Content uploaded by Thorleifur Agustsson
Author content
All content in this area was uploaded by Thorleifur Agustsson on May 22, 2019
Content may be subject to copyright.
Contents lists available at ScienceDirect
Ecotoxicology and Environmental Safety
journal homepage: www.elsevier.com/locate/ecoenv
Gill damage and delayed mortality of Northern shrimp (Pandalus borealis)
after short time exposure to anti-parasitic veterinary medicine containing
hydrogen peroxide
Renée Katrin Bechmann
a,∗
, Maj Arnberg
a,1
, Alessio Gomiero
a
, Stig Westerlund
a,2
, Emily Lyng
a
,
Mark Berry
a,3
, Thorleifur Agustsson
a
, Tjalling Jager
b
, Les E. Burridge
c
a
NORCE Norwegian Research Centre, Mekjarvik 12, 4072, Randaberg, Norway
b
DEBtox Research, NL-3731, DN De Bilt, Netherlands
c
Burridge Consulting Inc., 61 Emmalee Dr Stratford PE, Canada, C1B 0B5, Canada
ARTICLE INFO
Keywords:
Delayed effects
Histopathology
Aquaculture
Hydrogen peroxide
Toxicokinetic-toxicodynamic (TKTD) model
General unified threshold model for survival
(GUTS)
ABSTRACT
Hydrogen peroxide (H
2
O
2
) is used as anti-parasitic veterinary medicine in salmon farms worldwide. In the
period from 2009 to 2018 a total of 135 million kg of H
2
O
2
was used in Norway, the world's largest producer of
Atlantic salmon. Since the treatment water is discharged to the sea, concerns have been raised about effects of
H
2
O
2
on the coastal ecosystem. In the present study, Northern shrimp (Pandalus borealis) have been exposed to
short pulses of H
2
O
2
in the PARAMOVE
®
formulation, followed by a recovery period in clean seawater. The
exposure concentrations represented 100, 1000 and 10 000 times dilutions of the prescribed treatment con-
centration for salmon; 15 mg/L, 1.5 mg/L and 0.15 mg/L H
2
O
2
. Significantly increased mortality was observed
after 2 h exposure to 15 mg/L H
2
O
2
(50%) and after 2 h exposure to 1.5 mg/L H
2
O
2
on 3 consecutive days (33%),
but no mortality was observed after 2 h exposure to 0.15mg/L. The mortality occurred 2–4 days after the first
pulse of exposure. The patterns of acute effects (immobility and death) could be captured with a toxicokinetic-
toxicodynamic model (GUTS), which allows extrapolations to LC50s for constant exposure, or thresholds for
effects given untested exposure profiles. Effects of H
2
O
2
were also detected in shrimp that survived until the end
of the recovery period. The feeding rate was 66% lower than in the control after 12 days of recovery for the
three-pulse 1.5 mg/L exposure. Furthermore, dose dependent tissue damage was detected in the gills and evi-
dence of lipid peroxidation in the hepatopancreas in shrimp exposed for 1 h to 1.5 mg/L and 15mg/L and kept in
recovery for 8 days. Fluorescence intensity in the hepatopancreas of treated shrimp increased 47% and 157% at
1.5 mg/L and 15 mg/L, respectively, compared to the control. Local hydrodynamic conditions will determine
how fast the concentration of H
2
O
2
will be diluted and how far it will be transported horizontally and vertically.
Results from dispersion modelling (literature data) together with the current experiments indicate that treatment
water with toxic concentrations of H
2
O
2
(1.5 mg/L) could reach P. borealis living more than 1 km from a treated
salmon farm.
1. Introduction
Norway is the world's leading producer of Atlantic salmon, and the
Norwegian Ministry of Trade, Industry and Fisheries aims to increase
the production of salmon in Norway (Nærings- og fiskeridepartementet,
2015). Continued growth of salmon aquaculture may lead to increased
discharge of anti-salmon lice chemicals. Despite the increasing use of
non-chemical methods to remove sea lice from farmed salmon, several
chemicals (bath treatment and medicated feed) are still used ex-
tensively (Aaen et al., 2015;Burridge et al., 2014;Lillicrap et al., 2015).
https://doi.org/10.1016/j.ecoenv.2019.05.045
Received 6 February 2019; Received in revised form 10 May 2019; Accepted 13 May 2019
∗
Corresponding author.
E-mail addresses: rebe@norceresearch.no (R.K. Bechmann), maar@norceresearch.no,mar@akvaplan.niva.no (M. Arnberg), algo@norceresearch.no (A. Gomiero),
stig.westerlund@lyse.net (S. Westerlund), ely@norceresearch.no (E. Lyng), Mark.berry@marinemanagement.org.uk (M. Berry),
thag@norceresearch.no (T. Agustsson), tjalling@debtox.nl (T. Jager), leburridge@gmail.com (L.E. Burridge).
1
Present address (from 1 June 2019): Akvaplan-niva, Postboks 1268, 7462 Trondheim, Norway.
2
Present address: Vålandsbakken 23, 4010 Stavanger, Norway.
3
Present address: Marine Management Organisation (MMO), Room 13, Ground Floor, Crosskill House, Mill Lane, Beverley, HU17 9JB, UK.
Ecotoxicology and Environmental Safety 180 (2019) 473–482
0147-6513/ © 2019 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/BY-NC-ND/4.0/).
T
Updated information about the use of chemicals in Norwegian aqua-
culture is available on the web page of the Norwegian Institute of Public
Health.
4
Most fish farms depend on chemical treatments to keep the
parasite numbers below national maximum permitted levels (Helgesen
et al., 2015). The emerging occurrence of reduced sensitivity of salmon
lice towards other chemicals led to the re-introduction of hydrogen
peroxide (H
2
O
2
) formulations for salmon lice treatment in 2009. In the
period from 2009 to 2018 a total of 135 million kg of (100%) H
2
O
2
has
been used in Norwegian salmon farms, according to data available on
the web site of the Norwegian Institute of Public Health
1
. The use of
H
2
O
2
has, however, been reduced from 43 million kg in 2014 to 7
million kg in 2018. Chemicals used for bath treatments are either added
directly into the net pen surrounded by a tarpaulin or administered in
well-boats. After completed treatment, chemicals are released untreated
directly into the environment, either by release of the tarpaulin or from
the well-boat. During the production cycle, salmon can be treated
several times with chemicals, hence marine organisms, such as shrimp,
living in areas with aquaculture could be exposed repeatedly for a re-
latively short period each time. Limited data are available on pulsed
exposure and reports on repeated pulses or fluctuating exposures are
even rarer (Dennis et al., 2012). Repeated short-term exposure studies
have, however, been performed with American lobsters (Homarus
americanus) (Burridge et al., 2000,2008) and crab larvae (Metacarcinus
edwardsii) exposed to anti-parasitic veterinary medicines (Gebauer
et al., 2017). Langford et al. concluded that repeated short-term ex-
posures of coastal ecosystems are likely to occur in Norway due to the
number and proximity of fish farms (Langford et al., 2014). Norwegian
fishermen have observed that Northern shrimp (Pandalus borealis) have
disappeared from several shrimp fields in areas with salmon farms and
this observation was confirmed by the Institute of Marine Research for
the coast of Helgeland (Eraker, 2016;Steinhold and Thonhagen, 2017).
The fishermen claim that shrimp disappear from areas where chemicals
are used to treat salmon against lice. Concern has also been raised about
potential negative effects on the kelp community, due to the sensitivity
of sugar kelp (Saccharina latissimi) to H
2
O
2
(Haugland et al., 2019).
