Microorganism levels in spray from warm-water bidet toilet seats: factors affecting total viable and heterotrophic plate counts, and examination of the fluctuations and origins of Pseudomonas aeruginosa

Warm-water bidet toilet seats are equipped with a device that sprays warm water (spray water) on the external genitalia and anus after urination or defecation (Japan Monthly Web Magazine 2011). A consumer behavior survey conducted by the Japanese Cabinet Office showed that 79.1% of Japanese households have at least one such toilet seat, with every 100 households owning 110.9 units on average (effectively one per household) (Cabinet Office Government of Japan 2017). Broadly, two types of warm-water bidet toilet seats are available: the tank type and the on-demand type. With the tank type equipped with reservoir water heaters, the spray water is warmed to a suitable temperature in a tank with reservoir water heaters, whereas in the on-demand type tankless systems equipped with an instantaneous water heater, the water is warmed as needed inside a device with tankless, instantaneous water heaters (bidetsPLUS 2017) (Figure 1). Tank-type products are both cheaper and more common than on-demand types in Japan (Ministry of Economy Trade and Industry 2006).

Despite the reported health-related advantages (benefit to bowel movement, no clinical health risk for preterm birth and bacterial vaginosis, low incidence rates for bacterial vaginitis) of using warm water in bidets (Uchikawa et al. 2007; Asakura et al. 2013; Kiuchi et al. 2017), there are also health-related disadvantages (aggravation of vaginal microflora, correlation for an itch on the anus) associated with the use of these toilet seats (Ogino et al. 2010; Tsunoda et al. 2016). Although Katano et al. (2014) showed bacterial contamination (i.e., Pseudomonas aeruginosa and Escherichia coli isolated) of the spray water from the tank-type warm-water bidet toilet seats, research on both the amounts and the types of microorganisms present in the spray water is sparse.

Previously, we conducted a cross-sectional survey on the hygienic conditions of 127 warm-water bidet toilet seats in restrooms on a university campus, revealing relationships between the total viable count (TVC), heterotrophic plate count (HPC), and residual chlorine; the proportion of Pseudomonas aeruginosa in the spray water; and the composition of the microbiome in the spray water. We showed that the TVC, fecal indicator bacteria, and P. aeruginosa were present in the spray water at very low frequencies and concentrations, indicating that the overall hygienic safety was maintained (Iyo et al. 2016).

In this study, we conducted a long-term survey of the hygienic conditions of warm-water bidet toilet seats in a research building in a university campus (12 men's restrooms and one women's restroom), including restrooms in which P. aeruginosa was previously detected in the spray water (one men's restroom and one women's restroom). We thus empirically examined TVCs and HPCs, as well as P. aeruginosa behavior in the spray water; factors affecting these parameters (residual chlorine and differences between the tank and on-demand types); and the origins of P. aeruginosa from the viewpoint of preventing opportunistic infections due to P. aeruginosa. We also studied previously contaminated bidets for P. aeruginosa persistence, the age of the systems, variations over time, and the location of P. aeruginosa within the bidets. P. aeruginosa secretes a biopolymer that enables it to form biofilms and strong resistance to chlorine (Silva et al. 2008). The warm-water tanks in the bidets also promote biofilm formation and long-term survival of P. aeruginosa on the inner walls of these tanks.

Our aim was to conduct an appropriate microbial evaluation of the on-demand and tank-type warm-water bidet toilet seats in this study.

In a previous survey of 127 tank-type, warm-water bidet toilet seats, P. aeruginosa was detected in the spray water of only two (one men's and one women's restroom) in only one research building on a university campus (Iyo et al. 2016). The cause of the P. aeruginosa water contamination was not determined. Hence, focusing on the research building containing two warm-water bidet toilet seats where Pseudomonas aeruginosa was detected in the spray water, we surveyed warm-water bidet toilet seats in 12 men's and one women's restrooms in the research building (P. aeruginosa positive) on the same university campus, in this long-term field study. However, considering the difficulty for male researchers to collect the spray water from bidet toilet seats in women's restrooms, only one female bidet toilet seat in which P. aeruginosa was specifically detected in spray water was surveyed. The toilet seats were replaced with either tank-type or on-demand type seats in accordance with the study plan. Breakdown (disassembly) analysis was performed to detect P. aeruginosa, Escherichia coli, and enterococci in the piping of the toilet seats.

