Analytical framework to evaluate and optimize the use of imperfect diagnostics to inform outbreak response: Application to the 2017 plague epidemic in Madagascar.

During outbreaks, the lack of diagnostic "gold standard" can mask the true burden of infection in the population and hamper the allocation of resources required for control. Here, we present an analytical framework to evaluate and optimize the use of diagnostics when multiple yet imperfect diagnostic tests are available. We apply it to laboratory results of 2,136 samples, analyzed with 3 diagnostic tests (based on up to 7 diagnostic outcomes), collected during the 2017 pneumonic (PP) and bubonic plague (BP) outbreak in Madagascar, which was unprecedented both in the number of notified cases, clinical presentation, and spatial distribution. The extent of these outbreaks has however remained unclear due to nonoptimal assays. Using latent class methods, we estimate that 7% to 15% of notified cases were Yersinia pestis-infected. Overreporting was highest during the peak of the outbreak and lowest in the rural settings endemic to Y. pestis. Molecular biology methods offered the best compromise between sensitivity and specificity. The specificity of the rapid diagnostic test was relatively low (PP: 82%, BP: 85%), particularly for use in contexts with large quantities of misclassified cases. Comparison with data from a subsequent seasonal Y. pestis outbreak in 2018 reveal better test performance (BP: specificity 99%, sensitivity: 91%), indicating that factors related to the response to a large, explosive outbreak may well have affected test performance. We used our framework to optimize the case classification and derive consolidated epidemic trends. Our approach may help reduce uncertainties in other outbreaks where diagnostics are imperfect.

Multiple Introductions of Yersinia pestis during Urban Pneumonic Plague Epidemic, Madagascar, 2017.

Pneumonic plague (PP) is characterized by high infection rate, person-to-person transmission, and rapid progression to severe disease. In 2017, a PP epidemic occurred in 2 Madagascar urban areas, Antananarivo and Toamasina. We used epidemiologic data and Yersinia pestis genomic characterization to determine the sources of this epidemic. Human plague emerged independently from environmental reservoirs in rural endemic foci >20 times during August-November 2017. Confirmed cases from 5 emergences, including 4 PP cases, were documented in urban areas. Epidemiologic and genetic analyses of cases associated with the first emergence event to reach urban areas confirmed that transmission started in August; spread to Antananarivo, Toamasina, and other locations; and persisted in Antananarivo until at least mid-November. Two other Y. pestis lineages may have caused persistent PP transmission chains in Antananarivo. Multiple Y. pestis lineages were independently introduced to urban areas from several rural foci via travel of infected persons during the epidemic.

Human Exposure to Hantaviruses Associated with Rodents of the Murinae Subfamily, Madagascar.

We conducted a national human serologic study of a hantavirus detected in Madagascar rodents using a commercial kit and a new ELISA targeting the virus. Our results suggest a conservative estimate of 2.7% (46/1,680) IgG seroprevalence. A second single-district study using the new ELISA revealed a higher prevalence (7.2%; 10/139).

Leptospira in livestock in Madagascar: uncultured strains, mixed infections and small mammal-livestock transmission highlight challenges in controlling and diagnosing leptospirosis in the developing world.

In developing countries, estimates of the prevalence and diversity of Leptospira infections in livestock, an important but neglected zoonotic pathogen and cause of livestock productivity loss, are lacking. In Madagascar, abattoir sampling of cattle and pigs demonstrated a prevalence of infection of 20% in cattle and 5% in pigs by real-time PCR. In cattle, amplification and sequencing of the Leptospira-specific lfb1 gene revealed novel genotypes, mixed infections of two or more Leptospira species and evidence for potential transmission between small mammals and cattle. Sequencing of the secY gene demonstrated genetic similarities between Leptospira detected in Madagascar and, as yet, uncultured Leptospira strains identified in Tanzania, Reunion and Brazil. Detection of Leptospira DNA in the same animal was more likely in urine samples or pooled samples from four kidney lobes relative to samples collected from a single kidney lobe, suggesting an effect of sampling method on detection. In pigs, no molecular typing of positive samples was possible. Further research into the epidemiology of livestock leptospirosis in developing countries is needed to inform efforts to reduce human infections and to improve livestock productivity.

Mixed Leptospira Infections in a Diverse Reservoir Host Community, Madagascar, 2013-2015.

