At the surrender of Singapore on 15 February 1942, over 100,000 men became prisoners of the Japanese. This included many medical officers who, over the ensuing years, treated men (and some women) suffering the ravaging effects of disease, worsened by the inhumane conditions of captivity. Many medical officers stand out for their work. One in particular, Jacob Markowitz, developed a blood transfusion service, under the most extreme conditions, for the sick working as slave labour on the Burma Railway. Although he qualified 20 years before the outbreak of war, little has been written of Markowitz's early life, or of the impact of this on his war-time contributions.

This article details the remarkable life of Dr Thomas Somerville, who qualified both as a veterinary surgeon and medical practitioner, served in two world wars and was recommended for the nation’s highest award for gallantry. In doing so, it records the life of a man whose repeated gallantry on the battlefield has been overlooked.

The anti-globulin test was described in 1945, and ever since has been synonymous with the lead author, Robin Coombs, a young veterinary surgeon, at that time embarking on a career in immunological research. This was marked by a number of important contributions in the field, including the description and categorisation of hypersensitivity reactions, co-authored with Philip Gell. Together they wrote the classical text, Clinical Aspects of Immunology, which has been updated and republished over the ensuing 50 years. Although Robin Coombs is best remembered for his contributions to medical immunology, he made a number of significant early advances in the field of veterinary immunology.

The science of surgery is constantly evolving, based on an ever increasing and challenging research base. Readers will be aware of the increasing proportion of papers published in this journal reporting clinical trials, the results of which may quickly translate into surgical practice. However, it would be wrong to overlook the important contributions that have been, and continue to be, made in surgical research by experimental studies involving animals. Although the results may take time to translate into clinical practice, there can be no disputing that progress in areas such as organ transplantation, prosthetic materials and, more recently, immunotherapeutics are all the result of animal studies. The accurate and complete reporting of surgical research underpins the veracity of the research and reflects the ethical standards of this journal. Moreover, it allows both reviewers and readers fully to comprehend the reason for the research, how it was conducted and the relevance of the results. With the expansion of clinical trials during the past decade there has been the development of guidelines for publication, most notably the Consolidated Standards of Reporting Trials (CONSORT), adopted by this journal in 2010. At the same time, the reporting of animal-based studies has come under scrutiny, with many papers published in peer-reviewed journals found to lack details in a number of areas, ranging from the description of the strain, weight and/or sex of animals, to a lack of clarity in experimental design and statistical analysis. More recently guidelines have been published under the acronym ARRIVE (Animal Research: Reporting In Vivo Experiments)1. These represent a checklist of what is suggested as ‘minimum information’ to be contained in any publication reporting research involving experimental animals, ranging from a description of the animals used and the procedures they were subjected to, through to details of outcomes as well as any adverse events. These guidelines have been adopted by a number of international journals as well as major funding bodies in the UK. When reporting studies involving experimental surgery there is a need for a full description of the experimental procedures, including perioperative care. For both ethical and scientific reasons there is a need to prevent and/or alleviate pain in any experimental animal; furthermore, in many areas of the world there is a legal requirement to do so. Just as in human surgery, where there have been advances in pain control in recent years, the same is true in animals. It is now recognized that mammals, regardless of their size or position in the evolutionary tree, are able to experience pain2. It is, therefore, necessary in any publication reporting experimental surgery that authors clearly demonstrate that animals have been appropriately anaesthetized, and to do so researchers need to be aware of developments in the field of laboratory animal anaesthesia3; no longer can journals accept papers where there is evidence of inappropriate anaesthesia. Information is readily available in established texts, and researchers need to be aware that an anaesthetic regimen appropriate for human patients may not be so for animals. Monitoring of anaesthesia is also a necessary requirement during surgical procedures in the experimental animal, just as it is in humans. Appropriate intraoperative care is a further point for consideration. Readers will all be aware of the special requirements for paediatric surgery; many of those requirements are applicable to the anaesthetized rat, weighing some 250 g. Prevention of hypothermia both during and after surgery is critical to ensuring homeostasis4; failure to do so will readily contribute to poor recovery, and with it poor experimental outcomes. Perhaps the area most overlooked is that of analgesia. The ability of vertebrates to experience pain is now recognized from physiological studies, with these results influencing legislation controlling the use of animals in research in some countries5. There is no doubt that surgical procedures are painful; experimental studies have clearly demonstrated that a rat recovering from a laparotomy will experience pain that can be alleviated by analgesics6. Our knowledge of pain recognition and alleviation in animals has advanced considerably in recent years with the development of scoring systems based on objective measurements and assessments, such as weight change or locomotor activity, which have proven valuable in monitoring the response to analgesic administration7. Perhaps most illuminating in this area has been the clear demonstration that animals

propria and submucosa.Could it be that the higher rates of anastomotic dehiscence reported for ileal anastomoses are related to technical factors?The authors also give insufficient information on anaesthetic technique and monitoring.There is increasing awareness that intra-and perioperative factors, including mode of anaesthesia, can affect postoperative recovery and morbidity.Laboratory rodent work exists showing that, for intestinal anastomoses, periods of hypovolaemia and hypoxia may be detrimental to healing 4 .Similarly, pain and distress in animals should be minimized but de Hingh and workers provide no evidence concerning administration of analgesics, particularly to control animals not receiving celecoxib.Before attributing anastomotic failure directly to COX-2 inhibition, a more critical appraisal of the laboratory work is required in order to achieve watertight conclusions in an experiment, the results of which may eventually be extrapolated to man.

