Trends and status of liver neoplasms in flatfish

There has been a significant decrease in the prevalence of toxicopathic liver neoplasms (cancer) in flatfish from the Celtic Seas and Greater North Sea Marine Strategy Framework Directive sub-regions. Whilst Celtic Seas fish still exhibit a ‘significant response’, Greater North Sea fish exhibit an ‘elevated response’ and are marginally above background levels. The evidence provided in this assessment indicates that pollution effects in the marine environment, specifically concerning the formation of liver neoplasms, are decreasing. This finding informs good progress towards reducing chronic pollution effects.

Background

UK target on liver neoplasms in flatfish

This indicator is used to supplement the target covering the biological effects of contaminants set out in the Marine Strategy Part One (HM Government, 2012), which requires that “the intensity of those biological or ecological effects due to contaminants agreed by OSPAR as appropriate for Marine Strategy Framework Directive purposes are below the toxicologically-based standards”. Whilst this indicator has not yet been agreed by OSPAR for Marine Strategy Framework Directive (European Commission, 2008) purposes, it provides valuable information on the cause-effect relationship between environmental contaminants and the presence of liver lesions in fish and has, therefore, been included in the assessment of Good Environmental Status for contaminants.

Key pressures and impacts

The key pressure associated with high incidences of liver neoplasms in wild fish populations is chronic long-term exposure to anthropogenic contaminants including the organohalogens and metals. Fish age and sex also has an influence on the formation of liver neoplasms, therefore assessment tools with the ability to normalise for these factors are crucial for investigating the true effects of environmental contaminants.

Measures taken to address the impacts

There is a robust UK legislative framework in place for controlling pollution from the main sources of contaminants (energy production, transport, urban and industrial uses, including appropriate consenting and Water Framework Directive River Basin Management Plans (HM Government, 2015b) described in the Marine Strategy Part Three (HM Government, 2015a). However, genotoxic chemicals entered the sea before many of the controls were put in place and, due to their persistence, are still present.

Monitoring, assessment and regional co-operation

Areas that have been assessed

Status and trends assessments were conducted for the UK-portion of the Greater North Sea and Celtic Seas Marine Strategy Framework Directive sub-regions, and at the smaller scale of the eight UK biogeographic marine regions set out in the Charting Progress 2 (UKMMAS, 2010).

Monitoring and assessment methods

The prevalence of liver neoplasms in the flatfish dab (Limanda limanda;Figure 1), is monitored in UK waters as part of the UK Clean Seas Environment Monitoring Programme. The method used for this assessment is newly developed by the UK to complement the International Council for the Exploration of the Seas Fish Disease Index (Lang and Wosniok, 2008). The UK proposes this draft method as a tool for the specific assessment of fish liver neoplasms (cancers). This method can correct for the confounding factors of age and sex, thus allowing for a more accurate assessment of toxicopathic liver neoplasms.

Figure 1. The common dab (Limanda limanda) is the primary bioindicator species for coastal and offshore fish disease monitoring in the UK.

Assessment thresholds

The assessment thresholds were classified into three responses: background, elevated, and significant. The classification was achieved by separating the total observable range of liver neoplasm prevalence in an assessment region into thirds to inform progress towards achieving the target.

Regional co-operation

The Fish Disease Index was developed by the ICES Working Group on Pathology and Diseases in Marine Organisms (Lang and Wosniok, 2008), incorporating infectious external diseases to assess general fish health. This new assessment tool complements the Fish Disease Index and is specific to toxicopathic liver neoplasms only.

Further information

Fish disease monitoring has historically contributed to the OSPAR Co-ordinated Environmental Monitoring Programme of the Joint Assessment Monitoring Programme. This has been carried out in the UK under national monitoring activities since the 1980s as part of the National Marine Monitoring Programme and subsequently the Clean Seas Environment Monitoring Programme. The Clean Seas Environment Monitoring Programme monitors spatial and temporal trends of externally visible infectious diseases (including parasitic, bacterial and viral conditions) and toxicopathic pathological conditions (including reproductive disorders such as intersex and liver neoplasms). Monitoring of these diseases has been quality assured under the Biological Effects Quality Assurance in Marine Monitoring Programme since 2004. All UK data is submitted to the UK Marine Environment Monitoring and Assessment National database and the ICES International Databank.

Histopathology has previously been used to investigate the cause-effect relationship between environmental contaminants and the presence of toxicopathic liver lesions in numerous fish species (Stein and others, 1990; Myers and others, 1994; Stentiford and others, 2003; Stehr and others, 2004; Lang and others, 2017). Bottom-dwelling fish, particularly flatfish, are sensitive biomonitoring species resulting from their sedimentary habitat and a diet primarily consisting of benthic organisms. As such, in polluted environments they have a higher chance of being exposed to potentially harmful sediment-associated environmental contaminants. The liver is the major organ responsible for the absorption, metabolism and storage of nutrients. It is the primary barrier between the digestive system and the blood, therefore playing an important role in the detoxification of toxins, hazardous substances and their metabolites. Liver cells (hepatocytes) can store large quantities of glycogen or lipid (influenced by species, sex, maturity, physiological condition and temperature) resulting in differential sensitivity to hazardous lipophilic substances (Hinton and others, 2001). These substances can directly impact upon liver structure and function causing acute hepatocellular toxic injury and cell death following intracellular metabolism of parent compounds into highly reactive molecules. These molecules readily combine with proteins DNA, and RNA, impairing routine biological functions. They can also impart their effects indirectly by disrupting interactions and signalling between cells. The continued long-term assault of the liver by these mechanisms can result in the formation of chronic toxic liver injury, for example, neoplasia (carcinogenesis; Figures 2 and 3; Hinton and Couch, 1998; Hinton and others, 2008). For these reasons, the liver is a key target organ for investigating the toxic effects of contaminants in the aquatic environment.

Figure 2. Large cancerous neoplasm (*) on the liver from a dab sampled from the Greater North Sea.

 

Figure 3. Histology is used to confirm the diagnosis of externally visible neoplastic lesions, in addition to providing improved detection of lesions (arrow) that are not ordinarily visible to the naked eye.