We identified mixed infections of pathogenic Leptospira in small mammals across a landscape-scale study area in Madagascar by using primers targeting different Leptospira spp. Using targeted primers increased prevalence estimates and evidence for transmission between endemic and invasive hosts. Future studies should assess rodentborne transmission of Leptospira to humans.

Reproductive ecology of the black rat (Rattus rattus) in Madagascar: the influence of density-dependent and -independent effects.

The black rat (Rattus rattus) poses a severe threat to food security and public health in Madagascar, where it is a major cause of pre- and post-harvest crop losses and an important reservoir for many zoonotic diseases, including plague. Elsewhere, ecologically based rodent management (EBRM) strategies have been developed using ecological information to inform decisions on where and when to target control. EBRM could deliver improved health and well-being outcomes in Madagascar if adapted to the local ecological context. Using data collected from removal studies, we explored spatio-temporal patterns in the breeding activity of the black rat (R. rattus) in domestic and agricultural habitats across Madagascar and investigated to what extent these trends are influenced by rainfall and rat density. We identified clear spatio-temporal variation in the seasonality of R. rattus reproduction. Reproduction was highly seasonal both inside and outside of houses, but seasonal trends varied between these two habitats. Seasonal trends were explained, in part, by variation in rainfall; however, the effect of rainfall on reproductive rates did itself vary by season and habitat type. A decline in breeding intensity with increasing rat density was recorded outside of houses. This has important implications for control, as populations may compensate for removal through increased reproduction. We recommend that sustained control initiated before the main breeding season, combined with improved hygiene and adequate rodent-proofing in homes and grain stores, could curtail population growth and reduce pre- and post-harvest losses provided that these measures overcome the compensatory response of rodent populations.

Population dynamics of plague vector fleas in an endemic focus: implications for plague surveillance.

Plague is a zoonotic vector-borne disease caused by the bacterium Yersinia pestis. In Madagascar, it persists in identified foci, where it is a threat to public health generally from September to April. A more complete understanding of how the disease persists could guide control strategies. Fleas are the main vector for transmission between small mammal hosts and humans, and fleas likely play a role in the maintenance of plague. This study characterized the dynamics of flea populations in plague foci alongside the occurrence of human cases. From 2018 to 2020, small mammals were trapped at sites in the central Highlands of Madagascar. A total of 2,762 small mammals were captured and 5,295 fleas were collected. The analysis examines 2 plague vector species in Madagascar (Synopsyllus fonquerniei and Xenopsylla cheopis). Generalized linear models were used to relate flea abundance to abiotic factors, with adjustments for trap location and flea species. We observed significant effects of abiotic factors on the abundance, intensity, and infestation rate by the outdoor-associated flea species, S. fonquerniei, but weak seasonality for the indoor-associated flea species, X. cheopis. A difference in the timing of peak abundance was observed between the 2 flea species during and outside the plague season. While the present study did not identify a clear link between flea population dynamics and plague maintenance, as only one collected X. cheopis was infected, the results presented herein can be used by local health authorities to improve monitoring and control strategies of plague vector fleas in Madagascar.

Assessing the effectiveness of intervention to prevent plague through community and animal-based survey.

Bubonic plague, transmitted by infected flea bites, is the most common form of plague and, left untreated, can progress to the pneumonic form, which is highly contagious. Surveillance focusing on reservoir and vector is considered to be the main approach to prevent plague. Common rodent control methods include the use of rodenticide and snap traps but, in a plague context, the dispersal of fleas from killed animals may pose a serious health threat. Therefore, there is a need for strategies which address reservoir and vector control. The aim of this study was to assess the effects of combination of reservoir and vector control through community-based surveillance. Activities were implemented by local previously trained community agents in two active plague foci in Madagascar. Kartman bait stations containing rodenticide and insecticide were placed indoors while live traps were set outdoors. Small mammals were identified and killed with their fleas. Effectiveness of control measures was evaluated by comparison of plague incidence two years before and after intervention using data on reported human cases of plague from the Central Laboratory of Plague. A total of 4,302 small mammals were captured, with the predominance of the black rat Rattus rattus. Our results found a reduction in plague incidence in the treated site for at least two years after treatment. Community-based interventions played an important role in reducing contact between humans-rodents-fleas. Our study confirms the importance of animal surveillance during the low plague transmission season. The combination of reservoir and vector control with community involvement may be effective at reducing the risks of plague spillover to humans. The strategy of using Kartman bait stations indoors with live traps outdoors can be used to refine proactive plague prevention, however, due to the potential development of resistance to pesticides in flea and rat populations, overuse should be considered.

