Water and dialysate samples were obtained through a collaboration program between National Institute for Quality Control in Health (INCQS/FIOCRUZ), Sanitary Surveillance of the state of Rio de Janeiro, Superintendence of Zoonosis Control, and Sanitary Surveillance of the City of Rio de Janeiro. Data collection in mobile dialysis stations for research and monitoring purposes has started in July 2014.

Sample analyses were carried out according to standardized operational procedures (POPs) from INCQS: POP INCQS nº 65.3210.030 – (Contagem de viáveis totais e bactérias bile tolerantes em produtos farmacêuticos, matérias primas e água para diálise); POP INCQS nº 65.3210.010 – (Contagem de viáveis totais e bactérias bile tolerantes em produtos farmacêuticos, matérias primas e água para diálise); POP INCQS nº 65.3210.008 – (Pesquisa de patógenos em produtos não estéreis e matérias primas de uso em sua fabricação e água para diálise); POP INCQS nº 65.3330.006 – (Ensaio para endotoxinas bacterianas and POP INCQS nº 65.3120.030 - Determinação de alumínio em vacinas por espectrometria de absorção atômica por chama), procedures accredited by the Institute of Metrology and Technological Quality (INMETRO) and qualified by the World Health Organization (WHO).

Water (pre- and post-osmosis) and dialysate samples of approximately 200 mL were collected using sterile flasks. After collection, 1.8% of sodium thiosulfate were added to the solution for chloride removal. Samples were stored at temperatures lower than 10ºC and processed at the day of collection. Total or viable bacterial counting was accomplished using plate count agar (PCA). Samples were diluted in phosphate buffer saline (PBS) at pH 7.4, inoculating 1 mL of sample in 9 mL PBS (1:9 dilution). A volume of 1 mL was then added to 20 mL of PCA medium, melted and cooled to 45-50ºC. After solidification, the medium was incubated at 30-35ºC for 48 ± 3 hours. Counting of colony forming units (CFU) was carried out in plates containing up to 300 CFU.

After sample homogenization, a 100 mL volume were inoculated in 50 mL 3x-concentrated presence-absence (PA) broth and incubated at 30-35ºC for up to 48 hours. Confirmatory analysis was observed in the presence of positive PA indication of bacterial growth (i.e. yellow broth). Cultured cells were then inoculated in 10 mL brilliant green bile broth (BGBB) with Durham tube and incubated for 48 ± 3 hours at 30-35ºC. After incubation period, cultured cells were seeded in a plate containing eosin methylene blue agar (EMBA), MacConkey agar and tryptic soy agar (TSA) for posterior completion of biochemical proofs.

Biochemical identification was carried out according to POP INCQS/FIOCRUZ nº 65.3210.008. Using cells grown in soy casein broth, the samples were seeded in plates containing cetrimide agar, soy casein broth, mannitol salt agar, MacConkey agar and EMBA. After incubation period, the isolated colonies were submitted to biochemical identification tests.

Quantification of Al in samples were accomplished through emission spectrometry (ICP-OES), adapted from Medeiros et al. (2012) and POP INCQS nº 65.3120.030 - “Determinação de alumínio em vacinas por espectrometria de absorção atômica por chama”.

The presence of microorganisms was detected in all hospital units (1-12) and collection points, id est pre-osmosis (hospital’s water supply), post-osmosis (treated water, used in dialysis) and dialysate solution (electrolyte aqueous solution, with or without glucose). In the hospitals 8 and 12 (see Table 1), however, dialysate solutions were unavailable at the time of collection. Table 1 presents the total number of aerobic bacteria from collected samples at each acquisition point. Table 1 also presents quantification of Al levels, which were detected in only three pre- and post-osmosis samples, however, all three samples contained Al values above resolution (RDC 11/2004) limits of 0.01 mg L-1.

Detection of total coliforms and E. coli showed absence in 100 mL of samples from all hospital and collection points, thus indicating low coliform counting in waters used in these dialysis units.

After judicious colony isolation, the following microorganisms were detected and were common to all hospital units: Pseudomonas aeruginosa; P. stutzeri; Stenotrophomonas maltophilia; Acinetobacter anitratus; A. baumannii; A. calcoaceticos sp. Lowffii; Burkholderia cepacia; Sphingomonas paucimobilis; Brevundimonas diminuta; Moraxella osloensis; M. lacunata; M. phenylpyruvia; M. atlantae; Achronobacter xylosoxidans; Ralstonia pickettii. Also, biochemical analysis of endotoxin levels from all hospital samples was shown to be greater than upper limits preconized by the resolution RDC 11/2014 (i.e. < 0.25 EUmL-1).

Corroborating with the literature, it was observed the presence of Pseudomonas species at all hospitals and from at least one collection point (dialysis solution, pre- or post-osmosis water). Conversely, the analysis of total coliforms and E. coli indicated its absence in 100 mL; these bacteria, however, were previously found at dialysis centers from non-developed countries. While experimental conditions from our study and additional cultures (data not shown) were not growth-permissive or did not contain fungi, previous investigations have shown these microorganisms in distribution points from several hemodialysis centers (Ferreira et al., 2015).