Liver neoplasia (carcinogenesis) has been the subject of numerous studies investigating hazardous substances and their biological effects. Laboratory and mesocosm studies have successfully induced neoplastic lesions within fish livers following exposure to contaminants (Hawkins and others, 1990; Vethaak and others, 1996) and previous field studies have observed correlations between anthropogenic contaminants and neoplastic liver lesions. Several studies undertaken in North America provide strong evidence that polycyclic aromatic hydrocarbons and persistent organic pollutants (including polychlorinated biphenyls and organochloride pesticides) result in the formation of abnormal liver conditions (including neoplastic lesions), in several fish species (Myers and others, 1990; Myers and others, 1998; Stehr and others, 2004). Furthermore, polybrominated diphenyl ethers are structurally similar to polychlorinated biphenyls, persistent, and may also be implicated in the formation of liver neoplasms in fish. European field studies have since identified similar toxicopathic, pre-neoplastic and neoplastic lesions in the European flounder (Platichthys flesus) and common dab (Limanda limanda) that are considered the primary bioindicator species in the OSPAR North-east Atlantic region (Köhler, 1990; Lang and others, 2006; Vethaak and others, 2009; Stentiford and others, 2010).

The monitoring of liver neoplasms has been used successfully to demonstrate site recovery with regards to the biological effects of contaminants. Liver neoplasms decreased from 39% to 10% in local fish following the closure of a coking plant at a polycyclic aromatic hydrocarbon contaminated river in 1983 (Baumann and Harshbarger, 1995). Re-suspension of contaminated sediments following dredging in 1990, caused the prevalence to return to those levels previously seen during the 1980s. Similarly, remediation by sediment capping at a polycyclic aromatic hydrocarbon contaminated site (Myers and others, 2008) saw reductions in the prevalence of liver neoplasms. Therefore, monitoring of neoplastic lesions is useful for determining the indirect impact of human activities such as dredging, in addition to the direct toxicopathic effects of contaminants.

The assessment of general fish health using externally visible diseases is assessed using the ICES Fish Disease Index (Lang and Wosniok, 2008). The assessment of liver neoplasms in isolation requires a careful approach, not least due to the corresponding influence of fish age on their prevalence (Myers and others, 1994; Vethaak and others, 2009; Stentiford and others, 2010). Like other species, including humans, age has a positive influence on the formation of cancer in fish and is a potentially confounding factor for the interpretation of liver neoplasms in fish (Myers and others, 1994; Vethaak and others, 2009; Stentiford and others, 2010). Whether this influence in fish is simply the result of increasing age or the continued long-term chronic exposure to contaminants over time is difficult to determine and requires costly laboratory exposure studies. Nevertheless, it is crucial to determine the age of individual fish sampled if true like-for-like spatial and temporal comparisons are to be made. Previous comparisons of age-matched fish cohorts have revealed that whilst incidences of liver neoplasms increase with age, the age of onset is accelerated at those locations that also exhibit a high prevalence of liver neoplasms (Stentiford and others, 2010). Assessment tools that incorporate toxicopathic diseases in isolation, for example, liver neoplasms, whilst possessing the ability to normalise for age effects, will be increasingly valuable to investigate the true effects of environmental contaminants. The method used in this assessment has been developed to address this issue and is therefore able to discriminate between contaminant-related and age-related effects.

Assessment method

Field sample collection and analysis

Dab were collected from 43 fishing stations within English, Welsh and Scottish coastal and offshore waters during the Clean Seas Environment Monitoring Programme between 2004 and 2015, although data from Scotland was only collected between 2010 and 2015. The sampling strategy for England and Wales differed from Scotland in the number and size of fish sampled between years and sites. A sampling matrix is provided in Table 1.

Table 1. Sampling matrix (X indicates sampling) showing the fishing stations, UK marine regions and Marine Strategy Framework Directive (MSFD) sub-regions, sampled between 2004 and 2015. The England and Wales (E & W) Clean Seas Environment Monitoring Programme redesign in 2011 resulted in a biennial sampling of Greater North Seas and Celtic Seas. The Scotland Clean Seas Environment Monitoring Programme did not use the primary monitoring species (dab) prior to 2010.