Formulations with hydrogen peroxide as the active ingredient are
generally considered to be the bath treatment method with lowest en-
vironmental risk, because H
2
O
2
has a relatively short half-life in sea-
water and degrades to oxygen and water (Burridge et al., 2014;Haya
et al., 2005). The half-life of H
2
O
2
in seawater varies from 1 to 28 days,
depending on formulation, temperature, start concentration and or-
ganic content of the seawater (Bruno and Raynard, 1994;Fagereng,
2016;Haya et al., 2005;Lyons et al., 2014). Hydrogen peroxide is a
strong oxidizing agent, and the mechanism of action is lipid peroxida-
tion of cellular membranes and mechanical paralysis, in addition to
inactivation of enzymes and DNA replication (Cotran et al., 1989). In
sea lice, H
2
O
2
causes mechanical paralysis when air bubbles form in the
gut and hemolymph, causing the sea lice to detach and float to the
surface (Bruno and Raynard, 1994). Gill damage and mortality have
been observed for Atlantic salmon exposed to 1450–2580 mg/L H
2
O
2
for 10–20 min (Kiemer and Black, 1997). There is, however, limited
information published on the effects of the use of this chemical on non-
target marine crustaceans (Burridge et al., 2014;Urbina et al., 2019).
The main objective in this study was to investigate the effects in P.
borealis of exposure to short (1–2 h) pulses of diluted H
2
O
2
in the
PARAMOVE
®
formulation. A recovery period in clean seawater was
included in the experiments to study potential delayed effects of ex-
posure to PARAMOVE
®
. Mortality, feeding, swimming, gill histo-
pathology and lipid peroxidation of the hepatopancreas were in-
vestigated in order to assess the effects of exposure. The toxicokinetic-
toxicodynamic (TKTD) model GUTS (Jager et al., 2011) was used to
explain the dynamic mortality pattern of shrimp exposed to various
pulse treatments.
2. Materials and methods
2.1. Collection of shrimp and holding conditions
The experimental plan was approved by the Norwegian Animal
Research Authority (FOTS). Female shrimp (P. borealis) were collected
from Hillefjord (North of Åmøy, Rogaland County, Norway; 59° 04′00″
N, 5° 45′00′E) using a shrimp bottom trawl on the 10th of January
2017. Trawling lasted 40 min at 100 m depth. The animals were sorted
by hand, transferred to 50 L tanks that were regularly oxygenated to
maintain > 80% oxygen saturation, and transported to the laboratory
within 2 h of capture. Upon arrival at NORCE marine research facility
(Randaberg, Norway), shrimp were randomly divided between eight
500 L tanks (approximately 100 shrimp per tank). Each tank had a
continuous flow of sand filtered seawater adjusted to 7 °C, pumped from
75 m depth in the fjord close to the laboratory facilities. Shrimp were
fed 3 mm pellets of fish feed (Spirit supreme, Skretting, Norway) ad
libitum and acclimatized in laboratory conditions (7 °C, 34‰) for two
weeks before the first screening test was performed. All experiments
were performed in the period from 24th of January to 1st of March
2017. Female shrimp were used in all experiments, with a mean wet
weight of 8.3 g (SD 1.2 g), and mean carapace length 23 mm (SD 1 mm).
2.2. Experimental design, set-up and exposure concentrations
The experiments were performed in a continuous flow exposure
system with temperature regulated seawater. The mean temperature
was 6.81 ± 0.03 °C and the mean oxygen saturation in the tanks was
95.3 ± 0.8%. The mean salinity in the inlet water during the experi-
mental period was 34.3 ± 0.2‰. Female shrimp were exposed to one
or three pulses of diluted PARAMOVE
®
solution for either one or 2 h
followed by a recovery period in clean seawater to investigate delayed
effects of the short time exposure.
In Exp. 1 and Exp. 2. shrimp were exposed to short pulses of 15 mg/
L and 1.5 mg/L H
2
O
2
, respectively, to investigate effects on mortality,
immobilization, swimming activity and feeding rate (details below). Six
replicate tanks with 10 shrimp, were used for each treatment in both
experiments. The shrimp were exposed for 2 h to H
2
O
2
for three con-
secutive days, followed by two days of recovery in Exp. 1 and 12 days of
recovery in Exp. 2. In Exp. 3 shrimp were exposed to one pulse of
0.15 mg/L, 1.5 mg/L and 15 mg/L H
2
O
2
lasting either 1 h or 2 h fol-
lowed by 8 days of recovery. Effects of H
2
O
2
on gill histopathology, and
lipid peroxidation of the hepatopancreas were investigated in addition
to mortality, immobilization, swimming activity and feeding rate (de-
tails below). Four replicate tanks with 6 shrimp, were used for each
treatment. An overview of Exp. 1–3 is presented in Supplementary
Table 1 and the experimental set-up is illustrated in Supplementary
Figure 1.
The concentration of H
2
O
2
(PARAMOVE
®
; 49.5% w/w H
2
O
2
) used
to treat infestations of sea lice on Atlantic salmon is 1500 mg H
2
O
2
/L
for 20 min (Felleskatalogen, 2017). All experiments were conducted
using the PARAMOVE
®
formulation. During the experiments shrimp
were exposed to 0.15–15 mg/L H
2
O
2
, representing 10 000–100 times
dilution of the recommended treatment concentration for salmon. A
stock solution of PARAMOVE
®
in distilled water was prepared for each
test concentration. The density of PARAMOVE
®
was regularly measured
to ensure that PARAMOVE
®
still contained approximately 50% H
2
O
2
(density 1.2 g/cm
3
). To create the stock solutions for the 0.15 mg/L,
1.5 mg/L and 15 mg/L H
2
O
2
treatments, 0.255, 2.55 and 25.5 mL
Paramove
®
was added to 4 L distilled water in 5 L Schott bottles. The
bottles were agitated and placed on magnetic stirrers. A small amount
(3 mL/min) of stock solution was pumped by a peristaltic pump (model
520, Watson and Marlow, Cornwall, UK, Marprene pump tubing) into
the seawater inlets of each exposure tank (volume 46 L) to achieve the
4
https://www.fhi.no/nyheter/2019/2018-oppdrettsnaringen-bruker-stadig-
mindre-legemidler-mot-lakselus/.
R.K. Bechmann, et al. Ecotoxicology and Environmental Safety 180 (2019) 473–482
474
desired exposure concentrations of H
2
O
2
. The mean flow of seawater
into each exposure tank was 753 mL/min (Supplementary Figure 1).
The flow of stock solution and seawater into the tanks was monitored
during the experiments (Supplementary Figure 1). To verify the H
2
O
2
concentration in the exposure tanks and in diluted samples from the
header tanks a kit for semi-quantitative detection of H
2
O
2
was used
(Quantofix1–100 mg/L, Macherey-Nagel Germany supplied by Sigma
Aldrich). This semi-quantitative method could not be used in the
0.15 mg/L exposure which was below the detection limit for the
method.
2.3. Mortality, immobilization and swimming activity
The number of standing shrimp (normal control behaviour), swim-
ming shrimp (indication of stress), and immobilized shrimp in each
tank were recorded before and after each exposure pulse and once a day
in the recovery period in all three experiments. In Exp. 3, more frequent
observations of the shrimp were made during and right after the ex-
posure (0, 15, 30, 45, 60, 75, 90, 105, 120 and 240 min after start of the
experiment). Dead shrimp were recorded and removed daily.