The levels of residual chlorine (DPD method), TVCs (CompactDry^® TC; Nissui Pharmaceutical Co., Tokyo, Japan; growth at 37 °C for 2 days), HPCs (R2A agar culture medium; Eiken Chemical Co., Tokyo, Japan; growth at 20 °C for 7 days), P. aeruginosa (USEPA-Approved Defined Substrate Technologies Pseudalert^® and Quanti-Tray^® systems, IDEXX Laboratories Inc., Westbrook, ME, USA) and E. coli (USEPA-Approved Defined Substrate Technologies Colilert^® and Quanti-Tray^® systems, IDEXX Laboratories Inc., Westbrook, ME, USA) were measured in spray water and tap water (control), as described previously (Iyo et al. 2016). Samples of microbes in the internal piping and spray nozzles were collected by scraping the inside with a cotton swab or other similar tools. The collected samples were then inoculated into a nutrient broth (Eiken Chemical Co., Tokyo, Japan) and incubated at 37 °C for 48 h. The incubated solution was then smeared on agar plates. The media used for smear culture were nalidixic acid cetrimide (NAC) agar medium (Eiken Chemical Co., Tokyo, Japan), sodium dodecyl sulfate magenta-GAL–X-GLUC agar medium (ELMIEX Limited, Tokyo, Japan), and Chromocult^® enterococci-agar (Merck KGaA, Damstadt, Germany) for P. aeruginosa, E. coli, and enterococci, respectively. The culture temperature and duration used with these media were based on the manufacturers' instructions.

To investigate the origins of P. aeruginosa, the culture medium for P. aeruginosa was diluted appropriately and subjected to isolation culture to obtain a few single colonies on NAC agar plates. Subsequently, DNA was extracted from the isolated colonies with the CicaGeneus™ DNA Extraction Reagent (Kanto Chemical Co., Tokyo, Japan), and PCR-based ORF typing (POT) values (the combination of POT1 value and POT2 value) were measured using the CicaGeneus™ Pseudo POT (PCR-based ORF Typing) Kit (Kanto Chemical Co., Tokyo, Japan). Using the POT method with two sets of primer pairs, P. aeruginosa strains were classified by 10 ORFs from islets, 5 ORFs from genomic islands, and the metallo-beta-lactamases bla_IMP and bla_VIM (multidrug-resistant Pseudomonas aeruginosa gene) (Suzuki et al. 2015).

Figure 2 shows the timelines for the surveys of the warm-water bidet toilet seats, which were conducted between 2012 and 2016 (Surveys 1–6).

Survey 1: We followed up on the levels of P. aeruginosa in the spray water from a women's tank-type warm-water bidet toilet seat (TT-1f) and a men's tank-type warm-water bidet toilet seat (TT-1 m), wherein P. aeruginosa was detected in spray water from both in a previous survey. Samples of spray water were collected in the morning.

Survey 2: To eliminate the effects of ageing of the warm-water bidet toilet seats, two toilet seats (TT-1 m and TT-2 m) were replaced with new ones (TT-1 m-2 and TT-2 m-2). After approximately half a year of acclimation to stabilize the state of the biofilm in the toilet seats, spray water was surveyed in each toilet seat. Data from seats TT-1 m–TT-7 m were used as control for those from seats TT-1 m-2–TT-2 m-2. Spray water samples were collected in the morning.

Survey 3: We investigated daily (five times a day at 9:00, 12:00, 15:00, 18:00, and 21:00) and weekly variations (five weekdays from Monday to Friday) of residual chlorine, HPC, and P. aeruginosa levels without disinfecting the nozzle surface.

Survey 4: We examined the fluctuations in P. aeruginosa levels in spray water from seven men's tank-type warm-water bidet toilet seats (TT-1 m-2, TT-2 m-2, TT-3 m, TT-4 m, TT-5 m, TT-6 m, and TT-7 m) without disinfecting the nozzle surface.

Survey 5: In total, 12 men's tank-type warm-water bidet toilet seats (TT-1 m-2, TT-2 m-2, TT-3 m, TT-4 m, TT-5 m, TT-6 m, TT-7 m, TT-8 m, TT-9 m, TT-10 m, TT-11 m, and TT-12 m) were replaced with new, on-demand-type seats (OD-1 m, OD-2 m, OD-3 m, OD-4 m, OD-5 m, OD-6 m, OD-7 m, OD-8 m, OD-9 m, OD-10 m, OD-11 m, and OD-12 m). We surveyed the quality of spray water once a month between August and December. Spray water samples were collected in the morning. When comparing the spray water quality between the toilet seat types, we used data from November for both the tank-type and on-demand-type toilet seats because data for the tank type were collected in November.

