Investigators describe an outbreak of Legionnaires’ disease with 118 confirmed cases and seven deaths linked to cooling towers in Harlem, New York, last year, according to a Public Health Alerts report today.
The authors, with the New York City Health Department (NYCHD), quickly identified two cooling tower systems, both on the same block, as sources of infection with Legionella pneumophilia bacteria. One cooling tower was not regulated.
Public Health Alerts, a collaboration between NEJM Evidence and CIDRAP, fills a gap in reliable data, offering expert-reviewed reports that translate frontline observations into actionable public health evidence. An NEJM Evidence editorial explains the initiative further.
Two towers implicated, 1 unregistered
On July 25, 2025, NYCHD software detected eight reports of positive urine antigen test results for L pneumophila serogroup 1. All eight patients lived in Central Harlem within a 1-kilometer (roughly half-mile) radius. Five of the eight cases were rapidly confirmed.
Health department officials issued an advisory on July 28 for clinicians to remain alert for Legionnaires’ disease in patients who had pneumonia. Afterward, the NYCHD identified 129 reports and confirmed 118 cases. Median patient age was 65 years (range, 32 to 97 years).
Ninety-two patients (78.0%) required hospital care. Officials recorded seven outbreak-related deaths, with a median time of symptom onset to death of 8 days (range, 1 to 15 days).
Of 43 cooling tower systems sampled by the end of July, eight—including an unregistered tower—tested positive for L pneumophila or L pneumophila serogroup 1. Owners of the cooling towers were required to complete disinfection and full cleaning. The owner of the unregistered tower system was ordered to register. Cleaning was completed by August 15.
Based on whole-genome sequencing, two isolates from separate cooling tower systems appeared to be highly related to the original eight patients’ clinical isolates. One of these was the unregistered tower. This tower had major lapses in routine monitoring, lacked water treatment steps, and had missed required Legionella sampling. The other tower system met routine monitoring and sampling requirements. Both cooling tower systems were on the same block.
The authors write, “In the absence of defined thresholds on the amount of legionella bacteria required to infect humans and without a clearer understanding of air flow and exposure patterns, we were unable to conclude whether one cooling tower was the source of infection.”
They conclude, “This investigation shows the utility of a cooling tower registry, which enabled rapid identification and sampling of registered cooling towers in the outbreak zone.”