Knockdown resistance mutations are common and widely distributed in Xenopsylla cheopis fleas that transmit plague in Madagascar.

BACKGROUND: Plague, caused by the bacterium Yersinia pestis, remains an important disease in Madagascar, where the oriental rat flea, Xenopsylla cheopis, is a primary vector. To control fleas, synthetic pyrethroids (SPs) have been used for >20 years, resulting in resistance in many X. cheopis populations. The most common mechanisms of SP resistance are target site mutations in the voltage-gated sodium channel (VGSC) gene. METHODOLOGY/PRINCIPAL FINDINGS: We obtained 25 collections of X. cheopis from 22 locations across Madagascar and performed phenotypic tests to determine resistance to deltamethrin, permethrin, and/or dichlorodiphenyltrichloroethane (DDT). Most populations were resistant to all these insecticides. We sequenced a 535 bp segment of the VGSC gene and identified two different mutations encoding distinct substitutions at amino acid position 1014, which is associated with knockdown resistance (kdr) to SPs in insects. Kdr mutation L1014F occurred in all 25 collections; a rarer mutation, L1014H, was found in 12 collections. There was a significant positive relationship between the frequency of kdr alleles and the proportion of individuals surviving exposure to deltamethrin. Phylogenetic comparisons of 12 VGSC alleles in Madagascar suggested resistant alleles arose from susceptible lineages at least three times. Because genotype can reasonably predict resistance phenotype, we developed a TaqMan PCR assay for the rapid detection of kdr resistance alleles. CONCLUSIONS/SIGNIFICANCE: Our study provides new insights into VGSC mutations in Malagasy populations of X. cheopis and is the first to report a positive correlation between VGSC genotypes and SP resistance phenotypes in fleas. Widespread occurrence of these two SP resistance mutations in X. cheopis populations in Madagascar reduces the viability of these insecticides for flea control. However, the TaqMan assay described here facilitates rapid detection of kdr mutations to inform when use of these insecticides is still warranted to reduce transmission of plague.

Plague in Tanzania: first report of sylvatic plague in Morogoro region, persistence in Mbulu focus, and ongoing quiescence in Lushoto and Iringa foci.

Objectives: Plague has been a threat to human health in Tanzania since 1886. This zoonotic disease has established several endemic foci in the country, posing a risk of outbreaks. This study was conducted to investigate the presence of Yersinia pestis in small mammals in five districts. These districts were selected because of recent (Mbulu), past (40-18 years ago: Lushoto) and historic (>100 years ago: Iringa and Kilolo) human cases of plague. In addition, one region that has not had any reported human cases of plague was included (Morogoro-Mvomero). Methods: Blood from 645 captured small mammals was screened for antibodies against the fraction 1 (F1) antigen of Y. pestis using indirect enzyme-linked immunosorbent assay (ELISA) and competitive-blocking ELISA. Results: Specific antibodies against Y. pestis F1 antigens were detected in six (0.93%) animals belonging to Mastomys natalensis. Of these, four animals were captured in the active focus in Mbulu, and two animals were captured from an area with no history of human plague (Morogoro-Mvomero). Conclusion: These results provide evidence of the circulation of Y. pestis in small mammals in Tanzania. Furthermore, evidence of the circulation of Y. pestis in Morogoro-Mvomero highlights the importance of carrying out plague surveillance in areas with no history of human plague, which can help to predict areas where future outbreaks may occur.

Tracking of Mammals and Their Fleas for Plague Surveillance in Madagascar, 2018-2019.

Plague, a zoonotic disease caused by the bacterium Yersinia pestis, remains a major public health threat in Madagascar. To better understand the risk of transmission to humans and to guide targeted plague prevention and control measures, a survey of Y. pestis infection and exposure in mammals and their fleas was implemented. Small mammals were captured in five districts of Madagascar ranging in levels of plague endemicity, as measured by notified cases, from none to active foci. Blood and spleen samples and fleas were collected from small mammals for the detection of anti-Y. pestis F1 antibodies by ELISA, F1 antigens by rapid diagnostic tests, and pla, caf1, and inv genes by polymerase chain reaction. Some rodent fleas were kept alive and reared in the insectary to assess susceptibility to insecticides. Blood was also collected from 15 dogs and tested for anti-F1 antibodies. A total of 557 spleens, 484 sera, and 1,539 fleas were collected from 557 rodents and shrews. Nineteen (3.4%) spleens were positive for F1 antigen, most from Toamasina (N = 13), a historical plague focus. One dog was also found seropositive in Toamasina. Twenty-two (4.5%) serologic specimens from small mammals were positive for anti-F1 antibodies. The flea index was highest in the city of Antananarivo (8.8). No flea was positive for Y. pestis DNA. Flea populations exhibited resistance to various insecticides weakening the efficacy of vector control. This study highlights the potential use of animal-based surveillance to identify the risk of plague transmission in endemic and nonendemic foci for targeted prevention and control.

