Vital Signs: Deficiencies in Environmental Control Identified in Outbreaks of Legionnaires’ Disease — North America, 2000–2014

Introduction

Legionnaires’ disease, a severe, sometimes fatal pneumonia, can occur in persons who inhale aerosolized droplets of water contaminated with the bacterium Legionella. Exposure to Legionella in freshwater environments such as lakes and streams does not lead to disease; however, in manmade water systems, Legionella can grow and spread to susceptible hosts, including persons aged ≥50 years, smokers, and persons with underlying medical conditions such as chronic lung disease or immunosuppression.

CDC investigated the first outbreak of Legionnaires’ disease in 1976. Currently, approximately 5,000 cases of Legionnaires’ disease are reported to CDC each year; however, Legionnaires’ disease might be underdiagnosed. During 2000–2014, the rate of reported cases of legionellosis, which comprises both Legionnaires’ disease and Pontiac fever, a milder, self-limited, influenza-like illness, increased 286%, from 0.42 to 1.62 cases per 100,000 persons in the United States (1,2) (Figure 1). The reason for this increase is unknown but is likely multifactorial. The higher rates could represent a true increase in the frequency of disease related to several factors, such as a greater number of persons at risk for legionellosis because of underlying illness or immunocompromising medications, an aging U.S. population, aging plumbing infrastructure, or changes in the climate. Increased use of diagnostic testing because of greater awareness among clinicians and availability of diagnostic tests, as well as more reliable reporting to local and state health departments and CDC could also be playing a role. Approximately 9% of cases are fatal (3). Among 32 potable water–associated outbreaks reported in the United States during 2011–2012, legionellosis was implicated in 21 (66%) outbreaks and all 14 deaths (4). During 2000–2012, CDC’s Waterborne Disease and Outbreak Surveillance System received reports of approximately 160 legionellosis outbreaks (5).

Legionnaires’ disease outbreak investigations require an environmental assessment to identify potential sources of exposure. Environmental assessments are rarely conducted for Legionnaires’ disease cases that are not recognized as part of an outbreak; therefore, most of what is known about Legionella transmission has been learned from outbreak investigations. During 2005–2009, only 4% of confirmed legionellosis cases reported among U.S. residents were associated with a known outbreak or cluster (6), although some sporadic cases were likely associated with unrecognized outbreaks or clusters. Identified outbreaks generally are linked to environmental reservoirs in large or complex water systems,* such as those found in hotels or resorts, hospitals, long-term care facilities, and cruise ships. Transmission from these water systems to humans requires aerosol generation, as can occur from showerheads, cooling towers, hot tubs, and decorative fountains (7). Only one case of possible person-to-person transmission has been reported (8). Legionnaires’ disease is typically diagnosed by a Legionella urinary antigen test or culture of lower respiratory secretions using selective media; epidemiologic links to environmental sources can be confirmed when isolates from clinical and environmental specimens match by molecular typing (9). One species, Legionella pneumophila, accounts for approximately 90% of reported legionellosis cases in the United States (7).

Because Legionella transmission occurs from manmade environmental settings, the most effective strategy for prevention of Legionnaires’ disease is through control of Legionella in water systems in buildings. In 2015, ASHRAE (formerly known as the American Society of Heating, Refrigerating, and Air-Conditioning Engineers) published a consensus standard for the primary prevention of Legionnaires’ disease (10), which calls for the development and implementation of water management programs in large or complex water systems in buildings. The standard, which is based on best practices, focuses on identifying hazardous conditions and applying control measures to interrupt Legionella growth and transmission.

Outbreak investigations often find manmade water systems with maintenance gaps that permit the growth of Legionella. To identify opportunities for prevention, summaries of all CDC field investigations of outbreaks of Legionnaires’ disease during 2000–2014 were reviewed to characterize water system maintenance deficiencies leading to those outbreaks.

Methods

CDC offers assistance to health departments with field investigations of outbreaks of Legionnaires’ disease. After each investigation, CDC reviews and summarizes the field notes to understand conditions that led to the outbreak. These summaries highlight the main findings from each investigation, including the numbers of cases and deaths, clinical or environmental strains of Legionella identified, potential or confirmed environmental sources, and possible environmental factors that contributed to the outbreak, as well as recommended solutions for the management of current outbreaks and prevention of future outbreaks.

