Pre-anaesthetic metomidate sedation delays the stressresponse after caudal artery cannulation in Atlantic cod(Gadus morhua)

Anders Karlsson • Bjørn Olav Rosseland •

Jean-Charles Massabuau • Anders Kiessling

Received: 21 May 2010 / Accepted: 19 May 2011 / Published online: 3 June 2011

� Springer Science+Business Media B.V. 2011

Abstract Recovery from caudal artery cannulation

with and without pre-anaesthesia metomidate sedation

was assessed in Atlantic cod (Gadus morhua). The

levels of plasma cortisol, glucose, electrolytes and

acid–base parameters were compared between sedated

and unsedated cod and to those in uncannulated

individuals, where the samples were obtained by

sacrificial sampling (reference level). Metomidate

sedation delayed the stress response, causing sedated

cod plasma cortisol to return to the reference level more

slowly [day 4 post surgery (PS)] than in unsedated cod

(day 2 PS). Plasma glucose was elevated in both

sedated and unsedated cod up to and including day 5

PS. Plasma K? was lower and pH was higher in

cannulated cod than in the reference from 24 h PS until

the end of experimentation, indicating a stress effect of

sacrificial sampling on plasma K? and pH that was

likely caused by an acute stress response. Metomidate

sedation delayed the stress response following CA

cannulation and should therefore not be used as a pre-

anaesthetic sedation in Atlantic cod. The caudal artery

cannulation can be a useful tool in obtaining repeated

blood samples from Atlantic cod given an adequate

recovery time, which was determined to be 6 days

irrespective of pre-anaesthesia sedation status.

Keywords Metacain � Acute stress response � Fish �Gadoid � Blood physiology � Cortisol � Glucose

Introduction

Cannulation is the process of catheterising a blood

vessel and is usually performed in experimental

animals to draw blood samples, inject substances

directly into the blood stream or measure blood

pressure without handling the animal. The cannula-

tion techniques have proved to be valuable research

tools in physiological experiments with fish for

almost half a century (Conte et al. 1963). The various

cannulation methods allow repeated sampling of

blood from the same individual, providing an

A. Karlsson

Department of Plant and Environmental Sciences,

Norwegian University of Life Sciences, P.O. Box 5003,

1432 As, Norway

B. O. Rosseland

Department of Ecology and Natural Resource

Management, Norwegian University of Life Sciences,

P.O. Box 5003, 1432 As, Norway

J.-C. Massabuau

Station Marine, Universite Bordeaux 1, CNRS, UMR

5805 EPOC, Place du Dr Peyneau, 33120 Arcachon,

France

Present Address:A. Karlsson (&) � A. Kiessling

Aquaculture Protein Centre, Department of Animal

and Aquacultural Sciences, Norwegian University of Life

Sciences, P.O. Box 5003, 1432 As, Norway

e-mail: anders.karlsson@umb.no

123

Fish Physiol Biochem (2012) 38:401–411

DOI 10.1007/s10695-011-9516-x

alternative to sacrificial sampling when studying

changes in blood variables. Several blood vessels

have been targeted, e.g., the dorsal aorta (DA) (Conte

et al. 1963; Djordjevic et al. 2011; Kiessling et al.

2003; Smith and Bell 1964; Soivio et al. 1975; Sunde

et al. 2003), the ventral aorta (Axelsson et al. 1994),

the hepatic portal vein (Eliason et al. 2007; McLean

and Ash 1989) and the caudal artery (CA) (Forgue

et al. 1989; Karlsson et al. 2011) with duration of

experimental period varying from hours to several

weeks. Using cannulated fish in a repeated sampling

setup allows a reduction in the number of experi-

mental animals as all treatments can be applied to all

animals, thereby eliminating the need for one group

of animals per treatment. It can also reduce exper-

imental variability as it allows internal paired com-

parison as reference; control, test and recovery data

can be measured in the same individual. A final

advantage over sacrificial sampling is reduced vari-

ability due to sampling stress and handling; the fish

can be practically undisturbed by the sampling

procedure and experimental setup (Djordjevic et al.

2011). Therefore, when studying fish blood parame-

ters that may be affected by gill ventilation or

immediate stress responses, e.g., plasma pH, HCO3-

or partial pressures of respiratory gases, using

cannulated individuals and repeated sampling is a

necessity.

The DA cannulation, as it is described by Soivio

et al. (1975), has been one of the most frequently

used cannulation methods, possibly because of its

comparatively uncomplicated surgical procedure.

The Atlantic cod (Gadus morhua) has an anatomy

that complicates cannulation of the DA through the

roof of the mouth. The gill efferent blood vessels

merge to form the DA very close to the pharyngeal

sphincter (visual inspection by dissection). A catheter

piercing the skin in or close to the pharyngeal

sphincter would likely severely disturb the cod and

may cause stress or suffering. This would undermine

the justification for cannulation and produce unreli-

able and possibly ambiguous results.

Most previous studies involving cannulation of

Atlantic cod report occlusive cannulation of a gill

afferent and efferent blood vessel, effectively remov-

ing one gill arch (Axelsson and Nilsson 1986; Perry

et al. 1991; Smith et al. 1985). The occlusive and

invasive nature of these procedures may cause ambig-

uous results, especially so in long-term studies of

weeks or months, where tissue necrosis may be a

significant concern. An alternative artery to cannulate

in the Atlantic cod is the CA. The CA cannulation was

first described by Forgue et al. (1989) in Wels catfish

(Silurus glanis), using open surgery, and later, using

non-invasive methods, in other teleosts (Forsman et al.

