Harmful algal bloom and associated health risks among users of Lake Victoria freshwater: Ukerewe Island, Tanzania.

There is a global concern regarding the occurrences of harmful algal blooms (HABs) and their effects on human health. Lake Victoria (LV) has been reported to face eutrophication challenges, resulting in an increase of bloom-forming cyanobacteria. This study is aimed at understanding the association of HABs and health risks at Ukerewe Island. A cross-sectional study conducted on 432 study subjects and water samples for cyanobacteria species identification were collected at LV shores. The results reveal that concentrations of cyanobacteria cells are beyond (WHO) acceptable limits; species of Microcystis aeruginosa range from 90,361.63 to 3,032.031.65 cells/mL and Anabaena spp. range from 13,310.00 to 4,814,702 cells/mL. Water usage indicates that 31% use lake water, 53% well water and 16% treated supplied pipe water. Vomiting and throat irritation was highly reported by lake water users as compared to wells and pipe water (P < 0.001). Gastrointestinal illness (GI) was significantly elevated among lake water users as compared to pipe and well water users (P < 0.001). Visible blooms in lake water were associated with GI, skin irritation and vomiting as compared to water without visible blooms (P < 0.001). The concentration of cyanobacteria blooms poses greater risks when water is used without treatment.


INTRODUCTION
Cyanobacteria blooms have some toxin producing species which pose a potential health hazard to humans and other living organisms, and the risk is high when eutrophication occurs in fresh waters. Lake Victoria (LV), being a fresh water lake, has faced eutrophication challenges, resulting in an increase of bloom forming cyanobacteria throughout the year with the ability to produce hazardous cyanotoxins (Mbonde et al. ). Harmful algal blooms (HABs) can be visible as excessive scum of algal biomass on top of fresh water, which may be attributed to an increase in the levels of nitrogen and phosphorus in fresh water bodies, and this has been observed in Lake Victoria (Mchau et al. ). This increase in algal biomass causes low oxygen levels in the water which leads to the mortality of aquatic organisms, as well as reducing the quality of water for human consumption. Human diseases due to cyanobacteria vary based on related exposure such as swimming, skin contact, bathing, and drinking. Several acute symptoms have been reported in humans such as stomach upset, vomiting, skin irritation, nausea, diarrhea, fever, throat irritation, headache, mouth blisters, muscle and joint aches, eye irritation, and allergic reactions (Kibria ). Chronic health effects include the possible human carcinogen for liver and colorectal malignancies. The International Agency for Research in Cancer (IARC) has categorized MC-LR as a causative factor for chronic illness (IARC ). Water ingestion has been identified as the major route of exposure to harmful cyanobacteria toxins in human and animals, which causes human health effects and animal deaths (Codd et al. ; Hilborn et al. ). However, few epidemiological studies have been conducted to document acute and chronic human health effects caused by cyanobacteria and cyanotoxin with regards to the route and modes of exposure (Chorus & Bartram ).
Globally, few countries have set up surveillance systems that monitor occurrences of HAB related illnesses as well as establishing country specific acceptable limits for the concentration of cyanobacteria cell density in waters used for different human activities. Instead they use the WHO set general guidelines which are not country specific. The World Health Organization (WHO) has proposed guidelines for different cell densities of cyanobacteria bloom levels and related illnesses, such as: for recreational waters cell densities of 20,000 cells/mL are associated with risk of short-term adverse health outcomes; and at a higher cell density of 100,000 cells/mL an additional risk for long-term illness exists. Cyanobacteria scum formation in bathing areas is associated with the risk of potential severe health outcomes such as throat irritation, stomach upset, skin irritation, vomiting, nausea, diarrhea, fever, and headache (WHO ). Due to the non-specific nature of the symptoms of HAB related illnesses and lack of specific guidelines for surveillance and diagnosis, it has proven to be difficult for health-care providers to identify and document HAB associated illnesses. This poses a serious public health omission in the identification of HAB related effects in human. Therefore, health-care providers should be aware of the HAB related symptoms as illness may present differently from other recreational water-associated illnesses and onset may occur soon after exposure such as headache, throat irritation, stomach upset, diarrhea, vomiting, skin and eye irritation to name but a few.
This study aims to identify the different species of cyanobacteria and their potential health effects that may be attributed to the existing dominant HAB species in Lake Victoria in the Ukerewe districts. Most of the HAB associated water illnesses are underestimated and go unrecognized due to lack of water quality assessment and disease surveillance systems that can record the water quality parameters as recommended by WHO and monitor illness related to cyanobacteria blooms. This study will document health related effects of HABs on the population of Ukerewe Island which mostly depend on the lake water for their daily human consumption. The documented information can be used to develop evidence-based HAB related illness prevention strategies.

