We were aware of a number of small recreational lakes in Michigan, New York, Ohio, and other states that historically experienced microcystin-producing algal blooms in late summer, and we corresponded with colleagues conducting routine monitoring of these lakes. Based on previous experience with accidental exposures to water contaminated with microcystins, we expected that people involved in activities that involve ingesting water (i.e., swimming) or inhaling aerosols (i.e., jet-skiing, water skiing, or sailing a small boat) should receive enough exposure to allow us to detect microcystins in their blood. The minimum amount of microcystins detectable in blood was about 0. 9 ng microcystins in 10 ml of blood (approximately 0.1 ppb). This was very close to the 1 ng limit of detection for the ELISA assay.
Accordingly, we planned to conduct the field study within a week of receiving monitoring results that the microcystin concentrations in one of the lakes was at least 10 μg/L. Delays in conducting the tests and unpredictable changes in local weather can, however, rapidly change algal bloom characteristics. For this field study, water sample analyses completed after the study found that even with a significant, visible bloom of organisms capable of producing microcystins, the concentrations of microcystins were actually in the range of 2–5 ug/L. Thus people may have been exposed, but the waterborne microcystin concentrations were not high enough to detect microcystin in blood samples from our study subjects. Nevertheless, in this field study we verified that inhalation is a possible route of exposure to microcystins, and these data can be compared with data from studies that used the same protocol in lakes with potentially higher exposures.
We collected 24 water samples (four sample sites, each sampled in the morning and afternoon, on each of the 3 study days, August 4–6, 2006) for analysis of water quality, alga taxa, and microcystin concentrations from the lake with the algal bloom. Table 1
contains the ranges of water quality parameters tested over the 3-day study period as well as weather data collected during that same period.
In general, cyanobacterial populations are found and thrive in warm, calm waters with high concentrations of nutrients and characterized by consistent, relatively high pH and dissolved oxygen (DO) levels. High-pH (8–10) and DO (above saturation) would be a result of relatively high population densities and are not a prerequisite for growth. Still, if potentially toxigenic species were present, warm temperatures (25–30° C), low wind action (0–3 mph), and high nutrient loads (phosphorus and nitrogen), together with clear sunny skies, would stimulate growth and increase the potential risk for toxin production.
Several factors contributed to the weak bloom and low concentrations of microcystins in the water we tested. First, microcystins are endotoxins stored and isolated inside algal cells, thus they do not readily diffuse or leak into the aquatic environment. Under normal environmental conditions in a healthy, reproductively active toxin-producing bloom event, 95% to 98% of the toxins would be intracellular. Accordingly, we would expect to find toxin only in water containing cells. Second, during the sampling period for the lake with the bloom, winds generally increased in speed from morning to afternoon, thus disorganizing bloom accumulations and diluting microcystin concentrations by scattering algal cells. Third, Microcystis species are buoyant, surface-floating algae that tend to accumulate on or near shore as winds blow them landward. Three of the four samples sites where water samples were collected were open water—thus we would not expect high toxin concentrations unless the bloom was extensive. For example, early one morning before the winds had increased in strength, we sampled a small accumulation of cells and found microcystin concentrations of 43 ug/L. Later that morning and in the afternoon we did not observe cell accumulations, and in the water samples, toxin concentrations remained relatively low (3 ug/L – 5 ug/L). In addition, during the study period a storm with significant winds (> 5 mph) further dispersed the bloom.
