the clinic and household level from population-based surveillance in rural and urban Kenya.

BACKGROUND: Characterizing infectious disease burden in Africa is important for prioritizing and targeting limited resources for curative and preventive services and monitoring the impact of interventions. METHODS: From June 1, 2006 to May 31, 2008, we estimated rates of acute lower respiratory tract illness (ALRI), diarrhea and acute febrile illness (AFI) among >50,000 persons participating in population-based surveillance in impoverished, rural western Kenya (Asembo) and an informal settlement in Nairobi, Kenya (Kibera). Field workers visited households every two weeks, collecting recent illness information and performing limited exams. Participants could access free high-quality care in a designated referral clinic in each site. Incidence and longitudinal prevalence were calculated and compared using Poisson regression. RESULTS: INCIDENCE RATES RESULTING IN CLINIC VISITATION WERE THE FOLLOWING: ALRI--0.36 and 0.51 episodes per year for children <5 years and 0.067 and 0.026 for persons ≥ 5 years in Asembo and Kibera, respectively; diarrhea--0.40 and 0.71 episodes per year for children <5 years and 0.09 and 0.062 for persons ≥ 5 years in Asembo and Kibera, respectively; AFI--0.17 and 0.09 episodes per year for children <5 years and 0.03 and 0.015 for persons ≥ 5 years in Asembo and Kibera, respectively. Annually, based on household visits, children <5 years in Asembo and Kibera had 60 and 27 cough days, 10 and 8 diarrhea days, and 37 and 11 fever days, respectively. Household-based rates were higher than clinic rates for diarrhea and AFI, this difference being several-fold greater in the rural than urban site. CONCLUSIONS: Individuals in poor Kenyan communities still suffer from a high burden of infectious diseases, which likely hampers their development. Urban slum and rural disease incidence and clinic utilization are sufficiently disparate in Africa to warrant data from both settings for estimating burden and focusing interventions.

Health-seeking patterns among participants of population-based morbidity surveillance in rural western Kenya: implications for calculating disease rates.

BACKGROUND: Calculation of disease rates in developing countries using facility-based surveillance is affected by patterns of health utilization. We describe temporal patterns in health care seeking by syndrome as part of population-based morbidity surveillance in rural western Kenya. METHODS: From July 2006 to June 2008, health utilization data were collected from 27 171 participants at biweekly home visits and at Lwak Hospital, the designated referral clinic where free care provided by dedicated study clinical staff was available. Ill persons were asked if and where they sought care. Proportions seeking care for children and adults with fever, acute respiratory infection (ARI), acute lower respiratory infection (ALRI), and diarrhea were compared by Chi-square test. Care-seeking by distance was evaluated by logistic regression. RESULTS: While care-seeking outside the home was common for all syndromes (>50%), only 18-38% of care-seeking was to health facilities. Children were more likely than adults to visit health facilities for all syndromes. Of ill persons visiting Lwak Hospital, 45-54% had previously sought care elsewhere, mostly from informal drug sellers, and 11-24% with fever, ARI, or ALRI had already taken an antimalarial or antibiotic. The distance from the participant's home to Lwak Hospital was the most common reason (71%) for ill participants not seeking care there. The likelihood of visiting Lwak decreased with increasing distance of residence (p<0.001) and fluctuated significantly over the study period. CONCLUSIONS: Even in a study setting where free and reliable care is offered, health utilization is affected by other factors, such as distance. Health utilization data in population-based surveillance are important in adjusting disease rates.

High rate of pneumococcal bacteremia in a prospective cohort of older children and adults in an area of high HIV prevalence in rural western Kenya.

BACKGROUND: Although causing substantial morbidity, the burden of pneumococcal disease among older children and adults in Africa, particularly in rural settings, is not well-characterized. We evaluated pneumococcal bacteremia among 21,000 persons > or =5 years old in a prospective cohort as part of population-based infectious disease surveillance in rural western Kenya from October 2006-September 2008. METHODS: Blood cultures were done on patients meeting pre-defined criteria--severe acute respiratory illness (SARI), fever, and admission for any reason at a referral health facility within 5 kilometers of all 33 villages where surveillance took place. Serotyping of Streptococcus pneumoniae was done by latex agglutination and quellung reaction and antibiotic susceptibility testing was done using broth microdilution. We extrapolated incidence rates based on persons with compatible illnesses in the surveillance population who were not cultured. We estimated rates among HIV-infected persons based on community HIV prevalence. We projected the national burden of pneumococcal bacteremia cases based on these rates. RESULTS: Among 1,301 blood cultures among persons > or =5 years, 52 (4%) yielded pneumococcus, which was the most common bacteria isolated. The yield was higher among those > or =18 years than 5-17 years (6.9% versus 1.6%, p < 0.001). The highest yield was for inpatients with SARI (10%), compared with SARI outpatients (3%) and acute febrile outpatients (1%). Serotype 1 pneumococcus was most common (42% isolates) and 71% were serotypes included in the 10-valent pneumococcal conjugate vaccine (PCV10). Non-susceptibility to beta-lactam antibiotics was low (<5%), but to trimethoprim-sulfamethoxazole was high (>95%). The crude rate of pneumococcal bacteremia was 129/100,000 person-years, and the adjusted rate was 419/100,000 person-years. Nineteen (61%) of 31 patients with HIV results were HIV-positive. The adjusted rate among HIV-infected persons was 2,399/100,000 person-years (Rate ratio versus HIV-negative adults, 19.7, 95% CI 12.4-31.1). We project 58,483 cases of pneumococcal bacteremia will occur in Kenyan adults in 2010. CONCLUSIONS: Pneumococcal bacteremia rates were high among persons > or =5 years old, particularly among HIV-infected persons. Ongoing surveillance will document if expanded use of highly-active antiretroviral treatment for HIV and introduction of PCV10 for Kenyan children (anticipated in late 2010) result in substantial secondary benefits by reducing pneumococcal disease in adults.