Introduction

Anaesthetic agents have been used for fish since the beginning of the last century (McFarland 1959; McFarland and Klontz 1969; Schoettger and Julin 1967). They derive partly from human and veterinary medicine and partly from other sources and were introduced to fish through trial and error. They were initially employed in order to make handling easier but improved knowledge of fish physiology and a better understanding of the importance of anaesthetic treatment to maintain welfare have led to an increased use.

Anaesthesia (from Greek: an- ‘without’ and aisthesis ‘sensation’) comprises several components, including sedation, immobilisation, unconsciousness (narcosis), amnesia (loss of memory) and analgesia (pain relief). Sedation is a reduction in sensitivity, which results in tranquillity and calmness. Narcosis (general anaesthesia) causes a state of unconsciousness and amnesia and also includes immobilisation and pain relief (analgesia). The components of anaesthesia can be obtained by various anaesthetic agents, each of which gives rise to one or several of them. In this review, the agents such as benzocaine, metacaine (MS-222), metomidate hydrochloride, isoeugenol, 2-phenoxyethanol and quinaldine are discussed as anaesthetic agents for fish. Traditionally, these agents have been used and are still used in fish to obtain different components of anaesthesia irrespective of the actual effect of the specific agent. Agents with known analgesic effect, used in human and veterinary medicine as local analgesics, such as benzocaine, are administrated to fish for the purpose of sedation, immobilisation, analgesia as well as general anaesthesia (narcosis). Agents are also applied indiscriminately of species. Anaesthetic protocols of new species that are introduced to research or cultivation are generally based on drugs and dosages developed for the more established species. For example, in Norway, the anaesthetic protocols of Atlantic cod (Gadus morhua) and Atlantic halibut (Hippoglossus hippoglossus), which were introduced to fish farming in the 1980s, have thus been based on the protocols used for salmonids, which have been farmed since the 1960s and are the most important species in Norwegian fish farming.

Anaesthetic protocols of human and veterinary medicine comprise combinations of several drugs, each contributing to one or more of the effects needed in the anaesthesia. Different drugs are applied for induction and maintenance, and analgesic treatment is used both under and following surgical procedures. The drugs are selected on the basis of their properties, tailored both to the surgical intervention and to the physiological state of the patient or the animal. Pre-anaesthetic sedation used in order to avoid stress prior to anaesthesia is an integral part of veterinary protocols. By combining drugs of different properties, a more complete anaesthesia is obtained than what is possible with one single substance alone. Further, synergistic effects between different drugs may also permit a reduction in the dosage of each drug compared to individual administration. This may result in smoother induction and recovery and reduce the incidence of adverse effects (Rang et al. 2003).

Anaesthesia of fish

Anaesthetic agents are now used widely, ranging from light sedation that aims to reduce stress during handling and non-invasive procedures to full anaesthesia to avoid inflicting pain during surgery and larger interventions (Ackerman et al. 2005; Neiffer and Stamper 2009; Ross and Ross 2008; Summerfelt and Smith 1990). The most common route of administration is via inhalation. The anaesthetic agent is dispersed in the water and is absorbed across the gills. The effect is usually assessed by induction and recovery time, reflex reactions to external stimuli and responsiveness to handling.

The progress of induction and depth of anaesthesia are generally divided into distinct stages and ‘planes’ (Table 1). Common clinical indicators evaluated in order to determine stages and planes during anaesthesia of animals include behaviour, activity, corneal reflexes and pupil size, muscle tone, reflexes, respiratory rate, heart rate and blood pressure. As several of these indicators are difficult to assess in fish, it may be difficult to distinguish one stage or plane from the other, especially in situations where induction is rapid. In fish, the stages are therefore usually described by changes in swimming activity, balance, respiratory frequency and reactions to external stimuli (Table 1).

Table 1 Stages of anaesthesia in fish
Full size table

In order to obtain efficient surgical anaesthesia through inhalation using the agents described in this review, respiration will stop in spite of maintained heart rate and blood pressure (Kiessling et al. unpublished results). However, injecting a local anaesthetic at the site of incision will provide sufficient analgesic effect and allow surgical procedures at an anaesthetic depth where spontaneous respiration is maintained, corresponding to stage III, plane 2 (Table 1; Karlsson et al. 2011a; Kiessling et al. 2003, 2009).