2.4. Feeding rate
Feeding was determined as follows: new pellets of fish feed (one
pellet per shrimp) were added to the tanks and number of remaining
pellets were counted the following day. This gave information about
feed consumption which was expressed as number of pellets consumed
per shrimp per day. In Exp. 1 the feeding rate was determined once in
the exposure period (between first and second pulse) and the first day of
the recovery period. In Exp. 2 the feeding rate was determined once in
the exposure period (between second and third pulse), and every
second day of the recovery period. In Exp. 3 the feeding rate was de-
termined every second day of the experiment.
2.5. Gill histopathology
At the end of Exp. 3, gills from control shrimp and shrimp exposed
to 1.5 mg/L and 15 mg/L H
2
O
2
, were sampled for gill histopathology
(n = 15 samples per treatment). Gill samples were dissected out, fixed
in Davidson's fixative for 48 h and transferred to 10% buffered formalin
solution. Later, the samples were processed by serial alcohol dehydra-
tion, embedded in paraffin wax and cut into 6 μm transsections before
being stained by Mayer's hematoxylin and Eosin for observation by the
light microscope according to Boudet et al. (2015). The degree of his-
tological damage was scored in four fields on each of the 15 slides from
each treatment. Scores were based on the number of fields in which
histological changes were observed with (class 0) no histopathology in
any field, (class 1) = mild histopathology present in < 25% of the
fields, (class 2) = moderate histopathology present in 25%–75% of the
fields, and (class 3) = severe histopathology present in > 75% of the
field, following the scale suggested by Zodrow et al. (2004). Parasites
were scored as 0 = absent or 1 = present.
2.6. Lipid peroxidation in hepatopancreas
Hydrogen peroxide is a well-known oxidative stress inducer in
marine organisms. Among all toxicity mechanisms, one of the most
documented is Reactive Oxygen Species (ROS) induction in cells. ROS
are a major initiator of lipid peroxidation, and membrane bound
polyunsaturated fatty acids are their major targets. Linoleic acid (LA) is
the most abundant polyunsaturated fatty acid found in marine organ-
isms, including shrimp (Olsen et al., 2012;Ouraji et al., 2011). Fol-
lowing oxidative stress, lipids are oxidized by H
2
O
2
to organic per-
oxides (hydroperoxides), which further decompose to multiple α,β-
unsaturated aldehydes. The end-products of the peroxidation process
readily modify proteins at nucleophilic side chain sites. These are
insoluble and persistent alkyne-containing proteins accumulated and
stored in the lysosomal compartment. They can be selectively stained
and detected using 5-Carboxamido-(6-Azidohexanyl), Bis (Triethy-
lammonium Salt) 5-isomer based fluorescent probes in fixed cells and
tissues, using the Click-iT Lipid peroxidation detection kit (Life Tech-
nologies, Carsbad, USA).
At the end of Exp. 3, the hepatopancreas from control shrimp and
shrimp exposed to 1.5 mg/L and 15 mg/L H
2
O
2
, were sampled for his-
tochemical examination (n = 15 samples per treatment).
Hepatopancreas of each individual was quickly dissected out, snap-
frozen in liquid nitrogen and stored at −80 °C. The quantification of
lipid peroxidation was conducted using a modified protocol of Click-iT
Lipid Peroxidation Detection (catalog no. 10446, Life Technologies
Corporation) using the shrimp's natural linoleic acid levels as indicator
of no treatment control. The histochemical procedures were applied on
frozen tissue sections of hepatopancreas. Serial transsections of the
tissue (10 μm) were placed on glass slides, positioned in Hellendal's jars
and fixed in Baker's solution for 15 min at 4 °C. Sections were rinsed
twice with Hanks' saline solution to remove the fixative and incubated
with 0.5% TritonX-100 in 0.01M phosphate buffered saline (PBS), pH
7.4 at room temperature (RT). Cell membrane permeability reaction
was blocked by replacing the solution with 1% bovine serum albumin
(BSA) in PBS, pH 7.4 and incubating the sections at RT for 30 min. The
blocking solution was removed by washing tissue sections twice with
PBS before the fluorescent probe incubation, and then the slides were
placed in humidity chambers. Click-iT
®
reaction buffer, buffer additive
and CuSO
4
reaction promoter solution were prepared and mixed ac-
cording to the manufacturer's instructions. 200 μL of the reaction so-
lution was added to each slide, then left to react at RT in dark and
humid conditions for 30 min. The reaction solution was removed by
washing incubated tissue sections twice with 1% BSA in PBS and twice
with PBS. Sections were mounted in ProLong™Gold Antifade Mountant
(P36930, Invitrogen Corporation) and analysed in a Zeiss AxioVert 100
Inverted Fluorescence Phase Microscope with a FITC filter and coupled
to an Axiocam MRc5. Digital image analysis was carried out by ImageJ
1.51k software (NHI, USA). Each hepatopancreas section was divided
into quarters for statistical interpretation and scored for maximal
staining. The total area of digestive cells scanned in each measurement
was approximately 2000 μm
2
and five measurements were made per
section. Mean optical density was calculated for each sample and results
are reported as percentage of variation respect to control organisms
(Gomiero and Viarengo, 2014).
2.7. Toxicokinetic-toxicodynamic modelling
The complete data set for survival over time (Exp. 1–3) is analysed
with a toxicokinetic-toxicodynamic (TKTD) model, namely a special
case of the General Unified Threshold model for Survival (GUTS) (Jager
et al., 2011). As TKTD models explicitly deal with the mechanisms
underlying the effects over time, they offer the possibility to analyse the
survival data for the various treatments (different pulse length, dif-
ferent pulse intensity, and different number of pulses) in a single con-
sistent analysis. Thus, a single set of model parameters should be able to
capture the results from all of the treatments. For this model analysis,
we used the data for immobile (lying down) and dead animals com-
bined, as only two animals recovered from the immobile state, and in
the field, immobility likely is equivalent to death.
Details of the model and the modelling procedure are provided in
the supporting information. In short, the reduced GUTS model for
stochastic death was used, extended with a module to account for sa-
turation of the effect. This extension was needed as it was found that
increasing the exposure concentration (or the number of consecutive
pulses) led to less than the expected increase in the toxic effect. All
calculations were performed with the BYOM platform in Matlab (see
http://www.debtox.info/byom.html), maximising the likelihood func-
tion based on the multinomial distribution, corrected for animals
R.K. Bechmann, et al. Ecotoxicology and Environmental Safety 180 (2019) 473–482
475
sampled for histological analysis in the three-pulse experiment.
Confidence intervals were calculated by likelihood profiling.
2.8. Statistical analysis
All data analyses were performed in v 23 SPSS
®
(IBM, Chicago,
USA). The Kolmogorov-Smirnov test and Levens test was used to test for
normal distribution and variance. For normally distributed variables,
the General Linear Model (GLM) tests, two-way ANOVA was used to
test if significant the effect of Paramove
®
on biological parameters could
be found. The criterion for significance was set at p < 0.05. The
Wilcoxon Rank sum test was used to analyse the biological parameters
when normality or homogeneity criteria were not met. The criterion for
significance was set at p < 0.05.
3. Results
3.1. Chemical analyses
The semi-quantitative method showed that the concentration of
H
2
O
2
in the header tanks was stable for several days. The nominal
exposure concentration of H
2
O
2
was reached 30 min after starting the
pump delivering H
2
O
2
into the exposure tank. When the pump was
stopped 1 or 2 h later, the concentration was rapidly reduced
(Supplementary Figure 2).
3.2. Mortality, immobilization and swimming activity
There was no mortality of control shrimp in any of the experiments.
