In recent years, surveillance has demonstrated a high burden of influenza throughout Africa.1,2 Influenza vaccine is the most effective way to prevent influenza infection.3 Although influenza vaccine has been increasingly used in the United States and Europe, influenza vaccine is rarely used in most of the African continent.4 To date, there is no information on seasonal influenza vaccine effectiveness in tropical Africa. The increased incidence of helminthic infections, malaria, other chronic parasitic infections, HIV and malnutrition in some parts of the continent could lead to an altered immune response to vaccine.5
In Kenya, a country of 44 million people with a tropical climate and a per capita gross domestic product of $1800 in 2012,6 recent studies have shown that influenza circulates year-round and the incidence of influenza is at least as high as it is in other parts of the world and particularly high in children <2 years old.7,8 Influenza vaccine is not provided by the government, and <45,000 doses of influenza vaccine are purchased privately every year. A recent study found that Kenyans were aware of the seasonal influenza, although they were not aware of seasonal influenza vaccine.9 In 2010, nearly two-thirds of healthcare workers in 5 hospitals in Kenya chose to get a free pandemic H1N1 influenza vaccine.10 To assess the acceptability and effectiveness of seasonal influenza vaccine in Kenya, the Kenya Medical Research Institute/Centers for Disease Control-Kenya (KEMRI/CDC-K), together with the Kenyan Ministry of Health, offered free influenza vaccine to children 6 months to 10 years old in a low-income rural and urban population in Kenya during 3 consecutive years, 2010 to 2012.
MATERIALS AND METHODS
The International Emerging Infections Program of KEMRI/CDC-K has conducted population-based infectious disease surveillance (PBIDS) since 2006 in 2 sites in Kenya: Lwak, a rural location in western Kenya along Lake Victoria; and Gatwikira and Soweto villages within Kibera, an urban informal settlement in Nairobi.11 To be enrolled in PBIDS, individuals must have resided permanently in the area for 4 calendar months and have been registered into the KEMRI/CDC PBIDS system.11
In Lwak, from 2010 to 2012, the enrolled surveillance population ranged from 24,100 to 27,100, including 8000–8900 children ≤10 years old. This rural area comprises 100 km2, with an overall population density of about 325 persons per km2.11 Lwak area has perennial, high-level malaria transmission. In 2009, the estimated prevalence of HIV infection in adults was 17%.12 The daily mean temperature in Nyanza, the province where Lwak is located, ranges from a high of 31°C in February to a low of 28°C in July. Annual rainfall is approximately 1400 mm/yr, peaking in April–May and November–December.13 The latitude is −0°10′S, and the elevation is 1174 m above sea level.14
From 2010 to 2012, the urban site in Kibera included 27,200–28,800 people, including 9700–10,700 children ≤10 years old, within 2 villages in Kibera. The surveillance area covers 0.4 km2 and has a population density of about 77,000 persons per km2. Malaria is not endemic, although cases of malaria occur.11 In 2009, the estimated HIV prevalence for Kibera was 14%.12 Mean daily maximum temperature in Nairobi ranges from a high of 26°C in February and March to a low of 21°C in July. Rainfall averages approximately 1000 mm/yr, with peak rainfall in April–May and November–December.13 The latitude is −1°28′S, and the elevation is 1684 m above sea level.14
In the PBIDS sites in Kibera and Lwak, since 2007, residents have had access to free care for infectious disease-associated conditions. In Lwak, community residents can attend St. Elizabeth Lwak Mission Hospital, operated by the Franciscan Sisters of St. Anna, and in Kibera, residents are offered free care at Tabitha Clinic, operated by Carolina for Kibera.
Structured questionnaires were completed for all visits for medical care at both study clinics, capturing symptoms, comorbidities, prior health care-seeking activity, physical and laboratory examinations, diagnosis, treatment and outcome.
Clinic Surveillance Case Definitions
Patients who presented to either Lwak Mission Hospital or Tabitha Clinic and met a case definition for acute lower respiratory illness (ALRI) or influenza-like illness (ILI) were eligible to have a nasopharyngeal and an oropharyngeal swab collected. ALRI was defined in children <5 years old as a modified combination of the World Health Organization (WHO) Integrated Management for Childhood Illness case definitions for pneumonia and severe pneumonia.15 The ALRI case definition has been described previously.7 We defined ILI as documented temperature ≥38°C plus reported cough or sore throat.