Country

Fishing Station

UK marine region

MSFD sub-Region

2004

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

E&W

Amble

Northern North Sea

Greater North Sea

X

X

X

X

X

X

E&W

Farne Deeps

Northern North Sea

Greater North Sea

X

X

X

X

E&W

Off Flamborough

Northern North Sea

Greater North Sea

X

X

X

X

X

X

X

E&W

Tees Bay

Northern North Sea

Greater North Sea

X

X

X

X

X

X

X

E&W

Central Dogger

Southern North Sea

Greater North Sea

X

X

X

X

X

X

X

E&W

Indefatigable Bank

Southern North Sea

Greater North Sea

X

X

X

X

X

X

E&W

North Dogger

Southern North Sea

Greater North Sea

X

X

X

X

X

X

X

E&W

North East Dogger

Southern North Sea

Greater North Sea

X

X

X

X

X

E&W

Off Humber

Southern North Sea

Greater North Sea

X

X

X

X

X

X

E&W

Outer Humber

Southern North Sea

Greater North Sea

X

E&W

Outer Gabbard

Southern North Sea

Greater North Sea

X

X

E&W

West Dogger

Southern North Sea

Greater North Sea

X

X

X

X

X

X

X

E&W

Newhaven

Eastern Channel

Greater North Sea

X

X

X

X

X

E&W

Newhaven

Eastern Channel

Greater North Sea

X

X

X

X

X

E&W

Rye Bay

Eastern Channel

Greater North Sea

X

X

X

X

X

X

X

E&W

Lyme Bay

Eastern Channel

Greater North Sea

X

X

E&W

Camarthen Bay

W. Channel & Celtic Sea

Celtic Seas

X

X

X

X

X

E&W

Celtic Deep

W. Channel & Celtic Sea

Celtic Seas

X

E&W

West Lundy

W. Channel & Celtic Sea

Celtic Seas

X

X

E&W

Burbo Bight

Irish Sea

Celtic Seas

X

X

X

X

X

E&W

Inner Cardigan Bay

Irish Sea

Celtic Seas

X

X

X

X

E&W

Liverpool Bay

Irish Sea

Celtic Seas

X

X

X

X

X

X

E&W

Morecambe Bay

Irish Sea

Celtic Seas

X

X

X

X

X

E&W

North Cardigan Bay

Irish Sea

Celtic Seas

X

X

X

X

X

X

E&W

Red Wharf Bay

Irish Sea

Celtic Seas

X

X

X

X

X

E&W

St Bees Head

Irish Sea

Celtic Seas

X

X

X

X

X

X

Scotland

East of Balta Sound

Northern North Sea

Greater North Sea

X

X

X

Scotland

East Shetland Basin

Northern North Sea

Greater North Sea

X

Scotland

South East of Fair Isle

Northern North Sea

Greater North Sea

X

X

X

X

Scotland

Outer Moray Firth

Northern North Sea

Greater North Sea

X

X

Scotland

Southern Moray Firth

Northern North Sea

Greater North Sea

X

X

Scotland

Long Forties

Northern North Sea

Greater North Sea

X

Scotland

Montrose Bank

Northern North Sea

Greater North Sea

X

X

X

X

X

X

Scotland

Outer Firth of Forth

Northern North Sea

Greater North Sea

X

Scotland

Inner Firth of Forth

Northern North Sea

Greater North Sea

X

Scotland

Balcary Point

Irish Sea

Celtic Seas

X

X

X

Scotland

Holy Loch

Irish Sea

Celtic Seas

X

X

X

X

X

Scotland

North Minch

Minches & W Scotland

Celtic Seas

X

X

X

X

Scotland

Gallen Point

Scottish Continental Shelf

Celtic Seas

X

Scotland

Burra Haaf

Scottish Continental Shelf

Celtic Seas

X

Fish livers were processed for histological analysis using laboratory guidelines described by Feist and others (2004). All fish were analysed for neoplastic lesions using diagnostic criteria agreed under the Biological Effects Quality Assurance in Marine Monitoring Programme Fish Disease Measurement programme. Individual fish were given a non-accumulative score of 1, if at least one neoplastic lesion was detected (for example, the presence of two individual lesion types would result in a score of 1). Age determination of corresponding fish was carried out using otolith analysis (Easey and Milner, 2008) and used to determine the prevalence of liver neoplasms for each age class.

Preliminary data analysis and assessment

The England and Wales Clean Seas Environment Monitoring Programme was redesigned in 2011 to adopt a biennial sampling cycle to include the Greater North Sea one year, followed by the Celtic Seas the next. This redesign resulted in an inability to conduct a continuous year-on-year temporal trend assessment. A preliminary analysis was conducted to develop the assessment tool and to ensure the most appropriate use of the data collected. Specifically, this was to avoid any bias caused by potentially disproportionate numbers of fish within a given age class.

Data were subsequently split into two periods to include a baseline (2004 to 2010) and assessment (2011 to 2015) period. Scotland data was collected from 2010 to 2015, therefore the data for 2010 was removed from the analysis because it fell outside of the assessment period assigned to the England and Wales data (2011 to 2015). This was done because the analysis of data periods, comprised of different duration, are not directly comparable such as comparing 5 years of England and Wales data (2011 to 2015) to 6 years of Scotland data (2010 to 2015). The treatment of the data in this manner allowed for efficient use of UK data for the purposes of conducting a UK-wide assessment. Following the preliminary data analysis, a generalised linear model was used to normalise for the effects of sex and age, and determine any temporal trend increase or decrease in the prevalence of neoplasms between the baseline and assessment periods.

England and Wales data

Two levels of assessment were conducted using England and Wales data:

1. MSFD sub-regions (Greater North Sea and Celtic Seas)
2. UK biohydrographic marine regions (Eastern Channel, Irish Sea, Northern North Sea, Southern North Sea, Western Channel and Celtic Sea) that were previously assessed during the Charting Progress 2 report; UKMMAS, 2010)

The prevalence range of liver neoplasms in England and Wales waters is reportedly the highest observed within the OSPAR region of the North East Atlantic (Stentiford and others, 2009; Stentiford and others, 2010; Lang and others, 2017). The indicator response was classified into three categories of background, elevated and significant, by separating the total observable range of liver neoplasm prevalence, corresponding to Marine Strategy Framework Directive sub-region or UK marine region, into thirds.

Scotland data

Scotland data were treated independently of England and Wales data because of limitations in the number of fish and a lack of data from the baseline period (2004 to 2010). As a result, the Scotland and the England and Wales data were not directly comparable. Consequently, it was not possible to incorporate Scotland data into the generalised linear model, alongside England and Wales data, for an assessment at the Marine Strategy Framework Directive sub-regional level, or for the development of assessment criteria thresholds. However, an indicative assessment of Scotland data at the UK marine regional level was conducted against proposed assessment criteria derived from England and Wales marine regional data, although this data could not be normalised for age or sex.

One level of assessment was conducted using Scotland data: UK biohydrographic marine regions (Irish Sea, Minches and Western Scotland, Northern North Sea and the Scottish Continental Shelf) that were previously assessed during the Charting Progress 2 report (UKMMAS, 2010). Since age and the number of females have a positive influence (increasing liver neoplasms increase with both) a brief analysis of age and sex distribution using the Scotland data was conducted to inform any assessment as best as possible.

Results

Findings in the 2012 UK Initial Assessment

This assessment method is a newly developed approach; therefore, no direct comparison is possible to the UK Initial Assessment (HM Government, 2012). However, this indicator assessment does include data from 2004 to 2015, and therefore covers the period reported in the Initial Assessment.

Latest findings

The evidence provided in this indicator assessment demonstrates that the impact of contaminants in the marine environment, specifically concerning the formation of toxicopathic liver neoplasms (cancer), is decreasing. This finding is based on 12 years of monitoring data and informs good progress towards reducing chronic pollution effects.

Status assessment

There is a significant difference in the prevalence of liver neoplasms between Marine Strategy Framework Directive sub-regions. Liver neoplasm prevalence in the Greater North Sea was significantly lower (p ≤ 0.001) than in the Celtic Seas. This indicator assessment shows that the prevalence of liver neoplasms during the assessment period was 7.0% in Celtic Sea fish, thus exhibiting a ‘significant response’ (≥ 6.87% neoplasms in population), and 4.2% in Greater North Sea fish, thus exhibiting an ‘elevated response’ (3.44 to 6.86% neoplasms in population).

Trend assessment

The prevalence of liver neoplasms has decreased significantly, and at a similar rate, in both the Celtic Seas (from 10.3 to 7.0%) and Greater North Sea (from 6.4 to 4.2%), between the baseline (2004 to 2010) and assessment (2011 to 2015) periods.