2005; Karlsson et al. 2011). The present study is the

first evaluation and detailed description of a non-

invasive CA cannulation technique for use on gadoid

fishes.

Metomidate has previously been shown to prevent

cortisol release in Atlantic salmon (Olsen et al. 1995)

and has been used as a sedative prior to anaesthesia

and subsequent cannulation in several studies by

Kiessling and co-workers throughout the past decade

(Djordjevic et al. 2011; Kiessling et al. 2006, 2003;

Kristensen et al. 2010; Sunde et al. 2003). The pre-

anaesthesia sedation, in combination with other

improvements, has reduced the general stress level

in salmonids after DA cannulation to such an extent

that most blood variables stabilize within 1–3 h post

surgery (PS), and all parameters measured were

stable 24–72 h PS (Djordjevic et al. 2011). Atlantic

cod pre-anaesthesia sedated with metomidate was

recently reported to require a lower anaesthetic

concentration and have a faster recovery time from

anaesthesia than unsedated congeners (Zahl et al.

2009). However, it is not presently known whether

the beneficial effects of metomidate sedation

observed in cannulated salmonids and Atlantic cod

during recovery from anaesthesia are transferrable to

a quicker recovery from cannulation in Atlantic cod.

The objectives of this study were to evaluate the

non-invasive CA cannulation technique in terms of

response magnitude and recovery time of blood

variables and to evaluate the use of metomidate

sedation prior to anaesthesia and CA cannulation of

Atlantic cod.

Materials and methods

Animals and experimental procedure

Atlantic cod, average weight 700 g, were purchased

from a commercial cod farm (Profunda AS; Bardstad-

vik, Norway) in early February 2009 and transported

by truck to the Norwegian Institute for Water Research

(NIVA)—Marine Research Station at Solbergstrand

402 Fish Physiol Biochem (2012) 38:401–411

123

(NIVA-MFS), Norway. No cod died during or after

transportation, and the fish were in good health upon

arrival (veterinary approval). Finally, no signs of ill

health were observed prior to, or during, the experi-

ment. The cod were housed in a circular, 6-m diameter

holding tank at a density of up to 2.5 kg/M3, fed a

commercial cod diet (BioMar AS, Norway) to satia-

tion 5 days a week and visually inspected at least once

daily. All cod were kept in the holding tank for a

minimum of 5 weeks to acclimate to their new

surroundings before being transferred to the experi-

mental unit.

Cod were cannulated (as described below) and

recovery from cannulation followed in three trials. In

all three trials, the cod were transferred from the

holding tank to the experiment tanks (square 1 m

tanks, 500 L water volume) and given 7 days to

acclimate to the new environment in groups of 4–5

individuals per tank. The tanks were purposely

designed to hold individual, cannulated fish and

create a low stress environment as described in detail

by Djordjevic et al. (2011). At least 24 h prior to, and

throughout acclimation and experimentation, cod

were deprived of food. In trial 1, (Mar–Apr 2009)

all cod were sedated with metomidate in their tanks

(sedated) prior to anaesthesia with metacain in a

separate bath before cannulation. In trials 2 (Sep–Oct

2009) and 3 (Nov–Dec 2009), cod were either

sedated with metomidate in their tanks (sedated) or

left undisturbed (unsedated) prior to metacain anaes-

thesia and cannulation.

To follow recovery from cannulation, 22 cod,

average weight 1041 ± 288 g and length 44 ± 3 cm

[± standard deviation (SD)], were sampled for blood

through the CA cannula at the following times:

directly PS (0 h), one hour PS (1 h), three hours PS

(3 h), 24 h PS (24 h), daily after 24 h until day 7 PS

(Day 2–7) and at days 9, 11 and 14 PS. To establish a

comparison/reference level, 7 Atlantic cod, average

weight 734 ± 75 g and length 41 ± 2 cm (±SD),

from the same population as the experimental fish,

were placed in the tanks described above (1 cod per

tank). The fish were left undisturbed for 14 days and

then sacrificially sampled for blood from the caudal

artery/vein using aspiration with a disposable syringe

and needle.

Upon finishing the experiment, all cod with

cannulas still attached (20 of 22 fish, 2 fish lost their

cannulas during the experiment) were dissected to

determine the placement of the cannula. Sixteen cod

were cannulated in the CA, three cod were cannulated

in the caudal vein (CV) and cannula placement could

not be determined in one cod. Due to the possible

differences in pH, PCO2 and HCO3- between arterial

and venous blood, results from individuals cannulat-

ed in the CV or with unknown cannula placement (6

cod in total) were removed from statistical analysis of

pH, PCO2 and HCO3- data.

Throughout the experiments, full strength seawater

from a depth of 60 m and with a temperature of

7.2–10.6�C and a salinity of 31.8–34.5 parts per

thousand (ppt) was used. The water flow to the tanks

was kept at 2 L/min. The flow through system

secured good water quality throughout. Typical water

qualities in the facility include: oxygen tension

18.1 kPa; pH 7.84; alkalinity 2.41 mmol/L; turbidity

0.36 FNU; total nitrogen 173 lg/L; ammonium

13 lg/L; dissolved carbon dioxide 0.44 mg/L.