Specimen Collection and Processing
Trained laboratory technicians collected nasopharyngeal and oropharyngeal specimens only at the study clinics according to previously described procedures.16 Specimens were transported to the KEMRI/CDC-K laboratory for polymerase chain reaction (PCR) testing on a weekly basis.
Samples were aliquoted, and total RNA was extracted from 100 µL aliquots of each sample using Qiagen’s QIAamp Viral RNA Mini Kit (Qiagen Inc, Valencia, CA), according to the manufacturer’s instructions. An aliquot of each respiratory specimen was tested by real-time reverse transcription (rRT)-PCR for influenza A and influenza B after 1 freeze-thaw cycle, according to published methods.17,18 Specimens positive for influenza A were subtyped for seasonal H1, H3, H5 and H1pdm09 by rRT-PCR. One-step rRT-PCR was carried out using AgPath kits (Applied Biosystems, Foster City, CA). The primers, probes and positive controls for all influenza viruses were provided by CDC-Atlanta.17,18 Fluorescence was read at the combined annealing-extension step at 55°C and recorded as threshold cycle (Ct) values. A Ct value ≤39.9 was considered positive; Ct values ≥40.0 were considered negative. Influenza viruses were isolated from a subset of influenza-positive specimens with Ct values <35. Antigenic characterization was assessed using postinfection ferret antisera in hemagglutination inhibition assay by the WHO Influenza Collaborating Center at CDC, Atlanta, GA, using standard methods. An influenza virus was considered a close match to a reference or vaccine strain if its hemagglutination inhibition titer was equal to or within a 4-fold difference to the vaccine/reference strain.
During July–September 2010, April–June 2011 and March–May 2012, KEMRI/CDC-K offered free trivalent inactivated influenza vaccine (Vaxigrip, southern hemisphere composition, donated by Sanofi Pasteur S.A., France) to children 6 months to 10 years old who were residents of the 2 communities. Only residents enrolled in the PBIDS were offered vaccine. None of the children had been previously vaccinated with pandemic H1N1 vaccine. Pediatric doses (0.25 mL) were used for children <36 months old. We selected the vaccination period based on the availability of the vaccine, which varied each year, and previous surveillance data showing that peak influenza activity in both communities usually occurred during June–August.7 Vaccination was voluntary and required parental consent. Children <9 years old were offered 2 doses of vaccine if they had not been previously vaccinated.
We established 3 vaccination sites in each of the 2 communities. KEMRI/CDC staff met with community leaders to explain the goals of the project before the first vaccination campaign began. Vaccines were stored at a central location at 2°–8°C and distributed to sites in the morning in coolboxes; unused vaccines were returned to the storage site in the afternoon. In Kibera, vaccine was offered at the Tabitha study clinic, a church in the community and the Carolina for Kibera Community Center. In Lwak, vaccine was offered at the Lwak Hospital, Mahaya Health Center and Ong’ielo Subdistrict Hospital. Whenever a resident received a dose of vaccine, the resident’s PBIDS study identification number was recorded along with the date of vaccine administration, and this information was added to the resident’s general medical record. Children 6 months to 8 years old who received a first dose of vaccine and had not been fully vaccinated previously were instructed to return to the clinic after 28 days to receive a second dose of the vaccine.
Adverse Event Reporting
Patients were asked to remain at the vaccine centers for 15–20 minutes after administration of vaccine to allow for monitoring of any acute adverse events. In addition, patients were instructed to return to the free clinic for any acute illness within the first 48 hours after vaccine administration. All patients who had adverse events were seen by a medical doctor, who recorded information on patient demographics, symptoms, treatment and outcome.
Vaccinated children 9–10 years old and vaccinated children 6 months to 8 years old who had been fully vaccinated in a previous year were considered immunized beginning 14 days after the receipt of a single influenza vaccination. Vaccinated children 6 months to 8 years old who had not been fully vaccinated previously were considered fully immunized 14 days after the receipt of the second vaccine dose and partially immunized if they had received only 1 of the 2 vaccine doses.