Nociception in fish

Studies on rainbow trout and goldfish have demonstrated that fish possess the basic neural system necessary for nociception, i.e. perception of painful stimuli (Dunlop and Laming 2005; Sneddon 2002, 2003; Sneddon et al. 2003). Thus, to ensure the welfare of fish subjected to procedures that might inflict pain, anaesthetic agents with the ability to block nociceptive pathways are necessary. Of the agents assessed in this review, benzocaine and MS-222 possess this ability, as does isoeugenol. However, the route of administration may be of importance for the effect of this class of agents (local anaesthetics). Anaesthetic effect is produced locally at the site of administration, e.g. topically. Administration systemically has, on the other hand, been found inadequate for blocking peripheral nociceptive pathways in higher vertebrates (mice, rat, cat and man; Boas et al. 1982; Haegerstam 1979; Rigon and Takahashi 1996; Wiesenfeld-Hallin and Lindblom 1985; Woolf and Wiesenfeld-Hallin 1985). Haegerstam (1979) concluded that it was unlikely that peripheral pain pathways could be blocked by systemic i.v. injections of local anaesthetics. On the basis of these data, it is unclear whether a sufficient level of anaesthesia can be obtained peripherally when anaesthetic agents are administered to fish via inhalation. As the whole body surface of the fish is exposed to anaesthetics during bath immersions, i.e. topical administration in addition to inhalation, the agents will also have a peripheral effect. The absorption of small amounts of anaesthetic through the skin of the fish has been observed (Ferreira et al. 1984), but whether such amounts produce the peripheral analgesia needed for surgical procedures is not known. It is therefore important to determine whether a lack of response to nociceptive stimuli is due to a peripheral blockage of nociceptive afferent pathways or whether it is due to a general inhibition of the CNS. The latter may have severe welfare consequences as an inadequate blockage of nociceptive pathways may result in sensitisation and thus a reduction in threshold and an increased response to noxious stimuli. Work by Zahl et al. (2009, 2011) indicates that neither of the agents covered in this review administered via inhalation either individually or in combination blocks peripheral nociceptive pathways in fish. Kiessling et al. (1995, 2003) show that this is obtained when lidocaine is administered locally during combination anaesthesia with metomidate and MS-222 administered via inhalation. In the studies by Zahl et al. (2009, 2011), lower dosages in combination anaesthesia resulted in similar induction time as individually administrated agents. However, the fish displayed increased reactions to nociceptive stimulation although agents blocking nociception were used. This is possibly due to an enhanced central effect without a parallel increase in effect peripherally. Since all surgical procedures are associated with nociception, our conclusion must therefore be that any such procedure should never be conducted without local injections of anaesthetic agents with analgesic effect.

Anaesthetic agents

A wide range of anaesthetic agents is currently used for fish. Among the most common are MS-222, benzocaine, isoeugenol, metomidate, 2-phenoxyethanol and quinaldine (Ackerman et al. 2005; Neiffer and Stamper 2009; Ross and Ross 2008; Summerfelt and Smith 1990). Dosage recommendations are listed in Table 2. The sites of action in the nervous system, i.e. central or peripheral, for these drugs are not well known in fish, and most knowledge is based on that of higher vertebrates. There are many reasons for applying this knowledge in fish with caution. First, there are more than 30,000 different species of fish (Froese and Pauly 2011) that inhabit highly diverse environments, which vary in factors like salinity, temperature, chemistry and depth. Secondly, and possibly the most important and most often overlooked fact, the route of administration may differ both between different species and between different agents, i.m. for local anaesthetics vis-à-vis i.v. or inhalation for general anaesthetics. Finally, fish and higher vertebrates, particularly mammals, differ in the development of the central nervous system, while their peripheral nervous systems seem to be surprisingly similar (see e.g. Sneddon et al. 2003).

Table 2 Anaesthetic agents—dosage and exposure time recommendations by manufacturers/distributors
Full size table

Metacaine and benzocaine

Metacaine (ethyl 3-aminobenzoate, tricaine methanesulphonate and MS-222) and benzocaine (ethyl 4-aminobenzoate) are the two most common anaesthetics for fish used in Norway (Anon 2007). They are approved for use in food fish production with a withdrawal period of 21 days. MS-222 and benzocaine are local anaesthetics that act by blocking voltage-sensitive sodium channels (Frazier and Narahashi 1975; Neumcke et al. 1981), thereby preventing the voltage-dependent increase in sodium conductance. This inhibits the initiation and propagation of action potentials in excitable cells; thus, local anaesthetics block most neurons and muscle cells and may cause paralysis in addition to blocking nociception. This class of drugs is used in human and veterinary medicine for local analgesia. When administrated to fish via bath immersion, they enter the circulation and produce general anaesthesia by inhibiting neural signal transmission ranging from the periphery to higher parts of the nervous system. The precise mechanism of action in the central nervous system is not fully understood (Hara and Sata 2007; Hedrick and Winmill 2003; Ueta et al. 2007). Many of the adverse effects caused by this class of drugs in man are related to effects on the CNS and the cardiovascular system (Rang et al. 2003).

A range of adverse effects of MS-222 and benzocaine have been reported in fish. The initial response to these agents is characterised by a period of raised heart rate and respiration accompanied by elevated levels of blood glucose, followed by a depression of cardiovascular and respiratory function that eventually may come to a complete stop (Hill et al. 2002; Houston et al. 1971a; Lochowitz et al. 1974; Randall 1962; Ryan 1992). This results in hypoxaemia, observed as changes in the partial pressure of arterial O2 and CO2, accompanied by lower blood glucose levels and increased levels of lactate, hematocrit and haemoglobin, and erythrocyte swelling (Bourne 1984; Hill and Forster 2004; Holloway et al. 2004; Houston et al. 1971b; Iwama et al. 1989; Ortuno et al. 2002; Ryan 1992; Soivio et al. 1977; Thomas and Robertson 1991; Velisek et al. 2009). Elevated plasma catecholamine levels have also been observed (Gingerich and Drottar 1989; Iwama et al. 1989; Wedemeyer 1970) Benzocaine has also been found to have immunodepressant effects (Cuesta et al. 2004; Ortuno et al. 2002).

Metomidate

Metomidate hydrochloride (methyl 3-(1-phenylethyl) imidazole-4-carboxylate hydrochloride) is a methyl analogue of the imidazole derivate etomidate, which activates and modulates inhibitory gamma-aminobutyric acid type A (GABAA) receptors, thus affecting higher regions of the nervous system (Ashton and Wauquier 1985; Yang and Uchida 1996). Activation of the GABAA receptor triggers opening of a chloride ion-selective pore. The increased chloride conductance drives the membrane potential towards the reversal potential of the Cl¯ ion, thus inhibiting the firing of new action potentials. GABAA receptor activation produces sedation and hypnosis, but limited analgesia and immobilisation (Grasshoff et al. 2006).