Surgical procedure

The catheters were made from PE50 polyethylene

tubing (Intramedic�; Becton–Dickinson, New Jer-

sey, USA), and trocars/cannulas were made from

PL013 steel guitar wire. The PE50 was heated and

then stretched and narrowed at the penetrating end,

and a bubble was made by precision heating *5 cm

up the cannula; so that the catheter could be attached

to a suture without sliding. The catheter was then

flushed with heparinised physiological saline (NaCl,

9 g/L; Na-heparin, 150 IU/mL, injection quality)

(saline) and cut at the narrowing end to fit tightly

around the trocar. Finally, two small holes were made

on the catheter tip (one on each side) to prevent the

catheter from suctioning onto the vessel wall. The

catheters were stored immersed in 70% ethanol

without the trocar. Immediately prior to surgery, the

trocar was prepared by cutting it at an angle as low as

possible with a pair of wire cutters to produce a sharp

point with a cutting edge. Finally, the trocar was

inserted into the catheter so that only the cutting edge

and point was protruding from the PE50 tubing, and

the catheter was immersed in antiseptic fluid (Chlorh-

exidine, 0.5 g/L; Fresenius Kabi, Uppsala, Sweden).

Each fish was either sedated with metomidate

(0.5 mg/L) until it stopped responding to visual

stimuli, or not disturbed before it was transferred

from its’ respective tank to an anaesthetic bath

Fish Physiol Biochem (2012) 38:401–411 403

123

containing *30 L of aerated seawater and 0.08 g/L

metacain (Norwegian Medical Depot, Norway).

When the fish no longer responded to touch, weight

and length were measured before it was placed

upside-down in a purposely designed surgical cradle

and surrounding bath and covered with a wet cloth.

Throughout surgery, the gills were ventilated with

aerated seawater maintained at equal temperature as

in the tank (8–9�C) at *15 L/min containing a

maintenance dose of metacain (0.04 g/L).

Using a disposable insulin syringe with needle,

0.3–0.7 mL of lidocaine analgesia was injected at the

point of incision and suture placement through the

skin [10 or 20 g/L, with adrenaline (5 mg/L)]. An

#11-blade and a #3-scalpel was used to make a

horizontal incision (*5 mm) through the skin, ver-

tically positioned at approximately the same distance

from the anal fin and lateral line, and horizontally

approximately half way down the length of the anal

fin.

The ‘closed’ cannulation was performed by insert-

ing a pre-made catheter with trocar into the incision,

at approximately 45� horizontally, and pushed

through muscle and membrane tissue into the CA.

The trocar was then retracted, and the catheter was

securely placed inside the vessel by pushing it

2–3 cm into the vessel. To prevent clotting, the

catheter was filled with saline. The catheter was

secured to the fish using a single stitch of sterile, non-

absorbable suture (Supramid�, 3-0 USP; AgnTho’s

AB, Sweden) directly behind the insertion point of

the catheter. The catheter end was then melted shut,

and the surgery wound covered with Stomahesive�

paste (ConvaTec Norway AS, Oslo, Norway).

Finally, the fish received an injection of Oxytetracy-

cline antibiotic (100 g/L; Ceva Sante Animale,

Libourne, France) into the abdominal cavity to

prevent variation in condition caused by accidental

bacterial infection.

Sample collection, preparation and analysis

Prior to sampling blood from the CA catheter,

previously injected saline and a few drops of fresh

blood were discarded to ensure a pristine sample. By

light suction using a 1-mL disposable syringe with

blunted tip inserted into the catheter, 0.2–0.35 mL

blood samples were collected. The amount of sample

extracted was volume adjusted based on visual

observation of haematocrit, in order to obtain enough

plasma and draw as little blood as possible. After the

required amount of blood was sampled, saline was

injected into the catheter until blood was no longer

visible. To prevent clotting, another 0.1–0.2 mL of

saline was injected to ensure no blood was left in

catheter tip. The catheter end was sealed by melting

after each sampling.

Blood samples were immediately analysed for

glucose, PCO2, pH, HCO3- and ions (Na?, K?, Cl-)

using an i-STAT� Portable Clinical Analyzer (Med-

inor AS, Norway). Results for pH, PCO2 and HCO3-

were temperature corrected using formulas supplied

by the i-STAT� manufacturer (Abbott Point of Care

Inc.; Princeton, NJ, USA).

Blood samples were immediately centrifuged for

3 min at 20009g to separate red blood cells and

plasma. Plasma was immediately frozen at -20�C

and transferred to -80�C within 3 days. Plasma

samples (50 lL) were mixed with five times the

sample volume (250 lL) of ethyl acetate using a

vortex mixer. The mix was centrifuged for 2 min at

71559g in 4�C, and the resulting supernatant stored

at -80�C until analysis. The supernatant was later

analysed for cortisol using a Radioactive Immuno

Assay (RIA) kit (Spectria� Cortisol RIA; Orion

Diagnostica AS, Asker, Norway) and a NaI-gamma

counter (Wizard�; PerkinElmer Norge AS, Oslo,

Norway) according to instructions in the RIA kit

booklet. Samples determined to be below the detec-

tion limit of the RIA kit (5 ng/mL) were set to be

5 ng/mL before further analysis of the data.

Data and statistical analysis

Plasma Cl- was above the detection limit of the

instrument (140 mmol/L) in all samples and was

consequently excluded from further analysis. Two

samples were removed from the statistical analysis

for all blood variables due to extreme observations of

plasma cortisol; 267 and 315 ng/mL whilst no other

observations were [140 ng/mL. These samples orig-

inated from the same individual at 24 and 48 h PS.

All other samples from the same individual were

within normal range and were included in the

statistical analysis.

The data were analysed with SAS v. 9.13 (Statis-

tical Analysis Software), using the mixed procedure

for repeated measurements with a heterogeneous and

404 Fish Physiol Biochem (2012) 38:401–411

123

autoregressive covariance structure. Pre-anaesthetic

sedation status, recovery time and the interaction

between pre-anaesthetic sedation status and recovery

time were tested as class variables in the model.