We conducted a test-negative case-control study19,20 to evaluate the effectiveness of trivalent inactivated influenza vaccine during the 3 years. We compared the odds of vaccination among case-patients with clinically attended ILI or ALRI who tested positive for influenza to the odds of vaccination among controls—patients with clinically attended ILI or ALRI who tested negative for influenza. For each case-patient, 1 to 4 controls were enrolled, matched individually on age (6–23 months, 2–5 years and 5–10 years), date of sample collection (±14 days from the date the case-patient sample was collected) and site (Kibera vs. Lwak). For the primary analysis, we excluded cases and controls who were partially vaccinated. Because the controls were individually matched to the cases, we used a conditional logistic regression model that adjusted for the interval from symptom onset to sample collection and syndrome (ALRI vs. ILI).
Because influenza activity in Kenya occurs throughout the year with seasonal activity less pronounced than in temperate climates,7,21, and because influenza vaccine to our knowledge has not been evaluated in a field setting with continuous influenza activity over 12 months, we calculated vaccine effectiveness (VE) for 4 different evaluation periods: 2 weeks to 3 months, 2 weeks to 6 months, 2 weeks to 9 months and 2 weeks to 12 months. For each of the 3 years, we measured VE starting 2 weeks after the final dose was administered. We additionally evaluated the VE for 3 discrete time intervals: 2 weeks to ≤3 months; >3 to ≤6 months and >6 to ≤9 months. We estimated VE as [100 × (1 − adjusted odds ratio)]. For the 12-month analysis, for periods that overlapped with the beginning of the following year’s vaccination campaign, we only included cases and controls who had not been vaccinated with the following year’s vaccine. We performed stratified analyses by year, site and age group. The Mantel-Haenszel test was used to test for homogeneity of VE across the 3 discrete time intervals. We also performed a score test for trend of odds over time.
To evaluate the VE of partial vaccination, we performed a test-negative case-control analysis, similar to the analysis described above, which excluded cases and controls who were fully vaccinated.
The seasonal influenza VE study protocol was approved by the KEMRI Ethical Review Committee (SSC Protocol 1780). The CDC Institutional Review Board formally deferred to the KEMRI Ethical Review Committee for this review (CDC Institutional Review Board Protocol 5933). The study was registered on clinicaltrials.gov (NCT01432340). Written consent was obtained from all parents or guardians of all participants.
Of all eligible children in the 2 sites, 30%, 36% and 38% were fully vaccinated in 2010, 2011 and 2012, respectively; 11%, 12% and 13% were partially vaccinated (Table 1). Uptake was highest among children 5 to <10 years old (36.6%) and lowest among children 6 months to <2 years old (27.1%). Overall, the percentage of children who received the vaccine was slightly higher in Lwak compared with that in Kibera (37.0% vs. 32.9%; P < 0.001).
During the three 12-month follow-up periods after vaccination in 2010, 2011 and 2012, there were 156, 109 and 127 cases of influenza, respectively, among children aged 6 months to 10 years. After excluding 55 influenza-positive children who were partially vaccinated and 9 influenza-positive children for whom we could not find age-matched controls, we included 328 influenza test-positive cases (44% from Kibera and 56% from Lwak) who were matched to a total of 909 test-negative controls in our evaluation. The median age of cases was 5.2 years, and during all 3 years approximately half of the cases were <5 years old. Overall, 148 (56%) of the cases were male, and 1% had known HIV infection (Table 2). Approximately half of the cases had ALRI, and <5% of all cases were hospitalized. The number of cases varied during the 4 follow-up periods during the 3 years; we identified 51 (16%) cases during the first 3 months of follow-up, 77 (23%) cases during the second 3-month period of follow-up, 138 (42%) cases during the third 3-month period of follow-up and 62 (19%) cases during the final 3-month follow-up period. Although cases were relatively evenly divided between Kibera and Lwak in the first and third years of the study, during the second year of the study nearly three-quarters of the cases were from Lwak. Demographic and clinical information were similar between cases and controls (Table 2).
In the first 2 years, influenza A predominated (67% of cases in 2010 and 2011 and in 2011 and 2012) and influenza A composed 50% of influenza cases in 2012 and 2013. In all 3 years, influenza A (H1N1) pdm09, influenza A (H3N2) and influenza B were identified (Fig. 1). We cultured virus from 53 of 127 PCR-positive specimens in 2010 and 2011, 20 of 62 specimens in 2011 and 2012 and 20 of 77 specimens in 2012 and 2013 (Table 3); all isolates were submitted to WHO Influenza Collaborating Center at CDC for antigenic characterization and were reported as antigenically well matched to the vaccine strains (Table 3). Peak influenza activity varied during the 3 years. In Kibera, influenza activity peaked twice every year; the first peak usually preceded the period of vaccine administration (Fig. 1). In Lwak, there were fewer overall cases of influenza, and influenza activity tended to peak in the first few months of the year.