Mortality was significantly increased for shrimp exposed to 1.5 mg/L
and 15 mg/L H
2
O
2
(Wilcoxon, p < 0.05, significant difference from
control indicated by * in Fig. 1). In the three-pulse experiments, there
was no immediate effect on mortality after the first and second pulse of
exposure to 1.5 mg/L or 15 mg/L H
2
O
2
(Fig. 1, Exp. 1 and 2). However,
after the third pulse of exposure to 15 mg/L, 46% mortality was ob-
served, and two days later the mortality increased to 80%. Im-
mobilization of shrimp was observed two days earlier than mortality in
the three-pulse exposure to 15 mg/L (Supplementary Figure 3). Fur-
thermore, after the third pulse of exposure to 1.5 mg/L mortality started
to increase and mortality reached 33% after one day of recovery.
Delayed effect of PARAMOVE
®
was also observed in the one-pulse ex-
periment (Fig. 1, Exp. 3). There was no mortality in any treatment the
first two days after the pulse, but on the third day there was 38%
mortality in the 2-h 15 mg/L exposure, increasing to 50% later in the
recovery period (Fig. 1, Exp. 3). Although there was a tendency to a
dose- and time dependent increase in mortality also for the other
treatments in the one-pulse experiment, the difference from the control
was not statistically significant. Exposure to three pulses of 1.5 mg/L or
15 mg/L H
2
O
2
caused higher mortality than exposure to one pulse of
the same concentration (Wilcoxon, p < 0.05, Fig. 1). All the shrimp
exposed for 1 h or 2 h to 0.15 mg/L H
2
O
2
survived to the end of the
recovery period.
Swimming activity (i.e. percentage of shrimp that were swimming
at each observation point) was significantly increased during the 1 h
and 2 h exposure to 15 mg/L (significant difference from control
Fig. 1. Mortality of shrimp exposed to H
2
O
2
in the PARAMOVE
®
formulation (mean + SD). In Exp. 1 and 2 shrimp received 2 h exposure for 3 consecutive days. In
Exp. 3 shrimp were exposed once for 1 or 2 h. No mortality was observed in any of the controls in Exp. 1–3, nor in the 1 or 2 h exposure to 0.15 mg/L in Exp. 3.
Significant difference from control indicated by * (Wilcoxon, p < 0.05).
Fig. 2. Percent swimming P. borealis in 15 mg/L H
2
O
2
treatment compared to in
the control (white bars) before, during and after the 1 h (grey bars) or 2 h (black
bars) exposure in Exp. 3 (mean + SD). Significant difference from control in-
dicated by * (Wilcoxon, p < 0.05).
R.K. Bechmann, et al. Ecotoxicology and Environmental Safety 180 (2019) 473–482
476
indicated by * in Fig. 2, Wilcoxon, p < 0.05). The mean percentage of
swimming shrimp was 4–8% in the control compared to 29–38% during
exposure to 15 mg/L. The increased swimming activity lasted
throughout the pulse but went down to control level once the pulse had
stopped (Fig. 2). In the three-pulse experiments, no observations of
swimming activity were done during the pulse, and no significant dif-
ference in swimming activity was observed after each pulse
(F
1,10
= 3.35, p = 0.09, ANOVA).
3.3. Feeding rate
The feed intake of shrimp was significantly reduced after exposure
to 1.5 and 15 mg/L in Exp. 1 and Exp. 2 and remained so during the
recovery period (significant difference from control indicated by * in
Fig. 3, Wilcoxon, p < 0.05). After 12 days of recovery the feeding rate
was 66% lower for shrimp in the three-pulse 1.5 mg/L exposure than in
the control. Feed intake was documented for all 6 replicate control
tanks at all sampling days. At both sampling days in Exp. 1 and at 6 of
the 7 sampling days in Exp. 2 no feed intake was observed in five of the
six replicate tanks with shrimp exposed to H
2
O
2
. The feeding rate was
not consistently reduced in the recovery period after the one pulse
exposure in Exp. 3 (Supplementary Table 2).
3.4. Histopathology
Minor changes were observed in the gills of control shrimp (Figs. 4a
and 5a). Analysed individuals showed uniform arrangement of lamellae
with normal hemocoel and limited fusion of the lamellae (Fig. 4a).
Significant and progressive changes in the histoarchitecture of the gills
were observed in shrimp exposed to both 1.5 mg/L and 15 mg/L H
2
O
2
for 1 h followed by 8 days recovery in clean seawater (Wilcoxon, p-
values in Fig. 5). Swelling of gills, as well as the accumulation of he-
mocytes in the hemocoelic space of the gill lamellae, was the most re-
current observation in the 1.5 mg/L treatment (Figs. 4b and 5b). A si-
milar effect, but with more severe impairment was observed in the
15 mg/L treatment (Figs. 4c and 5c). Swelling and haemocyte infiltra-
tion phenomena induced the formation of diffuse disorganized tissues
masses in disrupted gills together with severe gill fusion and necrosis
phenomena, lifting of lamellar epithelium and hyperplasia in most in-
dividuals. The occurrence of parasites was generally low, and there was
no difference between treatments. Parasites were detected in two of the
15 shrimps from each treatment.
3.5. Histochemistry, lipid peroxidation in hepatopancreas
The 1 h H
2
O
2
exposure to both 1.5 mg/L and 15 mg/L induced
significant increase of the fluorescence intensity in the hepatopancreas
of treated shrimp compared to the control (+47% increment compared
to the control at 1.5 mg/L and +157% increment compared to the
control organisms at 15 mg/L, Kruskal–Wallis test, p < 0.05, Fig. 6).
Supplementary Figure 4 shows an example of the image analysis of
tissue damage in the hepatopancreas of shrimp.
3.6. Toxicokinetic-toxicodynamic modelling
The GUTS model fit is shown in Fig. 7, and the model parameters
are given in Supplementary Table 3. The model is able to explain all
data simultaneously with five parameters. In fact, one parameter
(background hazard rate) could have been excluded as there were no
deaths in the control over the entire duration of the tests. Note that for
this analysis, we treated the immobile animals (lying down) as dead.
The model cannot capture the slow onset of effects over the first days,
but overall provides a reasonable explanation for the overall pattern
across all treatments.
The estimated threshold for effects on mortality, as estimated using
the GUTS model, is extremely low at just 4.6 μg/L. This model para-
meter should be interpreted as the model prediction for the water
concentration that can be tolerated by the shrimp for an arbitrary
amount of time without any effect on survival. Clearly, for short ex-
posure pulses, higher concentrations can be tolerated, and the model
can be used to predict a threshold value for specific exposure scenarios.
Furthermore, the calibrated model can be used to predict an LC50 for
constant exposure concentrations, which facilitates comparison to more
standard laboratory tests. The LC50 depends on exposure time, and the
model predicts a 1-day LC50 of 2.7 (1.5–4.2) mg/L, a 2-day LC50 of
0.53 (0.32–0.88) mg/L and a 4-day LC50 of 0.14 (0.092–0.24) mg/L.
The threshold m
w
equals the incipient LC50.
4. Discussion
4.1. Mortality, immobilization, swimming activity and feeding rate
The effects documented in P. borealis exposed to one or three pulses
of 1.5 mg/L and 15 mg/L hydrogen peroxide in the present study de-
monstrate that P. borealis is more sensitive to H
2
O
2
in the PARAMOVE
®
formulation than the crustaceans tested by (Burridge et al., 2014) (see
Supplementary Table 4 for an overview of the main effects in P.
Fig. 3. Feeding rate for P. borealis exposed to 15 mg/L and 1.5mg/L H
2
O
2
. The shrimp were exposed for 2 h on day 1, 2 and 3 of each experiment. Significant
difference from control indicated by * (Wilcoxon, p < 0.05).