Etomidate is commonly used in human and veterinary medicine as a sedative and for inducing anaesthesia (Ching and Baum 2009; Darrouj et al. 2009; Falk and Zed 2004; Sams et al. 2008). One of the side effects of the drug in addition to respiratory and cardiovascular depression is suppression of adrenal steroidogenesis, thus leading to an inhibition of cortisol synthesis (Vanden Bossche et al. 1984; Wagner et al. 1984; man, cow and rat). Metomidate, in parity with etomidate in mammals, has also been found to have an inhibitory effect on cortisol in fish (Davis and Griffin 2004; Olsen et al. 1995; Small 2004; Thomas and Robertson 1991) but the efficiency and dose dependence of this inhibition are still unclear (Eliason et al. 2007; Olsen et al. 1995; Sandodden et al. 2001). However, a number of reports indicate cortisol release in spite of metomidate treatment (Kiessling, et al. 2009; Zahl et al. 2010). Other adverse effects of metomidate in fish are related to depressant effects on respiration and circulation that subsequently lead to hypoxaemia, observed as a decrease in partial pressure of O2, elevated partial pressure of CO2 and reduced pH of the blood (Hill and Forster 2004; Hill et al. 2002; Iwama et al. 1989).

2-Phenoxyethanol

2-Phenoxyethanol is a compound with antibacterial properties used as a preservative in vaccines, in dermatological products and as fixative for perfumes. To the best of our knowledge, the exact mechanism of its anaesthetic effect in fish has not been reported, but it has been suggested that it involves an expansion of neuronal cell membranes (Burka et al. 1997). 2-Phenoxyethanol has been found to inhibit the activity of excitatory N-methyl-d-aspartate (NMDA) receptors in Xenopus oocytes (Musshoff et al. 1999), and its anaesthetic effect may thus be related to the suppression of neural activity in higher regions of the nervous system. Activation of the NMDA receptors increases neuron excitability by lowering the threshold for firing of action potentials due to opening of ion channels that cause influx of positively charged ions, especially Ca2+, Na+ and K+. NMDA receptor inhibition has been linked to an analgesic effect (Grasshoff et al. 2006).

Adverse effects of 2-phenoxyethanol include reduced ventilation, decreased heart rate and blood pressure, reduced blood partial pressure of O2 accompanied by increased CO2 and decreased blood pH and elevated plasma levels of adrenaline and glucose (Fredricks et al. 1993; Iwama et al. 1989; Lambooij et al. 2009; Ortuno et al. 2002). 2-Phenoxyethanol has also been found to have immunodepressant effects (Cuesta et al. 2004; Ortuno et al. 2002).

Isoeugenol

Isoeugenol (2-methoxy-4-prop-1-enyl-phenol) is an oily liquid found in ylang–ylang and other essential oils. It is structurally similar to eugenol, a widely used analgesic in dentistry that inhibits sodium, potassium and calcium channels, inhibits NMDA receptors and potentiates GABAA receptors (Aoshima and Hamamoto 1999; Lee et al. 2005; Li et al. 2007; Park et al. 2006; Wie et al. 1997). Like eugenol, isoeugenol is one of the constituents of clove oil, another widely used fish anaesthetic. Isoeugenol is the active ingredient in Aqui-S®, which has approval for use in food fish production in New Zealand, Australia and Chile, Costa Rica and Republic of Korea with no withdrawal period. In Norway, Aqui-S® is used for brood stock and in research.

Adverse effects of isoeugenol (Aqui-S®) in fish include reduced ventilation and depression of the cardiovascular system, which result in slower heart rate and decreased cardiac output, in addition to reduced blood pressure and vascular tone (Hill and Forster 2004; Hill et al. 2002; Rothwell and Forster 2005). Elevated plasma levels of catecholamines have also been observed, as has increased haematocrit (Hill and Forster 2004). Eugenol works synergistically with d-tubocurarine; it is therefore believed to block nicotine2 receptors at the motor end plate in muscle (Brodin and Røed 1984). Nicotine2 receptor blockade causes paralysis without analgesic or hypnotic effect. This effect has since been confirmed for isoeugenol by Ingvast-Larsson et al. (2003), who used a rat phrenic nerve–diaphragm muscle preparation.

Quinaldine

Quinaldine (2-methylquinoline) is a colourless oily liquid that has been used to anaesthetise fish since early in the last century. To the best of our knowledge, its precise mechanism of action has not been reported.

Quinoline family compounds possess antiseptic and antipyretic properties. They are widely used as parent compounds to make drugs, especially anti-malarial medicines, and are also used in fungicides, biocides, alkaloids, dyes, rubber chemicals and flavouring agents.

Fish anaesthetised with quinaldine respond by increasing their heart rate, followed by bradycardia and impaired respiratory function (Lochowitz et al. 1974). Elevated concentrations of plasma cortisol and glucose have also been observed, as have increased levels of serum IgM (Cuesta et al. 2004; Davis and Griffin 2004; Ortuno et al. 2002).