VE among fully vaccinated children was 63% [95% confidence interval (CI): −5 to 87] for the 3-month follow-up period, 57% (95% CI: 29% to 74%) for the 6-month follow-up period, 39% (95% CI: 17% to 56%) for the 9-month follow-up period and 48% (95% CI: 32% to 61%) for the 12-month follow-up period (Table 4). When we analyzed data using 3 discrete time intervals, the VE was 42% (95% CI: 0.2 to 66) for the >3- to 6-month follow-up period, 42% (95% CI: 9 to 63) for the >6- to 9-month follow-up period and 75% (95% CI: 49 to 88) for the >9- to 12-month follow-up period. The VE was not significantly different from 0 for the 2-week to 3-month follow-up period [26% (95% CI: −71 to 68)] (Table 5). We found no evidence of heterogeneity of the VE (P = 0.2), and the score test did not show a trend (P = 0.2).
In the 12-month follow-up period, VE was statistically significant in children <5 years [50% (95% CI: 25 to 67)] and children 5 to <10 years old [46% (95% CI: 19 to 64)]. During the 3 years, for the 6-month and the 9-month follow-up period, VE was statistically significant in children <5 years old [60% (95% CI: 21 to 80) and 46% (95% CI: 16 to 66), respectively], but no statistically significant VE was found in children 5 to <10 years old (Table 6). VE was not significant in either age group for the 3-month follow-up period. The annual VE was only statistically significant in children <5 years during the 12-month follow-up period in 2010 and 2011 [55% (95% CI: 11 to 77)] and in both the 9-month and the 12-month follow-up periods in 2011 and 2012 [82% (95% CI: 37 to 95) and 71.6% (95% CI: 29 to 89), respectively].
For the 12-month follow-up period, influenza vaccine was effective in preventing influenza A-associated ILI and ALRI overall [55% (95% CI: 35 to 69)], during the 12-month follow-up period following vaccination in 2010 [48% (95% CI: 8 to 72)] and during the 12-month follow-up period after vaccination in 2011 [67% (95% CI: 36 to 83)]. The vaccine was also effective in preventing influenza A (H3)-associated ILI and ALRI overall [62% (95% CI: 25 to 80)] and during the 12-month follow-up period after vaccination in 2011 [69% (95% CI: 19 to 89)] but not in the other seasons. During the 12-month follow-up period overall, the vaccine was effective in preventing influenza B [38% (95% CI: 3 to 60)] and pdm09 [49% (95% CI: 8 to 72)]. Although the annual point estimates for the prevention of influenza B and pdm09 were positive, these values were not statistically significant in the individual seasons. Subtypes could not be determined for 76 influenza A cases likely because of low virus load (high Ct values).
We were able to find matched controls for 49 of the 55 cases of influenza among partially vaccinated children during the 3-year period. During the 12-month follow-up period, the overall VE for partially vaccinated children was statistically significant [39% (95% CI: 11 to 58)]. VE was not statistically significant for partial vaccination during any of the individual 3 years (data not shown).
During the 3 years, 20 medically attended suspected adverse events—acute illnesses that occurred within 48 hours of vaccine administration—were identified: 6 were from Kibera and 14 were from Lwak. Most (17) of the cases were treated as outpatients. Three children were hospitalized, one of whom died; none were judged to be the result of the vaccine. The 2 hospitalized children who survived tested positive for malaria, were treated and recovered. The 1 child who died received a clinical diagnosis of severe anemia.
To our knowledge, this is the first influenza VE study in tropical sub-Saharan Africa. We found that in an urban and rural community in Kenya, when offered a free influenza vaccine, parents of nearly half of eligible children <10 years old chose to get their children vaccinated at least once. In addition, the vaccine showed moderate effectiveness in preventing medically attended influenza-associated ALRI and ILI, and this effectiveness extended for 12 months after the vaccination period.