R.K. Bechmann, et al. Ecotoxicology and Environmental Safety 180 (2019) 473–482
477
borealis). These authors exposed three species of crustaceans to H
2
O
2
for
1 h followed by 95h recovery in clean seawater. The LC
50
was 973 mg/
L for Mysid sp., 3182 mg/L for Crangon septemspinosa and 1673 mg/L for
larvae of Homarus americanus. Crab larvae (Metacarcinus edwardsii) and
the shrimp Pandalus montagui were more sensitive to H
2
O
2
(Fagereng,
2016;Gebauer et al., 2017). High mortality of crab larvae was docu-
mented after multiple pulses (20 min per day for 7 days) of 188 mg/L
H
2
O
2
in the Hyperox
®
formulation (Gebauer et al., 2017), and the 6 h
LC50 for P. montagui was 170 mg/L H
2
O
2
(Fagereng, 2016). The co-
pepod Calanus finmarchicus also appeared to be more sensitive to H
2
O
2
than many other crustaceans (Escobar Lux, 2016;Hansen et al., 2017).
The 24 h LC
50
was 6 mg/L (Hansen et al., 2017) and increased mortality
of C. finmarchicus was observed after 1 h exposure 17–170 mg/L H
2
O
2
.
The 24 h LC
50
for P. borealis was recently determined to be 37 mg/L
H
2
O
2
(PARAMOVE
®
)(Refseth et al., 2016), but there are no literature
data on the effects of shorter H
2
O
2
exposure than 24 h for P. borealis. In
the present experiments, mortality did not occur during or right after
the first pulse of exposure, but 2–4 days later. However, in the three-
pulse experiment a higher percentage of immobilized shrimp were
observed early in the experiment, and increased mortality later in the
experiment. Immobilization of copepods has also been observed during
and after 1 h exposure to ≥10 mg/L H
2
O
2
(Interox
®
Paramove™50) (Van
Geest et al., 2014). Furthermore, reduced swimming activity of P.
montagui was observed after exposure to H
2
O
2
, and the 6 h EC
50
was
estimated to be 10 mg/L H
2
O
2
(Fagereng, 2016). Other studies may
have underestimated the effects of exposure to H
2
O
2
because delayed
effects were not investigated, i.e. the duration of these experiments was
too short (24 h or shorter) (e.g. (Brokke, 2015;Fagereng, 2016;Refseth
et al., 2016;Van Geest et al., 2014). The results show that standard
toxicity tests with constant exposure concentration and fixed (short)
duration may underestimate the risk to aquatic organisms as delayed
effects and carry-over toxicity due to multiple exposure pulses cannot
be detected (Ashauer et al., 2010).
During the one-pulse exposure to 15 mg/L H
2
O
2
there was a tran-
sient increase in swimming activity. This may indicate an escape re-
sponse or a stress response (Shumway et al., 1985), as the shrimp in
these treatments also showed increased mortality. The feeding rate of
shrimp exposed to three pulses of 1.5 mg/L was still significantly lower
than in the control 12 days after the last pulse. This indicates that the
shrimp were not able to recover from the exposure. Feeding inhibition
has also been documented for copepods after exposure for 1 h to
2.6–10 mg/L H
2
O
2
(Interox
®
Paramove™50) (Van Geest et al., 2014).
4.2. Gill damage
Ultra-structural changes in target organs of shrimps have been
shown to be useful tools to characterize both the health status of or-
ganisms in filed and laboratory exposure (Boudet et al., 2015;Sousa
et al., 2005). In this study, control shrimp showed a well-organized gill
structure, like that described for other species of shrimp (Bhavan and
Geraldine, 2000;Li et al., 2007;Wu et al., 2008), indicating good nu-
tritional status of individuals and normal function of the gill system.
Shrimp exposed to 1.5 mg/L of H
2
O
2
for 1 h showed clear evidence of
gill damage. Many adverse effects can be reversed by the immune
system reaction (hemocytes infiltration), tissue restructuration and
swelling of the gills. The exposure to a well-known oxidative stress
promoter such as hydrogen peroxide is expected to induce biochemical,
cellular and subcellular alterations in crustaceans, as already docu-
mented in other marine organisms (Kiemer and Black, 1997). Elevated
hydrogen peroxide concentrations increased the adverse effects, with
even more evident alterations in the gill tissue of shrimp exposed to
15 mg/L of H
2
O
2
for 1 h, indicating a dose-response relationship. Due to
the severe tissue destruction at this concentration, it is not likely that
the shrimp will recover, it is also likely that this is the cause of the
Fig. 4. Gills from control shrimp P. borealis
(a), shrimp exposed for 1 h to 1.5 mg/L (b)
or 15 mg/L H
2
O
2
(c) and sampled after 8
days recovery in clean seawater. Significant
and progressive structural alterations were
observed in the gill lamellae of shrimp ex-
posed to 1.5 mg/L and 15 mg/L. The accu-
mulation of hemocytes (HeM) in the he-
mocoelic space, swelling (SV) and lifting of
lamellar epithelium (Li), and marked hy-
pertrophy and hyperplasia in the gill epi-
thelium. Furthermore, necrotic (N), mod-
erate fusion of lamellae (FuS), clavate-
globate lamella and formation of a dis-
organized mass (DM) of disrupted gill la-
mellae are symptoms of necrosis and tissue
de-organisation phenomena.
R.K. Bechmann, et al. Ecotoxicology and Environmental Safety 180 (2019) 473–482
478
mortality in this treatment. For the 1 h 1.5 mg/L exposure to hydrogen
peroxide, the degree (and the pattern of alteration) indicates that very
little recovery and tissue restructuring is expected. This trend is further
supported by the results of the GUTS model pointing out an estimated
low damage-repair rate. However, in this context the ability of shrimp
to recover needs to be investigated in an experiment with even longer
recovery time than in the current experiment (i.e. more than 8 days).
Similar gill damage was observed for P. montagui exposed for 6 h to
3.5 mg/L H
2
O
2
(Fagereng, 2016).
4.3. Lipid peroxidation in hepatopancreas
Evidence of lipid peroxidation was observed in the hepatopancreas
already after 1 h exposure to 1.5 mg/L H
2
O
2
, and more evident in the
15 mg/L treatment. The analysis of the cryosections shows the in-
corporation of the peroxidation end-product in the lysosomes: this is a
well-known cell strategy to store unknown or potentially harmful
substances and make them less available in the cytosol. The analysis of
fluorescent marked peroxidation end-products (lipofuscins, carbony-
lated proteins etc.) shows a distribution indicating that lysosomes are
no longer able to store the large quantity of these oxidative stress de-
rived molecules in the study, and therefore they are more widespread in
the cytosol. A certain level of peroxidation is a naturally occurring ef-
fect of the biological ageing systems, and the end-products of such
Fig. 5. Gill histopathology in control shrimp (a) and shrimp exposed 1 h to 1.5 mg/L (b) or 15 mg/L H
2
O
2
(c) and sampled after 8 days recovery in clean seawater.
Score 0: no histopathology in any field, score 1: mild histopathology present in < 25% of the fields, class 2: moderate histopathology present in 25%–75% of the
fields, and score 3: severe histopathology present in > 75% of the field.
Fig. 6. Observed increment of membrane peroxidation in hepatopancreatic
cells of P. borealis exposed to 1.5 mg/L and 15 mg/L H
2
O
2
for 1 h followed by 8
days recovery in clean seawater. Average values as percentage increase with
respect to the control.
