Swimming in rivers and lakes was associated with a doubling of the risk of skin problems, and intestinal infections were better predicted by E. coli than enterococci.

Swimming in fresh open water can be associated with a significant, though relatively small, risk of health problems. A new study published in the International Journal of Hygiene and Environmental Health found that participants who waded into rivers and lakes were approximately twice as likely to report rashes, itching, or other skin problems as those who remained on the shore. The likelihood of gastrointestinal illness also increased with increasing concentrations of Escherichia coli and somatic coliphages in the water.

The study is interesting primarily for its design. The researchers analyzed data from 2,368 people who participated in four controlled experiments at freshwater beaches in Hungary. Of these, 1,133 were randomly assigned to swim for ten minutes, with at least three full facial submersions, while 1,235 remained on the shore. After a week, the scientists assessed whether gastrointestinal, skin, respiratory, ear, or eye symptoms developed.

But there's an unusual twist: the field experiments themselves were conducted back in 2006-2007 as part of the European Epibathe project. In 2026, researchers at the University of East Anglia conducted a new analysis of previously unpublished individual data, comparing each swimmer's symptoms with microbiological parameters of the water in their immediate area and around the time of their swim. This makes the study much more informative than typical surveys of open-water swimmers, although the age of the original data must be taken into account.

The most unexpected result was a comparison of pollution indicators. E. coli and somatic coliphages predicted gastrointestinal illnesses, while enterococci did not. This is potentially important for sanitary monitoring: current international approaches largely focus on enterococci, while the authors believe that optimal microbiological markers for freshwater may differ from those that work best on sea beaches.

Key points about the study

Parameter	Result
Study	Epibathe, a new analysis of previously unpublished data
Country	Hungary
Places	4 freshwater recreation areas
Participants	2368
Bathers	1133
Remaining on the shore	1235
Bathing	10 minutes, minimum 3 facial immersions
Observation	7 days
Gastrointestinal diseases	48 cases
Skin problems	47 cases
Skin risk for swimmers	adjusted RR 2.17
E. coli and gastrointestinal risk	RR 1.73 after multiple imputation
Somatic coliphages	RR 1.45-1.46
Enterococci	No significant association with gastrointestinal diseases was found.
DOI	10.1016/j.ijheh.2026.114888

Why test for bacteria in water at all if the indicator bacteria themselves don't necessarily cause disease?

The water of natural rivers and lakes is not sterile. It can contain microorganisms from sewage, agricultural runoff, animal and bird excrement, urban surface water, recreational areas, and other sources. Fecal contamination can also carry pathogens that cause intestinal infections, including bacteria, viruses, and protozoa. Ingestion of small amounts of contaminated water is one of the most obvious routes of infection when swimming.

However, it's impossible to test every river daily for hundreds of potential pathogens. Therefore, sanitary services use fecal indicator microorganisms. Their presence does not necessarily mean that a particular organism will cause disease. They act as a kind of "alarm signal": if they are present in large numbers, this may indicate the presence of fecal material and, therefore, a higher probability of the presence of true intestinal pathogens.

The most common indicators include E. coli and enterococci. Researchers also studied somatic coliphages—viruses that infect intestinal bacteria. These are not human viruses and are not considered a cause of swimmer's gastroenteritis, but they can serve as an indicator of fecal contamination and the behavior of viral particles in the aquatic environment.

The problem is that the same indicator can behave differently in seawater and freshwater. Sunlight, salinity, temperature, local microbiota, and the rate of microbial decay all influence how long bacteria persist after contamination. Therefore, a marker that effectively reflects sanitary risk on a sea beach may not be as reliable in a river or lake. This question is one of the central themes of the new analysis.

What was measured in the water?

Marker	What does it show?
E. coli	Fecal contamination
Intestinal enterococci	Fecal contamination
Somatic coliphages	Viral indicators of fecal contamination
Are they a list of all dangerous pathogens?	No
Main function	Indirect assessment of the probability of microbiological contamination

Scientists randomly assigned people to either swim or stay on the beach.