METHODS
A cross-sectional study was conducted at Ukerewe district (Island), Mwanza region in Tanzania. A total of 432 subjects were recruited in December 2018. The study subjects were individuals who have been residing in the district for more than five years and were above the age of 18 years. The study protocol was adopted and modified from a water quality and associated illness study carried out in the USA by Phytoplankton abundance was calculated using the formula (Greenberg ):

Health effect assessment
Health effects were self-reported by the study subjects, based on the following definition: Having diarrhea was considered when an individual reported having three or more loose stools in a 24-hour period; vomiting; ejection of stomach contents through the mouth, stomach upset; disorder of

Statistical methods
Data were collected through an open data kit (ODK) which was then exported to Microsoft Excel (MS). Analysis was carried out using Epi Info Version 7.2.1.0. The outcome of interest was the observed health effects such as skin irritation, vomiting, diarrhea, stomach upset, and throat and eye irritation. Each outcome of interest was independently tested in Univariate analysis using a 2 × 2 and Chi-square-Mantel-Haenszel (2-tailed P-value) was used to determine the level of significance for the outcome of interest response with less than five in the 2 × 2 table. The Fisher exact test was used to determine the level of significance. All statistically significant outcome variables in the univariate analysis were further put together in the multivariate analysis.
Multivariate analysis was carried out by a backward method where the outcome with the weakest association was removed in the model sequentially until the best-fit model was realized. In the univariate analysis crude odds ratio and their 95% confidence intervals were reported, while in the multivariate analysis adjusted odds ratio with their 95% confidence intervals were reported. The results were considered significant when the P-value was less than 0.05.

RESULTS
During this study, while collecting the samples the water colour was observed to be greenish with bloom on the surface, which was a clear indication that cyanobacteria member dominated the areas, four main groups (classes) of phytoplankton were found, namely Chlorophyceae, Cyanophyceae, Bacillariophyceae and Dinophyceae (Table 1). The results indicate that cyanobacteria (Cyanophyceae) were blooming in all eight sites and they formed more than 97% of all phytoplankton in the lake. The most common members were Microcystis, Anabaena, Merismopedia, the concentration of Microcystis aeruginosa ranged from 90,361.63 to 3,032,031.65 cells/mL and among the eight sample collection sites five (62.5%) had above 100,000 cells/mL and three (37.5%) had less than 50,000 cells/mL. Anabaena spp. cells concentration ranged from 13,310 to 4,814,702 cells/mL, of which seven (87.5%) sites had above 100,000 cells/mL and one (12.5%) had less than 100,000 cells/mL. Merismopedia spp. cells concentration ranged from 2,843.69 to 201,113.04 cells/mL where only two (25%) sites had cells above 100,000 cells/mL and six (75%) sites had less than 100,000 cells/mL (Table 1). Well and treated pipe water were not tested for availability of cyanobacteria.

