Andrew J. Mackay

Population Dynamics of Aedes aegypti and Dengue as Influenced by Weather and Human Behavior in San Juan, Puerto Rico

Previous studies on the influence of weather on Aedes aegypti dynamics in Puerto Rico suggested that rainfall was a significant driver of immature mosquito populations and dengue incidence, but mostly in the drier areas of the island. We conducted a longitudinal study of Ae. aegypti in two neighborhoods of the metropolitan area of San Juan city, Puerto Rico where rainfall is more uniformly distributed throughout the year. We assessed the impacts of rainfall, temperature, and human activities on the temporal dynamics of adult Ae. aegypti and oviposition. Changes in adult mosquitoes were monitored with BG-Sentinel traps and oviposition activity with CDC enhanced ovitraps. Pupal surveys were conducted during the drier and wetter parts of the year in both neighborhoods to determine the contribution of humans and rains to mosquito production. Mosquito dynamics in each neighborhood was compared with dengue incidence in their respective municipalities during the study. Our results showed that: 1. Most pupae were produced in containers managed by people, which explains the prevalence of adult mosquitoes at times when rainfall was scant; 2. Water meters were documented for the first time as productive habitats for Ae. aegypti; 3. Even though Puerto Rico has a reliable supply of tap water and an active tire recycling program, water storage containers and discarded tires were important mosquito producers; 4. Peaks in mosquito density preceded maximum dengue incidence; and 5. Ae. aegypti dynamics were driven by weather and human activity and oviposition was significantly correlated with dengue incidence.

Use of the CDC Autocidal Gravid Ovitrap to Control and Prevent Outbreaks of Aedes aegypti (Diptera: Culicidae)

ABSTRACT Populations of Aedes aegypti (L.) can be managed through reductions in adult mosquito survival, number of offspring produced, or both. Direct adult mortality can be caused by the use of space sprays or residual insecticides to mosquito resting sites, and with a variety of residual insecticideimpregnated surfaces that are being tested, such as curtains, covers for water-storage vessels, bednets, and ovitraps. The fertility of Ae. aegypti populations can be reduced by the use of autocidal oviposition cups that prevent the development of mosquitoes inside the trap by mechanical means or larvicides, as well as by releasing sterile, transgenic, and para-transgenic mosquitoes. Survival and fertility can be simultaneously reduced by capturing gravid female Ae. aegypti with sticky gravid traps. We tested the effectiveness of the novel Centers for Disease Control and Prevention autocidal gravid ovitrap (CDC-AGO trap) to control natural populations of Ae. aegypti under field conditions in two isolated urban areas (reference vs. intervention areas) in southern Puerto Rico for 1 yr. There were significant reductions in the captures of female Ae. aegypti (53–70%) in the intervention area. The presence of three to four AGO control traps per home in 81% of the houses prevented outbreaks of Ae. aegypti , which would be expected after rains. Mosquito captures in BG-Sentinel and AGO traps were significantly and positively correlated, showing that AGO traps are useful and inexpensive mosquito surveillance devices. The use of AGO traps to manage Ae. aegypti populations is compatible with other control means such as source reduction, larviciding, adulticiding, sterile insect techniques, induced cytoplasmic incompatibility, and dominant lethal gene systems.

An improved autocidal gravid ovitrap for the control and surveillance of Aedes aegypti

BackgroundLimited success has been achieved using traditional vector control methods to prevent the transmission of dengue viruses. Integrated control programs incorporating alternative tools, such as gravid ovitraps (lethal ovitraps and sticky ovitraps) may provide greater potential for monitoring and reducing vector populations and dengue virus transmission. We had developed an autocidal gravid ovitrap (AGO) as a simple, low-cost device for surveillance and control of Ae. aegypti without the use of pesticides that does not require servicing for an extended period of time. The purpose of our study was to improve the efficacy and efficiency of this device.MethodsCompetitive assays were performed in the laboratory and an outdoor cage to evaluate whether modifications to the structure and appearance of our original trap design (AGO-A), and the addition of an olfactory bait (hay infusion), improve trap function. The performance of a modified trap design (AGO-B) was then assessed and compared with conventional ovitraps in a series of field tests in San Juan City, Puerto Rico. Generalized linear mixed models were used to analyze adult Ae. aegypti capture data from the laboratory, outdoor cage and field experiments.ResultsIncreasing the size of the trap entrance, altering the color of trap components, and increasing the volume/surface area of the aqueous bait significantly improved the performance of the AGO in the outdoor cage. In a subsequent field comparison, captures of Ae. aegypti females were 3.7 fold greater in the improved trap (AGO-B), compared with the original design (AGO-A). An infusion bait produced “in situ” significantly improved capture rates of the improved trap under both semi-natural and field conditions. Semi-weekly collections of Ae. aegypti females in the AGO-B were significantly correlated with cumulative rainfall 8 to 28 days prior to sampling, whereas egg collections in paired conventional ovitraps were not. When vector abundance was low, the AGO-B provided greater sensitivity and precision as a surveillance device, compared with paired conventional ovitraps.ConclusionsThe AGO-B can be used to efficiently attract and capture gravid Ae. aegypti females for more than 8 weeks without the need for trap maintenance.