Individual was included as a random variable (subject

identification). No effect of pre-anaesthetic sedation

status (P = 0.6255) or interaction between recovery

time and pre-anaesthetic sedation status (P = 0.2085)

was found for plasma K?. Consequently, the two pre-

anaesthetic sedation statuses were considered to be the

same for statistical analysis of plasma K?.

An F test was used to determine statistical

significance of fixed effects. A t test with Tukey–

Kramer adjustment for multiple comparisons was

used to determine statistical differences between the

different sampling times and between sedation

statuses within sampling time. A t test with Dunnett

adjustment for multiple comparisons was used to

determine statistical difference between the reference

level and the different sampling times. Comparisons

yielding P values\0.05 after adjustment for multiple

comparisons were considered to be statistically

different. Unless otherwise stated, P \ 0.05 in all

comparisons where differences are indicated. All

values are presented as least squares mean (LS

mean) ± standard error (SE) unless stated otherwise.

Results

Plasma cortisol and glucose

Plasma cortisol and glucose were affected by recovery

time (P \ 0.0001) and the interaction between recov-

ery time and sedation status (P \ 0.05), but no effect

was found for sedation status alone (Table 1). Both

plasma cortisol and glucose levels increased initially,

peaked (at ca 90 ng/mL for cortisol and 9 mmol/L for

glucose), and then decreased with time PS irrespective

of sedation status (Fig. 1a, b). However, the unsedated

cod displayed peak plasma cortisol and glucose levels

earlier than did the sedated cod, 1–3 vs. 3–24 h PS for

cortisol and 1 vs. 24 h PS for glucose. In unsedated

cod, plasma cortisol was higher at 1 h and at 5 days

PS, and plasma glucose was lower at 24 h PS, than in

sedated cod. The sedated cod plasma cortisol levels

were elevated, compared to the reference level

(10.6 ± 9.3 ng/mL), from 1 h up to and including

72 h PS. In unsedated cod, plasma cortisol levels were

elevated compared to the reference level from directly

PS (0 h) up to and including 24 h PS, and at day 5 PS.

In both unsedated and sedated cod, plasma glu-

cose levels were elevated compared to the reference

level (2.38 ± 1.08 mmol/L), from 0 h (unsedated)

and 1 h (sedated) up to and including day 5 PS

(Fig. 1a, b).

Acid–base parameters

Plasma PCO2, HCO3- and pH were affected by

recovery time (P \ 0.0001), sedation status

(P \ 0.01) and their interaction (P \ 0.05) (Table 1).

No clear pattern could be seen for development of

plasma PCO2 PS. In sedated cod, plasma PCO2 was

lower than the reference level (0.52 ± 0.05 kPa) at

days 1, 2, 7 and 11 PS, whilst it never differed from

the reference level in unsedated cod. The sedated cod

plasma PCO2 levels were lower than in unsedated

cod at 1, 2, 7, 11 and 14 PS (Fig. 2a). In both sedated

and unsedated cod, plasma HCO3- increased from ca

4 mmol/L directly PS (not different from the refer-

ence level at 4.0 ± 0.5 mmol/L) to ca 6.5 mmol/L at

1 h PS. In unsedated cod, the plasma HCO3- did not

change from 1 h PS onwards and was higher than the

reference level until the end of experimentation.

However, in sedated cod, the plasma HCO3- dropped

after 3 h PS and was only higher than the reference

level at days 3, 4 and 9 PS thereafter. The sedated cod

plasma HCO3- concentration was lower than in

unsedated cod at days 1, 2, 7, 11 and 14 PS (Fig. 2b).

Table 1 P values for tests of fixed effects on blood variables

Cortisol Glucose Na? K? pH PCO2 HCO3-

Time post surgery \0.0001 \0.0001 0.0002 \0.0001 \0.0001 \0.0001 \0.0001

Sedation status 0.2826 0.7882 0.0465 0.6255 0.0035 0.0019 0.0076

Time 9 sedation 0.0015 0.0424 0.0366 0.2085 0.0127 \0.0001 0.0008

Fish Physiol Biochem (2012) 38:401–411 405

123

In both sedated and unsedated cod, the plasma pH

levels increased from near and not different from

reference (7.59 ± 0.04) levels directly PS (7.55 ±

0.02 in sedated and 7.67 ± 0.04 unsedated cod) to

stable levels at ca 7.9–8.0 in sedated and ca 7.8–7.9 in

unsedated cod in 3–24 h PS. In sedated cod, plasma

pH levels were higher than in unsedated cod at 0 and

3 h and at days 1, 2, 7, 11 and 14 PS. Plasma pH was

higher than the reference level from 1 h to day 14 PS

in sedated cod and from 3 h to day 14 PS in

unsedated cod (Fig. 2c).