The study recruited primarily healthy volunteers from communities located within approximately 15-50 kilometers of the bathing areas. Before the experiment, participants completed questionnaires on their health and behaviors that could influence the risk of gastrointestinal illness. Then, on the day of the study, they were assigned to one of the two groups using a computer-generated random sequence.

Participants in the main group were required to enter the water once for ten minutes and fully submerge their faces at least three times. This was an important element of the protocol: the researchers wanted to achieve a relatively standardized experience, rather than a situation where one person merely wet their feet and another swam for two hours and repeatedly swallowed water.

The control group remained on the shore for three hours and did not enter the water. Alternative entertainment was provided. Both groups received the same food, which helped reduce the likelihood that subsequent gastrointestinal symptoms were due to differences in diet on the day of the experiment. Seven days later, the volunteers were interviewed in person or by phone.

Intestinal illnesses were defined as diarrhea or vomiting, or a combination of nausea and fever, or indigestion and fever. Skin problems included rashes, ulcerative skin lesions, or itching. Acute febrile respiratory illness, ear infections, and eye infections were also recorded separately.

A particularly strong aspect of the study was the monitoring of the water itself. Each bathing area was divided into six sections, and samples were taken every 20 minutes for approximately three hours. This allowed each bather to be assigned a concentration of microbiological indicators that most closely matched the location and time of their actual contact with the water.

How the experiment was conducted

Stage	What did they do?
Kit	Local volunteers
Randomization	Swimming or being on the shore
Bathing	10 minutes
Face immersion	At least 3 times
Water zones	6 in each place
Sampling	Every 20 minutes
Samples per experimental session	54
Water analysis	E. coli, enterococci, somatic coliphages
Health check	In 7 days

Stomach upset occurred in only 2% of participants – but the risk increased with the amount of E. coli

A total of 48 of 2,368 participants (2.03%) met the criteria for gastrointestinal illness during the week. Among those remaining on the shore, illness was registered in 24 of 1,235 people (1.94%), and among those swimming, in 24 of 1,133 people (2.12%). A simple comparison of whether people swam or did not swim revealed no statistically significant increase in the risk of gastrointestinal illness.

In other words, this study does not support the claim that any ten-minute swim in freshwater increased the risk of gastroenteritis. The adjusted relative risk for swimmers compared to the shore group was 1.13, and after accounting for missing data, it was 1.12; confidence intervals included one, meaning the result could have been explained by chance.

A completely different picture emerged when the researchers looked not simply at the fact of swimming, but at the degree of microbiological contamination of the water in which a particular person had been. After adjusting for age, recent gastrointestinal illnesses, and other potentially influencing factors, an increased E. coli concentration was associated with a relative risk of 1.77; after multiple imputation for missing data, the relative risk was 1.73, with a 95% confidence interval of 1.13-2.65.

It's important to understand this number correctly. Bacterial concentrations were analyzed after natural logarithm transformation. Therefore, the RR of 1.73 corresponds to a one-unit increase in the logarithmic concentration—approximately a 2.7-fold increase in the baseline E. coli concentration. This change was associated with an approximately 73% increase in the relative risk of gastrointestinal illness, not with the fact that 73% of bathers became ill. The absolute number of cases in the entire study remained small—48.

Gastrointestinal diseases

Indicator	Meaning
All participants	48/2368 - 2.03%
We didn't swim.	24/1235 - 1.94%
We went swimming	24/1133 - 2.12%
Bathing on its own, adjusted RR	1.13
Statistically significant?	No
E. coli, adjusted RR	1.77
E. coli after imputation	1.73; 95% CI 1.13-2.65
Somatic coliphages	1.45; 95% CI 1.05-2.04
Enterococci	No connection found

Somatic coliphages also predicted stomach diseases

The second indicator found to be associated with gastrointestinal risk was somatic coliphages. After statistical adjustment, the relative risk was 1.48, and after imputing a small amount of missing data, it was 1.45, with a 95% confidence interval of 1.05–2.04. That is, a unit increase in the logarithmic phage concentration was associated with an approximately 45% increase in the relative risk.