Reported health effects from the study area
The study was conducted in December 2018, and a total of 432 subjects participated. The mean age was 42 years old with a range of 18-86 years old, 234 (54%) were male and 198 (46%) female. One hundred and thirty-four subjects (31%) reported using lake water as their main source of water, while 229 (53%) used well water and 69 (16%) used treated supplied pipe water as their main source of water for drinking and other domestic uses.
Reported health effects among the subjects were as follows: 78 (18.06%) did not report any health effects while 71 (16.44%) reported one health effect, 81 (18.75%) reported two health effects, 124 (28.70%) reported three health effects, 62 (14.35%) reported four health effects, 12 (2.78%) reported five health effects and four (0.98%) reported all six health effects. The study observed that there is a mixture of water uses and different exposure whereby a person can use pipe water for drinking purposes but use lake water for bathing and other water activities and vice versa. More than 50% of the subjects reported skin irritation, stomach upset and eye irritation followed by diarrhea (32%) while vomiting (9%) and throat irritation (10%) were least reported.
In univariate analysis, individuals who used the lake as their source of drinking water were six times more likely (OR 6.1, 95% CI 2.8-13.58) to have vomiting compared to those individuals who used wells as their source of drinking water and this was statistically significant (P < 0.001). The odds of getting throat irritation when using the lake as a source of drinking water were six times higher than when using wells as a source of drinking water OR ¼ 6.57 (95% CI ¼ 2.74-15.79, P < 0.001).
Those who had used the lake as their source of drinking water were observed to be associated with more stomach upsets (OR ¼ 8.4, 95% CI ¼ 4.49-14.78, P < 0.001) compared to those using wells as their source of drinking water. The risk of vomiting when an individual consumed lake water was observed to be 17-fold higher compared to when an individual consumed treated supplied piped water (95% CI ¼ 2.28-127.2) and this was statistically significant (P < 0.001). Bathing using treated supplied pipe water was protective against eye irritation as compared with the contaminated lake water source with cyanobacteria (OR ¼ 0.46, 95% CI ¼ 0.24-0.85) and this was statistically significant (P ¼ 0.01).
Getting a stomach upset when an individual used lake water for drinking was 45.2-fold higher than when they used treated supplied pipe water for drinking and this was statistically significant (95% CI ¼ 19.78-111.81, P < 0.001).
The odds of vomiting when consuming water with visible bloom was almost four times higher compared to drinking water without visible bloom, which was statistically significant (95% CI ¼ 1.49-10.44, P ¼ 0.003). The odds of getting diarrhea when drinking water with visible bloom is two-fold higher as compared with consuming water source with no visible bloom (95% CI ¼ 1.39-3.6 and P < 0.001). The odds of getting skin irritation among those individuals using water with visible bloom was almost two times higher compared to those who used water for bathing without visible bloom (95% CI ¼ 1.0-2.3, P ¼ 0.04). Study subjects who reported drinking water with visible bloom were four times more associated with throat irritation than those who consumed water without visible bloom (OR ¼ 4, 95% CI ¼ 1.37-11.8, P ¼ 0.006). The study showed that the likelihood of getting stomach upset when consuming water with visible bloom was three-fold higher than those who consumed water with no visible bloom and this was statistically significant (OR ¼ 3.4, CI ¼ 2.28-5.28, P < 0.001).
The study subjects' occupation was compared between fishermen and those employed with non-fishing activities.
Individuals who were in the fishing activities were more likely to have reported symptoms of vomiting as compared with their counterparts who were involved in non-fishing activities (OR ¼ 2, 95% CI ¼ 1-4.5, P ¼ 0.02). This study indicated that fishermen were twice as likely to report symptoms of diarrhea as compared to those in other non-fishing occupations (95% CI ¼ 1.2-3.3, P ¼ 0.003). The odds of getting skin irritation among fishermen were almost two-fold higher (95% CI ¼ 1-2.9, P ¼ 0.02) as compared with individuals in other non-fishing occupations.
Study subjects with fishing occupations were more than twice as likely to have throat irritation as compared to those with non-fishing occupations (OR ¼ 2.4, 95% 1.1-5.1, Diarrhea and throat irritation were reported to be mainly associated with the consumption of more than 1 L of lake water (OR ¼ 2, 95% CI ¼ 1-2.4, P ¼ 0.02 and OR ¼ 4, 95%CI ¼ 2.11-8.9, P < 0.001 respectively).

