**2. Results**

#### *2.1. Legionella Positivity Correlation*

A total of 269 samples were cultured for *Legionella* from domestic cold and hot water samples in 28 di fferent buildings and from 45 recirculating hot water systems. *Legionella* was cultured from 65/269 (24.2%) samples from 15/28 (53.6%) buildings sampled. The hot water return line sample was positive in 12/45 (26.7%) systems (Figure 1). Positive distal sites (faucets) were observed in 20/45 (44.4%) of the hot water systems. *L*. *pneumophila* was the only species of *Legionella* isolated from these water samples. There was a trend towards larger buildings having more distal site positivity, however this was not statistically significant (*p* = 0.06).

*Pathogens* **2019**, *8*, 295

**Figure 1.** Percent *Legionella* positivity and concentration was highest in distal site hot water and hot water return lines. The bars represent the standard error of the mean for the average *Legionella* concentration.

*Legionella* was isolated from at least one distal site in 91.7% (11/12) of the hot water systems that also had *Legionella* isolated from the hot water return line, with an average distal site positivity of 83.3% ± 8.7% (Figure 2). *Legionella* was isolated from at least one distal site in 27.3% (9/33) of the hot water systems with a *Legionella* negative hot water return line, with an average distal site positivity of 13.1% ± 4.3% (Figure 2). In 35 of the 45 sampled hot water systems, there was agreemen<sup>t</sup> between *Legionella* distal site positivity and hot water return line *Legionella* positivity, resulting in a 77.8% concordance rate.

We then analyzed if hot water return line *Legionella* positivity was able to correctly predict whether the distal sites would be positive or negative for *Legionella*. Hot water return line positivity was related to distal site positivity (*p* = 0.002), with a sensitivity of only 55% (11/20) and a specificity of 96% (24/25). However, when hot water return line positivity was used as a screening tool for distal site positivity, the positive predictive value was 91.7% and the negative predictive value was only 72.7%. The average distal site concentration of *Legionella* in systems with a positive hot water return was 483.5 ± 147.4 CFU/mL, versus 20.7 ± 8.4 CFU/mL in negative hot water returns (*p* < 0.003) (Figure 2).

**Figure 2.** *Legionella* distal site positivity and concentration was highest for hot water systems with *Legionella* positive hot water return lines. The bars represent the standard error of the mean for average *Legionella* concentration.

#### *2.2. Heterotrophic Plate Count (HPC) and Chemical Parameter Correlation*

All samples collected were also cultured for heterotrophic plate count (HPC) bacteria (Table 1). HPC concentrations in hot water samples ranged from 3 CFU/mL to 2,100,000 CFU/mL. Statistical analysis showed no correlation between distal site HPC concentration and *Legionella* distal site positivity (*p* = 0.788) (Figure 3). The best-fit linear regression line demonstrates that HPC concentration explains only 0.68% of the variance in *Legionella* distal site positivity (R<sup>2</sup> = 0.0068).

**Table 1.** Average concentrations (±standard deviation) of total bacteria (heterotrophic plate count—HPC) and physicochemical parameters.


NA: 48.7% samples were with Fe concentration below detection limit of 0.03 mg/L, detectable Fe concentration ranged from 0.03 to 4.6 mg/L; 76.2% samples were with Pb concentration below detection limit of 0.001 mg/L; Pb concentrations in detectable samples ranged from 0.001 to 0.63 mg/L. NT: not tested. TOC = Total Organic Carbon.

**Figure 3.** *Legionella* distal site positivity and average distal site HPC concentration are not significantly associated. Average HPC concentration is represented on a logarithmic scale x-axis. A line of best fit has been added to show the relationship between HPC concentration and distal site positivity.

The water from distal outlets was analyzed for pH, free chlorine, calcium (Ca), magnesium (Mg), zinc (Zn), manganese (Mn), copper (Cu), and TOC. These results were analyzed for correlation with distal site positivity (Table 1). None of the measured parameters were shown to have a correlation with *Legionella* distal site positivity (*p* values > 0.05). Hot water return line pH, free chlorine, calcium, magnesium, zinc, manganese, and copper also were not correlated with *Legionella* distal site positivity (*p* values > 0.05). No comparisons could be made between either distal site or hot water return line iron (Fe) or lead (Pb) concentrations and *Legionella* distal site positivity because the concentrations were below the lower detection limit of the test method in 49% and 77% of samples for iron and lead, respectively.