Survey 6: After confirming the P. aeruginosa levels and POT values in the spray and tank water of a women's tank-type warm-water bidet toilet seat (TT-1f) and in the spray water from a men's on-demand-type warm-water bidet toilet seat (OD-1 m), we examined the internal P. aeruginosa levels and POT values after dismantling these toilet seats. Subsequently, we prepared water piping from which tap water-derived P. aeruginosa was eliminated through microfiltration and then installed two on-demand-type warm-water bidet toilet seats (OD-1f and OD-1 m-2), using the piping. Spray water samples were collected from these seats and compared with two types of tap water as controls (Figure 3). Spray water and tap water samples were tested approximately once every 2 weeks for P. aeruginosa, E. coli, and enterococci. We also dismantled seats OD-1f and OD-1 m-2 thereafter to examine the presence and POT values of P. aeruginosa inside.

While calculating the geometric means (GMs) and geometric standard deviations (GSDs) of residual chlorine, TVC, HPC, and P. aeruginosa levels, we also determined the detection limits. Statistical analysis of factors affecting the quality of spray water was conducted by performing analysis of variance (ANOVA). The quality of spray water from tank and on-demand types was compared by performing the Mann–Whitney U test and the non-parametric Kruskal–Wallis test. Statistical analysis was performed using the Statistical Package for the Social Sciences, version 24 (IBM, Armonk, NY, USA). The significance level was set at P < 5%.

As indicated in Table 1, P. aeruginosa was detected in the same two tank-type warm-water bidet toilet seats (TT-1f and TT-1 m) that showed P. aeruginosa in a previous survey. In particular, P. aeruginosa was confirmed to have persisted for at least 2 years in the tank type warm-water bidet toilet seats (TT-1f) installed in a women's restroom (Survey 1). In seven of the tank-type seats (TT-1 m-2, TT-2 m-2, TT-3 m, TT-4 m, TT-5 m, TT-6 m, and TT-7 m), P. aeruginosa was detected in the spray water from two toilet seats (TT-1 m-2 and TT-2 m-2), albeit at low concentrations (Survey 4). These findings indicated the possibility of long-term survival of P. aeruginosa in warm-water bidet toilet seats (TT-1f), as well as the presence of viable P. aeruginosa attributable to use of the toilet seats (TT-1 m and TT-1 m-2). During the long-term survey of six warm-water bidet toilet seats, two (TT-3 m and TT-4 m) occasionally showed P. aeruginosa in the spray water. One of the on-demand-type seats (OD-1 m) consistently showed low levels of P. aeruginosa in its spray water. As also indicated in Table 1, although one of the tank-type seats (TT-2 m-2) carried P. aeruginosa in its spray water before replacement, the bacterium disappeared after the seat was replaced with a new on-demand-type seat (OD-2 m). Thus, the only toilet seats that consistently showed P. aeruginosa were TT-1f and TT-1 m through OD-1 m (Survey 5).

Following acclimation of two tank-type warm-water bidet toilet seats (TT-1 m-2 and TT-2 m-2), the residual chlorine, TVC, and HPC levels in the spray water were measured over approximately a half-year period. The mean values were 0.10–0.15 mg/L for residual chlorine, ≤1 CFU/mL for the TVC, and approximately 20,000 CFU/mL for the HPC, which were nearly identical with the mean values obtained with seats TT-1 m–TT-12 m (extracted from our previous data (Iyo et al. 2016); analyzed here as shown in Figure 4(a)). As shown in Figure 4(a), the TVC was nearly undetectable in the spray water from seats TT-1 m-2 and TT-2 m-2, and even when TVC was detected, the value did not exceed 10 CFU/mL. In addition, a weak negative correlation was observed between the residual chlorine and HPC levels (correlation coefficient [R]: 0.457, coefficient of determination [R²]: 0.209).

As shown in Figure 4(b), the TVCs were ≤1/mL in the spray water from the on-demand-type warm-water bidet toilet seats. When the TVCs were detected in the spray water, they tended to decrease as the residual chlorine levels increased. Most toilet seats (9 out of 11; 82%) that showed a TVC exceeding 1 CFU/mL were the on-demand-type seats (OD-1 m and OD-2 m) that replaced the previously P. aeruginosa-positive tank-type seats (TT-1 m-2 and TT-2 m-2). A weak negative correlation was also found between the residual chlorine and HPC levels (R: 0.491, R²: 0.241).