Plasma ions

The plasma Na? concentration was affected by

recovery time (P = 0.0002), sedation status and their

interaction, whilst plasma K? was affected by recov-

ery time only (Table 1). In both sedated and unsedated

cod, plasma Na? levels were higher than the refer-

ence level (150 ± 1 mmol/L) the first 24 h PS

(156–161 mmol/L). Between day 1 and 2 PS, both

sedated and unsedated cod plasma Na? levels dropped

to stable levels, neither of which were different from

0

2

4

6

8

10

12

14

0 48 96 144 192 240 288 336

Glu

cose

con

cent

ratio

n (m

mol

/L)

Time post surgery (hours)

Sedated

Unsedated

Refeff rence

B*###

a*##

BC##BC##

BCD#BCD#

DEDE

AAA

abaa ## ab#abaa #

ab#abaa abc

abcaac

b

0

20

40

60

80

100

0 48 96 144 192 240 288 336

Plas

ma

cort

isol

con

cent

ratio

n (n

g/m

/L)

Sedated

Unsedated

Refeff rence

a

acac##

ab**#

DE**

BC###

ABC##

ABD#bbc bc

AD

ABE

DE DEAE AE

bcbcbbbc

bb

bc

0

20

40

60

80

100

0 1 2 3 4

AD

ab##

a*### a###

C###

B*###

02468101214

0 1 2 3 4

ABE#

aabaa c#abaa c###

abaa ##

ABE#ACE

b

Fig. 1 Arterial blood plasma cortisol (a) and glucose (b) con-

centrations after CA cannulation surgery in Atlantic cod that

were either sedated with metomidate or not sedated prior to

anaesthesia and subsequent cannulation. A solid line represents

a reference level that consisted of cod sacrificially sampled

from the caudal vessels. The boxes display the development in

the first 4 h post surgery (PS). The results are presented as LS

means with S.E. bars. * or **Significant difference between

the sedation statuses at the respective sampling time at

P \ 0.05 or P \ 0.01, respectively. #, ## or ###Significant

difference between the relevant sampling time/sedation status

and the reference level at P \ 0.05, P \ 0.01 or P \ 0.001,

respectively. Different upper case letters indicate significant

differences between sampling times PS in sedated cod and

different lower case letters indicate significant differences

between sampling times PS in unsedated cod

406 Fish Physiol Biochem (2012) 38:401–411

123

the reference level. However, the sedated cod Na?

level (149–153 mmol/L) was lower than that of

unsedated cod (153–155 mmol/L) at days 3–7 PS

(Fig. 3a). In cannulated cod, the plasma K? concen-

tration dropped from 3.76 ± 0.22 to 2.62 ± 0.05

mmol/L in the first 24 h PS and was then lower than

in the reference cod (3.40 ± 0.13 mmol/L) until the

end of the experiment. From day 2 PS onwards, the

plasma K? concentration in cannulated cod did not

change throughout the experiment (Fig. 3b).

7.50

7.60

7.70

7.80

7.90

8.00

8.10

0 48 96 144 192 240 288 336

pH

Time post surgey (hours)

Sedated

Unsedated

Refefef rence

cd**######

BBE*###

CF**###

bc*###

CEF*###

c*##

c

BF*###

c*###

CDEF*###

cd*###

BF###

BF###

BF###

CEF#####

CEF###d###cd###

bcd###cd###bbbcd####

0 48 96 144 192 240 288 336

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Part

ial

pres

sure

ofC

O2

(kPa

)

Sedated

Unsedated

Reference

BCDE**#

ab*

BCDE**

ababa **

ab*ab**

BCDE*##

ab*

ab**

B**##

BCE*#

a

ABE

abab

abababa

abAE

ABE

ABE

ABE

0

1

2

3

4

5

6

7

8

9

0 48 96 144 192 240 288 336

HC

O3-

conc

entr

atio

n (m

mol

L/)

Sedated

Unsedated

b

AD*

b**###b*##

ACE*

b*##

ACE*

b*##

b*##

AD**

b##

b#b#b#b#b#b#b#b#b##

b##b##b##BDE####

AACD# ACEACE

ACE*

BDE#

7.57.67.77.87.98.0

0 3 4

a*

AA* c*##

BD*##B##

b

0 1 2 3 4

0.4

0.5

0.6

AC

A

AAAD

ab ab

345678

0

1 2

1 2 3 4

b# b##

a

A

BC## BC##

Reference

Fig. 2 Arterial blood

plasma partial pressure of

CO2 (a), bicarbonate

concentration (b) and pH

(c) after CA cannulation

surgery in Atlantic cod that

were either sedated with

metomidate or not sedated

prior to anaesthesia and

subsequent cannulation.

A solid line represents a

reference level that

consisted of cod

sacrificially sampled from

the caudal vessels. The

boxes display the

development in the first 4 h

post surgery (PS). The

results are presented as LS

means with S.E. bars. * or

**Significant difference

between the sedation

statuses at the respective

sampling time at P \ 0.05

or P \ 0.01, respectively.#, ## or ###Significant

difference between the

relevant sampling time/

sedation status and the

reference level at P \ 0.05,

P \ 0.01 or P \ 0.001,

respectively. Different

upper case letters indicate

significant differences

between sampling times PS

in sedated cod and different

lower case letters indicate

significant differences

between sampling times PS

in unsedated cod

Fish Physiol Biochem (2012) 38:401–411 407

123

Discussion

Plasma cortisol and glucose

Metomidate sedation delayed the plasma cortisol and

glucose responses to surgery in Atlantic cod but did

seemingly nothing to prevent the stress responses

themselves (Fig. 1). This effect of metomidate seda-

tion/anaesthesia has never before been reported in

Atlantic cod or any other species of fish and appears

to stand in contrast to the suppressing effect of

metomidate on cortisol release (Olsen et al. 1995).