Risk factors for intestinal parasite portage in an informal suburb on the West coast of Madagascar.

The deprived area of the Metzinger Valley in the city of Mahajanga has many healthcare concerns due to repeated flooding during the rainy season. Improving this health situation requires a better knowledge of the pathogens present in this area and of the risk factors favoring their propagation. The aim of this study was to analyze the relationship between the household socioeconomic status and the presence of parasites in the faeces of children between 1 and 10 years of age in order to determine the risk factors for intestinal parasitosis. The study included 746 children, of whom 30% were infected with parasites. Entamoeba coli, a good indicator of environmental fecal contamination, was the most prevalent parasite with an observation frequency of 16.7% followed by Giardia lamblia with a prevalence of 10%. For helminths, Trichuris and Ascaris were the most prevalent respectively 5.4% and 1.8%. A large heterogeneity in the prevalence of parasites was observed from one neighborhood to another. However, multivariate analysis showed that these differences were not related to environmental factors or household structure, but rather to the economic level of the family, the education level of the mother as well as the age of the child. For example, the prevalence of Giardia decreased from 23.5% to 8% for children of mothers with little education to those with higher education, respectively. For E. coli, the prevalence is higher among poor households and school-aged children. In the frame of IRCOD project, mothers are being sensitized to hygiene and risk factors for transmission by intestinal parasites and the present study proposes a multidimensional approach as an assessment tool.

Characterization of Klebsiella pneumoniae isolated from patients suspected of pulmonary or bubonic plague during the Madagascar epidemic in 2017.

Klebsiella pneumoniae can lead to a wide range of diseases including pneumonia, bloodstream and urinary tract infections. During a short period of a pulmonary plague epidemic in October 2017 in Madagascar, 12 K. pneumoniae isolates were identified in ten sputum and two buboes aspirate samples. These isolates were from 12 patients suspected of plague, without epidemiological relationships, but were negative for Yersinia pestis in culture. Data were collected from the plague national surveillance system. The isolates were characterized by antimicrobial susceptibility testing and whole-genome sequencing. Real-time PCR was performed to confirm the presence of K. pneumoniae DNA in buboes. All isolates were identified as K. pneumoniae sensu stricto. Five isolates were extended-spectrum ß-lactamases producers; eleven different sequence types were identified. Five isolates belonged to known hypervirulent sequence types. Our results demonstrate community-acquired pneumonia caused by K. pneumoniae isolates in patients suspected of plague stressing the importance of bed-side differential diagnosis.

Field assessment of dog as sentinel animal for plague in endemic foci of Madagascar.

The epidemiology of Yersinia pestis, the causative agent of plague, involves vectors and reservoirs in its transmission cycle. The passive plague surveillance in Madagascar targets mainly rodent and fleas. However, carnivores are routinely surveyed as sentinels of local plague activity in some countries. The aim of this study is to assess the use of domestic dog (Canis familiaris) as sentinel animal for field surveillance of plague in a highly endemic area in Madagascar. Cross-sectional surveys of plague antibody prevalence in C. familiaris were conducted in endemic areas with contrasting histories of plague cases in humans, as well as a plague free area. Rodent capture was done in parallel to evaluate evidence for Y. pestis circulation in the primary reservoirs. In 2 sites, dogs were later re-sampled to examine evidence of seroconversion and antibody persistence. Biological samplings were performed between March 2008 and February 2009. Plague antibody detection was assessed using anti-F1 ELISA. Our study showed a significant difference in dog prevalence rates between plague-endemic and plague-free areas, with no seropositive dogs detected in the plague free area. No correlation was found between rodents and dog prevalence rates, with an absence of seropositive rodents in some area where plague circulation was indicated by seropositive dogs. This is consistent with high mortality rates in rodents following infection. Re-sampling dogs identified individuals seropositive on both occasions, indicating high rates of re-exposure and/or persistence of plague antibodies for at least 9 months. Seroconversion or seropositive juvenile dogs indicated recent local plague circulation. In Madagascar, dog surveillance for plague antibody could be useful to identify plague circulation in new areas or quiescent areas within endemic zones. Within active endemic areas, monitoring of dog populations for seroconversion (negative to positive) or seropositive juvenile dogs could be useful for identifying areas at greatest risk of human outbreaks.