Combination anaesthesia

Current protocols of fish do not typically include administration of combinations of anaesthetics. Schoettger and Steucke (1970) examined the possible synergistic effects of combination anaesthesia consisting of a mixture of MS-222 and quinaldine in rainbow trout (Oncorhynchus mykiss) and northern pike (Esox lucius) and found that the mixture resulted in safer and more effective anaesthesia. The same mixture tested in 14 freshwater species showed that the dosages needed when administered in combination were considerably lower than when each drug was used alone (Gilderhus et al. 1973). However, it has also been found that not only efficacy but also toxicity increased when these drugs were combined (Dawson and Marking 1973). Synergistic effects have also been found in combination anaesthesia using a mixture of quinaldine sulphate and the benzodiazepine diazepam (Kumlu and Yanar 1999; Yanar and Kumlu 2001). Administered in combination with diazepam, quinaldine sulphate could be used in lower dosages in anaesthesia of both gilthead sea bream (Sparus aurata) and European sea bass (Dicentrarchus labrax). Hyperactivity and excitement, unwanted side effects of quinaldine sulphate, were reduced, and there were no mortalities (Kumlu and Yanar 1999; Yanar and Kumlu 2001). Combinations consisting of MS-222 anaesthesia and injections of local analgesics during surgery have been studied in koi carp (Cyprinus carpio; Harms et al. 2005). Injections of the opiate butorphanol during surgery reduced post-surgery behavioural changes, while the non-steroidal anti-inflammatory drug (NSAID) ketoprofen reduced muscle tissue damage. Protocols that include both pre-anaesthetic sedation and local analgesic treatment in addition to general anaesthesia have been successfully employed in anaesthesia of salmonids undergoing surgery for insertion of dorsal aorta and portal vein cannulae (Eliason et al. 2007; Karlsson et al. 2006; Kiessling et al. 1995, 2003). Pre-anaesthetic sedation with metomidate followed by full anaesthesia with MS-222 combined with injections of lidocaine at the site of incision improved recovery in comparison with MS-222 used individually, and fish quickly resumed their normal behaviour (Eliason et al. 2007; Karlsson et al. 2006; Kiessling et al. 1995, 2003). The response of Atlantic cod subjected to caudal artery cannulation using pre-anaesthetic sedation with metomidate has been found to be markedly different from that of salmonids, in that the former display a delay rather than a reduction in post-surgery cortisol release, and a prolonged recovery, measured as blood glucose, lasting from 3 to 5 days (Karlsson et al. 2011b). Zahl et al. (2009) found that pre-anaesthesia sedation of Atlantic cod with metomidate permitted a significant reduction in the dosage of anaesthetic to be made, with no change in induction or recovery time. However, the same authors report that Atlantic cod, subjected to pre-anaesthetic sedation with metomidate and the reduced dose of anaesthetic, displayed an increased response to external stimulation in the form of a caudal peduncle pinch. They also report that exposing Atlantic cod to pre-anaesthetic stress instead of pre-anaesthetic sedation resulted in the same induction time in spite of a lower dose of anaesthetic. These results clearly underline the fact that the effect of an anaesthetic agent varies with pre-anaesthetic stress and show that the use of an anaesthetic protocol that includes a combination of pre-sedation and a reduced anaesthetic dosage may produce an acceptable hypnotic effect without a sufficient analgesic effect.

Variations in response

There are substantial variations between fish species in response to anaesthetic agents, as well as large individual differences within each species. Variations may be the result of pharmacokinetic differences, usually described as what the body does to the drug, and by pharmacodynamic differences, usually described as what the drug does to the body. Both pharmacokinetic and pharmacodynamic differences in fish may be influenced by biological factors such as age, sex, life stage, body weight, growth rate, body composition, physiological condition and health status, as well as environmental factors such as water temperature, salinity, pH and oxygen level.

In the majority of assessments of anaesthesia in fish, the fish have been anaesthetised to a predetermined stage, often corresponding to loss of equilibrium, no response to external stimuli and reduced or even stopped respiration (Table 1, stage III, plane 3 to stage IV) and thereafter directly moved to clean water for recovery (Hoskonen and Pirhonen 2004; Mattson and Riple 1989; Mylonas et al. 2005; Stehly and Gingerich 1999; Sylvester and Holland 1982). As the induction time may differ between individual fish, exposure time has varied accordingly. Another approach is to impose an identical exposure time for all fish. When comparing the results of different studies, it is important to distinguish between these two approaches. The duration of the exposure should always be kept within an adequate margin of safety, as practical situations often involve anaesthesia of a large number of fish in one tank, with exposure time usually exceeding the induction time for individual fish.

Pharmacodynamics

The pharmacodynamic characteristics of a drug describe the consequences of the interaction between the drug and the receptor or other primary sites of action and include all effects and adverse effects. As described earlier, anaesthetic agents may act by binding to receptors or by interacting with ion channels. As several types of tissue may contain the target molecules of a drug, the primary responses may induce secondary responses or side effects that subsequently affect the primary response. Pharmacodynamic effects may also influence the pharmacokinetic processes. Thus, an effect on respiration and circulation will also be of importance for the absorption, distribution and elimination of a drug.