However, in the present study, even the unsedated

cod had a prolonged stress response PS when

145

150

155

160

165

0 48 96 144 192 240 288 336

Na+

conc

entr

atio

n(m

mol

/Ll/)Sedated

Unsedated

Refeff rence

CF* CDF*

ab*

BF*BF*

abaa *ab* ab*aa *

BF**

ab*

a

ababab

b

ab#aa #

E##

AFBFCDF

CDF

2.0

2.5

3.0

3.5

4.0

0 48 96 144 192 240 288 336

K+

conc

entr

atio

n(m

mol

/Ll/)

Time post surgeryrr (hours)

Cannulataa ed

Refeff renceA

A

BC###

D###

AB

BCD###D###D###CD###D####D###

D###BCD###

b

145

150

155

160

165

0 1 2 3 4

a##a #

ab##

ab###aa

ACACACACACE#

AE###

ADE##

Fig. 3 a Arterial blood plasma Na? concentration after CA

cannulation surgery in Atlantic cod that were either sedated

with metomidate or not sedated prior to anaesthesia and

subsequent cannulation. The solid line represents a reference

level that consisted of cod sacrificially sampled from the caudal

vessels. The box displays the development in the first 4 h post

surgery (PS). * or **Significant difference between the

sedation statuses at the respective sampling time at P \ 0.05

or P \ 0.01, respectively. #, ## or ###Significant difference

between the relevant sampling time/sedation status and the

reference level at P \ 0.05, P \ 0.01 or P \ 0.001, respec-

tively. Different upper case letters indicate significant

differences between sampling times PS in sedated cod and

different lower case letters indicate significant differences

between sampling times PS in unsedated cod. The results are

presented as LS means with S.E. bars. b Arterial blood plasma

K? concentration after CA cannulation surgery in Atlantic cod.

Sedation status did not affect the PS development of the plasma

K? concentration (see M&M for justification). Thus, the two

groups are combined in this figure. The solid line represents a

reference level that consisted of cod sacrificially sampled from

the caudal vessels. ###Significant difference between the

relevant sampling time and the reference level at P \ 0.001.

The results are presented as LS means with S.E. bars

408 Fish Physiol Biochem (2012) 38:401–411

123

compared to similar or more invasive surgeries

performed in Atlantic salmon (Salmo salar) (Djordj-

evic et al. 2011; Eliason et al. 2007), a species whose

plasma cortisol and glucose responses to stress have

similar time frames as Atlantic cod (Olsen et al. 2002,

2008). Due to the comparatively prolonged PS stress

response observed even in unsedated cod in the

present study, the cod likely experienced a significant

level of stress when the effects of metacain anaes-

thesia (and metomidate sedation) had worn off. In

sedated cod, this PS stress likely triggered a cortisol

release and subsequent plasma glucose increase upon

withdrawal from sedation/anaesthesia since metomi-

date had suppressed their initial cortisol response. In

unsedated cod, however, netting, anaesthesia and

cannulation likely induced a stress response that

elevated plasma cortisol and glucose above the

reference levels directly PS.

The recovery times (time PS where the parameter

is no longer elevated/lowered compared to the

reference level or has stabilized) of cod plasma

cortisol and glucose were longer than what has been

reported following stress in Atlantic cod (Olsen et al.

2008) and cannulation surgeries in Atlantic salmon

(Djordjevic et al. 2011; Eliason et al. 2007). The

comparatively long recovery times observed in the

present study is one example of how knowledge

cannot always be transferred between species or

between procedures within species. In both sedated

and unsedated cod, plasma glucose recovery time was

6 days PS. This recovery time is longer than in any

other measured parameter in the present study. Thus,

we recommend a recovery time of 6 days or more in

experiments utilising the CA cannulation technique

in Atlantic cod.

Acid–base parameters

Plasma pH, PCO2 and HCO3- concentrations dif-

fered intermittently between sedated and unsedated

cod and PCO2 and HCO3- concentrations varied

more in sedated than unsedated cod (Fig. 2). The

differences between sedation statuses and increased

variability in sedated cod suggest an effect of

metomidate sedation on respiration, metabolism or

red blood cell (RBC) ion transport in Atlantic cod

that continues to affect the cod even after 2 weeks of

recovery. However, the mechanism(s) behind these

differences remain uncertain and needs to be

examined further in order to fully understand the

effect of metomidate sedation on Atlantic cod.

Acute stress and the subsequent release of cate-

cholamines to the circulation are known to affect

several blood parameters in fishes, most of which are

directed towards securing oxygen uptake and delivery

to the tissues in stressful situations (Reid et al. 1998).

Amongst these responses are changes in RBC ion

movements directed at elevating RBC pH to increase

the oxygen carrying capacity of haemoglobin (Tho-

mas and Perry 1992). Upon stimulation from cate-

cholamines, the Na?/H? exchangers on the RBC

membrane increases Na? influx and H? efflux across

the RBC membrane, effectively increasing intracel-

lular pH and Na? concentration and reducing plasma

pH and Na? concentration (Fievet et al. 1987). In the

present study, plasma pH is reduced directly PS in

both groups of cannulated cod and in the reference

cod. This was likely caused by a release of catechol-

amines prior to or during cannulation surgery, or by

the sacrificial sampling procedure in the reference

cod. However, there was no concomitant decrease in

plasma Na? levels as could be expected if the RBC

membrane Na?/H? exchangers caused the drop in

plasma pH. In fact, the opposite result was observed

as plasma Na? levels were elevated the first 24 h PS

(see below for elaboration). Due to the plasma

acidification observed in the present study, net flux

through the Cl-/HCO3- exchanger would be

expected to decrease plasma Cl- and increase plasma

HCO3- (Fievet et al. 1988). However, the opposite

result was observed in the present study; in both

sedated and unsedated cod, plasma HCO3- levels

were lowered but not different from the reference

directly PS, compared to their stabilized level at 3 h

PS onwards.