The relatively high uptake we observed demonstrates that in 2 diverse socioeconomically deprived communities in Kenya, residents were interested in having their children vaccinated against influenza when offered a free vaccine. Despite the fact that none of the residents of the 2 communities had heard of influenza vaccine before it was introduced in 2010,9 residents sought out the vaccine in increasing numbers over the 3-year campaign. The rates of vaccine coverage in these 2 Kenyan communities were higher than the reported influenza vaccine coverage rates among children 6 months to 8 years in the United States (an age group where annual influenza vaccine is recommended3), which ranged from 22% in the 2008 and 2009 season to 27% in the 2011 and 2012 season,22 and higher than the reported coverage rates among children with chronic disease in Spain during the 2012 and 2013 season (16%).23 The vaccine in Kenya was free, which Kibera and Lwak residents said contributed to their decision to get the vaccine.9 In Kenya, the widespread familiarity with routine childhood vaccination may have made parents more likely to seek out the influenza vaccine for their children; routine national vaccination programs and one-off campaigns for polio and measles are heavily promoted throughout the country, and in Kenya in 2011, coverage for most routine childhood vaccinations was >85%.24
Because influenza is transmitted throughout the year in Kenya and seasonality can vary considerably from year to year,7 we were able to evaluate VE for up to 12 months after the vaccination period. This type of analysis would not be possible in temperate climates, where to date most VE studies have been conducted because the influenza season in these regions typically lasts <5 months.25 We observed a statistically significant VE of 39% and 48% for the 3-year period using a follow-up period of 9 months and 12 months, respectively. The combined results for the 3 years for the 12-month follow-up period also showed statistically significant VE against influenza A and B and influenza A subtypes. Although VE was not statistically significant from 0 for the 3-month follow-up period, overall numbers in this group were small.
Our findings of reduced but statistically significant VE at 9 and 12 months are consistent with previous immunologic studies of trivalent influenza vaccine in adults, which have described reduced but existent protective antibodies up to 40 weeks26 and 52 weeks27 after vaccination. Longer influenza vaccine protection would be particularly useful in tropical climates, where influenza seasonality is less discrete.28 The variable influenza seasonality we observed during the 3 years of our vaccine study illustrates some of the challenges of not just VE studies but of the use of influenza vaccine itself, in tropical countries such as Kenya. If additional influenza VE studies were to show protection during a 12-month period in tropical settings, a rolling approach to vaccination—where vaccine is offered through the year—could potentially be considered. One approach could be to alternate between Northern and Southern Hemisphere formulation throughout the year as each becomes available, as recently proposed.29
Our overall findings of moderate effectiveness are consistent with a number of previously reported results of observational and randomized controlled trials of influenza VE. A 2005 meta-analysis of VE studies of inactivated influenza vaccine reported moderate VE (66%) in children <6 years old using data from observational studies but no VE in the same age group using pooled data from randomized controlled trials.30,31 A 2011 meta-analysis of 10 randomized controlled trials reported a pooled efficacy for trivalent inactivated influenza vaccine of 59% (95% CI: 51 to 67) in adults aged 18–65 years.32 The only other VE study conducted in Africa, a 5-year study in a mostly adult population in temperate South Africa, found that VE ranged from −14.2% to 67.4% annually.33
In our study, influenza vaccine was moderately effective in preventing laboratory-confirmed influenza illness, despite the generally good match between circulating strains and vaccine strains. Similar findings have been reported previously and may be related to differences in strains that reflect mutations that are not routinely detected.34
Our study had some limitations. First, although using the test-negative case-control study approach controls for possible differences in health-seeking behavior, which varies considerably in the 2 populations in Kenya,35,36 test-negative control subjects are likely different from the general population in other ways.37 Vaccinated and unvaccinated children may differ in the extent they come in contact with people infected with influenza, and the test-negative control design can be subject to bias by including persons who seek care for acute respiratory infection when influenza circulation is minimal.20 Although the study clinics offered free care to all enrolled participants, ill patients may have sought care for respiratory illness at kiosks, unlicensed care providers and traditional healers—which we know occurs in both communities35,36—and this may have biased our results. We only evaluated influenza VE in preventing medically attended influenza-associated acute respiratory infections; other outcomes, such as nonmedically attended infection, influenza-related complications and influenza-associated death, were not evaluated. Because we wanted to offer the vaccine to all eligible participants to assess interest in vaccination in the 2 communities, we were not able to power our study for all possible outcomes because we did not know what level of uptake to expect. Because adverse event surveillance was passive, we may have missed some of these events. Finally, because few children had recorded health comorbidities, we were not able to evaluate VE in this subset.