Pharmacodynamic properties studied in Atlantic cod (Zahl et al. 2009) and Atlantic halibut (Zahl et al. 2011) include wide differences within and between these two species as well as between the agents. Induction time, defined as time of loss of equilibrium, occurred within 3 min for all agents tested, thus easily meeting one criterion of suitability for fish anaesthetic agents (Bell 1987; Marking and Meyer 1985). The recovery time differed widely between the agents. Metomidate produced the longest recovery time in Atlantic cod and was together with isoeugenol and 2-phenoxyethanol one of the agents that gave rise to the longest recovery time in Atlantic halibut. This agrees with pharmacokinetic data that indicate that a slower elimination and a prolonged recovery might be the expected outcome in agents of higher lipid solubility (Guenette et al. 2007; Hansen et al. 2003; Kiessling et al. 2009; Kildea et al. 2004). The recovery time in Atlantic cod reported by Zahl et al. (2009, 2011) was within the range defined for suitable anaesthetics, i.e. 5 min, (Bell 1987; Marking and Meyer 1985) for all agents except for metomidate, but exceeded this in Atlantic halibut. However, as the parameter ‘recovery time’ was recorded differently in Atlantic halibut and in Atlantic cod, values for recovery time need to be compared with caution. In pelagic fish such as Atlantic cod, recovery time is normally defined as time of recovery of equilibrium, whereas in Atlantic halibut, a bottom dwelling flatfish, recovery time is more difficult to assess. In the work by Zahl et al. (2011), recovery time was set at when the halibut managed to right itself after having been placed upside down in a recovery tank.

The agents studied by Zahl et al. (2009, 2011) differed greatly in their capacity to depress responses to external stimulation, assessed during handling of the anaesthetised fish and by a caudal peduncle pinch. Reflex reactions to the pinch and responsiveness to handling were most effectively reduced by two local anaesthetics MS-222 and benzocaine. This is probably related to the mode of action, as these substances suppress signal transmission in both central and peripheral nervous system. The result also tallies well with findings in higher vertebrates, in which systemic injections of local anaesthetics did not block peripheral nociceptive pathways, but merely increased the response threshold. Suppression of nerve signals at central level may have adverse effects. After being inhaled by the fish, the agents enter the circulation, where they may be distributed throughout the body. By passing the blood–brain barrier, these agents affect the brain and the centres that control respiration and circulation. A depression of respiration and circulation would produce hypoxaemia, metabolic acidosis and changes in blood electrolytes and lead to extended recovery time, thus enhancing the effects of the anaesthetic.

Pharmacokinetics

In order to induce a biological response, a drug must be present in a sufficient concentration at the receptor site for a certain length of time. The concentration achieved at the receptor site depends on the pharmacokinetic properties of the drug, i.e. how it is absorbed, distributed, metabolised and eliminated. In fish, drug administration via bath immersion is equivalent to inhalation anaesthesia in human and veterinary medicine. The processes of absorption and elimination are determined by the diffusion rate through the gill epithelium, which mainly depends on the lipid solubility, ionisation and polarity of the drug. Following absorption, a fraction of the drug binds to plasma protein, mainly albumin, while the rest remains unbound and pharmacologically active. The unbound drug is then distributed to various parts of the body, commonly separated into compartments, where it exerts its effect. The main compartments are plasma, interstitial fluid, intracellular fluid and adipose tissue. Lipid solubility and ionisation are important factors in determining the distribution of the drug. High lipid solubility results in rapid diffusion through the cell membranes and the possible accumulation of the drug within hydrophobic compartments such as the adipose tissue and the lipid bilayers of cell membranes, while ionised molecules and molecules with high polarity are less lipid-soluble and do not easily pass through biological membranes.

The ionisation of weak acids and bases is determined by the dissociation constant pKa, defined as the pH at which 50% of the acid or base is in ionic form (pKa, Table 3). When a compound is allowed to distribute itself between equal volumes of two immiscible liquids, the ratio of the concentration of the solute in two phases at equilibrium is called the partition coefficient (logP). For drugs, two phases used are octanol and an aquatic buffer with pH of 7.4 (Table 3). As the logP value describes the partition of neutral uncharged molecules, it is not an exact measure of the distribution of ionisable compounds such as many anaesthetic agents. Anaesthetic agents are usually associated with high lipid solubility, which may lead to accumulation in adipose tissue both during long exposures and following repeated exposures (Rang et al. 2003).

Table 3 log P (octanol:water), pKa and water solubility of anaesthetic agents
Full size table

In fish, anaesthetics are mainly eliminated via the gills in unchanged form (Hayton et al. 1996; Hunn and Allen 1974; Hunn et al. 1968; Meinertz et al. 1991; Stenger and Maren 1974). Some may also undergo metabolism before elimination (Hayton et al. 1996; Meinertz et al. 1991; Ryan 1992; Stenger and Maren 1974).

Models are used to describe the pharmacokinetic processes. Plasma concentration–time curves are employed to estimate the pharmacokinetic parameters and perform compartmental analyses to describe drug distribution. In the simplest compartmental pharmacokinetic model, the body is regarded as a single compartment throughout which the drug is distributed evenly following absorption. This model, called the one-compartment open model with first-order elimination, is a simplification used to illustrate basic pharmacokinetic principles. A two-compartment open model includes a peripheral compartment representing the tissue besides the central compartment, which represents the plasma, thereby introducing a distribution process into the model. From the central compartment, the drug can be removed both by distribution to the peripheral compartment and by elimination. The process of distribution is described by the initial rapid decrease in plasma concentration of the drug and is called the α–phase. When the distribution is complete, the elimination of the drug is determined by the slower β-phase.