WEST NILE VIRUS SURVEILLANCE IN EAST BATON ROUGE PARISH, LOUISIANA

ABSTRACT West Nile virus (WNV) was detected for the first time in Louisiana in the fall of 2001. Surveillance data collected from East Baton Rouge Parish in 2002 were examined to establish baseline data on WNV activity, to support the current design of disease surveillance programs, and to target vector control efforts in the parish. The first indications of WNV activity were from a dead Northern Cardinal collected in February and from a live male cardinal sampled on 14 March. In mosquito pools, WNV was first detected on June 11. The onset of the first human case and the first detection of WNV in sentinel chickens occurred concurrently on June 24. The number of reported human cases and minimum infection rates in mosquitoes peaked in July. WNV prevalence in wild birds increased in late August and was highest in December. WNV-positive wild birds and mosquito pools were detected an average of 31 and 59 days in advance of the onset date of reported human cases, respectively, within 5 km of the residence of a human case. Antibodies to WNV were detected in sera from 7 (Northern Cardinal, House Sparrow, Northern Mockingbird, Blue Jay, Hermit Thrush, Yellow-rumped Warbler, and White-throated Sparrow) of the 42 wild bird species tested. Wild bird serology indicated WNV activity during the winter. Out of 18 mosquito species tested, the only species found positive for WNV was Culex quinquefasciatus, a result suggesting that this species was the primary epizootic/epidemic vector.

Host Feeding Patterns of Culex Mosquitoes (Diptera: Culicidae) in East Baton Rouge Parish, Louisiana

ABSTRACT Host feeding patterns were examined for four species of Culex mosquitoes collected from 18 sites in or adjacent to East Baton Rouge Parish, LA, from November 2002 to October 2004. Host DNA from 37 bloodied Culex coronator Dyar and Knab, 67 bloodfed Cx. salinarius Coquillett, 112 bloodfed Cx. nigripalpus Theobald, and 684 bloodfed Cx. quinquefasciatus Say were identified. The percentages of bloodmeals containing mammalian DNA were 94.6% for Cx. coronator, 82.1% for Cx. salinarius, 66.1% for Cx. nigripalpus, and 40.1% for Cx. quinquefasciatus. Human DNA was detected in 7% of the bloodmeals from Cx. quinquefasciatus and 2.7% of the bloodmeals from Cx. nigripalpus. The northern cardinal was the most frequent avian host of Cx. quinquefasciatus and Cx. nigripalpus. In 2003 and 2004, there was no significant relationship from May through October between the proportion of Cx. quinquefasciatus feeding on mammalian hosts and the date of collection. Of the six avian species most frequently fed on by Cx. quinquefasciatus, the northern cardinal, northern mockingbird, common grackle, and brown thrasher were fed on more frequently than expected based on their abundance. House sparrows were fed on less frequently than expected based on their abundance. These data support the conclusions of previous studies that Cx. quinquefasciatus is the most important vector for both the enzootic amplification and transmission of West Nile virus to humans in southern Louisiana.

Dynamics of Aedes aegypti and Culex quinquefasciatus in Septic Tanks

Abstract Aedes aegypti and Culex quinquefasciatus were found in large numbers emerging from septic tanks in southern Puerto Rico during the dry season. Previous studies suggested that Ae. aegypti uses subterranean aquatic habitats only during dry periods when surface containers do not have water. This research investigated whether septic tanks are alternative aquatic habitats that this mosquito uses during unfavorable times of the year, or whether Ae. aegypti uses this aquatic habitat throughout the year. To assess temporal change, exit traps were used to collect mosquitoes emerging from septic tanks in Playa/Playita, southern Puerto Rico, from November 2006 to October 2007. We also investigated the hypotheses that 1) the production of Ae. aegypti in septic tanks was larger than in surface containers and 2) adult mosquitoes emerging from septic tanks were larger than those emerging from surface containers. This study demonstrated that unsealed septic tanks produced large numbers of Ae. aegypti and Cx. quinquefasciatus throughout the year, without any significant relationship with rainfall. The number of adult Ae. aegypti emerging per day from septic tanks in each community was 3 to 9 times larger than those produced in surface containers. It was also demonstrated that Ae. aegypti emerging from septic tanks were significantly larger than those emerging from surface container habitats. It is recommended that dengue prevention programs include regular inspection and maintenance of septic tanks in communities lacking sewerage.

Evidence of Vertical Transmission of West Nile Virus in Field-Collected Mosquitoes

ABSTRACT: Male and nulliparous female mosquitoes were surveyed for evidence of vertical WNV infection in East Baton Rouge Parish, Louisiana. Adult male mosquitoes collected by trapping and aspiration, and adult male and nulliparous female mosquitoes reared from field-collected larvae were tested. Adult male Culex spp., female Aedes albopictus (Skuse), and female Culex quinquifasciatus Say mosquitoes that were collected as larvae were test-positive for WNV RNA. Infectious WNV was detected using virus isolation in field-collected male Aedes triseriatus Say and Culex salinarius Coquillett; these data represent the first field evidence of vertical transmission of WNV in Ae. triseriatus and Cx. salinarius.

