The first LD case-patient on the merchant ship sought treatment at the nearest hospital emergency department, and provided no alternative exposure source. After laboratory confirmation of this case, the crew disembarked and the vessel was required to lie at anchor offshore. After emergency control measures by a private contractor, we obtained information on the ship’s plumbing, including potable, fresh, and hot water systems, water storage, air conditioning; food preparation areas, and sleeping quarters.

We then boarded the ship for environmental investigation on August 27, 2015, to collect samples from potential fomites around the vessel at 33 locations, including cabins and potable water outlets. We collected PCR swab samples in duplicate from inside showerheads and sink faucets (also known as mixer taps) aerators in sleeping quarters and food preparation areas, including those used by LD case-patients and their neighbors. The contractor disinfected the water system by using super chlorination the next day, and collected a second environmental sample series on September 4. Additional targeted control measures included replacement of showerheads and removal of faucet aerators from cabins.

Figure 1

Figure 1. Dismantled showerhead from nonpassenger merchant vessel showing multiple inner parts, including 7 O-rings, all of which were in contact with water passing through shower, Australia, 2015.

We collected a series of PCR swab samples from original test locations on October 12 to assess the residual health threat, and tested 24 of these samples on the ship (10). Duplicate samples were then tested in the reference laboratory (10). We analyzed showerheads removed from cabins (Figure 1). We tested samples of the inside surface of each showerhead and its O-ring gaskets by using PCR assays. We collected swab samples from potential reservoirs and tested for Legionella species: the O-rings; rinse samples from showerhead parts in sterile 0.08% NaCl solution for Legionella species; peptone water washings, showerhead contents, debris from a thermal mixing valve, fresh and pre–UV-treated water, showerheads, air conditioners, and faucets from cabins (11). We identified presumptive Legionella cultures on MWY and BMPA agars by using Legionella Latex Agglutination antisera (Oxoid; ThermoFisher Scientific Australia Pty Ltd, Scoresby, Melbourne, Victoria, Australia), and cultured for amoeba on showerhead rinse specimens. Detailed methods are provided in Technical Appendix Part 1.

Figure 2

Figure 2. Accommodation deck plan, Australia, 2015. Cabins (n = 10) and other rooms (ACU, air conditioning unit) from which environmental samples were collected on August 27, 2015 are indicated in dark gray....

In August 2015, the Western Australia Department of Health was notified of Legionnaires’ disease confirmed by L. pneumophila serogroup 1 urinary antigen test in a member of the vessel’s crew (case-patient 1), and was informed that other crew members had mild febrile respiratory illness (Table 1). Later that day, another crew member, who had symptoms of severe bilateral pneumonia and pneumothorax, arrived at the regional hospital and required aeromedical evacuation for intensive care (case-patient 2). LD was confirmed by urinary antigen testing and PCR assay on bronchial washings. Other crew members who had nonpneumonic respiratory and other symptoms were investigated for legionellosis by using urinary antigen tests and serologic tests which proved negative, except in case-patient 3, who had L. pneumophila seroconversion and Pontiac fever that did not require hospital admission. The 3 cases all satisfied Australian LD case definitions (12). Case-patients 1 and 2 occupied adjacent cabins and case-patient 3 was 2 cabins away from case-patient 2 (Figure 2).

L. pneumophila was not isolated from any environmental samples. Legionella PCR result was positive in 7/10 cabins tested (13/27 samples) (Table 2). A PCR result was positive for showerheads or residual water from sink faucets in the cabins of 2 LD cases. In 5 other cabins, only faucets were positive (Figure 2). Detection of sludge or biofilm in the shower heads and faucets prompted replacement with better-designed showerheads and removal of faucet aerators. Only 2/79 samples collected on the second visit on September 4 were Legionella PCR positive; a significant reduction (χ², Yates’ correction; 15.98, p<0.001). Only 1 of the 58 samples from the third series of samples was clearly PCR positive, from a faucet in a cabin unconnected to LD cases. The in-field PCR results were identical to the confirmatory reference laboratory replicate results. All 10 types of showerhead were rust-stained inside and smelled of chlorine. The most common showerhead types had either 7 silicone rubber O-rings or 1 complex silicone rubber gasket. Showerhead swabs and agar O-ring impressions grew profuse mixed bacteria, commonly Pseudomonas aeruginosa. Nonpneumophila Legionella sp. was isolated from 1 showerhead. Legionella PCR assays produced unambiguous positives in 13/16 showerheads (19/32 samples). Almost all O-rings from the common showerhead types were Legionella positive (Technical Appendix Part 2).

A recent study of nonpassenger merchant vessels (NPMVs) highlighted the risk for Legionella contamination of potable water systems (9), but did not establish a link with confirmed infections. Our investigation of L. pneumophila serogroup 1 infection in a merchant vessel’s crew highlights the need to control Legionella in NPMV water systems, and the challenge of using PCR assays, which do not detect viable bacteria. Culture-dependent methods did not contribute to determination of the environmental source or route of dissemination. Preliminary control measures by external contractors may have prevented Legionella isolation from our environmental samples, but have doubtful long-term preventive value without sustained control measures because environmental persistence occurs in ships despite biocide treatment (6).

The survey vessel had a gross tonnage of 2,620, was 64 m long, 16 m wide, a draft of 4.7 m, and a crew of 27. It had 2 water storage tanks with 60,000 L capacity, an ultraviolet water sterilization unit, and 2 hot water geysers. These tanks were refilled from bunkers while in port, and replenished at sea by reverse osmosis. Showers were highlighted in a previous study of NPMV potable water systems (9), and aerator devices have been implicated as bacterial amplification sites in tropical and nosocomial outbreaks (13,14).

Multiple positive PCR results from water outlets in the cabins implicated the showers and faucets as means of infection. All showerheads on the vessel had interior moving parts to control spray settings and were the leading PCR-positive location. A rust-colored biofilm inside most showerheads indicated possible deterioration of iron pipes in the ship’s distribution system, and persistence of Legionella in biofilms (15). The silicone rubber O-rings from the showerheads supported profuse growth of aquatic bacteria and were PCR positive for L. pneumophila. The O-rings formed a permanently wet niche for bacterial growth, and their movement will shear bacteria from biofilms. Faucet aerators also promote turbulent flow by mixing water and air under pressure. These results highlight the potential for Legionella aerosol generation. We recommended replacing the showerheads with a simpler plastic design, more suited to periodic removal, decontamination, and cleaning, and gravity drainage after daily use.

A cluster of L. pneumophila serogroup 1 infections in a vessel working in waters near Australia led to an environmental health assessment in which molecular methods enabled the field investigation team to implicate water outlets in crew quarters and tailor environmental controls. Deployment of quantitative PCR assays extended our investigative reach offshore, enabling faster return of the vessel to active service. The leadership and crew of nonpassenger merchant vessels operating in tropical waters need heightened Legionella awareness and require control measures more stringent than those applied in passenger vessels.