The distribution of a drug between the plasma and the rest of the body is described by the pharmacokinetic parameter volume of distribution (Vd). The Vd indicates how a drug relocates into body compartments relative to the blood, and is a non-physiological volume that corresponds to the volume that would be necessary to dilute the drug to produce the concentration present in the blood. Lipophilic drugs that have a tendency to accumulate in body fat thus have high Vd values, while drugs with a high degree of plasma protein binding have low Vd values. The rate at which the drug is removed from the blood is described by the parameter clearance (C), which is defined as the volume of blood or plasma that is cleared per unit time. Together with clearance, the volume of distribution determines the plasma half-life (t 1/2) of a drug. A study by Kiessling et al. (2009) revealed large differences in the pharmacokinetics of various anaesthetics. MS-222, the most hydrophilic of the agents tested (Table 3), was distributed rapidly, followed by rapid clearance and elimination. The plasma data were best described by a one-compartment open model with first-order elimination, which indicates that distribution was homogenous and elimination proportional to plasma concentration. The plasma data of benzocaine and isoeugenol were best described by a two-compartment open model, indicating a rapid transfer of drug from plasma to tissues followed by the slower process of elimination from the plasma. Both benzocaine and isoeugenol were found by Kiessling et al. (2009) to be rapidly distributed, but their rates of clearance and elimination were much slower than for MS-222. An initial phase of rapid distribution followed by a period of slow elimination has also been found in rainbow trout anaesthetised with benzocaine (Meinertz et al. 1996; Stehly et al. 1998). The pharmacokinetic properties of benzocaine in rainbow trout were best described by a two-compartment model when benzocaine was administered through bath immersion, while a three-compartment model best described the properties of the distribution and clearance process following intra-arterial injection (Meinertz et al. 1996; Stehly et al. 1998). Guenette et al. (2007) found that eugenol, which is similar to isolegenol, was well absorbed and eliminated by rainbow trout. The plasma concentration versus time curve was biphasic, which indicates two compartments. A biphasic plasma elimination curve was also found in the elasmobranch dogfish (Squalus acanthias) following arterial injection of MS-222 (Stenger and Maren 1974). MS-222 was excreted rapidly across the gills, mainly in unchanged form. Kiessling et al. (2009) found that artificial ventilation had no influence on the compartment analysis. However, elimination rates were faster in the groups of fish that received artificial ventilation, demonstrating the importance of the gills as a pathway for drug elimination. This is in agreement with the previous reports in other species (Hayton et al. 1996; Hunn and Allen 1974; Meinertz et al. 1991) and emphasises the importance of respiratory function, i.e. exchange of water over the gills, during recovery (Table 4).

Table 4 Effects of different anaesthetic agents in fish
Full size table

Water temperature

Temperature is an important factor in determining the rate of physiological processes in ectothermic animals such as fish and thus plays an important role in the processes related to the uptake and elimination of drugs. A 10°C rise in water temperature (Q10) typically doubles the basal metabolic rate of teleost fish (Clarke and Johnston 1999). This leads to a higher oxygen demand that is met by enhanced respiration, increased cardiac output and increased blood flow through the gills (Graham and Farrell 1989; Nilsson and Sundin 1998; Webber et al. 1998). Poorer oxygen solubility due to rising water temperature leads to an additional need to enhance respiration and blood flow.

Temperature increases have been reported to shorten induction and recovery time for a number of anaesthetic agents in several teleost species, including benzocaine in striped bass (Morone saxatilis; Gilderhus et al. 1991), 2-phenoxyethanol and clove oil in European sea bass and gilthead sea bream (Mylonas et al. 2005), isoeugenol in rainbow trout (Stehly and Gingerich 1999), clove oil in rainbow trout, brown trout (Salmo trutta), Atlantic salmon, whitefish (Coregonus lavaretus), perch (Perca fluviatilis) and roach (Rutilus rutilus) (Hoskonen and Pirhonen 2004; Woolsey et al. 2004), and benzocaine, MS-222, metomidate and 2-phenoxyethanol in Atlantic cod (Zahl et al. 2009). Reduced induction time at higher water temperatures has also been demonstrated in common carp (Cyprinus carpio), rainbow trout, fathead minnows (Pimephales promelas) and Atlantic halibut anaesthetised with MS-222 and in Atlantic halibut anaesthetised with benzocaine (Hikasa et al. 1986; Houston and Woods 1976; Sylvester and Holland 1982; Zahl et al. 2011).

The rapid induction time seen at higher water temperature may be related to higher basal metabolic rate at higher temperatures (Clarke and Johnston 1999; Saunders 1963; Schurmann and Steffensen 1997) and the corresponding increase in oxygen demand leading to increased respiration and circulation. This facilitates a rapid absorption and distribution of anaesthetic as well as faster clearance rates but may also lead to a greater amount being absorbed. While faster clearance rates of both benzocaine and isoeugenol at higher temperature were found in rainbow trout and silver perch (Bidyanus bidyanus) (Kildea et al. 2004; Stehly et al. 1998), significantly higher concentrations of isoeugenol were found in rainbow trout fillets after exposure at 17°C than at 7°C (Meinertz et al. 2006). Zahl et al. (2009) report that Atlantic cod made a quicker recovery at higher water temperature and had a wider therapeutic window, indicating a more rapid clearance and elimination. Atlantic halibut, on the other hand, seem to recover more slowly at higher temperatures, but in spite of longer recovery time the fish displayed stronger reactions to handling, indicating a lack of anaesthetic effect (Zahl et al. 2011). In large Atlantic cod (1,000 g) anaesthetised at a low water temperature (8°C), the dosage of both metomidate and MS-222 had to be reduced, while for benzocaine no suitable dosage was found (Zahl et al. 2009). This was due to the fact that after five minutes of exposure to benzocaine the fish reacted both to handling and to the caudal peduncle pinch, but suffered respiratory arrest and died in the recovery bath. The findings in Atlantic cod agree with those in brown trout and goldfish, in which higher sensitivity was seen at lower water temperature during anaesthesia with benzocaine and 2-phenoxyethanol, respectively (Dawson and Gilderhus 1979; Weyl et al. 1996). However, the findings contradict with observations in striped bass, where higher dosages of benzocaine were used at lower water temperature for fish of large body weight (Gilderhus et al. 1991). In an evaluation of Aqui-S® in rainbow trout, reduced toxicity was found at higher temperature (Stehly and Gingerich 1999).

