Morbidity scoring after abdominal surgery

Abstract

Postoperative monitoring of pain and distress in small rodents is not standardized, and widely accepted score sheets are not available. Here we describe a score sheet used in abdominal surgery of rodents, with particular reference to procedures involving the liver.

The worldwide use of animals for medical research has resulted in the development of an enormous number of different ‘disease models’ in a few select species, primarily rodents. The focus of these animal models is on the simulation of a particular disease, either through genetic manipulation, application of drugs or surgical interventions. These experimentations can be a potential source of pain and suffering. The protection of animals used for scientific purposes is regulated by specific national laws, but the way these laws are implemented shows considerable variation among countries.

Despite the lack of standardized animal protection laws, the development and use of animal models should not only focus on the simulation of disease but also include an evaluation of proper pain management and surveillance. Much effort has been invested in the past in developing reliable protocols for the assessment of pain and analgesic treatments of laboratory animals.

Nevertheless, due mainly to uncertainty regarding pain and stress detection, and inconsistencies in the data obtained, no final protocols have been developed yet for appropriate and practicable monitoring and management of pain and suffering.

Current efforts into improvements in behavioral monitoring and pain assessment clearly indicate a need for action. As an example, the use of the buprenorphine opioid for postoperative analgesia is widely accepted. ¹ Although its half-life is longer than in other opioid–analgesics, it is less than 4 h, requiring a tight and repetitive application every 4–6 h by injection. It is obvious that the investigator would not like to follow such a schedule as it would involve applying the analgesia twice during the night. Also, repeated injections at short intervals may be stressful for the animals.^2–4 One alternative is to provide the drug through drinking water which requires a 10-fold higher amount of buprenorphine, due to the level of consumption and to the first-pass hepatic extraction of orally-applied buprenorphine. Furthermore, there is also the possibility of spillage of the drinking water with most of the current caging systems. In addition, it remains unclear, whether individual animals voluntarily consume enough of the medicated water to ensure permanent pain alleviation, especially during the daylight phase, when mice do not drink regularly. Nevertheless, providing analgesia with drinking water has become popular, particularly for short half-life opioids (e.g. tramadol) or non-steroidal anti-inflammatory drugs (NSAIDs, e.g. acetaminophen and paracetamol). ⁵ Thus, long-lasting forms of analgesia are badly needed. However, sustained–release formulations of painkillers are not commercially available to date in Europe, although these drugs have been developed for rodents, particularly buprenorphine which provides 1–3 days of continuous pain relief.^3,6,7

NSAIDs (e.g. meloxicam and carprofen) are often administered in mice; some of which are injected only 1–2 times per day based on the assumption that they provide pain relief for up to one day. ¹ However current publications have raised doubts about the analgesic efficacy of some NSAIDs. In particular the efficacy of commonly used dosages and the suggested length of action in each case remain unclear.^8–10 One of the problems arising from provision of pain-relief drugs to experimental animals is that these can interfere with experimental results. For example, in studies evaluating nociception or the effect of pain on a disease model, treatment with an analgesic may counter the purpose of the experiment. Similarly, when studying inflammatory diseases, medication including analgesics may affect the inflammatory process itself and hence may partially interfere with the experimental goal. ⁸ These examples indicate that an evaluation of efficient pain management is difficult and may require certain compromises.

