Ionization Failure Not Due to Resistance

Sir—Rohr et al. [1] report their experience with copper-silver ionization in eradicating Legionella in the hot water plumbing systems of a German university hospital. They report that the percentage of 1-L samples of water from distal sites that were positive for Legionella (the detection limit was 1 cfu/L) were as follows: before installation of the ionization unit, 100%; in year 1 after installation, 55%; in year 2, 76%; in year 3, 78%; and in year 4, 75%. From this data it appears that the copper-silver ionization system installed in this hospital did not effectively control Legionella in the water plumbing system, even in the first year. Although the number of cfu of Legionella detected decreased in the first year, the percentage of samples positive for Legionella remained as high as 55%. Cases of hospital-acquired legionnaires disease correlate directly with percentage of samples that are positive for Legionella [2, 3–4], but not with the number of Legionella organisms detected at each distal site. We suspect that the copper-silver ionization system could not eradicate Legionella because the concentration of ions in the water system was inadequate (table 1). The efficacy of copper-silver ionization depends on maintaining adequate concentrations of both copper and silver ions in the water system. Legionella positivity was significantly reduced (from 70% to 0%) only after the copper and silver ion concentrations reached 400 and 40 µg/L, respectively [5]. Other studies have also shown that maintaining ion concentrations between 200–400 μg/L of copper and 20–40 µg/L of silver was crucial [6–8]. Copper and silver ion concentrations in the water of a German hospital in the 4 years after an ionization system was installed, as reported by Rohr at al. [1]. In this hospital, the background copper ion concentration was 200 µg/L, and the average copper ion concentration after installation of the ionization system was also 200 µg/L (range, 131–1159 µg/L). This suggests that insufficient copper ions were released into the plumbing system by the ionization system. In addition, the silver ion concentration applied in the water system (5µg/L) was far below the effective concentration of 20–40 µg/L recommended previously. In vitro results from our investigations [9] differ from those of Rohr et al. and show that copper and silver have synergistic activity against Legionella [10]; silver ions alone were inferior to the combination of copper and silver. In hospitals that use silver ions as the only disinfectant, the recommended silver concentration is 60–100 µg/L (Gunner Lyslo, personal communication), which is much higher than the 2–44.6 µg/L achieved in the hospital discussed by Rohr et al. One possible explanation for the low concentrations of copper and silver ions in the water system has to do with the distribution of the ions within it. The ionization systems were installed on the feed line of the main hot water station and on the hot water feed line of the peripheral building (400 m away from the central hot water station). It appears that the ionization system operates on what is basically a “pass-through” mode: incoming water flows through the ionization flow cell and ions are released into the water. This system has been described in a previous article by Rohr et al. [11]. The disadvantage of this type of installation is that the ion concentrations will fluctuate depending on the flow rate of incoming water. If a large volume of water flows through the flow cell, copper and silver ions are diluted and the concentrations will be low. I suggest the authors collect water samples at the distal sites 10–30 s after flushing and 5 min after. Collecting water samples after a 5–10 min flush, as described in their article, may actually overestimate the concentrations in the water plumbing system, because the ion concentrations recorded are not the ion concentrations already in the hot water system, but the concentration just released from the generator: the ions generated from the feed line will probably take less than 5–10 minutes to get to the outlet for sample collection. Regardless of the timing of sample collection, the ion concentrations are too low to effectively kill Legionella. Ionization systems installed in US hospitals generally place the flow cells on the hot water recirculating lines. The advantage of this approach is that it recirculates the hot water, so the ion concentrations can achieve a steady level adequate to control Legionella. Given the German drinking water regulation that maximum silver concentration cannot exceed 10μg/L, electrodes of 60% silver and 40% copper will not be effective. Ionization systems installed in the United States (manufactured by Tarnpure, Pittsburgh and LiquiTech USA, Willowbrook, IL) use 30% silver and 70% copper electrodes. A solution to Germany's 10µg/L limit of silver may be to increase the copper concentration to ⩾400 µg/L while maintaining a low silver concentration of <10 µg/L. The claim of Rohr at al. that Legionella developed resistance to silver is unsupported by any data in their report. However, their data do document that the failure to eradicate Legionella is likely due to insufficient copper and silver ions in the water. A possible solution might be to increase copper ion concentration to 400 µg/L and maintain sustained copper and silver ions in the water system.