CDC reviewed all investigation summaries and associated publications describing Legionnaires’ disease outbreak investigations conducted during 2000–2014. Investigations involving cruise ships were excluded, because their water systems are managed differently from land-based water systems. Two investigators used a standard abstraction form to review the relevant materials. Confirmed and suspected Legionnaires’ disease cases were defined using each outbreak’s case definitions; thus, slight variations in case definition among outbreaks were possible. Investigation summaries were reviewed to identify possible root causes that could facilitate Legionella growth and transmission. Each reviewer independently assigned findings to one or more of four categories: 1) process failures, in which a process, such as a water management program, was missing or inadequate; 2) human errors, in which a person did not perform as expected, such as not replacing hot tub filters according to manufacturer’s recommendations; 3) equipment failures, in which a piece of equipment did not operate as expected, such as a malfunctioning disinfectant delivery system; and 4) unmanaged external changes, in which adjustments were not made to account for events outside a building water system, such as nearby construction leading to changes in potable water quality. Discrepant categorizations were resolved through consultation with a third reviewer.

Results

During 2000–2014, CDC participated in 38 field investigations of Legionnaires’ disease. Three investigations, determined not to be outbreaks because of lack of sufficient clinical or epidemiologic evidence, were excluded. Eight investigations involving cruise ships, associated with 19 confirmed and 17 suspected cases of Legionnaires’ disease, including two deaths, were also excluded. Among the remaining 27 investigations, 24 occurred in U.S. states and territories, two in Mexico, and one in Canada. The most frequent outbreak settings were hotels and resorts (n = 12, 44%), long-term care facilities (5, 19%), and hospitals (4, 15%) (Table 1). The remaining six outbreaks were evenly distributed among senior living facilities (n = 2, 7%), workplaces (2, 7%), and the community (2, 7%). Potable water was the most frequent source of exposure (n = 15, 56%), followed by cooling towers (6, 22%), hot tubs (2, 7%), industrial equipment (1, 4%), and a decorative fountain (1, 4%); for two outbreaks (7%), sources were not identified (Figure 2). Potable water sources accounted for 58% of travel-associated outbreaks (in hotels and resorts) and 67% of health care–associated outbreaks (in hospitals and long-term care facilities).

All 27 outbreaks were caused by Legionella pneumophila serogroup 1. Among 13 (48%) investigations (Table 1), links between human cases and water sources were established through DNA-sequence–based typing that identified indistinguishable clinical and environmental isolates. No clinical isolate was available for nine (33%) outbreaks, no environmental isolate was available for one (4%), and neither a clinical nor environmental isolate was available for two (7%); the clinical and environmental isolates did not match for the remaining two (7%) outbreaks. All available outbreak strains reacted with monoclonal antibody 2 of the international L. pneumophila serogroup 1 panel, a potential marker of increased virulence (11).

The 27 outbreaks included 415 cases, 323 (78%) of which were confirmed^† and 92 (22%) suspected (Table 1). A median of 10 confirmed and suspected cases occurred in each outbreak (range = 3–82). The median number of cases in cooling tower outbreaks was 22, and in potable water outbreaks was 10. Health care–associated outbreaks accounted for 57% of all 415 cases, with a median of 19 cases per health care–associated outbreak; travel-associated outbreaks accounted for 25% of cases, with a median of seven cases per travel-associated outbreak.

Among confirmed and suspected Legionnaires’ disease cases, 65 deaths occurred; the median outbreak case fatality rate was 7% (range = 0%–80%). Health care–associated outbreaks accounted for 85% of deaths (median health care–associated outbreak case fatality rate = 24%, range = 6%–80%); travel-associated outbreaks accounted for 6% of deaths (median travel-associated outbreak case fatality rate = 0%, range = 0%–17%). Patients in seven of the nine health care–associated outbreaks included persons who were employees, visitors, or outpatients who did not stay overnight at the facility. No transplant patients were among the health care–associated cases. In 23 investigations for which the outbreak duration could be determined, the median interval from onset of the first to last cases was 49 days. Median outbreak duration was longer for potable water outbreaks (98 days) than for outbreaks linked to other sources (28 days).

Twenty-three (85%) investigation summaries had sufficient information to evaluate the contribution of deficiencies in water system maintenance to the outbreak (Table 2). The most frequent deficiencies noted were categorized as process failures (n = 15, 65%), followed by human errors (12, 52%), equipment failures (8, 35%), and unmanaged external changes (8, 35%). For 11 (48%) outbreaks, deficiencies in more than one category were reported. Sixteen (70%) investigations reported inadequate water disinfectant levels and 12 (52%) reported water temperatures in the optimal range for Legionella growth (12).^§ Indications of inadequate maintenance of hot tubs and decorative fountains were almost always noted. Among the seven investigations where outbreaks were believed to be associated with unmanaged external changes, nearby construction (n = 3, 43%) and problems with water mains (3, 43%) were most frequently noted. Three buildings had water management programs (all developed before the publication of ASHRAE’s standard in 2015); however, the occurrence of outbreaks suggests that the existing water management programs were inadequate (13,14).

