Waterborne disease strikes back
Since the work of Dr. John Snow on waterborne disease in the 1850s, important advancements in and widespread implementation of drinking water treatment and disinfection have dramatically improved public health. Disinfection technologies such as chlorination and UV have reduced the risk of enteric pathogens to a level where water is safe to drink directly from the tap in many places. In the case of the United States, while total waterborne disease outbreaks have decreased over the last decades, there has been an increase in the annual proportion of outbreaks reported in individual water systems and the proportion of outbreaks associated with premise plumbing deficiencies in public water systems. Furthermore, there has been an increasing importance of Legionella (for which US EPA has regulated) and other opportunistic premise plumbing pathogens (OPPPs).
Opportunistic premise plumbing pathogens: choking on our aspirations
OPPPs generally include Nontuberculous Mycobacteria (NTM), Pseudomonas spp., and Legionella spp. They generally cause infection through the inhalation of OPPP-laden aerosols and especially affect immunosuppressed or elderly people. In contrast to the acute gastroenteric infections caused by traditional waterborne pathogens, OPPPs mainly cause chronic pulmonary disease. In a study of waterborne disease burden in the United States in 2014, cases of cryptosporidiosis, giardiasis, norovirus, and otitis externa were far more numerous than OPPP-related diseases, but OPPPs were estimated to account for the majority of hospital stays, deaths, and direct healthcare costs ($2.4 billion out of a total of $3.3 billion). NTM disease was particularly burdensome even within the OPPP-related diseases, causing $1.5 billion in direct healthcare costs. Indeed, NTM disease requires prolonged (12-18 months) treatment, often involving multi-drug therapy, and treatments are insufficient due to the growing antibiotic resistance and diversity of NTM infections. In the context of drinking water, we looked to the US EPA's recent Contaminant Candidate List 5 (CCL 5), which newly added Mycobacterium abscessus (other OPPPs, namely M. avium, Legionella pneumophila, and Pseudomonas aeruginosa, were already included since earlier editions of the CCL), and chose M. abscessus as the focus of our study.
UV-LEDs: a new hope
In drinking water systems, OPPPs are resistant to chemical disinfectants (e.g., chlorine), form biofilms in pipes, and can colonize premise plumbing systems. Since OPPPs can be selected for (especially in chlorinated waters, which suppress bacteria that would have otherwise outcompeted OPPPs) and regrow throughout water distribution systems, point-of-use (POU) treatment may be a more effective strategy than centralized treatment. Recent advances in ultraviolet light-emitting diode (UV-LED) technology have enabled high UV fluences (~10-40 mJ/cm2) to be delivered even in compact, POU water disinfection devices. However, the fluence-response of the important M. abscessus (among other OPPPs) to UV-LED radiation remains unclear. Thus, we reported in this study the fluence-response curve of M. abscessus to 280 nm UV-LED radiation, a popular wavelength used in current UV-LED POU disinfection devices.
Our results showed a sigmoidal fluence-response curve (Figure 1) exhibiting shoulder and tailing regions, which may be related to clumping as well as various phenomena common to bacteria. Disinfection units that can deliver a fluence of 10 mJ/cm2 are expected to achieve nearly 2 log (99%) inactivation of M. abscessus. Furthermore, 4 log inactivation can be expected at a fluence of about 15 mJ/cm2. Considering that up to 3500 CFU/L of total culturable Mycobacteria has been reported in distal sites of a water distribution system, we recommend that disinfection devices be designed to deliver >~15 mJ/cm2 fluence to reduce the M. abscessus fraction to a trace or negative level.
Figure 1. Fluence-response profile for M. abscessus in PBS solution exposed to 280 nm UV-LED radiation. Regression curves based on a log-linear model (Red dotted line) and Geeraerd’s model (Green line) are shown. Each data point is the mean (n = 3), and error bars represent the range of values.
Log-linear and Geeraerd's models were fitted to the curve, and a fluence-based inactivation rate constant, k, of 0.356 cm2/mJ was derived from the log-linear model and compared with those of other reference waterborne pathogens (Table 1.). While 280 nm UV-LED data was not available in all the compared studies, the overall trend suggests that M. abscessus is among the most UV-resistant OPPPs. Further research might lead to appropriate surrogate microorganisms for reactor validation.
Table 1. Comparison of fluence-based inactivation rate constants from recent UV inactivation studies for selected NTM and microorganisms relevant to drinking water disinfection. M. Mycobacterium.
It is important to note that Mycobacteria are known to clump ubiquitously, and we observed the presence of clumps (data not published) in our washed resuspension. We also expect clumped Mycobacteria to exist in real-world drinking waters, whether as a result of clumped growth or the sloughing of biofilms. Clumping can impede the penetration of UV radiation and may be at least partially responsible for the shouldering and/or tailing exhibited in this study. Future studies should shed more light on the degree of clumping of Mycobacteria in real-world waters and consider how to account for the clumping effect in bench-scale experiments. For instance, clumping can be largely reduced by supplementing media with a small amount of Tween-80, which promotes dispersed growth of Mycobacterium spp., and Corynebacterium spp. with waxy surfaces.
We concluded that 280 nm UV-LEDs are effective for the disinfection of M. abscessus in water, with 2 log inactivation expected at a fluence of 10 mJ/cm2, which is technically possible to deliver even with commercially-available compact apparatuses for POU applications. Also, M. abscessus is more resistant than P. aeruginosa and L. pneumophila, suggesting that NTM are among the most UV-resistant OPPPs. Our results can be used to inform UV-LED disinfection unit design against these important pathogens in drinking water systems. While numerous strategies have been proposed for controlling NTM at the POU (briefly summarized in our paper), we consider UV- and filtration-based methods to be the most facile, effective, and verifiable. Considering that NTM and possibly other OPPPs likely exhibit a considerable degree of clumping in the environment, a multi-barrier approach combining UV and filtration may offer a more resilient, comprehensive solution versus using either method in isolation.