Abstract The Editors welcome topical correspondence from readers relating to articles published in the Journal. Responses can be sent electronically via the BJS website (www.bjs.co.uk) or by post. All letters will be reviewed and, if approved, appear on the website. A selection of these will be edited and published in the Journal. Letters must be no more than 250 words in length. Letters submitted by post should be typed on A4-sized paper in double spacing and should be accompanied by a disk.

This study aimed to develop a reproducible model of phosgene-induced lung injury in the pig to facilitate the future development of therapeutic strategies. Ten female young adult large white pigs were used. Following induction of anaesthesia using a halothane/oxygen/nitrous oxide mixture, arterial and venous catheters were inserted together with a pulmonary artery thermodilution catheter, and a suprapubic urinary catheter by laparotomy. Anaesthesia was maintained throughout the experiment by intravenous infusion of ketamine, midazolam and alfentanil. On completion of surgery the animals were allowed to equilibrate for 1 h and then were divided into two groups. Group 1 (n = 5) was exposed to phosgene for 10 min (mean Ct = 2443 +/- 35 mg min m(-3)) while spontaneously breathing, whereas control animals (Group 2 n = 5) were exposed to air. At 30 min post-exposure, anaesthesia was deepened in order to allow the initiation of intermittent positive pressure ventilation and the animals were monitored for up to 24 h. Cardiovascular and respiratory parameters were monitored every 30 min and blood samples were taken for arterial and mixed venous blood gas analysis and clinical chemistry. A detailed post-mortem and histopathology was carried out on all animals following death or euthanasia at the end of the 24-h monitoring period. Control animals (Group 2) all survived until the end of the 24-h monitoring period with normal pathophysiological parameters. Histopathology showed only minimal passive congestion of the lung. Following exposure to phosgene (Group 1) there was one survivor to 24 h, with the remainder dying between 16.5 and 23 h (mean = 20 h). Histopathology from these animals showed areas of widespread pulmonary oedema, petechial haemorrhage and bronchial epithelial necrosis. There was also a significant increase in lung wet weight/body weight ratio (P < 0.001). During and immediately following exposure, a transient decrease in oxygen saturation and stroke volume index was observed. From 6 h there were significant decreases in arterial pH (P < 0.01), P(a)O(2) (P < 0.01) and lung compliance (P < 0.01), whereas oxygen delivery and consumption was reduced from 15 h onwards in phosgene-exposed animals. Mean pulmonary artery pressure of phosgene-exposed animals was increased from 15 h post-exposure, with periods of increased pulmonary vascular resistance index being recorded from 9 h onwards. We have developed a reproducible model of phosgene-induced lung injury in the anaesthetized pig. We have followed changes in cardiovascular and pulmonary dynamics for up to 24 h after exposure in order to demonstrate evidence of primary acute lung injury from 16 h post-exposure. Histopathology showed evidence of widespread damage to the lung and there was also a significant increase in lung wet weight/body weight ratio (P < 0.001).

The use of hypertonic saline Dextran (HSD) for resuscitation following trauma has many potential benefits, especially for the treatment of military casualties, but there is very limited data on the responses following multiple dosing with this fluid. The effects of a second dose of hypertonic saline dextran for resuscitation of uncontrolled intra-abdominal haemorrhage were studied in an experimental model. Under general anaesthesia, 17 large white pigs (weight range, 48-67 kg) were subjected to uncontrolled hemorrhage produced by a tear in the common iliac artery. This resulted in a significant (P < 0.01) reduction of both mean arterial pressure (MAP) and cardiac index. One hour after injury, animals were assigned to one of three groups, receiving either no resuscitation (Group A), or two doses of HSD (at a dose rate of 4 mL/kg), administered either 1 and 4 h after injury (Group B) or 1 and 7 h after injury (Group C). Animals were monitored for 12 h post-injury. A second infusion of HSD caused a significant hypernatremia and diuresis (P < 0.01) in both Groups B and C. There was a non-significant rise in MAP in both treatment groups, and in Group B only, there was a significant increase in cardiac index (P = 0.014). It is concluded that repeat administration of HSD for the resuscitation of uncontrolled hemorrhage results in limited cardiovascular improvements, but that the metabolic sequelae are potentially detrimental to survival.

The management of open fractures requires excision of all devitalised tissues, both bony and soft tissue, and failure to do so is likely to increase the risk of infection. This study evaluated the applicability of laser Doppler flowmetry for the objective evaluation of fracture fragment viability in an experimental open ballistic fracture over a period of 12 h. The results indicate that this technique could not be used to distinguish between vascularised and non-vascularised fragments at any time, and did not aid the surgeon in their decision making at the time of wound excision. Subjective evaluation, based upon the degree of soft tissue attachment of fragments, was a far better indicator of fragment vascularity, although it had a relatively low specificity. There remains the need for education and training for trauma surgeons in the evaluation of fragment viability to ensure adequate wound excision as part of fracture management.