Conclusions and Comments

The number of cases of Legionnaires’ disease in the United States is increasing, and associated mortality is substantial. Identifying ways to reduce environmental transmission of Legionella is crucial to reducing morbidity and mortality associated with Legionnaires’ disease. The new ASHRAE standard establishes minimum requirements for management of the risk for Legionella growth and transmission in building water systems. Gaps in maintenance that could be addressed with a water management program to prevent Legionnaires’ disease outbreaks were described in 23 (85%) of 27 investigated outbreaks. Outbreaks resulted from a combination of deficiencies, most frequently classified as process failures and human errors. In the majority of outbreaks, inadequate water disinfectant levels and temperatures in the optimal range for Legionella growth were observed; implementing a functional water management program could address these deficiencies through routine monitoring of disinfectant levels and water temperature (10). Deficiencies related to equipment failures and unmanaged external changes were less common but are also remediable through preventive measures, such as flushing of potable water systems after water main breaks. Although approximately half the outbreaks included in this analysis resulted from multiple deficiencies, approximately half resulted from a single deficiency, suggesting that even a single deficiency can be sufficient to cause an outbreak; thus, all deficiencies should be addressed.

The most frequent outbreak settings in this analysis were hotels and resorts, long-term care facilities, and hospitals. Although 44% of the outbreaks were travel-associated and 33% were health care–associated, health care–associated outbreaks were larger and resulted in more deaths than travel-associated outbreaks. Potable water was the most frequent source of exposure; however, outbreaks related to cooling tower outbreaks were associated with larger numbers of cases. This finding is consistent with the outdoor location of cooling towers and their ability to create plumes of potentially contaminated water that can expose larger numbers of persons than potable water outbreaks. Potable water outbreaks are usually associated with cases among building occupants, such as hospital patients and hotel guests. Hot tubs have been reported to be an important cause of outbreaks in hotels and cruise ships (15). Regardless of setting or source, a comprehensive approach to prevention requires an understanding of the mechanisms by which Legionella growth and transmission can occur in any building water system. Understanding the nature of deficiencies in water system maintenance using a categorization scheme such as the one described in this report can help inform plans for remediation and prevention following outbreaks of Legionnaires’ disease.

Until published by ASHRAE in 2015, consensus recommendations regarding the development of water management programs to reduce transmission of Legionella were unavailable; thus, ASHRAE’s approach to developing and implementing Legionella water management programs for water systems in buildings might be unfamiliar to building owners and managers (10). CDC and its partners have developed a toolkit (http://www.cdc.gov/legionella/WMPtoolkit) to facilitate implementation of this new standard. The multistep process begins by determining if a building is at increased risk for growth and transmission of Legionella, in which case the formation of a specialized management team is required. The toolkit guides the team through the process of identifying and controlling conditions that can permit Legionella growth and transmission in their building water systems. The process requires careful planning, frequent communication, consistent implementation, and regular review. Taking these steps should reduce the risk for Legionella growth and transmission.

The findings in this report are subject to at least three limitations. First, the scope of legionellosis encompasses Legionnaires’ disease and Pontiac fever. Because fatality is only associated with Legionnaires’ disease, prevention messages are generally targeted toward preventing Legionnaires’ disease and not Pontiac fever; therefore, the Pontiac fever cases in the five outbreaks reporting Pontiac fever (range = 1–101 cases) were excluded. Second, although understanding the clinical aspects of Legionnaires’ disease is an essential step in addressing the increasing number of reported cases, these aspects have been reported elsewhere (6,16) and are not discussed here. Finally, this analysis might not capture all possible gaps in maintenance for several reasons. CDC typically completes investigation summaries within a few weeks of an investigation and investigators might not have had access to environmental information that might have become available later. In addition, CDC does not participate in all investigations of outbreaks of Legionnaires’ disease; thus, these findings might not represent all root causes associated with outbreaks. Moreover, the outbreak-associated deficiencies described in this report might not represent root causes associated with sporadic Legionnaires’ disease cases.

Missed prevention opportunities can lead to outbreaks of Legionnaires’ disease. Making water management programs a routine part of building ownership and management will require education and reinforcement. Environmental and public health professionals can help by incorporating the ASHRAE standard into licensing and accreditation requirements, modifying building and public health codes to include water management programs, and providing tools and information to help local building owners and managers implement water management programs. Future studies should evaluate the implementation and effectiveness of water management programs in buildings with large or complex water systems. Widespread use of such programs might reduce the growth and transmission of Legionella, which, in addition to early diagnosis with appropriate clinical testing, might reduce the number and size of Legionnaires’ disease outbreaks and help reduce the occurrence of Legionnaires’ disease in the United States.