Rodent control to fight plague: field assessment of methods based on rat density reduction.

Rodents represent a serious threat to food security and public health. The extent to which rodent control can mitigate the risk from rodent-borne disease depends on both the effectiveness of control in reducing rodent abundance and the impact on disease epidemiology. Focusing on a plague-endemic region of Madagascar, this study compared the effectiveness of 3 methods: live-traps, snap-traps, and rodenticides. Control interventions were implemented inside houses between May and October 2019. Tracking tiles monitored rodent abundance. Rodent fleas, the vector involved in plague transmission, were collected. Rodent populations consisted of Rattus rattus and Mus musculus. In terms of trap success, we found that our live-trap regime was more effective than snap-traps. While all 3 control strategies appeared to reduce in-house rodent activity in the short term, we found no evidence of a longer-term effect, with in-house rodent abundance in treated sites comparable to non-treatment sites by the following month. Endemic flea, Synopsyllus fonquerniei, is a key plague vector usually found on rats living outdoors. Although we found no evidence that its abundance inside houses increased following control, this may have been due to a lack of power caused by significant variation in S. fonquerniei abundance. The presence of S. fonquerniei in houses was more likely when S. fonquerniei abundance on outdoor rats was higher, which in turn correlated with high rat abundance. Our results emphasize that control strategies need to consider this connectivity between in-house rat-flea populations and the outdoor populations, and any potential consequences for plague transmission.

Presence of the Oriental Rat Flea (Siphonaptera: Pulicidae) Infesting an Endemic Mammal and Confirmed Plague Circulation in a Forest Area of Madagascar.

The Oriental rat flea, Xenopsylla cheopis (Rothschild 1903), is a cosmopolitan flea usually found infesting domestic rats. This flea is a well-known major human plague vector in Madagascar. As part of field sampling, fleas and small mammals were collected in the village of South Andranofeno and the natural reserve of Sohisika, two sites of the district of Ankazobe, located in the Central Highlands of Madagascar. Rats inside houses and forest small mammals were trapped using Besancon Technical Services and pitfall traps, respectively. Their fleas were collected and preserved for laboratory works. Collected fleas from the village and forest belonged to five species, which were X. cheopis, Synopsyllus fonquerniei (Wagner and Roubaud 1932) (Siphonaptera: Pulicidae), Echidnophaga gallinacea (Westwood 1875) (Siphonaptera: Pulicidae), Ctenocephalides felisstrongylus (Jordan 1925) (Siphonaptera: Pulicidae), Pulex irritans (Linnaeus 1758) (Siphonaptera: Pulicidae). After sampling in the forest zone, one specimen of X. cheopis was unexpectedly collected while infesting an endemic tenrec Setifer setosus (Schreber 1777) (Afrosoricida: Tenrecidae). Polymerase chain reaction (PCR) diagnosis on all collected fleas allowed detecting plague bacterium Yersinia pestis (Lehmann and Neumann 1896) (Enterobacterales: Yersiniaceae) on nine specimens of the endemic flea S. fonquerniei collected inside forest. The presence of the oriental rat flea in forest highlights the connection between human and wild environments due to animal movements and the fact that the rat flea can infest various hosts. As only one specimen of X. cheopis was collected on S. setosus, we hypothesize that flea was carried from the village to forest. Yersinia pestis infection of forest fleas outlines plague circulation in this sylvatic area.

High Prevalence of Leptospira spp. in Rodents in an Urban Setting in Madagascar.

Leptospirosis is a neglected zoonotic bacterial disease caused by pathogenic Leptospira spp. Only limited studies have been conducted on the presence of Leptospira spp. in rats in Antananarivo, the capital city of Madagascar. We assessed Leptospira prevalence in small mammals in urban areas of Antananarivo where sanitation is inadequate and there is risk of flooding during the rainy season. We captured rodents and shrews at two sites and examined kidney samples from 114 animals using culture and a real-time polymerase chain reaction (PCR) assay specific to pathogenic Leptospira spp. We identified 23 positive samples containing Leptospira interrogans and Leptospira borgpetersenii, with a high prevalence in Rattus norvegicus (44.9%). Our results indicate that small mammals, in particular R. norvegicus, present a major public health risk for acquiring leptospirosis in Antananarivo.