This study allowed the development of an in vitro model of a high-energy ballistic fracture. Direct fractures are more heavily contaminated than indirect fractures, and the spread of contamination is more extensive than examination of the wounds, particularly the entry wounds, would suggest. The major spread of contamination is along tissue planes, and we would recommend that these are thoroughly lavaged in the management of the fracture. The fracture site is heavily contaminated, with endosteal spread in direct fractures. However, there appears to be relatively little bony contamination beyond the fracture site.

An ovine model was developed to study the outcome following intramedullary nailing of a heavily contaminated fracture. Animals in the control group received no treatment directed at controlling infection, whereas those in the treatment group received wound debridement, lavage and the use of appropriate systemic antibiotics. Despite this, infection developed at the osteotomy site and along the entire length of the implant in all animals in both groups. It was felt that standard methods of wound management might be inadequate when applied at the time of intramedullary nailing since the spread of contamination following the use of reamers was demonstrated in all animals. Results from this study support the clinical impression that heavily contaminated fractures should not be treated by primary intramedullary nailing.

Implant sepsis, due to previous external fixator pin track infection, is the most common complication of secondary intramedullary (IM) nailing of the tibia. We have developed an animal model, which allows different treatment methods to be studied. Using an established ovine model of a pin track infection, Staphylococcus aureus was used to infect the external fixator pins, two weeks prior to reamed IM nailing. In the control group, the animals were killed at a mean of 10.5 days following nailing, when widespread infection was evident, with septic arthritis, abscess formation, and infection of the entire length of the tibia in all six animals. In the treatment group, before IM nailing, the pin sites were débrided, and both local and systemic antibiotics were administered. All surgical wounds healed without evidence of infection, 4 of the 6 animals survived for 28 days, and bacteria were only isolated from 1 of the 6 implants. Treatment was successful at reducing, but not eliminating, infection after secondary nailing.

Data from civilian trauma practice would support the non-operative management of low-energy transfer ballistic wounds to bone. However, there is a paucity of data to indicate whether this would be an effective treatment for similar injuries in military casualties, in whom wounds are heavily contaminated and there is a delay to receipt of medical treatment. We created a model of ballistic wounding in the proximal tibia of pigs. Fractures were contaminated withStaphylococcus aureus. Animals were divided into two groups; Control (n = 8), received no treatment, Treatment (n = 6), received 600 mg benzylpenicillin and 500 mg flucloxacillin by intramuscular injection for seven days, commencing six hours after injury. Animals were killed 14 days after injury. Treatment with antibiotics caused a statistically significant reduction in the incidence of osteomyelitis assessed by bacteriological examination. The results have important implications for the management of low-energy transfer wounds of bone in both civilian and military casualties.

BACKGROUND:The mortality rate from sepsis has improved little over the past two decades. One reason for this has been the use of flawed or inappropriate experimental models in preclinical sepsis studies.METHODS:A literature review of animal models of sepsis was performed following a Medline search based on the following medical subject headings: disease models, endotoxin, inflammation, peritonitis and sepsis. Additional references were identified from the papers identified in the search.RESULTS AND CONCLUSION:Many animal models of sepsis have been described but none has proved to be superior. Extrapolation of results from endotoxicosis or bacterial infusion models should be regarded with caution. Peritonitis models should be accepted as the 'gold standard' but the use of appropriate virulent bacterial species needs to be ensured. A standardized panel of animal models for the preclinical assessment of immunomodulatory agents should be established, including at least one immuno- suppressed model to simulate the immunocompromised patient with sepsis. A uniform and valid definition of sepsis applicable to both small and large animal species is required.

An animal model was used to study the pathogenesis of pin track infection. The roles of fluid accumulation around the pin/bone interface and mechanical loosening of the pin were specifically studied. In addition, the spread of bacteria in relation to the clinical appearance of the pin track was assessed. This study demonstrated that fluid accumulation around the interface is an important factor in the spread of infection from the superficial wound track to the medulla of the bone. Spread occurs despite the absence of mechanical pin loosening, and can occur before any clinical features of infection are apparent.

We used laser Doppler flowmetry (LDF) to measure flux in cortical bone fragments as a method of determining their vascular status and viability. In an experimental tibial osteotomy measurements of flux were made from specific cortical sites both before and after osteotomy. Flux levels fell rapidly in non-vascularised fragments and remained significantly reduced throughout the experiment. By contrast, those in vascularised fragments were significantly reduced one and two hours after the osteotomy but then increased. From three hours after the osteotomy there was no significant difference in flux levels between the vascularised fragments and proximal bone stock. We conclude that measurement of bone flux by LDF may have a role in the objective evaluation of the viability of bone fragments, but that further studies are required to validate the technique before its adoption in the management of the injured patient.