This is interesting from a public health perspective, as coliphages can behave more closely in the environment than traditional bacterial indicators. They do not prove the presence of a specific human virus, but they can provide additional information about fecal contamination and the potential viral hazard of water. The authors therefore consider them a promising additional marker of freshwater quality.

The result wasn't solely dependent on the most contaminated of the four sites. One of the sites yielded the majority of the highest concentrations of E. coli and coliphages, so the researchers repeated the analysis without it. The association remained: for E. coli, the relative risk was 1.77, and for somatic coliphages, 1.58. This increases the robustness of the observation, although with the site excluded, the statistical model was based on only three geographic clusters.

However, microbial concentrations were not randomly assigned to participants. Swimming was randomized, and the level of E. coli or coliphages was determined by the natural state of the water in the specific area and time. Therefore, the "higher indicator = higher intestinal illness" relationship is a strong epidemiological association, but does not in itself prove that the measured organism was the cause of the symptoms.

Three microbiological indicators

Marker	Association with gastrointestinal diseases
E. coli	Yes
Adjusted RR after imputation	1.73
Somatic coliphages	Yes
Adjusted RR after imputation	1.45
Intestinal enterococci	There is no statistically significant relationship
The authors' main conclusion	E. coli and coliphages may better reflect GI risk in freshwater

The most pronounced effect was found for the skin: problems occurred more than twice as often

Skin symptoms were observed almost as frequently as gastrointestinal symptoms: 47 cases among 2,368 people. However, here the difference between groups was much more pronounced. Among non-swimmers, skin problems occurred in 16 of 1,235 people (1.30%), while among swimmers, 31 of 1,133 people (2.74%) experienced skin problems.

After accounting for potential confounding factors, swimming was associated with a relative risk of 2.17 with a confidence interval of 1.55–3.03. In an analysis with multiple imputation, the estimate was even 2.30. Because participants were randomly assigned to swimming or staying on the beach, this observation has a stronger causal basis than a simple observational comparison of people who self-selected to swim or not.

Neither the concentration of E. coli, nor enterococci, nor somatic coliphages explained the increased skin risk. This is an important distinction: fecal contamination indicators worked well for gastrointestinal symptoms, but the increase in rashes and itching could likely be related to other characteristics of the natural water.

The authors propose several possible mechanisms. One is microorganisms naturally present in freshwater and not associated with human feces. For example, avian schistosome larvae can penetrate human skin and cause cercarial dermatitis, or "swimmer's itch." Another possibility is chemical irritants from agricultural or industrial runoff. The study did not identify a specific cause for the skin symptoms, so these remain hypotheses.

Skin problems

Group	Cases
All participants	47/2368 - 1.98%
We didn't swim.	16/1235 - 1.30%
We went swimming	31/1133 - 2.74%
Adjusted RR	2.17
95% confidence interval	1.55-3.03
Imputation analysis	RR 2.30
E. coli association	No
Association with enterococci	No
Relationship with coliphages	No

Why did enterococci unexpectedly "lose" to E. coli?

One of the most important scientific findings of the study is the absence of a significant relationship between intestinal enterococci concentrations and gastrointestinal diseases. In the adjusted model, the relative risk was 1.20, and after multiple imputation, it was 1.18, with confidence intervals crossing unity. Therefore, the authors did not find a statistically convincing association.

This contrasts with seawater, where enterococci are one of the most reliable indicators of sanitary risk. One possible explanation relates to differences in the survival rates of the organisms. Environmental studies show that E. coli and somatic coliphages can survive longer in freshwater, while enterococci are relatively resistant in marine environments.

This is why the authors question the idea of a single microbiological indicator for both marine and freshwater. The 2021 World Health Organization guidelines excluded E. coli from the recommended indicators for recreational freshwater in favor of enterococci. The new study, on the contrary, found a reproducible association between E. coli and gastrointestinal risk.

But one publication doesn't mean international standards should be immediately changed. The authors themselves point to the need for research in other climates, water systems, and countries. The results demonstrate a reason to reconsider the scientific basis of the standards, rather than establishing a new, safe bacterial concentration for every lake or river.