Further information

Preliminary data analysis and development of assessment

A preliminary analysis was conducted prior to the assessment being undertaken, to ensure the most appropriate use of the data collected. Specifically, this was to avoid any bias caused by potentially disproportionate numbers of fish within a given age class.

A summary of England and Wales fish that were positive or negative for liver neoplasms in each age class during the entire sampling period of 2004 to 2015 is shown in Table 2. This shows a very low incidence of neoplasms in young fish aged 1 to 2 years. Conversely, older fish generally have increased incidences of liver neoplasms, although these were not caught in sufficient numbers. Therefore, the high incidence of liver neoplasms in relatively low numbers of older fish has considerable impact on the data.

Table 2. The number of dab sampled from each age class with (positive) and without (negative) liver neoplasms in England and Wales sampling regions for the period 2004 to 2015.

Age (years)	1	2	3	4	5	6	7	8	9	10	11	12	13
negative	562	1338	1809	1436	1262	1112	532	209	61	18	2	1	0
positive	0	6	30	58	123	193	151	105	56	28	5	0	1

Similar numbers of fish were sampled between the baseline (2004 to 2010) and assessment period (2011 to 2015) across all regions (Table 2). Examination of the summary statistics for both the baseline and assessment periods demonstrated that the age distribution of fish was very similar, with the interquartile range comprised of 3 to 6-year-old fish (Table 4). Based on this preliminary analysis, the final indicator assessment only incorporated 3 to 6-year-old fish to avoid bias whilst maintaining statistical rigour.

Table 3. The number of dab sampled from each age class between the baseline (2004 to 2010) and assessment (2011 to 2015) periods from all sampling regions.

Age (years	1	2	3	4	5	6	7	8	9	10	11	12	13
baseline period	392	603	973	785	806	686	323	142	60	20	3	1	0
assessment period	170	741	866	709	579	619	360	172	57	26	4	0	1

Table 4. Summary statistics showing the age distribution (in years) of dab sampled during the baseline (2004 to 2010), assessment (2011 to 2015) and Scottish assessment periods (2011 to 2015).

	minimum	1st Quartile	median	mean	3rd Quartile	maximum
baseline period	1	3	4	4.169	6	12
assessment period	1	3	4	4.280	6	13
Scottish assessment period	1	3	5	4.731	6	10

A further breakdown of the number of fish sampled within the 3 to 6 year old age range from all Marine Strategy Framework Directive sub-regions and UK marine regions is provided in Table 5. In addition to providing the most reliable data source due to their abundance, using 3 to 6-year-old fish in combination with five-year assessment periods has the advantage of using independent fish cohorts between the baseline and assessment periods.

Table 5. The number of 36-year-old dab sampled in Marine Strategy Framework Directive sub-regions (*) and UK marine regions during the baseline (2004 to 2010) and assessment (2011 to 2015) periods. For each region, the baseline and assessment periods correspond to the left and right columns respectively.

Age	Celtic Seas*		Greater North Sea		Eastern Channel		Irish Sea		Northern North Sea		Southern North Sea		Western Channel and Celtic Sea
3	400	355	573	511	205	164	250	291	121	187	247	160	150	64
4	308	253	477	456	36	102	257	195	145	149	296	205	51	58
5	354	200	452	379	10	29	309	177	87	126	355	224	45	23
6	277	147	409	472	2	17	257	129	83	132	324	323	20	18

The preliminary analysis revealed that female fish were more susceptible to liver neoplasms compared to males. Therefore, sex was included in the generalised linear model to adjust for the differences in the ratio of male to female fish caught between sites, regions and years. Age was also incorporated into the model to adjust for any age differences between individual age classes. This allowed for an effective comparison of liver tumour prevalence between regions where age and sex distribution may differ. Comparisons of neoplasm prevalence were made between baseline (2004 to 2010) and assessment (2011 to 2015) periods for Marine Strategy Framework Directive sub-region and UK marine regions. Statistical comparisons were made using logistic regression to determine spatial and temporal differences, including their statistical significance. The indicator response was classified into three categories of background, elevated and significant, by separating the total observable range of liver neoplasm prevalence, corresponding to Marine Strategy Framework Directive sub-region or UK marine region, into thirds (Table 6).

Table 6. Thresholds for background, elevated and significant responses for Marine Strategy Framework Directive sub-regions and UK marine regions.

Region	Background Response	Elevated response	Significant Response
MSFD Region	≤3.43%	3.44-6.86%	≥6.87%
UK Marine Region	≤3.87%	3.88-7.74%	≥7.75%

Assessment of liver neoplasms by UK marine and Marine Strategy Framework Directive sub-regions (England & Wales data)

 

There is a significant difference in the prevalence of liver neoplasms between individual UK marine regions. This assessment does not incorporate data from sampling sites in Scotland that also share the same UK marine regions as England and Wales such as Northern North Sea and Celtic Sea. Logistic regression demonstrated that there has been a statistically significant reduction in the prevalence of liver neoplasms between the baseline and assessment periods for all UK marine regions (p ≤ 0.01), except for Eastern Channel (Table 7).

Table 7. The percentage prevalence and assessment status of liver neoplasms for Marine Strategy Framework Directive sub-regions and UK marine regions in England, Wales and Scotland. Data for Scotland were not normalised for age or sex during this assessment. The increasing trend denoted with * is not statistically significant.

MSFD REGION	Baseline Period			Assessment Period
	percentage prevalence	assessment status	percentage prevalence		assessment status	trend
CELTIC SEAS	10.3	significant	7.0		significant	decreasing
GREATER NORTH SEA	6.4	elevated	4.2		elevated	decreasing
UK MARINE REGION (ENGLAND AND WALES DATA)
EASTERN CHANNEL	0.4	background	3.2		background	increasing*
IRISH SEA	11.6	significant	7.8		elevated	decreasing
NORTHERN NORTH SEA	2.5	background	1.9		background	decreasing
SOUTHERN SOUTH SEA	9.1	significant	6.0		elevated	decreasing
WESTERN CHANNEL & CELTIC SEA	4.9	elevated	3.1		background	decreasing
UK MARINE REGION (SCOTLAND DATA)
IRISH SEA	no data		4.4		elevated	unknown
MINCHES AND WESTERN SCOTLAND	no data		2.4		background	unknown
NORTHERN NORTH SEA	no data		3.4		background	unknown
SCOTTISH CONTINENTAL SHELF	no data		1.1		background	unknown