New indicators of pain are necessary to improve the quality of pain assessment. This is not simple to achieve; and probably only the identification of behavioral traits can provide sufficient information during the simple cage-side monitoring, as is performed when score sheets are used. For example, audible vocalizations cannot be unambiguously interpreted: are they due to real pain or distress? Recent studies, aimed at the identification of objective parameters for the assessment of pain, analyzed the link between behavioral changes and alterations in heart rate and heart rate variability using sensors implanted to continuously monitor physiological parameters.^11,12 These experiments show a good correlation between cardiovascular reactions and signs of pain and distress. Monitoring long-term physiological parameters and associated behaviors by continuous telemetric or video recordings (lasting for hours or days) can only provide evidence of pain retrospectively. From a practical point of view such sensors cannot be implanted into each and every experimental animal, first because this would involve additional unnecessary stress and would undoubtedly be painful for some time after surgical implantation of the transmitter, but also because not all experimental situations would allow such a sophisticated operation, e.g. in very young animals. Additionally, it has recently been established that when alterations of specific spontaneous home-cage behavior such as burrowing and nest-building are observed, these are not only signs of disturbed well-being ¹³ or indications of the progression of neurodegenerative diseases,^14–16 but may also be signs of postoperative pain. However, performing these observations for the detection of postoperative pain may be time-consuming.^16,17 Thus short cage-side observations taken immediately after a painful insult have been used for many years, and allow the immediate treatment for pain. In practical terms, individual animals are monitored for approximately 10 min either in the home cage or at an observation area for very short-term aberrations of bodily appearance. These aberrations, such as press, stagger, stretch and twitch, are known to be indicative of specific types of pain, and are thought to be typical signs of abdominal pain in mice. ¹⁸ Such very specific symptoms are often combined into composite scorings, summarizing the most frequent and indicative alterations. ¹⁹ However, a well-known drawback of such cage-side observations is that the occurrence of aberrations does provide some evidence of pain, but absence of these during the short observation period does not necessarily prove that animals are pain-free.

Another approach to assess pain is the grimace scale as facial changes and contortions may indicate pain. ²⁰ This approach, although plausible, requires videotaping of individual animals, and and its use during routine monitoring is not yet validated. Such assessment requires adaptation to the experimental condition, animal strain, sex, etc. and the observer needs training and experience in order to recognize and categorize facial features. ²¹

Thus, methods that can reliably indicate pain are quite complicated, and need intensive analysis of individual animals and additional technical devices. Behavioral science has implemented quantitative approaches, based on statistical analysis of animal cohorts. Assessment of individual animals, however, is more subjective and may be biased due to the wish of investigators to continue without interfering with the experiment. Hence, implementation of robust surveillance protocols, which are quantitative, reproducible and allow longitudinal follow-up, are needed. Many investigators have established their own ‘score’ sheet based on the requirements of their local animal ethics committee. These score sheets are usually not included or detailed in publications. Here we propose a score sheet for postoperative surveillance that includes a quantitative approach with cut-offs defining termination of the experiment and euthanasia of the animal in case of unacceptable suffering. We are well aware that this score sheet is not exclusively based on scientific evidence but also on experience with several decades of animal experiments. The score sheet has been developed along with animal welfare officers and authorities in charge of regulations for animal experimentation.

Animal models in abdominal surgery

Surgical modifications of organs in the abdominal cavity include operations on the liver, e.g. partial liver resection, ²² liver transplantation, ²³ and bile duct ligation. ²⁴ Partial or full splenectomies, kidney transplantation, ²⁵ induction of strictures in the intestine, pancreatectomies, ²⁶ pancreatic duct ligation, ²⁷ etc. are other examples of intra-abdominal interventions. All these models have one thing in common, namely access to the abdominal cavity is via a laparotomy (incision through the abdomen). The postoperative time frame of these experiments can be between a few hours and several months, whereby pain is experienced within the first 2–3 days.

The most common procedure (e.g. partial hepatectomy) usually consists of a preoperative intraperitoneal injection of buprenorphine (0.1 mg/kg body weight), approximately 30 min before the operation. Subsequently, the animal is shaved and immobilized under inhalation anesthesia using isoflurane. The animal is fixed by tape on a stainless steel plate in a supine position. The plate is warmed with a heating pad, maintaining a 37–40°C temperature. Eye ointment is applied and the abdomen is opened by a sagittal incision after disinfection with alcohol. To access the liver, the sternum is pulled upwards and fixed with a 6-0 suture. The liver ligaments are then freed followed by individual ligation of vessels to liver lobes, which can then be dissected. After completion, the abdomen is then closed with a running suture of the peritoneal muscle/fascia followed by closure of the skin. During recuperation, the animal is placed in a warming cabinet (28–30°C) for a few hours or overnight, prior to its return to the regular cage.

As in human surgery, the impact of a specific type of intervention (e.g. partial hepatectomy) may be moderate to severe; and thus access to the abdomen, i.e. the laparotomy itself, may have a variable impact on the overall procedure. Furthermore, many procedures in human patients can be performed by laparoscopy (minimally invasive), which strongly reduces postoperative morbidity. Although minimally invasive approaches have been tested in rodents, routine use is not feasible.

