On November 9, 2005, a 17-year-old man who arrived at the airport in Newark, New Jersey, United States, had symptoms consistent with measles. The man was part of a group of 148 refugees from the Eastleigh community in Nairobi, Kenya, who arrived in the United States from November 3 through 15. Genotype B3 (subgroup B3.1) was identified from virus samples from this patient; the sequence was identical to sequences from measles viruses collected in Nairobi and Machakos, Kenya, in October 2005 (Figure). All but 1 of the 6 viruses collected from Nairobi (Figure, MVi/Nairobi.KEN.xx.05) were from patients from the Eastleigh area of Nairobi, where an outbreak of measles had been reported in the Somali and Ethiopian communities (6).

Also in November 2005, a single case of measles was reported in the Netherlands. This patient had visited New York City, returned to the Netherlands on November 15, noted a rash on November 23, and was hospitalized with pneumonia and fever on November 24. The initial investigation focused on potential settings where exposure may have occurred in New York City. The source of infection was traced to an unrecognized exposure to the patient in New Jersey only after analysis of the Netherlands viral sequence demonstrated complete identity with the New Jersey genotype B3 virus. The possibility of an epidemiologic link between the 2 cases led to the discovery that the Dutch visitor had arrived at the Newark airport on November 9 and waited in the arrival area for 1 h, along with the group of refugees from Nairobi.

Subsequently, genotype B3 was identified from patients who had had measles during December 2005 in Texas, Canada, and Mexico. In Texas, during the first 2 weeks of December, 3 cases of measles were reported in members of a family from Houston. The patients had flown directly from Houston to the resort area of Cabo San Lucas, Mexico, where they stayed from November 22 through 27. In Mexico, health authorities reported 5 cases of measles beginning on December 12 among baggage handlers and other airport workers at the Mexico City airport. In New Brunswick, Canada, a patient developed a rash on December 19.

Although the earlier cases in New Jersey and in the Netherlands could be traced directly to the outbreak in Kenya, the sources of the cases in Texas, Canada, and Mexico were unknown. However, the sequences from Texas, Canada, and Mexico were identical to the sequences directly linked to the outbreak in Kenya (Figure). Measles viruses in the same chain of transmission have identical sequences (7,8), which indicates that the source of the virus for the cases in Texas, Mexico, and Canada was likely to have been the outbreak in Kenya.

Virus transmission may have occurred through contact with international travelers in airports or during transit because epidemiologic investigations did not detect other measles cases in Cabo San Lucas or Texas. The exception to possible air travel–related exposure was the single case that occurred in Canada. This patient had traveled by car, although investigations found no measles cases in the areas visited: Bangor, Maine (December 2); Boston, Massachusetts (December 3–6); and Portsmouth, New Hampshire (December 6). Two refugees from Eastleigh settled in a state visited by the Canadian patient. They entered Massachusetts on November 10, 2005, and by 21 days after arrival, measles had not developed. However, a measles outbreak was detected in southern Germany in January 2006 (9), and the viral sequence matched that of the Kenya outbreak virus (Figure; MVs/Stuttgart.DEU/4.06), which indicates that the source of this outbreak was also likely to have been Kenya. Therefore, B3 viruses with identical sequences could have been introduced into Texas, Mexico, or Canada by travelers infected with B3 virus in Europe.

Although genotype B3 has been the most frequently detected measles genotype in western and central Africa (4,10–12), ours is the first report of the detection of genotype B3 in Kenya. Moreover, the sequence from a virus isolated in Nigeria in June 2005 (Figure; MVs/Yola.NIE/25.05) was identical to the sequences in the Kenya outbreak group. Although a link has not been established between Nigeria and Kenya, a survey of measles genotypes in Kenya in 2002 detected only genotype D4 viruses (13). Sequences of viruses isolated during 2004 and 2005 from Côte d'Ivoire and Benin (this study) were included in our analysis (Figure) because these B3 sequences represent closely related viruses from western Africa.

The analysis and dissemination of viral sequences from measles cases led to the identification of an unrecognized epidemiologic link at an airport and linked sporadic cases in 4 countries that do not have endemic measles to an ongoing outbreak in Kenya. Investigators in the field need to collect adequate specimens for virus isolation. Timely communication of sequence data among epidemiologists and microbiologists is critical for identifying possible links among sporadic cases of measles. The potential for rapid transmission of measles during brief encounters with international travelers underscores the importance of global surveillance of measles virus.