Field assessment of insecticide dusting and bait station treatment impact against rodent flea and house flea species in the Madagascar plague context.

Bubonic is the most prevalent plague form in Madagascar. Indoor ground application of insecticide dust is the conventional method used to control potentially infected rodent fleas that transmit the plague bacterium from rodents to humans. The use of bait stations is an alternative approach for vector control during plague epidemics, as well as a preventive control method during non-epidemic seasons. Bait stations have many advantages, principally by reducing the amount of insecticide used, lowering the cost of the treatment and minimizing insecticide exposure in the environment. A previous study reported promising results on controlling simultaneously the reservoir and vectors, when slow-acting rodenticide was incorporated in bait stations called "Boîtes de Kartman". However, little evidence of an effective control of the fleas prior to the elimination of rodents was found. In this study, we evaluated bait stations containing insecticide powder and non-toxic attractive rodent bait for their potential to control rat fleas. Its efficacy was compared to the standard method. The impact of both methods on indoor and outdoor rodent fleas, as well as the human household flea Pulex irritans were analyzed at different time points after treatments. Bait stations did not cause any significant immediate or delayed reduction of rat fleas and increasing the number of operational bait stations per household did not significantly improve their efficacy. Insecticide ground dusting appeared to be the most efficient method to control indoor rat fleas. Both methods appeared to have little impact on the density of outdoor rat fleas and human fleas. These results demonstrate limited effectiveness for bait stations and encourage the maintenance of insecticide dusting as a first-line control strategy in case of epidemic emergence of plague, when immediate effect on rodent fleas is needed. Recommendations are given to improve the efficacy of the bait station method.

Epidemiological characteristics of an urban plague epidemic in Madagascar, August-November, 2017: an outbreak report.

BACKGROUND: Madagascar accounts for 75% of global plague cases reported to WHO, with an annual incidence of 200-700 suspected cases (mainly bubonic plague). In 2017, a pneumonic plague epidemic of unusual size occurred. The extent of this epidemic provides a unique opportunity to better understand the epidemiology of pneumonic plagues, particularly in urban settings. METHODS: Clinically suspected plague cases were notified to the Central Laboratory for Plague at Institut Pasteur de Madagascar (Antananarivo, Madagascar), where biological samples were tested. Based on cases recorded between Aug 1, and Nov 26, 2017, we assessed the epidemiological characteristics of this epidemic. Cases were classified as suspected, probable, or confirmed based on the results of three types of diagnostic tests (rapid diagnostic test, molecular methods, and culture) according to 2006 WHO recommendations. FINDINGS: 2414 clinically suspected plague cases were reported, including 1878 (78%) pneumonic plague cases, 395 (16%) bubonic plague cases, one (<1%) septicaemic case, and 140 (6%) cases with unspecified clinical form. 386 (21%) of 1878 notified pneumonic plague cases were probable and 32 (2%) were confirmed. 73 (18%) of 395 notified bubonic plague cases were probable and 66 (17%) were confirmed. The case fatality ratio was higher among confirmed cases (eight [25%] of 32 cases) than probable (27 [8%] of 360 cases) or suspected pneumonic plague cases (74 [5%] of 1358 cases) and a similar trend was seen for bubonic plague cases (16 [24%] of 66 confirmed cases, four [6%] of 68 probable cases, and six [2%] of 243 suspected cases). 351 (84%) of 418 confirmed or probable pneumonic plague cases were concentrated in Antananarivo, the capital city, and Toamasina, the main seaport. All 50 isolated Yersinia pestis strains were susceptible to the tested antibiotics. INTERPRETATION: This predominantly urban plague epidemic was characterised by a large number of notifications in two major urban areas and an unusually high proportion of pneumonic forms, with only 23% having one or more positive laboratory tests. Lessons about clinical and biological diagnosis, case definition, surveillance, and the logistical management of the response identified in this epidemic are crucial to improve the response to future plague outbreaks. FUNDING: US Agency for International Development, WHO, Institut Pasteur, US Department of Health and Human Services, Laboratoire d'Excellence Integrative Biology of Emerging Infectious Diseases, Models of Infectious Disease Agent Study of the National Institute of General Medical Sciences, AXA Research Fund, and the INCEPTION programme.