In the current report, we focus predominantly on experiments concerning liver surgery. A classic example is a 70% resection of liver lobes. ²² This procedure is fairly standardized and benefits from the anatomy of the mouse or rat liver being quite lobulated. Therefore, parenchymal transections are not required because the lobes can be individually removed by ligation of the vessels. This procedure is well tolerated. The liver usually regains its original size within a week. In addition, extended hepatectomies (e.g. removal of 86% and 91% of liver tissue) have been established to explore the effects of insufficient liver volume. ²⁸ This situation can lead to a so-called small-for-size syndrome, the small liver is unable to regenerate, most probably due to the metabolic overload and concurrent accumulation of toxic products. ²⁹

The small-for-size syndrome, particularly with a full blown liver failure, results in reduced activity, apathy and encephalopathy (‘hepatic coma’). Pain assessment of these animals has been debated, as reports from patients with a similar syndrome do not complain of pain. In addition, analgesia may further impact on hepatic function, and is given with caution.

Assessment of welfare

Usually, pain is assessed by an investigator who observes the animals for a short period of time, looking predominantly for active behavior, ruffling of the fur and hunchback position. In a cage with several animals an individual repeated assessment can be inaccurate over time. Once an animal is identified that is clearly in pain, it should properly be treated with analgesics. This leads to the following questions: Should all animals in the experiment receive the same treatment to ensure consistency? Would it be better to treat all animals from the start with analgesics to avoid having to change the experimental protocol based on an individual animal? Alternatively, if a single animal shows abnormal behavior and pain, would it be better to eliminate the animal from the experiment? These are questions that should be clarified before beginning an experiment, particularly if the animals are valuable due to complex genetics or require special treatment, e.g. diabetic mice receiving insulin.

Below we describe a score sheet used for monitoring animals after abdominal operations, particularly on the liver (see the score sheet, which is available online with this article LAN.sagepub.com).

Postoperative score sheet

For every animal undergoing abdominal surgery, a score sheet is prepared which shows animal ID, protocol ID, date and prospective severity degree (CH 0–III). The type of intervention should also be indicated. This basic information will allow a replacement investigator to take over surveillance, particularly if an animal caretaker or veterinarian has reported that an animal has taken ill.

Depending on the protocol, the severity or expected health issue, the investigator may be asked to supervise the animals very closely, i.e. twice a day. For proper observation, it is advised to take each animal out of its cage and place it on top of the grid or lid, as is deemed best for observation of the animal. If there are only one or two animals per cage, observation may be performed without touching the animals, however, they must be freely visible. Overall, a solid knowledge of the normal behavior and appearance of a mouse of the respective strain, sex and age is a prerequisite for proper monitoring.

Activity

In the following, scored behaviors and signs of distress are described. Activity may be judged taking into account the time of day, which affects the diurnal activity pattern. Absence of activity is certainly an indicator of impaired condition and health. However, normal or even increased activity is not necessarily an indicator of the absence of pain or suffering, and it is mandatory to observe other parameters or symptoms to assess the whole picture. Aggressive animals may be observed as being very active, and aggression is a potential indicator of pain. Reduced activity is given one point while immobility is a strong factor and should be given three points. Increased activity is very subjective and might be induced by buprenorphine treatment ³ and therefore is not given any points in our score sheet.

Breathing

Breathing rate and depth are not easy to assess in mice without specialized equipment. However, animals with flat breathing but normal frequency are given one point; while shallow or rapid breathing is an indicator for more severe physical impairment (two points).

Fur

The appearance of fur (piloerection and ruffling) is perhaps the easiest parameter to judge. Ruffled, unclean or matted fur indicates absence of normal grooming behavior. This is either based on physical impairment (through the experiment) or a stressful situation in the cage, particularly, if several males share a cage. In this case, presence of bites, wounds or aggressive behavior may indicate social imbalance. Such an imbalance can be prompted if one or two animals of the cohort are operated on and may be weakened, resulting in a change of the social hierarchy. It is therefore recommended that operated animals should not be mixed with untreated, healthy littermates during the recovery period, i.e. during the first one to maximally four postoperative days only operated animals should be kept together. Ruffled fur receives one point.