During 3 years in an urban and rural community in Kenya, wherein residents had previously never heard of influenza vaccine, parents of over one-third of eligible children aged 6 months to 10 years chose to have their children fully vaccinated with a free trivalent inactivated influenza vaccine. Fully vaccinated children were approximately 40% less likely to have an influenza-associated medically attended respiratory illness overall. These results are encouraging as far as the acceptability and effectiveness of the vaccine in 2 vaccine-naive communities in sub-Saharan Africa. However, in this area of the world, where the burden of other vaccine-preventable diarrheal and respiratory diseases is high, more work is needed to characterize the cost-effectiveness of an annual influenza vaccine. Consideration of the targeted use of the vaccine in high-risk populations, such as young children or pregnant women, where VE has been shown to protect both mothers and young infants,38 could be a reasonable starting point.
The authors thank David Shay and Mark Thompson for their help in the design of the study and the review of the manuscript; Jackie Mwendwa, Rachel Ochola and William Mwiti for their contributions in the field; Stella Gikunju, Gilbert Kikwai, Eunice Rwamba and Wycliffe Mwika for their work in the laboratory; Geoffrey Arunga and Bryan Nyawanda for statistical support. They are grateful to Sanofi Pasteur for donating the influenza vaccines. They thank the entire KEMRI/CDC field staff for their work in support of this study in Lwak and Kibera. They also thank Zohar Mor for his review of this manuscript.
1. Radin JM, Katz MA, Tempia S, et al. Influenza
surveillance in 15 countries in Africa, 2006–2010. J Infect Dis. 2012(suppl 1):206–S21:S14–S21
2. Katz MA, Schoub BD, Heraud JM, et al. Influenza
in Africa: uncovering the epidemiology of a long-overlooked disease. J Infect Dis. 2012;206(suppl 1):S1–S4
3. Centers for Disease Control and Prevention. . Prevention and control of influenza
with vaccines: recommendations of the Advisory Committee on Immunization Practices (ACIP)—United States, 2012–13 influenza
season. MMWR Morb Mortal Wkly Rep. 2012;61:613–618
4. Duque J, McMorrow ML, Cohen AL. Influenza
vaccines and influenza
antiviral drugs in Africa: are they available and do guidelines for their use exist? BMC Public Health. 2014;14:41
5. Yazdanbakhsh M, Kremsner PG. Influenza
in Africa. PLoS Med. 2009;6:e1000182
6. CIA World Factbook. 2013 Kenya
Available at: https://www.cia.gov/library/publications/the-world-factbook/geos/ke.html
. Accessed September 3, 2015.
7. Katz MA, Lebo E, Emukule G, et al. Epidemiology, seasonality, and burden of influenza
-like illness in Urban and Rural Kenya
, 2007–2010. J Infect Dis. 2012;206(suppl 1):S53–S60
8. Onyango CO, Njeru R, Kazungu S, et al. Influenza
surveillance among children with pneumonia admitted to a district hospital in coastal Kenya
, 2007–2010. J Infect Dis. 2012;206(suppl 1):S61–S67
9. Oria PA, Arunga G, Lebo E, et al. Assessing parents’ knowledge and attitudes towards seasonal influenza
vaccination of children before and after a seasonal influenza
study in low-income urban and rural Kenya
, 2010–2011. BMC Public Health. 2013;13:391
10. Njuguna H, Ahmed J, Oria PA, et al. Uptake and effectiveness
of monovalent influenza
A (H1N1) pandemic 2009 vaccine
among healthcare personnel in Kenya
, 2010. Vaccine
11. Feikin DR, Audi A, Olack B, et al. Evaluation of the optimal recall period for disease symptoms in home-based morbidity surveillance in rural and urban Kenya
. Int J Epidemiol. 2010;39:450–458
12. Dalal W, Feikin DR, Amolloh M, et al. Home-based HIV testing and counseling in rural and urban Kenyan communities. J Acquir Immune Defic Syndr. 2013;62:e47–e54
13. World Weather Information Service. 2010 Kenya
Available at: http://www.worldweather.org/071/c00251f.htm#climate
. Accessed September 3, 2015