We assessed algal taxonomy in 12 samples collected during the morning of each of the 3 study days from the exposure lake at the four sampling stations. The range of phytoplankton concentrations was 175,000 cells/mL to 688,000 cells/mL. Over 95% of the cells were cyanobacteria.
contains the densities of potentially toxigenic cyanobacteria (PTOX-C). During the sampling period, densities ranged from 54,000 total cells/mL to 144,000 total cells/mL. The dominant genera of PTOX-C reported in water samples were Anabaena
, and Microcystis
. The two documented microcystin-producing genera present were Anabaena
. No Planktothrix
species were observed. The dominant species present were Anabaena macrospora
, Aphanizomenon aphanizomenoides/gracile
, and Microcystis cf. botrys. Aphanizomenon aphanizomenonoides/gracile
is, however, a suspected microcystin producer. Because the toxin levels were very low, we could not examine in the collected water samples any potential relationship between the number of microcystin-producing PTOX-C and the concentrations of microcystins.
On each of the three study days we detected microcystins at levels near the level of detection (LOD) (0.0037 ng/m3) on each stage of the high volume, airborne particle impactor. The flow rate, sampling time, and total concentration of microcystins from the five stages and the back-up filters, respectively, were as follows: day 1, 1.17 m3/min, 360 min, 0.050 ng/m3; day 2, 1.15 m3/min, 386 min, 23 ng/m3; and day 3, 1.22 m3/min, 484 min, 0.057 ng/m3.
We deployed 10 personal air samplers on study days 1 and 2, and 4 personal air samplers on study day 3. A summary of the mass and air concentration of microcystins as determined on each study day using the personal samplers is presented in Table 3
. We were able to detect microcystins in 9 of 10 samples collected on day 1 and 6 of 9 samples on day 2. On day 3 we did not collect measurable concentrations of microcystins using personal air samplers.
Stewart et al.
] found an increase in respiratory symptoms in people exposed to higher levels of cyanobacteria (surface area > 12.0 mm2
/mL) than in people exposed to lower levels of cyanobacteria (surface area < 2.4 mm2
/mL). It is possible that respiratory symptoms could be caused by infections from viruses rather than cyanobacteria. Thus, for our study, we added the analysis of the presence of adenoviruses, as well as enteroviruses in the lake water.
Each 10-L water sample was tested for adenoviruses using real-time PCR and enteroviruses using real-time reverse-transcription (RT)-PCR. Adenoviruses and enteroviruses were not detected in any water sample. Based on published detection limits for these assays, the testing protocol used in this study resulted in detection limits of approximately 1,250 adenovirus gene equivalent copies (GEC) per 10-L water sample and 200 enterovirus plaque-forming units (PFU) per 10-L sample. Thus, these data indicate that adenoviruses and enteroviruses were not present at substantial concentrations in the lake with the Microcystis
bloom. In some samples, however, PCR inhibition was observed based on the delays in CT values measured for water samples seeded with an inhibitor standard versus CT values for seeded reagent-grade water controls. The range of CT value differences between seeded samples and controls was 0.6 to 9.3, with an average difference of 2.8 ± 2.0 CT values. This inhibition was likely due to the presence of bloom-associated constituents in the water samples—such as algal matter and other organic material (including humic and fulvic acids)—that are generally reflected by the relatively high turbidity data reported in Table 1
The positive real-time PCR tests for Escherichia coli
(data not shown), indicated that the organism was present at low levels (10 colony-forming units [cfu]/mL to 100 cfu/100 mL) in water from the lake with the bloom. E. coli
is an established, fecal-specific water quality indicator that has been shown to correlate with reports of gastrointestinal illness in people exposed to the water during recreational activities [21
]. Thus despite the absence of any fecal-associated virus contamination in the lake with the algal bloom, we did find evidence of fecal contamination (i.e., the presence of E. coli
We recruited 104 study participants from lake visitors planning recreational activities that would generate aerosols, such as boating and using personal watercraft. Ninety-seven participants planned to use the lake with the bloom (the exposed group) and 7 planned to use a nearby lake with no bloom (the unexposed group). We obtained complete questionnaire data from 96 people in the exposed group and 7 in the unexposed group.