How the indicators may differ

Wednesday	The most informative indicator based on available data
Sea water	Often intestinal enterococci
Fresh water in new study	E. coli
Additional promising indicator	Somatic coliphages
Enterococci in this study	GI disease was not predicted
Is it necessary to change international norms?	It's too early

Even water that officially meets standards is not sterile.

At the time of the study, all four Hungarian resorts complied with the mandatory standards of the then-current European Bathing Water Directive. This does not mean that their indicators were always identical: there were significant differences in microbial concentrations between the sites, and one site demonstrated particularly high and variable values of E. coli and coliphages.

Two sites, Dömsöd and Fadd, were located on former Danube riverbeds and, in terms of hydrological characteristics, were more reminiscent of lakes. The other two, Tiszakécske and Csongrád, were riverine areas on the Tisza. Potential pollution sources included settlements, recreational activities, agriculture, livestock, and, in some areas, treated wastewater flowing upstream.

Water temperatures ranged from 18 to 26°C on different days of the experiment, and weather conditions ranged from cool and cloudy to hot and sunny. This is significant because temperature and sunlight influence the survival of bacteria and viral indicators in water. However, the study was conducted at each site for only one day, so the effect of temperature could not be separated from all other differences between sites.

Therefore, the study's conclusion is not that "officially approved beaches are dangerous." A sanitary standard always refers to an acceptable level of risk, not the absence of microorganisms. Furthermore, the field data was collected almost twenty years ago and relates to the previous European regulatory regime. The current state of a specific body of water must be assessed based on up-to-date monitoring data.

Where the study took place

Plot	Water type/features
Dömsöd	The old Danube riverbed, a lake-like environment
Fadd	Isolated old bed of the Danube
Tiszakécske	The Tisza River
Csongrád	The Tisza River
Water temperature	18-26 °C
All sites complied with the then mandatory EU standard	Yes
Are the bacteria values the same?	No, they differed significantly.

There were too few cases for the eyes, ears and respiratory tract

Those who swam were numerically more likely to experience not only skin symptoms but also respiratory, ear, and eye symptoms. For example, respiratory illness was recorded in 7 of 1,133 swamers versus 2 of 1,235 non-swamers, eye infections in 9 versus 4, and ear infections in 8 versus 5. However, the absolute number of events was extremely small.

There were only 9 respiratory, 13 ear, and 13 eye cases. This was insufficient for a stable multivariate statistical model. Therefore, the authors deliberately do not draw conclusions about whether swimming increases the risk of such diseases and which microbiological indicators might predict them.

The researchers calculated that to achieve comparable statistical power for such rare outcomes, a sample size approximately five times larger—approximately 13,000 people—as well as a larger number of water bodies would be needed. This is a good example of why the absence of a statistically significant result does not always mean the absence of an effect: sometimes, the data are simply insufficient.

Age may also influence susceptibility to gastrointestinal illnesses. Younger participants, particularly those aged 11-20 and 21-30, demonstrated a slightly higher predicted risk with increasing E. coli levels. However, the overall test of the interaction between age and bacterial concentration yielded a p=0.06, meaning it did not formally reach the accepted level of statistical significance. Therefore, it is not yet possible to conclude that younger people are definitively more vulnerable.

Other outcomes

Disease	Total	We didn't swim.	We went swimming
Respiratory	9 (0.38%)	2	7
Ear	13 (0.55%)	5	8
Eye	13 (0.55%)	4	9
Enough data for a reliable RR?	No	-	-
Indicative sampling for more reliable analysis	≈13,000 people

Why this study is stronger than a typical survey of swimming enthusiasts

Most open-water studies compare people who voluntarily swim with those who don't. This introduces numerous potential confounds. For example, swimmers may be younger, more physically active, spend more time at the beach, eat certain foods more often, or have different baseline disease risks.

Here, experimental group assignment was random. Baseline characteristics between bathers and non-bathers were well balanced, and after the intervention, almost all participants successfully completed the week-long survey. This significantly strengthens the conclusion about the link between water exposure and skin problems.

Another advantage was the individual measurement of water quality. Instead of using, for example, a single sample taken in the morning for the entire beach, the researchers took dozens of samples, taking into account specific areas and times. This reduces the error in estimating the actual microbiological exposure of each swimmer.