Figure 5 shows the daily and weekly variations of the residual chlorine, P. aeruginosa, and HPC levels. The only source of chlorine was the tap water, and the concentration of chlorine decreased due to heating and remaining stagnant for a long period. The graph shows an inverted U-shaped pattern in which the residual chlorine levels tended to be relatively low in the morning and peak around noon before dropping towards the evening. The HPC and P. aeruginosa levels, however, tended to be higher in the morning before decreasing in the evening. This result may reflect nocturnal growth of bacteria within the water tank, with subsequent washout and dilution of bacteria using spray water tank throughout the day. Alternatively, this result may reflect an inverse relationship between bacterial and chlorine levels. However, no weekly variations were observed. The difference in residual chlorine levels between TT-1m2 and TT-2 m-2 was likely due to the frequency of use.

After the tank-type toilet seats were replaced with on-demand-type seats, we compared the residual chlorine, TVC, and HPC levels in the spray water. Residual chlorine level increased more in the on-demand type than in the tank type, and HPCs decreased more in the on-demand type than in the tank type (P < 001; Figure 6). TVCs decreased more in the on-demand type than in the tank type, but the difference was not statistically significant (P = 0.148; Figure 6). The on-demand-type of bidet toilet seat does not have a tank with reservoir water heaters, and the residual chlorine in tap water scarcely decreases because of the low degree of residual chlorine evaporation. Consequently, the residual chlorine level in tap water was highly correlated with that in spray water (R: 0.963, R²: 0.926; Figure 7).

We dismantled tank-type toilet seat TT-1f and on-demand-type toilet OD-1 m, both of which were consistently P. aeruginosa-positive, and analyzed the spray water (for both) and the tank water (for TT-1f). P. aeruginosa was detected near the filter in the piping and in the spray water in seat TT-1f, as well as in the piping within the nozzle and on the nozzle surface in seat OD-1 m (Figure 8). The P. aeruginosa POT values (i.e., POT1 and POT2) differed between TT-1f and OD-1 m.

Table 2 shows the results of an experiment performed to determine the origins of P. aeruginosa, as described in Figure 3. With seat OD-1f, only the tap water sample upstream of the microfilter (tap water 1) consistently showed high P. aeruginosa levels. In contrast, seat OD-1 m-2 occasionally showed P. aeruginosa in the spray water. The POT values in TT-1f were identical with those in OD-1f, but the POT values in OD-1 m and OD-1 m-2 were similar but not identical to each other. P. aeruginosa was not detected in the water from the washbasins of OD-1f and OD-1 m-2 in the restroom where these seats were installed (tap water 2). When OD-1f and OD-1 m-2 were dismantled for further investigation, P. aeruginosa was detected in the inlet of the microfilter for OD-1f and in the spray nozzle for OD-1 m-2 (Figure 8). The P. aeruginosa POT values differed between OD-1f and OD-1 m-2; in the case of the former, the POT values at the inlet of the microfilter were identical to those in tap water 1 (Table 2). With OD-1 m-2, P. aeruginosa was detected at the tip of the spray nozzle, but the POT values differed from those in the OD-1 m-2 spray water and tap water 1 (Figure 8 and Table 2).

The ability to maintain and regulate residual chlorine is an important built-in factor for maintaining the hygiene and disinfection of spray water. Since there was no obvious change in the toilet seats surveyed over time, the observed decrease in residual chlorine was not due to deterioration of the toilet seats themselves. In this study, we only surveyed warm-water bidet toilet seats in 12 men's and one women's restrooms in the research building in our university. Further studies are needed on warm-water bidet toilet seats by age, location, and time, increasing the number of samples.

When P. aeruginosa was not detected in the spray water over an extended duration, the probability of a measurable TVC was low in both the tank and on-demand types (TT-1 m-2, TT-2 m-2, OD-3 m, OD-4 m, OD-5 m, OD-6 m, OD-7 m, OD-8 m, OD-9 m, OD-10 m, OD-11 m, and OD-12 m). The overall HPC levels tended to be lower in the on-demand type (3 log₁₀) than in the tank type (4 log₁₀). The lack of internal parts, such as a storage tank to heat the tap water in the on-demand type, may account for this tendency. The results of our survey clearly differed from those reported by Katano et al. (2014), who showed bacterial contamination in the spray water from tank-type warm-water bidet toilet seats.