Plasma ions

The blood plasma ions measured in the present study,

Na? and K?, stabilized within 48 h PS, and sedation

status did not affect the recovery time of these

parameters. However, in sedated cod, the plasma Na?

concentration stabilized at a lower level than in

unsedated cod from day 3 to day 7 PS (Fig. 3). The

elevated plasma Na? concentrations observed in both

sedated and unsedated cod the first 24 h PS is

opposite to what can be expected based on the plasma

acidification observed the first 24 h PS. The high

Fish Physiol Biochem (2012) 38:401–411 409

123

plasma Na? levels 0–24 h PS may have been caused

by a reduced ability to maintain osmotic balance

during and directly after surgery, causing plasma Na?

to increase, despite Na? influx to the RBC via the

Na?/H? exchanger. Another possibility is mis-

matched activity levels of RBC ion exchangers,

e.g., between the Na?/K? ATPase, the Na?/H?

exchanger and the Na?/K?/Cl- co-transporters dur-

ing cannulation. Catecholamines are known to

increase the activity of RBC Na?/K? ATPase

(Bourne and Cossins 1982) and Na?/K?/Cl- co-

transporters (Russell 2000), either of which may be

the reason for the observed high plasma Na? levels

the first 24 h PS. In cannulated cod, there was a

significant drop in plasma K? concentration from

directly PS until it stabilized at day 2 PS. The initial

(0–24 h PS) high plasma K? concentration was likely

caused by a net efflux of K? from the RBC during

cannulation, likely due to increased activity of K?/

Cl- or Na?/K?/Cl- co-transporters on the RBC

membrane (Nikinmaa 2006; Russell 2000).

Cannulated cod versus sacrificially sampled

reference cod

The cod used as reference in the present study were

netted, euthanized by a blow to the head and finally

sampled for blood from the caudal vessels. Although

the procedure lasted only a few seconds, it likely

produced an acute stress response and subsequent

catecholamine release in the cod (Perry and Bernier

1999), as plasma pH, HCO3- and K? were all

elevated or lowered compared to the stable levels

observed at day 2 PS onwards in cannulated cod

(Figs. 2b, c and 3b). These findings highlight the

importance of using undisturbed fish when studying

biological parameters that may change rapidly due to

stress and indicate that sacrificially sampled fish are,

for several blood variables, not a good point of

reference. In spite of the suggested acute stress

response and subsequent catecholamine release, the

plasma cortisol and glucose levels are low in the

reference cod. This is most likely an effect of the

short time frame of the sacrificial sampling procedure

(\1 min), as plasma cortisol and glucose levels are

known to peak 1–3 h after acute stress in Atlantic cod

(Olsen et al. 2008).

The cod used as reference in the present study

were smaller than the cannulated cod. The difference

in body weight between the cannulated and the

reference cod could potentially have affected

the results. However, in the present study, none of

the regressions between plasma cortisol or glucose (at

day 14 in cannulated cod) and body weight or length

were significant. Thus, the difference in body weight

between cannulated and reference cod most likely did

not affect the results in the present study.

Conclusions and general remarks

The delayed stress response in metomidate sedated

cod observed in the present study has many possible

implications for aquaculture operating procedures

and experiments with fish. Therefore, the effects of

sedation or anaesthesia using metomidate should be

carefully evaluated for each specific application prior

to use, in order to prevent stressed and possibly

suffering fish and the possibility of producing unre-

liable or ambiguous results from experiments. In both

sedated and unsedated cod, all measured parameters

had stabilized or returned to the reference level at day

6 PS. Thus, we conclude that the CA cannulation is a

valid technique for obtaining arterial blood samples

from undisturbed Atlantic cod, given a recovery time

of 6 days or more is applied. Due to the absence of an

improved recovery from CA cannulation when using

metomidate, we recommend that metomidate not be

used as a sedative prior to anaesthesia and subsequent

handling of Atlantic cod.

Acknowledgments This study was funded by the Norwegian

Research Council (NFR) through the research projects

PROCOD (NFR project number 172263) and MODSMO

(NFR project number 172514). We would like to thank the

staff at NIVA-MFS for their technical assistance and care for

the fish prior to and during the experiment, and an anonymous

reviewer for the critical comments and suggestions on how to

improve the manuscript.

References

Axelsson M, Nilsson S (1986) Blood pressure control during

exercise in the Atlantic cod, Gadus morhua. J Exp Biol

126:225–236

Axelsson M, Davison B, Forster M, Nilsson S (1994) Blood

pressure control in the Antarctic fish Pagothenia bor-chgrevinki. J Exp Biol 190:265–279

