Our data indicate that at the time of the study, 66.7% of the water distribution systems of ferries were positive for Legionella spp. with concentrations of ≥500 CFU/L in at least one sample and 33.3% were heavily colonized. Extremely high concentrations (>3 × 105 CFU/L) were found in one ferry and specifically in samples collected from the third class cabins with common showers that were rarely used by passengers. According to the European Surveillance Scheme for Travel Associated Legionnaires' Disease guidelines  concerning action levels following Legionella sampling in hot and cold water systems, 13 ferries hot water distribution systems (61.9%) needed remedial actions and thermal disinfection was applied. About 38% of the WDS disinfected needed second disinfection application and 15% needed third. This shows that thermal disinfection is not totally effective in eradicating Legionella spp. Similar were the findings of a previous study conducted in WDS of buildings (hotels, hospitals etc.) .
A similar prevalence study was carried out on 2 cruise ships and 7 ferries docked at the seaports of northern Sardinia in 2004, indicated that 6/7 ferries were positive, 42% (38/90) of the water samples examined were contaminated by Legionella spp. and 77.8% of the water samples contained ≥104 CFU/L . In 1993, a survey carried out in 33 water tanks of yachts in Athens showed that approximately 40% of them were positive for Legionella spp. . In our study, we observed a lower frequency of L. pneumophila sg 1 (in 43 samples), while 87 of the total 96 isolates from the cold and hot water samples were L. pneumophila. In the study carried out in the northern Sardinia, no sample contained L. pneumophila sg 1, however L. pneumophila was isolated in 42/44 samples (95.5%), followed by L. micdadei (4.5%). The strains that were identified in that particular study were L. pneumophila sg 6 (45.2%; 19 samples), sg 2 to 14 (42.9%), sg 5 (7.1%) and sg 3 (4.8%) .
We observed high colonization (17.5%) of Legionella spp. in the cold WDS, compared to the contamination of other buildings such as hotels, where only 6.4% of the cold samples were positive . This may be attributed to the relatively high temperature (median = 23°C, IQR = 20.30–26.00) of cold water aboard ferries. Pipes of cold water pass through the high temperature engine rooms of the ships. Controlling the water temperature in order to prevent Legionella spp. colonization of WDS will be effective only if pipes are insulated. Cold water should be maintained at temperatures below . Where this temperature cannot be achieved due to local conditions, suitable alternative residual disinfection procedures must be used and supported by regular (at least quarterly) testing for legionellae .
In a previous study aboard 2 cruise ships no water system including potable water distribution system, spa or swimming pool, was positive for Legionella spp. . In our study, the relatively high temperature of hot water aboard cruise ships in comparison to the temperatures found in ferries, may partially explain the great difference in Legionella spp. colonization between cruise ships and ferries. In addition, the ten cruise ships that were investigated applied the requirements of the United States Vessel Sanitation Program Manual (VSP) regarding potable water safety and swimming and spa pool management. The legionellae-negative microbiological results of the samples collected from the cruise ships indicate that correct compliance with the requirements that were included in the VSP manual were effective to maintain legionellae-free water systems. These guidelines involve continuous halogenation to at least 2.0 mg/L free residual halogen at the time of bunkering or production of potable water, maintenance of a free residual halogen concentration of ≥0.2 mg/L and ≤5.0 mg/L throughout the distribution system and periodical disinfection of the distribution system and water tanks, and other control measures regarding spa pools . Other additional measures may have been applied aboard the cruise ships.
No significant association was found between L. pneumophila sg 1 concentrations and free chlorine concentration of ≥0.2 mg/L. Contrary, free chlorine of ≥0.2 mg/L inhibited colonization of L. pneumophila sg 2–14, Legionella non pneumophila and Legionella spp. These findings may imply that L. pneumophila sg 1 is more resistant in higher chlorine concentrations. This was also observed in a another study in hotels .
The high levels of Legionella spp. and the high number of positive samples for L. pneumophila found aboard ferries, present a risk for infection, since most ship associated cases have been ascribed to infection by L. pneumophila sg 1 . However, L. pneumophila sg 5 [5, 6], 3, 4, and L. longbeachae sg 1  have been isolated from clinical specimens of passengers with LD. Unfortunately, we did not monitor the development of pneumonia symptoms among passengers and crew members during the study period. Practical difficulties such as the short stay period aboard ferries ranging from a few hours to one day maximum, limits the evidence for linking the source of infection. In addition, many of the passengers are travelling and may be exposed to legionellae in other locales as they travel (e.g., hotels, restaurants, etc.). However, the limited stay period aboard ferries limits the risk of infection.
Since water systems of ferries could be a possible source for infection, during LD case investigation, exposure history should include information on travelling by ships. This will help to identify passenger cases that usually go undetected, to timely implement environmental control measures and to prevent possible future outbreaks in subsequent voyages.
The temperatures of the hot water samples were negatively associated with contamination with L. pneumophila sg 1 and with sg 2 to 14. Similar were the findings of a previous study in hotels WDS . In our study, increases in pH correlated positively with the counts of L. pneumophila serogroups 2 to 14. These findings confirm other reports, which indicated a positive association of L. pneumophila with pH .
The limited number of legionellae-positive WDS of ferries (only 14 positive of the total 21 examined), influence the results of association between colonization and ship factors such as capacity and number of passengers, and therefore we could not draw any conclusions.
LD outbreaks have occurred aboard passenger ships in recent years [8, 22], and even fatal cases have been reported . In many cases the source of the organism was clearly demonstrated [9, 10]to be the ship's water system. LD outbreaks cause negative economic impact on the cruise ship industry and may result in removing a vessel from service, or delay the sailing from a port to allow control measures to be carried out . Three ships are known to have changed owners and names after being linked with cases of legionellosis . During 2004, there were 205 cruise ships with itineraries in Europe, which represents 87% of all cruise ships in operation http://www.europeancruisecouncil.com/faq.html. Measures for the prevention and control of the disease should be further taken to a European or even to an International level. In 2005, the European Commission funded the SHIPSAN project which aims at assessing the usefulness of an EU ship sanitation program and coordinated action for the control of communicable diseases in cruise ships and ferries http://www.eu-shipsan.gr/. We believe that a future European ship sanitation program should include LD prevention and control guidelines.
Apart from the EWGLI guidelines which apply in accommodation sites in general, specific measures for passenger ships have been published by the WHO in the revised draft for review of the Guide to Ship Sanitation and in the book entitled "Legionella and the prevention of Legionellosis" . CDC published recommendations to minimize transmission of Legionnaires' disease from whirlpool spas on cruise ship in 1997 .