R.K. Bechmann, et al. Ecotoxicology and Environmental Safety 180 (2019) 473–482
479
process are accumulated over the lifespan of the organisms. As no ef-
ficient homeostatic mechanisms are present in the cells of organisms the
continuous accumulation of peroxided material is considered a per-
manent effect.
4.4. Toxicokinetic-toxicodynamic modelling
The GUTS model provided a very reasonable explanation for the
patterns of effects across all treatments but could not capture the slow
onset of effects over the first days. It is conceivable that the mechanism
of damage accrual and repair is more complex than simple one-com-
partment first-order kinetics. The estimated threshold for effects on
mortality (m
w
) is very low (4.6 μg/L), and within the range of reported
background concentrations for H
2
0
2
in sea water. This indicates that
even very low exposure concentrations (over a long period of time) may
have some additional effect on mortality. However, this prediction rests
on the assumption that the model is true and cannot be verified as we
have not tested the consequences of very low constant exposure con-
centrations. Mortality among the individuals in the tested cohorts
continues for a long time after the last exposure event, which is re-
flected in the rather low estimate for the dominant rate constant (k
d
).
This rate constant is the one-compartment approximation of a more
complex chain of damage processes, likely reflecting the slowest pro-
cess in the chain. This rate constant indicates that it takes 17 days for
the survivors of a pulse event to recover to 5% of the maximum damage
level. In other words, only if the interval between two pulses is at least
17 days, we could consider the two events to be ‘toxicologically in-
dependent’. For shorter intervals, carry-over toxicity (Ashauer et al.,
2010) would occur, which is reflected in the much stronger effects in
Exp. 1–2 with three pulses (separated by 1 day), compared to the effects
observed due to a single pulse. Interestingly, in a similar analysis by
Smit et al. (2008) for other crustaceans, a misfit of the initial response
was not observed, and there was also no indication of saturating effects.
However, those experiments were done with constant exposure. Smit
et al. report 4-day LC50s of 46 mg/L for Corophium volutator and
168 mg/L for Artemia salina, which are very high compared to the value
of 0.14 mg/L we could estimate from the GUTS parameters. Since Smit
and co-workers essentially applied the same model (apart from the
saturation constant C
K
), we can compare the model parameters between
the different species (Supplementary Table 3). The most striking dif-
ference is that the shrimp have, compared to the other crustaceans
tested, a much lower threshold and a much higher killing rate. This is a
clear indication that the shrimp are intrinsically more sensitive to H
2
O
2
in the PARAMOVE
®
formulation than some other crustaceans. Com-
paring our LC
50
-values to those reported for the copepod Calanus fin-
marchicus (Hansen et al., 2017), the 1-day LC50 for the shrimp is
roughly half that for the copepod. However, the 4-day LC50 is almost a
factor of 20 less. This indicates that the shrimp are more sensitive than
the copepods, but also that the dynamics of the response may be quite
different for both species. At the moment, we have no explanation for
these inter-species differences in response to H
2
O
2
; since species seem
to differ in more than one parameter of the TKTD model
(Supplementary Table 3), our analysis does not suggest a simple me-
chanistic hypothesis. However, it must be stressed that the LC50 only
considers the mortality at the end of the exposure period, and not the
delayed mortality that is observed after exposure has ended. The data in
Fig. 7 clearly show that one should not assess the effect of a 2 h pulse of
H
2
O
2
using a 2 h LC
50
value. The parameterised model can be used to
make predictions (with confidence interval) for untested exposure
profiles, or derive scenario-specific no-effect thresholds, which are
more meaningful exercises for risk-assessment purposes than focussing
on LC
50
s.
4.5. Possible consequences in the field
To estimate the potential consequences of discharging treatment
water containing H
2
O
2
on populations of P. borealis living in areas with
salmon farms, the dispersion and degradation of H
2
O
2
need to be
considered. Although H
2
O
2
degrades relatively rapidly to water and
oxygen, the half-life of H
2
O
2
has been reported to range from 1 day up
to 28 days (Bruno and Raynard, 1994;Fagereng, 2016;Haya et al.,
2005;Lyons et al., 2014) considerably longer than the time (1–2h)
needed to cause mortality and gill damage in P. borealis. All results
reported here are for exposures to the commercial formulation and,
while results are reported on the basis of hydrogen peroxide exposure
and concentration, the potential for formulation ingredients to play a
role cannot be discounted.
It is the transport and dilution of the treatment water in the hours
after the discharge that are the most relevant factors for estimating the
risk for P. borealis and other non-target crustaceans. There are few field
studies available where the concentration of H
2
O
2
has been measured
in surrounding water after bath treatment of salmonids. In a recent
study, more than 600 water samples taken at 5–60 m depth 0–25 min
after treatment with H
2
O
2
were analysed (Andersen and Hagen, 2016;
Fagereng, 2016). The raw data from this field study shows that the
concentration of H
2
O
2
was above 1.5 mg/L in 20% of the samples, and
that concentrations above 1.5 mg/L were detected in some samples
from all depths (5–60 m). High measured concentrations of H
2
O
2
(> 700 mg/L) in some samples from 60 m were most likely caused by
“plume sinking”that can happen when the water column is well mixed
in the period from October to May (Refseth et al., 2016). Most of the
measured values were, however, below 1.5 mg/L indicating that po-
tential exposure of P. borealis will depend on the transport and dilution
of water containing H
2
O
2
after the first 25 min, as studied by Fagereng
Fig. 7. Observed survival probability over time as a result of various pulse treatments (animals lying down were considered to be dead). Solid lines are the
simultaneous model fit on all data, and dotted lines connect symbols to the corresponding model curve. The right panel has been truncated for readability; ob-
servations in the control and the 3 × 2 h 1.5 mg/L continue till day 19 with only one additional death in the treatment.
R.K. Bechmann, et al. Ecotoxicology and Environmental Safety 180 (2019) 473–482
480
(2016).
Hydrodynamic dispersion modelling showed that release of chemi-
cals in an open location could lead to a dispersion of several kilometres
(Brokke, 2015). Release in a sheltered location usually led to a lower
dilution and a shorter dispersion distance. Some areas may contain
concentrations of 1–3% of the initial dosages used 24 h after release
(Brokke, 2015). Page et al. (2014) added a fluorescein dye into the
treatment volume along with the chemical to track the transport and
dispersal of the released chemicals following tarped net-pen or well-
boat application. Data from the dye experiment was combined with
modelling to estimate how far the chemicals were transported and the
dilution with time after release. Page et al. (2014) showed that releases
from tarp treatments are diluted by several orders of magnitude as they
are advected from their point of release. The dilution, however, may
take several hours to occur, thus increasing the opportunity for ex-
posure of non-target organisms (Page et al., 2014). More specifically,
the estimated dilution factor was 100–1000 after 3 h (i.e. 15–1.5 mg/L
H
2
O
2
) and the discharge could be transported several kilometres away,
depending on the water current speed (see (Page et al., 2014) for more
details). Dispersion modelling for an area on the west coast of Norway
estimated that concentrations as high as 10 mg/L H
2
O
2
could occur
several kilometres away from the treated farm up to 3 h after the dis-
charge (Refseth et al., 2016). Local hydrodynamic conditions will de-
termine how fast the concentration of H
2
O
2
will be diluted and how far
it will be transported horizontally and vertically. Results from disper-
sion modelling (Brokke, 2015;Page et al., 2014;Refseth et al., 2016)
and the current experiments indicate that treatment water with toxic
concentrations of H
2
O
2
(1.5 mg/L) could reach P. borealis living more
than 1 km from a treated salmon farm.