A reduction in the prevalence of liver neoplasms was observed between the baseline and assessment periods in the Irish Sea (down from 11.6 to 7.8%), Northern North Sea (from 2.5 to 1.9%), Southern North Sea (from 9.1 to 6.0%), and Western Channel and Celtic Sea (from 4.9 to 3.1%) regions. An increase was observed in the Eastern Channel (up from 0.4 to 3.2%), although this was statistically insignificant. The inclusion of an interaction term between region and sampling period was not significant (p ≥ 0.05), suggesting that the prevalence of neoplasms at all sites (except Eastern channel) have reduced at a similar rate. The application of indicator response thresholds to the assessment period data (2010 to 2015), revealed the following (Figures 4 and 5):

• Irish Sea fish exhibit a ‘significant response’ (≥ 7.7% neoplasms in population)
• Southern North Sea fish exhibit an ‘elevated response’ (3.88 to 7.74% neoplasms in population)

Eastern Channel, Northern North Sea, and the Western Channel and Celtic Seas fish exhibit a ‘background response’ (≤ 3.87% neoplasms in population)

Figure 4. Relative position of UK marine regions (English and Welsh data) to background, elevated (green) and significant (orange) response thresholds (dashed lines) during the baseline (2004-10, diamonds) and assessment (2011-15, circles) periods. Arrows denote direction of trend (diagonal arrows indicate where trends are statistically significant).

 

Figure 5. UK marine regional status for liver neoplasms in dab. Regions coloured amber indicate ‘significant response’; green regions indicate ‘elevated response’; and the blue regions indicate a ‘background response’. The grey regions had no data collected. Status assessments in regions 6 and 7 used un-normalised data against proposed assessment criteria and are shown as an “indicative assessment” only. The Northern North Sea and Irish Sea UK marine regions (Regions 1 and 5) are coloured based on England and Wales data only due to the absence of normalised data from Scotland.

There is a significant difference in the prevalence of liver neoplasms between Marine Strategy Framework Directive sub-regions. Logistic regression demonstrated that liver neoplasm prevalence in the Greater North Sea was significantly lower (p≤ 0.001) than in the Celtic Seas. The data also show that the prevalence has reduced significantly (p ≤ 0.001) and at a similar rate, in both the Celtic Seas (down from 10.3 to 7.0%) and Greater North Sea (from 6.4 to 4.2%; Table 7 and Figures 6 and 7). The application of indicator response thresholds to the assessment period data (2010 to 2015), revealed that Celtic Sea fish exhibit a ‘significant response’ (≥ 6.87% neoplasms in population) and Greater North Sea fish exhibit an ‘elevated response’ (3.44 to 6.86% neoplasms in population).

Figure 6. The relative position of Celtic Seas and Greater North Sea Marine Strategy Framework Directive sub-regions (England and Wales data) to elevated (green) and significant (orange) response thresholds (dashed lines) during the baseline (2004 to 2010, diamonds) and assessment (2011 to 2015, circles) periods. Arrows denote direction of trend (diagonal arrows indicate where trends are statistically significant).

 

Figure 7. Marine Strategy Framework Directive sub-regional status for liver neoplasms in dab. Regions coloured amber indicate a ‘significant response’ and the green regions indicate a ‘background response’. The grey regions in Scotland had insufficient data for incorporation into assessment model for subsequent age and sex normalisation. The Greater North Sea region is coloured resulting from the use of England and Wales data only.

Assessment of liver neoplasms by UK marine regions (Scotland data)

A preliminary analysis was conducted prior to the assessment being undertaken, to ensure the most appropriate use of the data collected. Analysis of age distribution between the England and Wales baseline period (2004 to 2010), England and Wales (2011 to 2015) and the Scotland assessment period (2010 and 2015), indicated that the mean age was 4.2, 4.3 and 4.7 respectively (Table 4). The standard deviation was calculated to be 1.93, 1.97 and 1.69 respectively. Further analyses of Scotland data indicated a higher proportion of female fish (n=523) were sampled compared to male fish (n=265). Females generally exhibit a higher prevalence of liver neoplasms compared to males. Whilst the ratio of females to males was considerably higher, any assessment using assessment criteria derived from England and Wales data would, therefore, represent a “worst case scenario” in the absence of any normalisation. We concluded that a rudimentary assessment could be conducted using Scotland data and assessment criteria derived from England and Wales data. This assessment does not incorporate data from sampling sites in England and Wales that also share the same UK marine regions as Scotland such as Northern North Sea and Celtic Sea.

The prevalence of liver neoplasms (Table 7) from Irish Sea, Minches and Western Scotland, Northern North Sea and the Scottish Continental Shelf regions was 4.4%, 3.4%, 2.4% and 1.1% respectively. The application of England and Wales indicator response thresholds to Scotland assessment period data (2010 to 2015) revealed that fish from all Scottish UK marine regions, except for the Irish Sea, exhibit a ‘background response’ (≤ 3.87% neoplasms in population), and Irish Sea fish, exhibit an ‘elevated response’ (≥ 7.7 neoplasms in population). Based on data limitations, the Scotland assessment should be viewed as indicative only. Scotland data (un-normalised) could not be combined with England and Wales data (normalised with respect to age and sex) for assessment of Marine Strategy Framework Directive sub-regions.

Conclusions

This Marine Strategy Framework Directive regional assessment indicates that there has been a statistically significant decrease in the prevalence of toxicopathic liver neoplasms (cancer) in UK waters. The Marine Strategy Framework Directive regional assessment of liver neoplasms indicate that Greater North Sea and Celtic Seas fish exhibit an ‘elevated’ and ‘significant’ response respectively. The ‘significant response’ observed in the Celtics Seas was largely caused by the disproportionately high number of sampling sites situated within the Irish Sea UK marine area. Furthermore, whilst Greater North Sea fish exhibit an “elevated response”, liver neoplasm prevalence has decreased and are now marginally above background levels.

Overall, this assessment shows that all UK marine and Marine Strategy Framework Directive regions (except for the increase observed at Eastern Channel) exhibited a statistically significant decrease in the prevalence of toxicopathic liver neoplasms. This indicates that pollution effects in the marine environment are decreasing, which informs good progress to reducing chronic pollution effects and establishing Good Environmental Status.