Vertebrate Hosts of Aedes aegypti and Aedes mediovittatus (Diptera: Culicidae) in Rural Puerto Rico

ABSTRACT The distribution of Aedes (Stegomyia) aegypti (L.), the main vector of dengue viruses (DENV) worldwide, overlaps with Aedes (Gymnometopa) mediovittatus (Coquillett), the Caribbean treehole mosquito, in urban, suburban, and rural areas. Ae. mediovittatus is a competent vector of DENV with high rates of vertical DENV transmission in the laboratory. This study determined whether Ae. mediovittatus feeds on humans and compared its feeding patterns with co-occurring Ae. aegypti in two rural communities of Puerto Rico. Adult mosquitoes were captured for three consecutive days every week from July 2009 to May 2010 using BG-Sentinel traps with skin lures that were placed in the front yard of houses in both communities. Three methods were used to identify the 756 bloodmeals obtained in this study: a multiplex polymerase chain reaction (PCR) for humans and dogs targeting cytochrome b; a PCR targeting the 16S rRNA; and a nested PCR targeting cytochrome b. Ae. mediovittatus fed mostly on humans (45–52%) and dogs (28–32%) but also on cats, cows, horses, rats, pigs, goats, sheep, and chickens. Ae. aegypti fed mostly on humans (76–79%) and dogs (18–21%) but also on cats, horses, and chickens. Our results indicate that Ae. mediovittatus may have a relatively high rate of vector—human contact, which might facilitate virus transmission or harborage in rural areas of Puerto Rico.

West Nile Virus Detection in Mosquitoes in East Baton Rouge Parish, Louisiana, from November 2002 to October 2004

ABSTRACT The prevalence of West Nile virus (WNV) was determined in mosquitoes between November 2002 and October 2004 in East Baton Rouge Parish, LA. A total of 244,374 female mosquitoes were collected and tested by viral isolation. Additionally, 131,896 female mosquitoes were collected in 2003 and tested by VecTest and 167,175 female mosquitoes were collected in 2004 and tested by reverse transcriptase–polymerase chain reaction (RT-PCR). West Nile virus was isolated by cell culture from 17 (47.2%) out of 36 mosquito species collected over the study period. In 2003, WNV was detected in 9 (33.3%) out of 27 species tested by VecTest. In 2004, 14 (50%) out of the 28 mosquito species tested by RT-PCR were positive for WNV. The species with the greatest number of WNV-positive pools detected by all 3 testing methods was Culex quinquefasciatus. A significantly greater proportion of Cx. salinarius pools collected in light traps placed at a 3-m height were positive for WNV by viral isolation than in pools collected in light traps placed at a 1.5-m height.

Mosquito vectors of West Nile virus during an epizootic outbreak in Puerto Rico.

ABSTRACT The purpose of this investigation was to identify the mosquito (Diptera: Culicidae) vectors of West Nile virus (WNV; family Flaviviridae, genus Flavivirus) during an epizootic WNV outbreak in eastern Puerto Rico in 2007. In June 2006, 12 sentinel chicken pens with five chickens per pen were deployed in six types of habitats: herbaceous wetlands, mangrove forests, deciduous forests, evergreen forests, rural areas, and urban areas. Once WNV seroconversion in chickens was detected in June 2007, we began trapping mosquitoes using Centers for Disease Control and Prevention (CDC) miniature (light/CO2-baited) traps, CMT-20 collapsible mosquito (CO2- and ISCA SkinLure-baited) traps, and CDC gravid (hay infusion-baited) traps. We placed the CDC miniature traps both 2–4 m and >30 m from the chicken pens, the collapsible traps 2–4 m from the pens, and the gravid traps in backyards of houses with sentinel chicken pens and in a wetland adjacent to an urban area. We found numerous blood-engorged mosquitoes in the traps nearest to the sentinel chickens and reasoned that any such mosquitoes with a disseminated WNV infection likely served as vectors for the transmission of WNV to the sentinels. We used reverse transcriptase-polymerase chain reaction and isolation (C636) on pools of heads, thoraxes/abdomens, and legs of collected blood-engorged mosquitoes to determine whether the mosquitoes carried WNV. We detected WNV-disseminated infections in and obtained WNV isolates from Culex nigripalpus Theo (minimum infection rate [MIR] 1.1–9.7/1,000), Culex bahamensis Dyar and Knab (MIR 1.8–6.0/1,000), and Aedes taeniorhynchus (Wied.) (MIR 0.34–0.36/1,000). WNV was also identified in and isolated from the pool of thoraxes and abdomens of Culex quinquefasciatus Say (4.17/1,000) and identified in one pool of thoraxes and abdomens of Culex habilitator Dyar and Knab (13.39/1,000). Accumulated evidence since 2002 suggests that WNV has not become endemic in Puerto Rico.