Pandalus borealis are found throughout the North Atlantic and
Pacific. They have a circumpolar distribution and are commercially
fished, not only in Norway but in the cold waters of the North Atlantic,
including the Bay of Fundy and Newfoundland in Canada (Bergström,
2000). Shrimp, in Norway, are often found in waters close to aqua-
culture activity (Olsen et al., 2012). If multiple farms are using the
PARAMOVE
®
product in the same area, or even in multiple cages at a
single site, there is a potential for negative consequences for shrimp
populations in the field.
Contributors
RKB designed the study, participated in the practical work in the lab
and is the main writer of this manuscript with the help of MA who
participated in the design, practical work, analyses of the data and
writing. SW was responsible for the exposure system including the
chemical measurements and input to writing. AG was responsible for
performing and writing about the histology. TJ was responsible for
GUTS modelling work including writing of the modelling sections. EL
and MB contributed to the daily practical work with the lab experi-
ments. TA and LB contributed to reading and commenting on the
manuscript. TA is WP leader in the ECOAST project #258833 COFASP
that co funded the experiments, LB is scientific advisor in the
PestPuls#267746/E40 project funded by the Research Council of
Norway, and RKB is project leader for PestPuls.
Declarations of interest
None.
Acknowledgements
We thank Frederike Keitel-Gröner for technical assistance in the lab.
Financial support was received from EU through the ECOAST project
#258833 COFASP, the Research Council of Norway through the
PestPuls #267746/E40 project, and from Solvay Interox. Ltd.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.ecoenv.2019.05.045.
References
Aaen, S.M., Helgesen, K.O., Bakke, M.J., Kaur, K., Horsberg, T.E., 2015. Drug resistance in
sea lice: A threat to salmonid aquaculture. Trends Parasitol. 31, 72–81. http://doi.
org/10.1016/j.pt.2014.12.006.
Andersen, P.A., Hagen, L., 2016. Fortynningsstudier - Hydrogenperoksid, September
2016. Aqua Kompetanse A7S, Flatanger, pp. 30. https://docplayer.me/34408749-
Fortynningsstudier-hydrogenperoksid-september-2016.html.
Ashauer, R., Hintermeister, A., Caravatti, I., Kretschmann, A., Escher, B.I., 2010.
Toxicokinetic and toxicodynamic modeling explains carry-over toxicity from ex-
posure to Diazinon by slow organism recovery. Enivron. Sci. Technol. 44. https://
pubs.acs.org/doi/abs/10.1021/es903478b.
Bergström, B.I., 2000. The biology of Pandalus. Adv. Mar. Biol. 38, 55–245. https://doi.
org/10.1016/S0065-2881(00)38003-8.
Bhavan, P.S., Geraldine, P., 2000. Histopathology of the hepatopancreas and gills of the
prawn Macrobrachium malcolmosonii exposed to endosulfan. Aquat. Toxicol. 50,
331–339. https://doi.org/10.1016/S0166-445X(00)00096-5.
Boudet, L.N.C., Polizzi, P., Romero, M.S., Robles, A., Marcovecchio, J.E., Gerpe, M.S.,
2015. Histopathological and biochemical evidence of hepatopancreatic toxicity
caused by cadmium in white shrimp, Palaemonetes argentinus. Ecotoxicol. Environ.
Saf. 113, 231–240. https://doi.org/10.1016/j.ecoenv.2014.11.019.
Brokke, K.E., 2015. Mortality caused by de-licing agents on the non-target organisms
chameleon shrimp (Praunus flexuosus) and grass prawns (Palaemon elegans). MSc
thesis., Department of Biology. University of Bergen, Norway, pp. 105.
Bruno, D.W., Raynard, R.S., 1994. Studies on the use of hydrogen peroxide as a method
for the control of sea lice on Atlantic salmon. Aquacult. Int. 2, 10–18. https://doi.org/
10.1007/bf00118529.
Burridge, L.E., Haya, K., Waddy, S.L., 2008. The effect of repeated exposure to aza-
methiphos on survival and spawning in the American lobster (Homarus americanus).
Ecotoxicol. Environ. Saf. 69, 411–415. https://doi.org/10.1016/j.ecoenv.2007.05.
001.
Burridge, L.E., Haya, K., Waddy, S.L., Wade, J., 2000. The lethality of anti-sea lice for-
mulations Salmosan®(Azamethiphos) and Excis®(Cypermethrin) to stage IV and
adult lobsters (Homarus americanus) during repeated short-term exposures.
Aquaculture 182, 27–35. https://doi.org/10.1016/s0044-8486(99)00251-3.
Burridge, L.E., Lyons, M.C., Wong, D.K.H., MacKeigan, K., VanGeest, J.L., 2014. The acute
lethality of three anti-sea lice formulations: AlphaMax®, Salmosan®, and
Interox®Paramove™50 to lobster and shrimp. Aquaculture 420–421, 180–186.
https://doi.org/10.1016/j.aquaculture.2013.10.041.
Cotran, R.S., Kumar, V., Robbins, S.L., 1989. Pathological Basis of Disease, 4th ed.
Saunders, Philadelphia.
Dennis, N., Tiede, K., Thompson, H., 2012. Repeated and Multiple Stress (Exposure to
Pesticides) on Aquatic Organisms. Food and Environment Research Agency (FERA),
Sand Hutton, York, UK, pp. 147. Supporting Publications 2012:EN-347. https://efsa.
onlinelibrary.wiley.com/doi/10.2903/sp.efsa.2012.EN-347.
Eraker, H., 2016. Advart Mot Rekedød Siden 1998. Norwegian Media Article (nrk.No).
https://www.nrk.no/dokumentar/xl/advart-mot-rekedod-siden-1998-1.13223172.
Escobar Lux, H.R., 2016. The effects of an anti-sea lice chemotherapeutant, hydrogen
peroxide, on mortality, escape response and oxygen consumption of Calanus spp.
Master thesis. Universite Pierre et Marie Curie and the Institute of Marine Research
42 Norway. (Key data/results from this MSc thesis is included in the open report
Refseth et al. 2016 refered to below).
Fagereng, M.B., 2016. Bruk Av Hydrogenperoksid i Oppdrettsanlegg; fortynningstudier
Og Effekter På blomsterreke (Pandalus montagui), Senter for farmasi. University of
Bergen, Norway, pp. 104. http://bora.uib.no/handle/1956/13008.
Felleskatalogen, 2017. Paramove. Solvay. Ektoparasittmiddel. ATCvet-Nr. QD08A X01.
https://www.felleskatalogen.no/medisin-vet/paramove-solvay-595323.
Gebauer, P., Paschke, K., Vera, C., Toro, J.E., Pardo, M., Urbina, M., 2017. Lethal and sub-
lethal effects of commonly used anti-sea lice formulations on non-target crab
Metacarcinus edwardsii larvae. Chemosphere 185, 1019–1029. https://doi.org/10.
1016/j.chemosphere.2017.07.108.
Gomiero, A., Viarengo, A., 2014. Effects of elevated temperature on the toxicity of copper
and oxytetracycline in the marine model, Euplotes crassus: A climate change per-
spective. Environ. Pollut. 194, 262–271. https://doi.org/10.1016/j.envpol.2014.07.
035.
Hansen, B.H., Hallmann, A., Altin, D., Jenssen, B.M., Ciesielski, T.M., 2017. Acute hy-
drogen peroxide (H
2
O
2
) exposure does not cause oxidative stress in late-copepodite
stage of Calanus finmarchicus. J. Toxicol. Environ. Health A Curr. Issues 80, 820–829.
https://doi.org/10.1080/15287394.2017.1352182.