In the study on Atlantic halibut (Zahl et al. 2011), only MS-222 and benzocaine were tested at higher water temperature (15°C), but even so, the data indicate that in this species the central and peripheral effects of the anaesthetics are differently influenced by the temperature increase in comparison with Atlantic cod (Zahl et al. 2009). This might be explained in part by central effects in Atlantic halibut being more pronounced at higher water temperature, while the peripheral effects are less so, resulting in greater responsiveness and fish that regain equilibrium more slowly. Zahl et al. (2011) report that this effect is most prominent with benzocaine, which may reflect higher lipid solubility of the drug.

Anaesthetic agents are associated with high lipid solubility and a tendency to accumulate in tissues with a high content of fat (Rang et al. 2003). This might be a problem following repeated or long-duration exposures, for instance during larger surgical procedures.

Uptake of fat from the diet, storage in adipose tissue and the incorporation of lipids into membranes all vary with temperature. There is greater incorporation of unsaturated fatty acids and increased aggregation of hydrophobic substances in membranes as temperature falls (Berge et al. 1980; Cossins 1983; Ruyter et al. 2006; Sellner and Hazel 1982). Substances with higher lipid solubility such as benzocaine, isoeugenol and metomidate may therefore penetrate and possibly also accumulate in the lipid bilayer of cell membranes more readily at lower temperatures than more hydrophilic substances such as MS-222. Any accumulation would then alter the characteristics of the membranes and affect normal cell function by influencing transport mechanisms, stabilising the membranes and possibly inhibiting nervous signal propagation. If so, exposure time would influence this effect of membrane lipid changes and could explain some of the contradictory effects of temperature found in the literature. Anaesthetic depth has been associated with the concentration of anaesthetic agent in the brain. This has been reported for snapper (Pagrus auratus) and several freshwater species anaesthetised with MS-222, as well as sting ray (Dasyatis sabina) and lemon shark (Negaprion brevirostris) anaesthetised with quinaldine (Brown et al. 1972; Hunn 1970; Ryan 1992). The brain receives a large supply of oxygen-rich blood from the gills, and the blood–brain barrier is permeable to anaesthetic agents. Accumulation in the brain probably leads to an enhanced central effect, with depression of brain centres controlling respiratory and cardiovascular function as a consequence.

Body weight

The relationship between body weight and sensitivity to anaesthetics may be influenced by several characteristics related to body weight, such as age, body composition, growth rate and sexual maturity, as all of these characteristics influence the physiology of the fish. Since the major route of both uptake and elimination is through the gills, the rate of oxygen consumption, the ratio of body volume to gill surface area and the rate of gill perfusion are of great importance. These factors vary within and between species and with life stage and life style as well as with body size. The gill surface area decreases in relation to increased body weight (Muir 1969; Oikawa and Itazawa 1985; Oikawa et al. 1999). The basal metabolic rate falls relative to increasing body size, and larger animals thus have lower oxygen consumption relative to body size than smaller animals (Clarke and Johnston 1999; Schmidt-Nielsen 1984). The scaling relationship between metabolic rate and body weight is usually described by the equation Vo2 = a M 0.75b , where the metabolic rate (Vo2) scales as the 0.75 power of body mass (M b) and the constant a indicates the level of the resting metabolic rate (Schmidt-Nielsen 1984). The constant a may differ between species and varies in relation to lifestyle, and in ectotherm animals such as fish, it is highly dependent on ambient temperature. For teleost fish, a scaling exponent of 0.8 has been found (Clarke and Johnston 1999). Sedate and less active fish species have a lower resting metabolic rate than more active species (Morris and North 1984). Clarke and Johnston (1999) found that the resting metabolic rate of fish varied between taxonomic groups, gadoids having higher resting metabolic rates than pleuronectiformes and salmoniformes, although there was a large degree of overlap.

Reports regarding the importance of body weight for the effect of anaesthetic agents in fish are contradictory. Some studies show no relationship between body size and induction and recovery time (Houston and Woods 1972; Stehly and Gingerich 1999), while others indicate that such a relationship exists (Gilderhus and Marking 1987; Houston et al. 1976; Olsen et al. 1995). Hoskonen and Pirhonen (2004) observed a shortening of induction time with increasing body size in whitefish, but observed the opposite relationship in rainbow trout, and found no size-related differences in Atlantic salmon or brown trout. Weber et al. (2009) found that induction time increased with increasing body weight in Senegalese sole (Solea senegalensis) anaesthetised with 2-phenoxyethanol, metomidate and clove oil, while this was not found for MS-222. The dynamics of the recovery time were more complex and were only weight-related for MS-222 (Weber et al. 2009).