Further information

The ‘significant response’ observed in the Celtics Seas was largely caused by the disproportionately high number of sampling sites situated within the Irish Sea UK marine area (n = 9), compared to the Western Channel and Celtic Sea UK marine area (n = 4). Since these two UK marine areas collectively make up the Celtic Seas Marine Strategy Framework Directive sub-region, the high incidences of liver neoplasms in the Irish Sea has a substantial influence on the overall indicator response for this Marine Strategy Framework Directive region. Consequently, the Celtic Seas exhibits a ‘significant response’. However, the indicative assessment of Scotland data in isolation suggests that the future inclusion of this data into the assessment model could further reduce the indicator response in the Celtic Seas region.

This indicator assessment shows that the assessment of smaller UK marine areas provides increased data resolution and allows identification of where Marine Strategy Framework Directive sub-regions may not be reaching specified targets. All UK marine areas, except the Eastern Channel, showed significant decreases in liver neoplasm prevalence. Only two UK marine areas (Irish Sea and Southern North Sea) of the 7 that were assessed, demonstrated an above background prevalence.

Knowledge gaps

The assessments for Scotland are indicative only. Insufficient numbers of fish and corresponding age data from Scotland resulted in an inability to normalise data for age. As such, Scotland could not be incorporated into the wider Marine Strategy Framework Directive regional assessment. Both England and Wales and Scotland possess sampling sites within the Northern North Sea and Irish Sea UK marine regions. Assessments within these regions did not include Scotland data because they could not be normalised alongside England and Wales data within the assessment model.

The response thresholds presented in this assessment are preliminary and further refinement will be required via the appropriate ICES working groups.

Further information

• Age determination should be undertaken for all samples to allow data to be incorporated into the assessment model.
• Assessment criteria were derived from England and Wales data only, although it is understood to represent the largest observable range concerning the prevalence of liver neoplasms in the OSPAR region.
• Further analysis of data held in the ICES databank should be conducted to further refine thresholds and ensure they are applicable to the entire OSPAR area.

References

Baumann PC, Harshbarger JC (1995) ‘Decline in liver neoplasms in wild brown bullhead catfish after coking plant closes and environmental PAHs plummet’ Environmental Health Perspectives, 103(2):168 (viewed on 14 December 2018)

Easey W, Millner RS (2008) ‘Improved Methods for the Preparation and Staining of Thin Sections of Fish Otoliths for Age Determination’ Science Series Technical Report Cefas Lowestoft 143:12pp (viewed on 14 December 2018)

European Commission (2008) ‘Directive 2008/56/EC of the European Parliament and of the Council of 17 June 2008 establishing a framework for community action in the field of marine environmental policy (Marine Strategy Framework Directive)’ Official Journal of the European Union L 164, 25.6.2008, pages 19-40 (viewed 21 September 2018)

Feist SW, Lang T, Stentiford GD, Köhler A, (2004) ‘Biological effects of contaminants: Use of liver pathology of the European flatfish dab (Limanda limanda L) and flounder (Platichthys flesus L) for monitoring’ ICES Techniques in Marine Environmental Sciences, 1-43 (viewed on 14 December 2018)

Hawkins WE, Walker WW, Overstreet RM, Lytle JS, Lytle TF (1990) ‘Carcinogenic effects of some polycyclic aromatic hydrocarbons on the Japanese medaka and guppy in waterborne exposures’ Science of the Total Environment 94(1-2): pages 155-167 (viewed on 14 December 2018)

Hinton DE, Couch JA (1998) ‘Architectural pattern, tissue and cellular morphology in livers of fishes: relationship to experimentally-induced neoplastic responses’ In Fish Ecotoxicology, Braunbeck T, Hinton DE, Streit B (eds), pages 141-64 Birkhäuser Verlag, Basel, Switzerland (viewed on 14 December 2018)

Hinton DE, Segner H, Braunbeck T (2001) ‘Toxic responses of the liver In Target Organ Toxicity in Marine and Freshwater teleosts’, Volume 1 Schlenk D, Benson WH (eds), pages 225 - 266 CRC Press, Boca Raton

Hinton DE, Segner H, Au DW, Kullman SW, Hardman RC (2008) ‘Liver toxicity’ In The Toxicology of Fishes Giulo, RT, Hinton, DE (eds), pages 327 - 400 CRC Press, Boca Raton

HM Government (2012) ‘Marine Strategy Part One: UK Initial Assessment and Good Environmental Status’ (viewed on 16 November 2018)

HM Government (2015a) ‘Marine Strategy Part Three: UK Programme of Measures’ December 2015 (viewed on 5 July 2018)

HM Government (2015b) ‘River basin management plans: 2015’ (viewed on 15 November 2018)

Köhler A, (1990) ‘Identification of contaminant-induced cellular and subcellular lesions in the liver of flounder (Platichthys flesus L) caught at differently polluted estuaries’ Aquatic Toxicology 16:271-293 (viewed on 14 December 2018)

Lang T, Wosniok W, Barsiene J, Broeg K, Kopecka J, Parkkonen J (2006) ‘Liver histopathology in Baltic flounder (Platichthys flesus) as indicator of biological effects of contaminants’ Marine Pollution Bulletin 53:488-496 (viewed on 14 December 2018)

Lang T, Wosniok W (2008) ‘The Fish Disease Index: a method to assess wild fish disease data  in the context of marine environmental monitoring’ ICES CM 2008/D:01, page 13 (viewed on 12 December 2018)

Lang T, Feist SW, Stentiford GD, Bignell JP, Vethaak AD, Wosniok W (2017) ‘Diseases of dab (Limanda limanda): Analysis and assessment of data on externally visible diseases, macroscopic liver neoplasms and liver histopathology in the North Sea, Baltic Sea and off Iceland Marine environmental research’ 124:61-69 (viewed on 14 December 2018)

Myers MS, Anulacion BF, French BL, Reichert WL, Laetz CA, Buzitis J, Olson OP, Sol S , Collier, TK (2008) ‘Improved flatfish health following remediation of a PAH-contaminated site in Eagle Harbor’, Washington Aquatic Toxicology, 88(4):277-288 (viewed on 14 December 2018)