Health effects reported from different water sources utilization
The assessment of water consumption among study subjects shows that 31% use lake water as the main source of water, 53% use well water and 16% used treated supplied pipe water as the main water source for drinking and domestic purposes. The subjects reported vomiting as one of the health effects occurring after drinking water from the lake source, which is infested with cyanobacteria. The risk of vomiting is 6.1-fold higher compared to those accessing drinking water from wells (P < 0.001). The risk of vomiting when consuming bloom contaminated water is 17-fold compared to treated supplied pipe water (P < 0.001; see Table 2).
The same finding that cyanobacteria infested water can cause vomiting was reported in the UK (Turner et al. ) and in Australia (Pilotto et al. ). Vomiting was still statistically significant in multivariate analysis (P < 0.05; Table 2). The odds of getting throat irritation when drinking cyanobacteria contaminated water from the lake source is 6.57 higher than when drinking from the wells (P < 0.001), and this association was also strong in multivariate analysis (P < 0.05; Table 2). Throat irritation is one of the documented health effects after ingestion of algal bloom contaminated water in most recreational activities or water related exposure (Kibria ).
Gastrointestinal illness (GI) was significantly elevated among study subjects using water from the lake as compared to pipe and well water users. Getting stomach upsets when drinking cyanobacteria contaminated water from the lake source is 45.2 and 8.4-fold higher than drinking treated supplied pipe and well water respectively (P < 0.001).
This finding was also reported by Hilborn et al. () and Collier et al. (). The association between stomach upset and contaminated water was observed to be strong in multivariate analysis (P < 0.001). Bathing using treated supplied pipe water is protective against eye irritation as compared to contaminated lake water source with cyano- is almost two and four times higher than those using water without visible bloom respectively (P < 0.05; see Table 3). The

Occupation risk and amount of water consumption
Occupation of the study subjects was compared between fisherman and other non-fishing activities, which includes employed, non-employed and farmers. Fishing is reported to be one of the risk activities of contracting health effects related to cyanobacteria bloom and cyanotoxin due to the frequency of water contact and usage during fishing activities. GI was strongly linked with the fishing occupation whereby vomiting and diarrhea were reported among fishermen as compared to non-fishermen (OR ¼ 2, P < 0.05), vomiting and diarrhea shows a strong association in multivariate also (P < 0.05). Throat irritation was reported more by fishermen as compared to other occupations (P < 0.05; see Table 4). Studies conducted in China report the same higher risk for fishermen (Chen et al. ). The amount of water consumption was reported to be one of the factors that may contribute to reported health risk, whereby drinking more than 1 L is more of a risk than less than 1 L. Diarrhea and throat irritation were reported to be associated with the amount of water consumption where the likelihoods were two and four times more than these reported to consume less than 1 L respectively (P < 0.05); throat irritation had a strong link with the amount of water intake in multivariate analysis (P < 0.05;   and Anabaena spp. cells concentration range from 13,310.00 to 4,814,702 cells/mL. At this level water use from the lake source was found to be associated with acute health outcomes. The result indicates the potential health risk of using lake water without any treatment for human consumption. It is therefore advised to continue to monitor the water quality at Ukerewe area in order to understand its spatial and temporal dynamism of phytoplankton.
The long-term study of Phytoplankton helps to understand the nature of nutrients or pollution entering the water body because phytoplanktons are a good and cheaper indicator of environmental change as compared to chemical indicators. The documented illness associated with cyanobacteria infested water can be used as a baseline to improve case detection in the district and contribute to the development of evidence-based prevention strategies to mitigate adverse health outcomes that may result in long-term exposure to HABs. It remains important to be able to identify and quantify the different types of cyanotoxin produced which exist in lake water. This will help to establish a long-term study that determines the chronic effect of persistence exposure of cyanotoxin.