A missed plague case seeded a Madagascar cluster that killed 12 of 21

A newly published outbreak investigation shows how a single misdiagnosed case can turn plague — a disease that is curable with prompt antibiotics — into a lethal family cluster. Writing in the US Centers for Disease Control and Prevention journal Emerging Infectious Diseases, investigators reconstruct a 2022 pneumonic plague outbreak in the district of Fenoarivobe, Madagascar, in which 21 people were identified as cases and 12 died — a case-fatality rate of 57.1% [s1].

Madagascar remains the world's plague hotspot. The bacterium Yersinia pestis is endemic there, transmission typically runs from August to April, and most cases occur in the central highlands above 800 metres [s1]. Bubonic plague is the usual form, but pneumonic outbreaks — the form spread person to person through respiratory droplets — occur every year [s1]. The World Health Organization notes that while bubonic plague carries a case-fatality ratio of 30% to 60% untreated, the pneumonic and septicaemic forms are always fatal without treatment, which is why speed of diagnosis governs survival [s2]. Plague is a zoonotic bacterium, usually carried by small mammals and their fleas, and it reaches people through flea bites, direct contact with infected tissues, or inhalation of respiratory particles — the last being what sustains person-to-person pneumonic transmission once a case is established [s2]. The WHO puts the incubation period at one to seven days, a span that leaves little room for diagnostic detours [s2].

A chain that started with malaria

The outbreak surfaced on 18 February 2022, when the Bongolava regional health directorate reported eight suspected pneumonic plague deaths from Fenoarivobe [s1]. Investigators ultimately identified 21 suspected cases between 28 January and 1 March 2022 — nine in people still alive and 12 in people already dead, with all deaths occurring at the community level across multiple households [s1]. The outbreak lasted 33 days and produced an attack rate of 4.6% among the 45,636 inhabitants [s1].

The presumed index patient fell ill on 28 January with fever, headache, weakness and vomiting [s1]. He sought care the same day, was diagnosed with malaria by rapid test, and was given antimalarial treatment [s1]. On 31 January he died at home with severe haemorrhagic symptoms [s1]. From that household, transmission moved outward in primary, secondary and tertiary waves through intrahousehold contact and gatherings [s1]. The investigators judged the index patient to have had either secondary pneumonic plague following a neglected bubonic infection or primary pneumonic plague from an unknown source; for the first two weeks the chain ran largely within his household before spreading more widely [s1]. Cases were spread across eight households, and the case-patients had a median age of 28 years; 11 were women and girls and 10 were men and boys [s1].

Treatment arrived late

The investigation's central finding is one of timing. Anti-plague treatment was not administered to patients seeking care until the fourth week of the outbreak, when health workers began to suspect plague rather than malaria [s1]. By then much of the damage was done. Twelve case-patients received antibiotics against plague — the nine who were still alive plus three of the 12 who died [s1].

Laboratory confirmation came through the Central Laboratory for Plague at the Institut Pasteur in Antananarivo, where 12 people had specimens tested: six were PCR-confirmed, split evenly between living and deceased patients, and one was classified as a probable case [s1]. Among direct contacts of the case-patients, 14% — 14 of 100 tested — were positive for F1 IgG antibodies, a marker of exposure [s1].

The clinical picture was consistent with pneumonic plague throughout. All case-patients had respiratory symptoms; the most common were cough in 81% (17 of 21), chest pain in 71.4% (15 of 21), bloody sputum in 66.7% (14 of 21) and fever in 61.9% (13 of 21) [s1]. The incubation period ranged from one to 13 days and the median illness lasted three days [s1] — a narrow window in which recognition and treatment have to happen.

The lesson is operational, not biological

Nothing about this outbreak points to a more dangerous pathogen. Y. pestis responds to antibiotics, and the WHO stresses that early diagnosis and prompt treatment can prevent the severe forms from turning fatal [s2]. What the Fenoarivobe cluster documents is the cost of diagnostic delay in an endemic setting where a febrile, haemorrhagic presentation was first read as malaria. In districts where plague circulates seasonally, the investigators' implicit prescription is to keep plague in the differential diagnosis early — before a single untreated case becomes a household chain, and a household chain becomes a community outbreak.