Myers, MS, Johnson, LL, Hom, T, Collier, TK, Stein, JE and Varanasi, U (1998) ‘Toxicopathic hepatic lesions in subadult English sole (Pleuronectes vetuls) from Puget Sound, Washington, USA: relationships with other biomarkers of contaminant exposure’ Marine Environmental Research, 45(1):pp47-67 (viewed on 14 December 2018)

Myers MS, Landahl JT, Krahn MM, Johnson LL , McCain BB (1990) ‘Overview of studies on liver carcinogenesis in English sole from Puget Sound; evidence for a xenobiotic chemical etiology I: pathology and epizootiology’ Science of the Total Environment, 94(1):33-50 (viewed on 14 December 2018)

Myers MS, Stehr CM, Olson OP, Johnson LL, McCain BB, Chan SL , Varanasi U (1994) ‘Relationships between toxicopathic hepatic lesions and exposure to chemical contaminants in English sole (Pleuronectes vetulus), starry flounder (Platichthys stellatus), and white croaker (Genyonemus lineatus) from selected marine sites on the Pacific Coast’, USA Environmental Health Perspectives, 102(2):200 (viewed on 14 December 2018)

Stehr CM, Myers MS, Johnson LL, Spencer S, Stein JE (2004) ‘Toxicopathic liver lesions in English sole and chemical contaminant exposure in Vancouver Harbour’ Canada Marine Environmental Research 57:55-74 (viewed on 14 December 2018)

Stein JE, Reichert WL, Nishimoto M , Varanasi U (1990) ‘Overview of studies on liver carcinogenesis in English sole from Puget Sound; evidence for a xenobiotic chemical etiology II: biochemical studies’ Science of the total environment, 94(1), pp51-69 (viewed on 14 December 2018)

Stentiford GD, Bignell JP, Lyons BP, Feist SW (2009) ‘Site-specific disease profiles in fish and their use in environmental monitoring’ Marine Ecology Progress Series, 381:1-15 (viewed on 14 December 2018)

Stentiford GD, Bignell JP, Lyons BP, Thain JE, Feist SW (2010) ‘Effect of age on liver pathology and other diseases in flatfish: implications for assessment of marine ecological health status’ Marine Ecology Progress Series 411, 215-230 (viewed on 14 December 2018)

Stentiford GD, Longshaw M, Lyons BP, Jones G, Green M, Feist SW (2003) ‘Histopathological biomarkers in estuarine fish species for the assessment of biological effects of contaminants’ Marine Environmental Research 55:137-159 (viewed on 14 December 2018)

UKMMAS (2010) ‘Charting Progress 2: An assessment of the state of the UK seas’ Published by Defra on behalf of the UK Marine Monitoring and Assessment Strategy community (viewed on 4 January 2018)

Vethaak AD, Jol JG (1996) ‘Diseases of flounder Platichthys flesus in Dutch coastal and estuarine waters, with particular reference to environmental stress factors I Epizootiology of gross lesions’ Diseases of Aquatic Organisms, 26(2):81-97 (viewed on 14 December 2018)

Vethaak, AD, Jol, JG and Pieters, JP (2009) ‘Long-term trends in the prevalence of cancer and other major diseases among flatfish in the southeastern North Sea as indicators of changing ecosystem health’ Environmental Science , Technology, 43(6):2151-2158 (viewed on 14 December 2018)