In their statistical analysis, the scientists also took into account whether individuals lived in the same body of water, their age, recent illnesses, and some behavioral factors. The results for E. coli and coliphages remained virtually unchanged after multiple imputation of missing values: information was missing for only approximately 1.65% of observations in the gastrointestinal disease model.

Strengths of the job

Peculiarity	Why is it important?
Randomization	Reduces systematic differences between bathers and controls
2368 participants	A large sample for such a field experiment
Standardized bathing	Similar duration of exposure
3 face dives	Provides real contact of water with the face
Frequent sampling	Better characterizes individual exposure
Four different bodies of water	Expands the range of pollution
Adjustment for risk factors	Reduces confounding of results

But the study has an important limitation: the experiment is almost twenty years old.

A key factor, easily missed by the 2026 publication date, is that the water and participants were studied in 2006 and 2007. This is a secondary analysis of previously unpublished data from the Epibathe project. Wastewater treatment systems, agriculture, climate, recreational intensity, and sanitary standards may have changed over the past two decades. Therefore, the quantitative data cannot be directly applied to a specific river or lake today.

The second limitation is only four water bodies and essentially one experimental day at each. Despite the large number of participants, from an environmental perspective, the sample consists of only four geographic clusters. The authors themselves note that research in other climatic and environmental conditions is necessary to confirm the generalizability of the results.

The third issue is partial non-compliance with the rules outside the experimental day. Despite being asked not to swim in fresh water for a week before and after the study, 25% of participants in the control group and 27% of those in the study group reported additional swimming. This could have blurred the difference in gastrointestinal illnesses between those who swam and those who did not. The authors believe that such non-compliance likely biased the effect toward no difference.

The fourth limitation is that E. coli, enterococci, and coliphages are indicators, not direct measures of all pathogens. High concentrations could have come from both human and animal sources. Therefore, the study demonstrates an epidemiological link between the indicator and the disease, but does not establish which specific bacteria, virus, or protozoan caused each individual episode.

Finally, symptoms were determined by participant interviews, not by laboratory confirmation of infection. Diarrhea or itchy skin after swimming are not necessarily infectious. This is especially true for skin reactions, which could be caused by parasites, irritants, algae, or other factors not measured in the study.

Restrictions

Limitation	Possible influence
Data from 2006-2007	Modern reservoirs may differ
Only 4 seats	Limited geographic portability
One experimental day per object	Poorly reflects seasonal changes
Bathing outside the protocol	Could reduce differences between groups
Self-reported symptoms	There is no laboratory confirmation of the cause.
FIO - indirect indicators	The specific pathogen is not identified
Rare ear/eye/respiratory diseases	Insufficient statistical power

What the study really says about swimming safety

The study does not demonstrate that swimming in rivers and lakes should be considered a generally dangerous activity. The absolute incidence of illnesses was relatively low: gastrointestinal symptoms of the appropriate definition occurred in approximately 2% of participants, and skin symptoms also occurred in approximately 2%. However, for skin symptoms, the difference between the randomized groups was significant: approximately 1.3% on the shore versus 2.74% after swimming.

For intestinal diseases, it was not the bathing itself that was more important, but the quality of the water itself. At low concentrations of fecal indicators, the risk was lower, but as E. coli and coliphage levels increased, it increased. This supports the idea of regular and timely microbiological monitoring of outdoor swimming areas.

The study also shows why water shouldn't be assessed solely by its appearance. Clear, odorless water may contain microscopic indicators of fecal contamination, while cloudy water doesn't necessarily indicate a high infection risk. Laboratory analysis is necessary for a sanitary assessment, and the authors believe that E. coli may be a particularly useful indicator for freshwater.

The main practical conclusion for public health systems, therefore, is not a ban on open swimming, but a more precise selection of risk indicators. If this result is confirmed by modern studies in other countries, standards for marine and freshwater recreational areas may indeed need to be further differentiated, and somatic coliphages could complement traditional bacterial indicators.