For the exposed group, the age range was 12 years to 67 years. Ninety-three (97%) were white, and 4 (4%) were of another race—no Hispanics were in the study. Forty-five (47%) of the exposed participants were female. Of the 96 exposed participants, fifty-seven (59%) reported that during the 7 days before the study they had participated in water-related recreational activities on the study lake. In addition, these exposed participants reported that in the year before the study, 75 (77%) had boated on, 34 (35%) had fished in, and 72 (74%) had swum in the study lake. Two (2%) reported using dietary supplements containing blue-green algae.
Study participants in the exposed group reported the following activities during the study period: swimming (73 [76%]), boating (34 [51%]), tubing (20 [30%]), riding personal water craft (13 [14%]), wake-boarding (8 [12%]), wading (3 [4%]), and fishing (2 [3%]). Of the study participants who swam, the amount of time ranged from 5 min to 180 min. Of those who rode personal watercraft, the amount of time ranged from 15 min to 120 min. Of those who did the other activities, the amount of time ranged from 10 min to 300 min Sixty-one (64%) of the exposed group reported that during the study period they had put their head underwater, and 39 (41%) reported they had swallowed water.
For the unexposed group, the age range was from 15 years to 58 years. All of the unexposed participants were white and non-Hispanic, and 4 (57%) were female. Of the 7 unexposed participants, 6 (86%) reported they participated in water-related recreational activities on the study lake (exposed site) during the 7 days before the study and in the year before the study. No one reported using dietary supplements containing blue-green algae.
The seven participants in the unexposed group reported only swimming during the study period. Six (86%) were in the water for 60 min, and one (14%) was in the water for 90 min. Six (86%) reported that during the study period they had put their head underwater, and 1 (14%) reported swallowing water.
contains a summary of symptoms reported by both exposed and unexposed participants during the interviews. As a group, study participants reported more symptoms in the 7 days before the study than during study or during the 7 to 10 days after the study. In addition, as a group, study participants tended to report more symptoms (e.g., cough, sore throat for the exposed group, skin complaints for the unexposed group) immediately before doing study activities than immediately after such activities. A single participant in the control group reported five skin-related symptoms before going into the water but not after coming out of the water. No differences appeared in the frequency of reported symptoms between the exposed and control groups during the 7 days before the study, immediately before or immediately after doing study activities, or during the follow-up period.
We hypothesized that exposure to aerosolized microcystins would result in acute dermal and respiratory symptoms; thus, we expected people to report more symptoms immediately after study activities than during the week before the study. However, more people reported symptoms before doing study activities than after. In addition, participants reported more symptoms for the period of 7 days before the study than either immediately before or after doing study activities. Also, more than half of the participants in the exposed group and all but one of the participants in the control group had participated in water-related recreational activities on the study lake in the week before the study. Finally, the number of people reporting symptoms during the seven to 10 days after the exposure increased to levels that were more consistent with the number of people reporting symptoms for the period of 7 days before the study than with the number of people reporting symptoms either immediately before or immediately after doing study activities. While these differences in symptom reporting are not significantly different, the trend suggests that exposure to aerosolized microcystins may result in symptoms with onset a day or few days after the exposure, i.e., during the week after weekend activities on the blooming lake. This possibility could be addressed in a future study by identifying a study population that uses a lake that historically develops a summer bloom of microcystin-producing cyanobacteria and conducting symptom surveys before their first visit to the lake and then periodically throughout the summer.
We collected usable blood samples from 96 exposed and 6 unexposed participants. Of the 102 blood samples, 101 samples tested below the limit of detection according to the limit of detection described in the Envirologix Kit (0.147 μg/L). Only one blood sample showed detectable levels of microcystins—about 1.0 μg/L. LC/MS analysis of this sample showed the absence of microcystin-LR, -RR, and -YR. Given the LC/MS results, we concluded that this sample represented either a false positive or exposure from microcystin(s) not identified as microcystin-LR, -RR, or -YR. We did not identify anything unusual about this particular study participant (e.g., they did not use blue-green algae food supplements).