Acknowledgements

Assessment metadata

Assessment Type	UK Marine Strategy Framework Directive Indicator Assessment
	D8
	D8.2 Effects of Contaminants
	Marine Strategy Part One
Point of contact email	marinestrategy@defra.gov.uk
Metadata date	Wednesday, August 1, 2018
Title	Trends and status of liver neoplasms in flatfish
Resource abstract	There has been a significant decrease in the prevalence of toxicopathic liver neoplasms (cancer) in flatfish from the Celtic Seas and Greater North Sea Marine Strategy Framework Directive sub regions. Whilst Celtic Seas fish still exhibit a ‘significant response’, Greater North Sea fish exhibit an ‘elevated response’ and are marginally above background levels. The evidence provided in this assessment indicates that pollution effects in the marine environment are decreasing, specifically concerning the formation of liver neoplasms. This finding informs good progress towards reducing chronic pollution effects.
Linkage	Please see links provided in ‘References’ section above.
Conditions applying to access and use	© Crown copyright, licenced under the Open Government Licence (OGL).
Assessment Lineage	The 2016 assessment of the UK's Clean Seas Environment Monitoring Programme described the status and trends of contaminant concentrations and biological effects measurements in biota and sediment at monitoring stations in waters around the UK. Liver neoplasm assessment was conducted on simplified data previously submitted to the MERMAN database between 2004 – 2015. All data were quality assured in accordance to the Biological Effects Quality Assurance in Monitoring Programmes (BEQUALM) Fish Disease Measurement scheme. The E&W CSEMP programme was redesigned in 2011 to adopt a biennial sampling cycle to include the Greater North Sea one year, followed by the Celtic Seas the next. This redesign resulted in an inability to conduct a continuous year-on-year temporal trend assessment. A preliminary analysis was conducted to ensure the most appropriate use of the data collected. Specifically, this was to avoid any bias caused by potentially disproportionate numbers of fish within a given age class.  Data were subsequently split into two periods to include a baseline (2004-10) and assessment (2011-15) period.   Furthermore, the Scotland data were collected from 2010-15, therefore the Scotland data for 2010 was removed from the analysis because it fell outside of the assessment period assigned to the E&W data (2011-15). This was done because the analysis of data periods, comprised of different durations, are not directly comparable i.e. comparing 5 years of E&W data (2011-15) to 6 years of Scotland data (2010-15). The treatment of the data in this manner allowed for efficient use of UK data for the purposes of conducting a UK-wide assessment. The results of the preliminary data analysis, used to develop the assessment tool, are located within the Results (extended) section of the indicator assessment. Following the preliminary data analysis, a generalised linear model (GLM) was used to normalise for the effects of sex and age, and determine any temporal trend increase or decrease in the prevalence of neoplasms between the baseline and assessment periods. Two levels of assessment were conducted using E&W data: (1) Marine Strategy Framework Directive sub-Regions and (2) UK biohydrographic marine regions (UKMMAS, 2010). The prevalence range of liver neoplasms in E&W waters is reportedly the highest observed within the OSPAR region of the North East Atlantic (Stentiford et al., 2009; Stentiford et al., 2010; Lang et al., 2015). Thus, the indicator response was classified into three categories of background, elevated and significant, by separating the total observable range of liver neoplasm prevalence, corresponding to Marine Strategy Framework Directive sub-Region or UK marine region, into thirds. Scotland data were treated independently of E&W data resulting from (a) limitations in the number of fish and (b) lack of data from baseline period (2004-2010). As a result, the E&W and Scotland data were not directly comparable. Consequently, it was not possible to incorporate Scotland data into the GLM, alongside E&W data, for an assessment at the Marine Strategy Framework Directive sub-regional level, or for the development of assessment criteria thresholds. However, an indicative assessment of Scotland data at the UK biohydrographic regional level was conducted against proposed assessment criteria derived from E&W marine regional data, although this data could not be normalised for age or sex. One level of assessment was conducted using Scotland data: UK biohydrographic marine regions (UKMMAS, 2010). Since older age and the number of females has a positive influence i.e. liver neoplasms increase; a brief analysis of age and sex distribution using the Scotland data was conducted to inform any assessment as best as possible. The results of this analysis are located within the extended results section of this indicator assessment.
Dataset metadata	RA: The Research Accession (RA) number is the unique laboratory identifier. Sub Sample Number: The sub sample number represents the individual fish number corresponding to a specific RA number. Location: The name provided to a specific sampling station visited during an annual CSEMP monitoring survey. Region: The name provided to a specific UK biogeographic region. Marine Strategy Framework Directive Region: The name provided to a specific Marine Strategy Framework Directive assessment region. Length (cm): The length is the total fish length in centimetres recorded to one decimal place. NR= not recorded. Weight (g): The weight is the whole fish weight in grams recorded to one decimal place. NR= not recorded. Sex: M= male, F= female, NR= not recorded. Liver Nodule (LN): Liver nodules are macroscopic liver tumours seen on the surface of the liver during tissue sampling. The number represents the maximum diameter in millimetres (mm). ccFCA: Clear-cell Foci of Cellular Alteration (FCA) are pre-neoplastic microscopic liver lesions. The pre-fix indicates their histological staining appearance i.e. clear-cell (cc). 1= present 0= absent. vFCA: Vacuolated Foci of Cellular Alteration (FCA) are pre-neoplastic microscopic liver lesions. The pre-fix indicates their histological staining appearance i.e. vacuolated (v). 1= present 0= absent. eFCA: Eosinophilic Foci of Cellular Alteration (FCA) are pre-neoplastic microscopic liver lesions. The pre-fix indicates their histological staining appearance i.e. eosinophilic (e). 1= present 0= absent. bFCA: Basophilic Foci of Cellular Alteration (FCA) are pre-neoplastic microscopic liver lesions. The pre-fix indicates their histological staining appearance i.e. basophilic (b). 1= present 0= absent. mFCA: Mixed-cell Foci of Cellular Alteration (FCA) are pre-neoplastic microscopic liver lesions. The pre-fix indicates their histological staining appearance i.e. mixed-cell (m). 1= present 0= absent. FCA: Indicates whether an FCA sub-type (above) was observed histologically, irrespective of frequency and type. 1= present 0= absent. Hepatocellular Adenoma (HCA): Indicates presence of one or more benign hepatocellular adenoma neoplasms. 1= present 0= absent. Cholangioma: Indicates presence of one or more benign cholangioma neoplasms. 1= present 0= absent. Hemangioma: Indicates presence of one or more benign hemangioma neoplasms. 1= present 0= absent. Pancreatic acinar cell adenoma: Indicates presence of one or more benign pancreatic acinar cell adenoma neoplasms. 1= present 0= absent. Hepatocellular Carcinoma (HCC): Indicates presence of one or more malignant hepatocellular carcinoma neoplasms. 1= present 0= absent. Cholangiocarcinoma: Indicates presence of one or more malignant cholangiocarcinoma neoplasms. 1= present 0= absent. Pancreatic acinar cell carcinoma: Indicates presence of one or more malignant pancreatic acinar cell carcinoma neoplasms. 1= present 0= absent. Mixed hepatobiliary carcinoma: Indicates presence of one or more malignant mixed hepatobiliary carcinoma neoplasms. 1= present 0= absent. Hemangiosarcoma: Indicates presence of one or more malignant hemangiosarcoma neoplasms. 1= present 0= absent. Hemangiopericytic sarcoma: Indicates presence of one or more malignant hemangiopericytic sarcoma neoplasms. 1= present 0= absent. Cancer: Indicates whether a liver neoplasm sub-type (above) was observed histologically, irrespective of frequency and type. 1= present 0= absent. Age: Age in years determined by otolith analysis. Reporting Year: The year of annual CSEMP Monitoring Survey in which samples were obtained.
Dataset DOI	Bignell and others (2019). Trends and status of liver neoplasms in flatfish from UK waters collected between 2004 and 2015. Cefas, UK. V1.doi: https://doi.org/10.14466/CefasDataHub.89

The Metadata are “data about the content, quality, condition, and other characteristics of data” (FGDC Content Standard for Digital Geospatial Metadata Workbook, Ver 2.0, May 1, 2000).

Metadata definitions

Assessment Lineage - description of data sets and method used to obtain the results of the assessment

Dataset – The datasets included in the assessment should be accessible, and reflect the exact copies or versions of the data used in the assessment. This means that if extracts from existing data were modified, filtered, or otherwise altered, then the modified data should be separately accessible, and described by metadata (acknowledging the originators of the raw data).

Dataset metadata – information on the data sources and characteristics of data sets used in the assessment (MEDIN and INSPIRE compliance).

Digital Object Identifier (DOI) – a persistent identifier to provide a link to a dataset (or other resource) on digital networks. Please note that persistent identifiers can be created/minted, even if a dataset is not directly available online.

Indicator assessment metadata – data and information about the content, quality, condition, and other characteristics of an indicator assessment.

MEDIN discovery metadata - a list of standardized information that accompanies a marine dataset and allows other people to find out what the dataset contains, where it was collected and how they can get hold of it.

Recommended reference for this indicator assessment

John P. Bignell¹, Craig Robinson² and Nick Taylor¹ 2018. Trends and status of liver neoplasms in flatfish. UK Marine Online Assessment Tool, available at: https://moat.cefas.co.uk/pressures-from-human-activities/contaminants/liver-neoplasms/

¹Centre for Environment, Fisheries and Aquaculture Science

²Marine Scotland