14. worldatlas. 2015 Available at: http://www.worldatlas.com/af/ke/17/where-is-kisumu.html
. Accessed September 3, 2015
15. World Health Organization. Integrated Management of Childhood Illness. 2008
16. Kim C, Ahmed JA, Eidex RB, et al. Comparison of nasopharyngeal and oropharyngeal swabs for the diagnosis of eight respiratory
viruses by real-time reverse transcription-PCR assays. PLoS One. 2011;6:e21610
17. Kodani M, Yang G, Conklin LM, et al. Application of TaqMan low-density arrays for simultaneous detection of multiple respiratory
pathogens. J Clin Microbiol. 2011;49:2175–2182
19. Foppa IM, Haber M, Ferdinands JM, et al. The case test-negative design for studies of the effectiveness
of influenza vaccine
20. Jackson ML, Nelson JC. The test-negative design for estimating influenza vaccine effectiveness
21. Katz MA, Muthoka P, Emukule GO, et al. Results from the first six years of national sentinel surveillance for influenza
, July 2007-June 2013. PLoS One. 2014;9:e98615
22. Lu P, Santibanez TA, Williams WW, et al. Surveillance of influenza
vaccination coverage—United States, 2007–08 through 2011–12 influenza
seasons. MMWR Surveill Summ. 2013;62:1–28
23. Jimenez-Garcia R, Esteban-Vasallo MD, Rodríguez-Rieiro C, et al. Coverage and predictors of vaccination against 2012/13 seasonal influenza
in Madrid, Spain: analysis of population-based computerized immunization registries and clinical records. Hum Vaccin Immunother. 2014;10:449–455
25. Finkelman BS, Viboud C, Koelle K, et al. Global patterns in seasonal activity of influenza
A/H3N2, A/H1N1, and B from 1997 to 2005: viral coexistence and latitudinal gradients. PLoS One. 2007;2:e1296
26. Künzel W, Glathe H, Engelmann H, et al. Kinetics of humoral antibody response to trivalent inactivated split influenza vaccine
in subjects previously vaccinated or vaccinated for the first time. Vaccine
27. MacKenzie JS. Influenza
: antibody responses to one and two doses of vaccine
and length of response, with particular reference to the elderly. Br Med J. 1977;1:200–202
28. Viboud C, Alonso WJ, Simonsen L. Influenza
in tropical regions. PLoS Med. 2006;3:468–471
29. Lambach P, Alvarez AM, Hirve S, et al. Considerations of strategies to provide influenza vaccine
year round. Vaccine
30. Jefferson T, Smith S, Demicheli V, et al. Assessment of the efficacy and effectiveness
vaccines in healthy children: systematic review. Lancet. 2005;365:773–780
31. Fu C, He Q, Li Z, et al. Seasonal influenza vaccine effectiveness
among children, 2010–2012. Influenza
Other Respi Viruses. 2013;7:1168–1174
32. Osterholm MT, Kelley NS, Sommer A, et al. Efficacy and effectiveness
vaccines: a systematic review and meta-analysis. Lancet Infect Dis. 2012;12:36–44
33. Ntshoe GM, McAnerney JM, Tempia S, et al. Influenza
epidemiology and vaccine effectiveness
among patients with influenza
-like illness, viral watch sentinel sites, South Africa, 2005–2009. PLoS One. 2014;9:e94681
34. Skowronski DM, Janjua NZ, De Serres G, et al. Low 2012-13 influenza vaccine effectiveness
associated with mutation in the egg-adapted H3N2 vaccine
strain not antigenic drift in circulating viruses. PLoS One. 2014;9:e92153
35. Burton DC, Flannery B, Onyango B, et al. Healthcare-seeking behaviour for common infectious disease-related illnesses in rural Kenya
: a community-based house-to-house survey. J Heal Popul Nutr. 2011;29:61–70
36. Breiman RF, Olack B, Shultz A, et al. Healthcare-use for major infectious disease syndromes in an informal settlement in Nairobi, Kenya
. J Health Popul Nutr. 2011;29:123–133
37. Belongia EA, Kieke BA, Donahue JG, et al. Effectiveness
of inactivated influenza
vaccines varied substantially with antigenic match from the 2004–2005 season to the 2006–2007 season. J Infect Dis. 2009;199:159–167
38. Zaman K, Roy E, Arifeen SE, et al. Effectiveness
of maternal influenza
immunization in mothers and infants. N Engl J Med. 2008;359:1555–1564