Seasonal temperature fluctuations in tap water cause the residual chlorine levels to fluctuate as well (Futigami 2011; Fisher et al. 2012). As mentioned above, the structure of the on-demand-type warm-water bidet toilet seat makes it susceptible to the residual chlorine levels in tap water (Figure 7). Therefore, as shown in Figures 4(b) and 7, the residual chlorine levels also fluctuated seasonally in the spray water from the on-demand-type toilet seat. An inverse relationship was observed between the residual chlorine levels and the HPCs (Figure 4(a) and 4(b)). The HPC is a widely used indicator of the bacterial count and biofilm growth inside pipes and tanks, which represent an aquatic environment with low nutrition (Ndiongue et al. 2005; Berry et al. 2006). The residual chlorine levels tended to be relatively low in the morning and peaked around noon before dropping towards the evening. This was probably due to heating of the tank water, coupled with a prolongation of the water-retention time from evening to dawn, which decreased the residual chlorine level in the tank. When variance analysis of the HPCs was performed with nozzle disinfection and temporal variation in sample collection studied as experimental parameters, the HPC values were significantly associated with the temporal variation. This result suggested that the HPCs of the spray water were more likely to be affected by the time of collection than by contamination on the nozzle surface (Iyo et al. 2015).

Furthermore, P. aeruginosa can remain inside biofilms on pipes for long periods (Mena & Gerba 2009). P. aeruginosa is not pathogenic in immunocompetent, healthy individuals; therefore, there is no need to worry about the risk of infection from this organism from warm-water bidet toilet seats in regular households. However, it can cause opportunistic infections in individuals with reduced immunity (Reuter et al. 2002; Aumeran et al. 2007; Trautmann et al. 2008). P. aeruginosa can easily develop resistance to antibiotics (Bert et al. 1998; Ferroni et al. 1998; Durojaiye et al. 2011). Therefore, when warm-water bidet toilet seats are placed in hospital wards, sufficient attention needs to be paid to the possibility of P. aeruginosa contamination and propagation, which should include regular testing for P. aeruginosa.

Although tank-type toilet seats have lower concentrations of residual chlorine in the tank compared with on-demand-type toilet seats, the former has a longer hydraulic retention time, which can influence the residual chlorine values (Zamyadi et al. 2012).

Empirical investigation into the origins of P. aeruginosa revealed the presence of both feces-derived and tap water-derived strains. The genotypes obtained using POT are described with two types of POT values (POT1 and POT2). The POT1 and POT2 values for a presumed tap water-derived strain (detected in TT-1f and OD-1f, and compared with detection point of both TT-1f and OD-1f in Figure 8) were only 201 and 20, respectively, whereas those values for presumed feces-derived strains (detected in OD-1 m and OD-1 m-2) were [POT1: 840 and POT2: 0], [POT1: 876 and POT2: 0], [POT1: 620 and POT2: 1] and [POT1: 636 and POT2: 0], respectively. These variations in POT values probably reflect the fact that contamination of the nozzle was caused by feces from different users. As none of the POT2 values of the detected strains exceeded 64 (metallo beta-lactamase-negative) (Suzuki et al. 2015), these were most unlikely to be multidrug-resistant P. aeruginosa (MDRP) strains. From the viewpoint of preventing opportunistic infections due to P. aeruginosa, reducing the presence of MDRP to undetectable levels should reduce the infection risk.

Thus, here, we reported a long-term survey of the hygienic conditions of warm-water bidet toilet seats in a university research building. Microorganism levels in spray water from warm-water bidet toilet seats showed that the overall hygienic safety was maintained. In a future study, an increased number of samples from warm-water bidet toilet seats is needed, especially for bidet toilet seats installed in female restrooms. In addition, to improve the hygienic safety of spray water from warm-water bidet toilet seats in their many different locations, self-cleaning spray nozzle mechanisms should be improved and structural changes should be made to prevent the growth of biofilms inside both warm-water tanks and pipes.

We performed a long-term survey of the behavior of microorganisms within warm-water bidet toilet seats. The TVCs were ≤1/mL in spray water samples from both on-demand and tank-type warm-water bidet toilet seats. The HPCs in the spray water were 4 log₁₀ in the tank type and 3 log₁₀ in the on-demand type. However, the residual chlorine and HPC levels in the spray water were only weakly correlated. When Pseudomonas aeruginosa was detected in spray water over an extended duration, the bacterial origin was likely feces or tap water. The POT values of presumably tap water-derived P. aeruginosa were uniform, whereas those of presumably feces-derived P. aeruginosa varied. The isolated P. aeruginosa strain was unlikely to possess enzymes conferring multidrug resistance. Microorganism levels in spray water from warm-water bidet toilet seats showed that the overall hygienic safety was maintained in these facilities.

We would like to thank Editage (www.editage.com) for English language editing and publication support. This work was supported by the Japan Sanitary Equipment Industry Association (JSEIA). JSEIA had no role in the design of the study; in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the article for publication. The authors declare no conflicts of interest.