Posture

Another parameter which is quite easy to gauge is posture, such as hunchback or arching behavior. This is seen when the animals are in strong abdominal pain and is found in combination with ruffled fur. Contradictory reports exist on the interpretation of hunched posture in mice,^18,30–33 and it was recently described as an unreliable indicator of pain.^18,33 Based on our previous results on laparotomy ¹¹ we have assigned one point for hunchback or arching behavior.

Jaundice

The presence of jaundice indicates loss of liver function, and this may be due to a direct interference with the liver during surgery or pharmacological treatments targeting the liver. It appears as a yellowing of the skin. For mice with black or dark fur, the eyes or the footpads easily allow an assessment. If jaundice is part of the pathophysiology of the model, bilirubin should be measured regularly by blood analysis. If another organ is targeted, blood parameters can be assessed, creatinine for kidney, amylase for pancreas, T3/4 for thyroid, etc. This parameter receives one point.

Operative site: assessment of the wound

Wound infections are rare in rodents. In contrast to humans and larger animal species, particularly pets, the wound cannot be covered by a wound dressing. It is therefore crucial that the wound be properly closed to avoid entry of foreign material, or worse, extrusion of an intestinal loop. For proper postoperative examination, particularly during the first two days, mice have to be removed from the cage and individually inspected.

Mice tend to bite and gnaw at their wounds particularly when pain medication is insufficient, further enhancing chances for infection. However, overdosing with buprenorphine might also lead to aberrant behavior where the animals mutilate the wound (e.g. removal of sutures).

An open or infected wound receives two points. Additional criteria specific to the experiment can be introduced and scored.

Overall score

At conclusion of an assessment, points should be added up. A mouse having two or more points should receive pain treatment. Where it is given five points, a critical physical impairment has been reached which should receive particular attention. Adequate pain management should be provided in both cases, i.e. a protocol (giving detailed advice on drug, dosage, administration route and interval) must be in place, which has been developed and tailored for the model (e.g. interference with experimental goals) and the type of pain. Typically, we administer a bolus injection of buprenorphine (0.1 mg/kg subcutaneously) two or three times a day. If no improvement is observed within the next 24 h, the animal should be euthanized. Below is a summary of the criteria leading to termination. Some are independent criteria, e.g. cachexia (short- and long-term weight loss) or if the animal is self-mutilated.

Summary of termination criteria:

The animal does not recuperate from the intervention (immobile, unresponsive to stimuli 2 h after the operation).

Score ≥ 6 points.

Score = 5 points: plus no improvement within the next 24 h.

The animal loses >15% of body weight within 12 h after operation (short-term weight loss).

The animal loses >20% of body weight compared with the start of experiment (long-term weight loss).

Animal is self-mutilated.

Animal exhibits signs of pain despite buprenorphine treatment for three days.

Animal does not respond to pain treatment within 24 h.

Refinement opportunities

An animal is transiently incapacitated and may lose its social status following an invasive procedure of the abdominal cavity. It is imperative that the animal has opportunities to create a refuge in the form of nest building or burrowing material or shelters. Particularly with male animals, short-term isolation during intervention and physical impairment may severely disturb the social structure. It should be evaluated whether a single individual animal should be placed back in the cage with its healthy littermates. It is reasonable to presume that if all animals in a cage are operated on at the same time, the individual animal will suffer less from the possibly disrupted social structure, or may even profit from the social support. ³

Recommendations and further 3R research

Historically, surgeons have preferred to operate on male mice. However with their tendency to aggressive behavior, increasing with age, female mice might be more amenable. In particular, studies addressing age-related topics would greatly benefit from the inclusion of female mice. The social structure of the rat is different and males are less aggressive. Here, the need to focus on one sex is not as crucial.

In both species, it might be advisable to use both genders in future experiments. First, to exclude conclusions that may or may not apply to both sexes, and second, the use of all animals reduces the overall number of animals to be produced. Currently, a vast number of animals are still killed since the demand for one sex is much greater for certain animal models.

Conclusions

Proper animal surveillance in routine settings is only possible if the process is easy and quick. Otherwise the likelihood is that investigators will not assess the well-being of the animals thoroughly. This is to the detriment of the animals. It may ultimately affect the experimental outcome negatively and prevent relevant conclusions. The lack of diligence may also be a result of mice or rats being observed simply as ‘happily’ and actively running around the cage, which may imply that these animals are not in pain. This misconception is still commonly made by many investigators and should be rectified through education and training.