Haugland, B.T., Rastrick, S.P.S., Agnalt, A.L., Husa, V., Kutti, T., Samuelsen, O.B., 2019.
Mortality and reduced photosynthetic performance in sugar kelp Saccharina latissima
caused by the salmon-lice therapeutant hydrogen peroxide. Aquacult. Environ.
Interact. 11, 1–17. https://doi.org/10.3354/aei00292.
Haya, K., Burridge, L.E., Davies, I.M., Ervik, A., 2005. A review and assessment of en-
vironmental risk of chemicals used for the treatment of sea lice infestations of cul-
tured salmon. In: B.T.,H. (Ed.), Environmental Effects of Marine Finfish Aquaculture.
Handbook of Environmental Chemistry. Springer, Berlin, Heidelberg. https://doi.
org/10.1007/b136016.
Helgesen, K.O., Romstad, H., Aaen, S.M., Horsberg, T.E., 2015. First report of reduced
R.K. Bechmann, et al. Ecotoxicology and Environmental Safety 180 (2019) 473–482
481
sensitivity towards hydrogen peroxide found in the salmon louse Lepeophtheirus sal-
monis in Norway. Aquaculture Reports 1, 37–42. https://doi.org/10.1016/j.aqrep.
2015.01.001.
Jager, T., Albert, C., Preuss, T.G., Ashauer, R., 2011. General unified threshold model of
survival - a toxicokinetic-toxicodynamic framework for ecotoxicology. Enivron. Sci.
Technol. 45, 2529–2540. https://doi.org/10.1021/es103092a.
Kiemer, M.C.B., Black, K.D., 1997. The effects of hydrogen peroxide on the gill tissues of
Atlantic salmon. Salmo salar L. Aquaculture 153, 181–189. https://doi.org/10.1016/
S0044-8486(97)00037-9.
Langford, K.H., Oxnevad, S., Schoyen, M., Thomas, K.V., 2014. Do antiparasitic medicines
used in aquaculture pose a risk to the Norwegian aquatic environment? Enivron. Sci.
Technol. 48, 7774–7780. https://doi.org/10.1021/es5005329.
Li, N., Zhao, Y.L., Yang, J., 2007. Impact of waterborne copper on the structure of gills
and hepatopancreas and its impact on the content of metallothionein in juvenile giant
freshwater prawn Macrobrachium rosenbergii (Crustacea : Decapoda). Arch. Environ.
Contam. Toxicol. 52, 73–79. https://doi.org/10.1007/s00244-005-0214-5.
Lillicrap, A., Macken, A., Thomas, K.V., 2015. Recommendations for the inclusion of
targeted testing to improve the regulatory environmental risk assessment of veter-
inary medicines used in aquaculture. Environ. Int. 85, 1. https://doi.org/10.1016/j.
envint.2015.07.019.
Lyons, M.C., Wong, D.K.H., Page, F.H., 2014. Degradation of hydrogen peroxide in sea-
water using the anti-sea louse formulation Interox®Paramove™50. Can. Tech. Rep.
Fish. Aquat. Sci. 3080, 15. http://publications.gc.ca/site/eng/463870/publication.
html.
Nærings- og fiskeridepartementet, 2015. Meld. St. 16 (2014-2015). Melding til Stortinget.
Forutsigbar og miljømessig bærekraftig vekst i norsk lakse- og ørretoppdrett. https://
www.regjeringen.no/no/dokumenter/meld.-st.-16-2014-2015/id2401865/.
Olsen, S.A., Ervik, A., Grahl-Nielsen, O., 2012. Tracing fish farm waste in the northern
shrimp Pandalus borealis (Kroyer, 1838) using lipid biomarkers. Aquacult. Environ.
Interact. 2, 133–144. https://doi.org/10.3354/aei00036.
Ouraji, H., Fereidoni, A.E., Shayegan, M., Asil, S.M., 2011. Comparison of fatty acid
composition between farmed and wild Indian white shrimps, Fenneropenaeus indicus.
Food Nutr. Sci. 824–829. https://doi.org/10.4236/fns.2011.28113.
Page, C., Chang, B.D., Beattie, M., Losier, R., McCurdy, P., Bakker, J., Haughn, K., Thorpe,
B., Fife, J., Scouten, S., G.,B., Ernst, B., 2014. Transport and Dispersal of Sea lice bath
therapeutants from Salmon farm Net-Pens and Well-boats Operated in Southwest
NewBrunswick: a Mid-project Perspective and Perspective for Discussion. Doc. 2014/
102, DFO Can. Sci. Advis. Sec. Res.. http://publications.gc.ca/collections/collection_
2015/mpo-dfo/Fs70-5-2014-102-eng.pdf.
Refseth, G.H., Sæther, K., Drivdal, M., Nøst, O.A., Augustine, S., Camus, L., Tassara, L.,
Agnalt, A.-L., Samuelsen, O.B., 2016. Miljørisiko ved bruk Av Hydrogenperoksid.
Økotoksikologisk vurdering Og Grenseverdi for Effekt. https://www.fhf.no/
prosjekter/prosjektbasen/901249/.
Shumway, S.E., Perkins, H.C., Schick, D.F., Stickney, A.P., 1985. Synopsis of biological
Data of the Pink Shrimp Pandalus borealis (Krøyer, 1838). FAO Fisheries Synopsis No.
144; NOAA Technical Report of the National Marine Fisheries Service. NOAA, pp. 57.
http://www.fao.org/3/a-ap949e.pdf.
Smit, M.G.D., Ebbens, E., Jak, R.G., Huijbregts, M.A.J., 2008. Time and concentration
dependency in the potentially affected fraction of species: The case of hydrogen
peroxide treatment of ballast water. Environ. Toxicol. Chem. 27, 746–753. https://
doi.org/10.1897/07-343.1.
Sousa, L.G., Cuartas, E.I., Petriella, A.M., 2005. Fine structural analysis of the epithelial
cells in the hepatopancreas of Palaemonetes argentinus (Crustacea, Decapoda, Caridea)
in intermoult. Biocell 29, 25–31.
Steinhold, M., Thonhagen, M., 2017. Ingen vet Hva Som Har Skjedd Med Alle Rekene –
Nå forsøker forskere å løse Mysteriet. Norwegian media article published on nrk.no:
https://www.nrk.no/nordland/ingen-vet-hva-som-har-skjedd-med-alle-rekene-_-na-
forsoker-forskere-a-lose-mysteriet-1.13750961.
Urbina, M.A., Cumillaf, J.P., Paschke, K., Gebauer, P., 2019. Effects of pharmaceuticals
used to treat salmon lice on non-target species: Evidence from a systematic review.
Sci. Total Environ. 649, 1124–1136. https://doi.org/10.1016/j.scitotenv.2018.08.
334.
Van Geest, J.L., Burridge, L.E., Fife, F.J., Kidd, K.A., 2014. Feeding response in marine
copepods as a measure of acute toxicity of four anti-sea lice pesticides. Mar. Environ.
Res. 101, 145–152. https://doi.org/10.1016/j.marenvres.2014.09.011.
Wu, J.P., Chen, H.C., Huang, D.J., 2008. Histopathological and biochemical evidence of
hepatopancreatic toxicity caused by cadmium and zinc in the white shrimp,
Litopenaeus vannamei. Chemosphere 7.
Zodrow, J.M., Stegeman, J.J., Tanguay, R.L., 2004. Histological analysis of acute toxicity
of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in zebrafish. Aquat. Toxicol. 66,
25–38. https://doi.org/10.1016/j.aquatox.2003.07.002.
R.K. Bechmann, et al. Ecotoxicology and Environmental Safety 180 (2019) 473–482
482