Zahl et al. (2009, 2011) found no uniform relationship between the effects of anaesthetic agents and the body weight of the fish in either Atlantic cod or Atlantic halibut. For Atlantic cod, the results suggest that the body weight of the fish, or factors related to body weight, is less important for the differences in effect than the properties of the anaesthetic agents themselves.

Stress

When fish are subjected to handling, such as crowding or dip-netting, they respond by attempting to escape, which increases oxygen consumption. Paralleling the alterations in behaviour, a cascade of physiological reactions makes up the physiological stress response, first termed the General Adaptation Syndrome (see e.g. Selye 1985). The physiological stress response in fish is similar to that described in other vertebrates and is characterised by three phases (see Wendelaar Bonga 1997, for details). Large species differences in the magnitude of the cortisol response to stress have been found (Barton 2002; Barton and Iwama 1991), and high and low responders within species have also been identified (Fevolden et al. 1999; Mommsen et al. 1999; Pottinger and Carrick 1999; Pottinger et al. 1992, 1994; Sumpter et al. 1986; Tanck et al. 2001; Trenzado et al. 2003).

Stressed animals display abnormal reactions to anaesthesia and may require larger doses for both induction and maintenance (Hall et al. 2001). In fish, increased physical activity while attempting to escape handling, followed by the cascade of physiological alterations of the stress response, contributes to facilitating anaesthetic uptake. The elevated levels of catecholamines and corticosteroids further stimulate ventilation and enhance cardiac output, in turn increasing gill blood flow (Farrell 1984; Graham and Farrell 1989; Nilsson and Sundin 1998). The perfusion of the gill lamellae increases and lamellae that in a normal resting state are barely or not perfused at all are recruited, expanding the respiratory surface area and facilitating diffusion, and thereby affecting the uptake of substances across the gill.

Several reports show that stress associated with handling of fish can be effectively reduced by anaesthetic agents. Metomidate has been found to reduce handling stress in Atlantic salmon, Chinook salmon (Oncorhynchus tshawytscha), hybrid striped bass (Morone chrysops x Morone saxatilis), channel catfish (Ictalurus punctatus) and red drum (Sciaenops ocellatus), but this may be related to the cortisol blocking side effects of this drug. Isoeugenol has been shown to be effective in reducing stress in Atlantic salmon and channel catfish and MS-222 inhibits the stress response to handling in rainbow trout and channel catfish (Davis and Griffin 2004; Iversen et al. 2003; King et al. 2005; Kreiberg and Powell 1991; Olsen et al. 1995; Small 2003, 2004; Small and Chatakondi 2005; Thomas and Robertson 1991). However, it has also been found that exposure to anaesthetics may itself induce a stress response. Of the agents discussed in this review, MS-222 has been found to induce stress in Atlantic salmon, rainbow trout, Atlantic cod, Atlantic halibut, channel catfish, striped bass, hybrid striped bass, gilthead sea bream and red drum, while isoeugenol and eugenol induce stress in Atlantic salmon, rainbow trout and hybrid striped bass, and benzocaine and metomidate have a similar effect in Atlantic salmon, Atlantic cod and Atlantic halibut (Barton and Peter 1982; Davidson et al. 2000; Davis and Griffin 2004; Davis et al. 1982; Iversen et al. 2003; Kiessling et al. 2009; Molinero and Gonzalez 1995; Olsen et al. 1995; Small 2003; Thomas and Robertson 1991, Zahl et al. 2010). However, confinement and relocation in connection with anaesthesia influence the responses to anaesthesia and may also mask the stress-inducing effects of the anaesthetic agents themselves (Hill and Forster 2004; Molinero and Gonzalez 1995; Thomas and Robertson 1991).

In a practical situation, stress, accompanied by the cascade of the physiological stress response, facilitates the uptake of anaesthetic by stimulating respiration and circulation and increasing the area of the blood/water interface. A doubling of oxygen consumption has been observed in coho salmon (Oncorhynchus kisutch), rainbow trout and Atlantic cod following handling, returning to basal levels within 1 h in coho salmon and 3 to 5 h in Atlantic cod (Barton and Schreck 1987; Davis and Schreck 1997; Saunders 1963). This probably results in an immediate and excessive uptake of anaesthetic, leading to faster induction and possibly also more deeply anaesthetised fish. In the study by Zahl et al. (2009), in which Atlantic cod were subjected to stress in the form of 30 s in air in a dip-net prior to anaesthesia, the fish displayed shorter induction time, became more deeply anaesthetised and recovered more slowly than unstressed fish. Moreover, the dosage of MS-222 had to be reduced in order to avoid mortality in the stressed group, but was too low to provide sufficient anaesthetic effect in the unstressed group. Both stress and sedation prior to anaesthesia may have marked practical consequences. This was well demonstrated by Kiessling and co-workers in a series of studies on feeding and growth after vaccination in salmon smolts anaesthetised with benzocaine with and without pre-anaesthetic metomidate sedation (Kiessling et al. 2001; Oppedal et al. 2000). They found that fish vaccinated using combination anaesthesia resumed normal feeding behaviour within a week, while fish subjected to the ordinary procedure with netting directly into the benzocaine bath without pre-sedation needed more than 3 weeks to resume normal feeding behaviour. At that time, the weights of the fish in the pre-sedated group were 22 per cent higher than the fish in the non-pre-sedated group.