Bourne PK, Cossins AR (1982) On the instability of K? influx

in erythrocytes of the rainbow trout, Salmo gairdneri, and

410 Fish Physiol Biochem (2012) 38:401–411

123

the role of catecholamine hormones in maintaining in vivo

influx activity. J Exp Biol 101:93–104

Conte FP, Wagner HH, Harris TO (1963) Measurement of

blood volume in the fish (Salmo gairdneri gairdneri). Am

J Physiol 205(3):533–540

Djordjevic B, Kristensen T, Øverli Ø, Rosseland BO, Kiessling

A (2011) Effect of nutritional status and sampling inten-

sity on recovery after dorsal aorta cannulation in free-

swimming Atlantic salmon (Salmo salar L.). Fish Physiol

Biochem Published online. doi:10.1007/s10695-009-

9362-2

Eliason EJ, Kiessling A, Karlsson A, Djordjevic B, Farrell AP

(2007) Validation of the hepatic portal vein cannulation

technique using Atlantic salmon Salmo salar L. J Fish

Biol 71(1):290–297. doi:10.1111/j.1095-8649.2007.014

89.x

Fievet B, Motais R, Thomas S (1987) Role of adrenergic

dependent H? release from red cells in acidosis induced

by hypoxia in trout. Am J Physiol 252(2):R269–R275

Fievet B, Claireaux G, Thomas S, Motais R (1988) Adaptive

respiratory responses of trout to acute hypoxia. III. Ion

movements and pH changes in the red blood cell. Respir

Physiol 74(1):99–113

Forgue J, Burtin B, Massabuau JC (1989) Maintenance of

oxygen consumption in resting Silurus glanis at different

levels of ambient oxygenation. J Exp Biol 143:305–319

Forsman A, Kiessling A, Ruohonen K, Vasquez J, Grove D

(2005) EU-project report ‘‘Selfish’’ Q5CA-2001-00989

Karlsson A, Sørlie Heier L, Rosseland BO, Salbu B, Kiessling

A (2011) Changes in arterial PO2, physiological blood

parameters and intracellular antioxidants in free-swim-

ming Atlantic cod (Gadus morhua) exposed to varying

levels of hyperoxia. Fish Physiol Biochem 37(1):249–258.

doi:10.1007/s10695-010-9438-z

Kiessling A, Olsen RE, Buttle L (2003) Given the same dietary

carotenoid inclusion, Atlantic salmon, Salmo salar (L.)

display higher blood levels of canthaxanthin than asta-

xanthin. Aquacult Nutr 9(4):253–261

Kiessling A, Dosanjh B, Koppe W, Higgs D (2006) Relation-

ship between blood and muscle levels of astaxanthin in

dorsal aorta cannulated Atlantic salmon. Aquaculture

254(1–4):653–657

Kristensen T, Rosseland BO, Kiessling A, Djordjevic B,

Massabau JC (2010) Lack of arterial PO2 downregulation

in Atlantic salmon (Salmo salar L.) during long-term

normoxia and hyperoxia. Fish Physiol Biochem 36(4):

1087–1095. doi:10.1007/s10695-010-9386-7

McLean E, Ash R (1989) Chronic cannulation of the hepatic

portal vein in rainbow trout, Salmo gairdneri: a pre-

requisite to net absorption studies. Aquaculture 78(2):

195–205

Nikinmaa M (2006) Gas transport. In: Evans DH, Claiborne JB

(eds) The physiology of fishes, 3rd edn. CRC Press, Boca

Raton, FL, p 601

Olsen YA, Einarsdottir IE, Nilssen KJ (1995) Metomidate

anesthesia in Atlantic salmon, Salmo salar, prevents

plasma cortisol increase during stress. Aquaculture

134(1–2):155–168

Olsen RE, Sundell K, Hansen T, Hemre GI, Myklebust R,

Mayhew TM, Ringo E (2002) Acute stress alters the

intestinal lining of Atlantic salmon, Salmo salar L.: An

electron microscopical study. Fish Physiol Biochem

26(3):211–221

Olsen RE, Sundell K, Ringø E, Myklebust R, Hemre GI, Hansen

T, Karlsen O (2008) The acute stress response in fed and

food deprived Atlantic cod, Gadus morhua L. Aquaculture

280(1–4):232–241. doi:10.1016/j.aquaculture.2008.05.006

Perry SF, Bernier NJ (1999) The acute humoral adrenergic

stress response in fish: facts and fiction. Aquaculture

177(1–4):285–295

Perry SF, Fritsche R, Kinkead R, Nilsson S (1991) Control of

catecholamine release in vivo and in situ in the Atlantic

cod (Gadus morhua) during hypoxia. J Exp Biol 155:

549–566

Reid SG, Bernier NJ, Perry SF (1998) The adrenergic stress

response in fish: control of catecholamine storage and

release. Comp Biochem Physiol C: Toxicol Pharmacol

120(1):1–27

Russell JM (2000) Sodium-potassium-chloride cotransport.

Physiol Rev 80(1):211–276

Smith LS, Bell GR (1964) A technique for prolonged blood

sampling in free-swimming salmon. J Fish Res Board Can

21(4):711–717

Smith DG, Nilsson S, Wahlqvist I, Eriksson BM (1985) Ner-

vous control of the blood pressure in the Atlantic cod,

Gadus morhua. J Exp Biol 117:335–347

Soivio A, Nyholm K, Westman K (1975) A technique for

repeated sampling of the blood of individual resting fish.

J Exp Biol 62:207–217

Sunde J, Kiessling A, Higgs D, Opstvedt J, Venturini G,

Rungruangsak-Torrissen K (2003) Evaluation of feed

protein quality by measuring plasma free amino acids in

Atlantic salmon (Salmo salar L.) after dorsal aorta can-

nulation. Aquacult Nutr 9(6):351–360

Thomas S, Perry SF (1992) Control and consequences of

adrenergic activation of red blood cell Na?/H? exchange

on blood oxygen and carbon dioxide transport in fish.

J Exp Zool 263(2):160–175

Zahl IH, Kiessling A, Samuelsen OB, Hansen MK (2009)

Anaesthesia of Atlantic cod (Gadus morhua)—Effect of

pre-anaesthetic sedation, and importance of body weight,

temperature and stress. Aquaculture 295(1–2):52–59. doi:

10.1016/j.aquaculture.2009.06.019

Fish Physiol Biochem (2012) 38:401–411 411

123