Aluminum toxicity may affect bone health and induce encephalopathic syndromes, and previous works have evaluated the effect of Al present in haemodialysis waters on patient’s health (Charhon, Pascale, Meunier, & Accominotti, 1985). From the 12 mobile dialysis services, from pre- and post-osmosis water samples, only three contained Al levels above the Brazilian limits of 0.01 mg L-1 (RDC 11/2004). Aluminum concentrations from these samples, however, were found to be ten to twenty-fold (~ 0.215 mg L-1) higher than legally preconized. It’s important to question whether these Al levels could affect patients if continuously present in the dialysis solutions

Table 1.

Counting of heterotrophic bacteria and quantification of aluminum levels.

*UFC = colony forming units. AI = unavaliable sample. TR = traces/undetectable

Al-Naseri, Mahdi, and Hashim (2013) found that in 60% of hemodialysis centers from Bagdad (Iraq), the bacterial counting was between 50 and 100 CFU mL-1. Also, the presence of P. aeruginosa was correlated with higher average endotoxin levels. Here, it was found relatively low but significant levels of heterotrophic bacteria (~ 2.00 CFU mL-1). Also, these microorganisms were isolated from all mobile services (hospitals) and points of collection (dialysis solution, pre- and post-osmosis water); only samples 7, 9 and 10 showed <10 CFU mL-1 on at least one collection point, however, samples 7 and 9 were among those that had higher Al levels.

Biochemical identification (see Table 2) of isolated bacteria shows the presence of indicator microorganisms (Pseudomonas, especially P. aeruginosa), which were previously shown to negatively affect dialysis patients (Ferreira et al., 2015). Also, S. maltophilia was shown to induce bacteremia in a dialysis patient with long-term central venous catheter (Kara, YilmazM, Şit, Kadiroğlu, & Kökoğlu, 2006) and in end-stage renal disease patients receiving maintenance hemodialysis (Gnanasekaran & Bajaj, 2009). Strateva, Kostyanev, and Setchanova (2012) reported a case of sepsis induced by R. pickettii in a hemodialysis patient from Bulgaria. Kaitwatcharachai, Silpapojakul, SirojJitsurong, & Kalnauwakul (2000) reported a subclavian catheter-related B. cepacia outbreak in nine patients undergoing hemodialysis; authors also observed that B. cepacia formed biofilms while growing in the catheters.

Sphingomonas. paucimobilis, also detected in the water samples of this study, is an uncommon cause of peritonitis and other dialysis-related infections. Recent reports, however, have found this microorganism in co-infections caused by opportunistic catheter growth. One case of S. paucimobilis­-induced peritonitis was reported by Lee et al. (2013). There are little or no evidence of B. diminuta infection in dialysis patients; other species, B. vesicularis, was previously shown to cause peritonitis (Choi et al., 2006). Badrising, Bakker, Lobatto, and Van Es (2014) reported a peritonitis case due to a co-infection by R. radiobacter and M. osleonsis. Other Moraxella species have also been shown to cause dialysis-related infections.

In 2009, A. xylosoxidans was shown, in a case report, to induce prosthetic valve endocarditis, where it was also characterized as an emerging pathogen in catheter-related infections used in dialysis (Ahmed, Nistal, Jayan, Kuduvalli, & Anijeet, 2009). Finally, Acinetobacter species are among the main opportunistic microorganisms in nosocomial infections, and A. baumannii strains are usually characterized as highly dangerous pathogens due to its mechanisms of antibiotic resistance, clinical impact and co-infection potential (Peleg, Seifert, & Paterson, 2008).

The monitoring of clinics and mobile dialysis units by sanitary surveillance, in collaboration to INCQS, demonstrates that lowering tolerance levels for chemical and microbiological quality of HD fluids is effective in optimizing quality standards within those dialysis services. Medical centers should, however, try to surpass such minimum standards, thus ensuring patient safety and accordance to legal determinations. Application of ultrapure dialysates, for instance, may represent a significant advance in the treatment of patients with kidney disease.

Although a great variety of microorganisms have been detected in this study, P. aeruginosa was predominant (data not shown), corroborating with previous results (Favero et al., 1992; Morin, 2000). This species produces stable biofilms, resistant to several disinfection treatments, thus it’s important to consider determining a legal parameter (e.g. presence/absence in 10 mL) for dialysates. After exhaustive analysis (2014-2016), data strongly indicates that microbiological contamination found in samples from the hospital water supplies (i.e. pre-osmosis) are resistant to disinfection and filtration procedures, allowing the presence of pyrogens (i.e. lipopolysaccharides) and live bacteria throughout the dialytic process.

Our data indicate the need for establishing novel quality standards within national legislation, determining more restrict parameters for evaluation and analysis of the water used in mobile dialysis in the state of Rio de Janeiro. In addition, it is important to highlight, as previous reviewed in Ferreira et al. (2015), that bacterial contamination in HD centers is a worldwide reality, therefore mobile dialysis units should also be verified for such parameters in other Brazilian States and overseas.

freitashr@biof.ufrj.br