How to interpret the result correctly

Statement	Right?
Any swimming in the lake causes gastroenteritis	No
GI diseases were significantly more common among all swimmers	No
Higher E. coli was associated with greater GI risk	Yes
Coliphages also predicted GI risk	Yes
Enterococci proved to be a good predictor	No
Skin problems were approximately twice as common after bathing.	Yes
It is known what exactly caused the skin symptoms.	No
The results may impact monitoring standards.	Potentially yes

Why the results may be important for sanitary standards

The authors emphasize that international practice remains uncertain regarding which fecal indicator is most appropriate for freshwater. European standards traditionally use both E. coli and enterococci, while the 2021 World Health Organization recommendations focus on intestinal enterococci.

The new study challenges the idea of complete standardization. If E. coli and coliphages truly disappear more slowly in freshwater and better preserve information about recent fecal contamination, using only enterococci could potentially underrepresent intestinal risk specifically in rivers and lakes.

This finding is supported by the fact that a previously published randomized trial of freshwater bathing in Germany also found a link between E. coli and gastrointestinal illness. According to the authors, the current analysis is only the second published controlled experiment of this type in freshwater, demonstrating both the strength of the result and how sparse the evidence base remains.

Therefore, the most likely impact of the study is not an immediate change in regulations, but rather a new argument in favor of modern, large-scale studies. These should include more water bodies, different climate zones, a broader range of indicators, and, ideally, direct detection of human pathogens. Only then can we reliably determine whether sanitary thresholds for rivers and lakes should be revised.

What might change in the future

Question	Current output
Do we need the same indicators for the sea and rivers?	Maybe not.
Is E. coli good for fresh water?	Research data says yes.
Are somatic coliphages promising?	Yes
Are enterococci informative enough on their own?	This work raises doubts
Can we set a new safe limit already?	No
What's next?	Modern research in a larger number of water bodies

The main conclusion

This study provides two separate but related pieces of information. First, in a randomized comparison, ten-minute bathing more than doubled the relative risk of skin problems—2.74% of bathers versus 1.30% of those remaining on the shore. The reason for this was unrelated to the measured indicators of faecal contamination and remains unknown.

Second, swimming itself did not lead to a statistically significant increase in gastrointestinal illnesses, but within the group of swimmers, the risk increased with water pollution. E. coli showed a RR of 1.73, and somatic coliphages approximately 1.45 per unit natural logarithm of concentration, while no significant relationship was found for intestinal enterococci.

This makes the study primarily about water quality, not the dangers of swimming per se. If a river or lake has low levels of fecal contamination, the results cannot be interpreted as evidence of a high intestinal risk simply due to contact with fresh water. On the contrary, an increase in E. coli provides information about when such a risk becomes higher.

And the most important scientific conclusion may be regulatory: freshwater and seawater likely require different microbiological control systems. In this experiment, E. coli and somatic coliphages were more informative for predicting gastrointestinal problems in freshwater than enterococci. But before changing international standards, this result needs to be confirmed using modern data and a significantly larger number of water bodies.

News source

Edward KS Lam, Márta Vargha, Daniel Thomas, Roland Salmon, Mihály Kádár, Maria José Figueras Salvat, Julii Brainard, Jamie Bartram, Zsofia Barna, Paul R. Hunter. “A Randomized Controlled Trial Assessing Infectious Disease Risks from Bathing in Inland Recreational Waters.” International Journal of Hygiene and Environmental Health, 2026;277:114888. The article was published online August 22, 2026. This is a secondary analysis of previously unpublished individual data from the freshwater part of the European Epibathe study. DOI: 10.1016/j.ijheh.2026.114888.

The study was funded by the European Union's Sixth Framework Programme, and the new analysis was funded by the UK National Institute for Health and Care Research's Gastrointestinal Infections Health Protection Unit in partnership with the UK Health Security Agency. The authors declare no conflicts of interest.

The main result: skin problems were approximately twice as common among swimmers, while the level of microbial contamination was the determining factor for gastrointestinal illnesses: higher concentrations of E. coli and somatic coliphages were associated with an increased risk, while intestinal enterococci showed no such association. The study raises the question of whether sanitary standards for freshwater should differ from those for sea beaches.