The spleen was a source of intrigue to ancient physicians and philosophers alike. Galen described the spleen as an organ of mystery, and believed that the spleen extracted ‘melancholy’ from the circulating blood and from the liver, and returned it to the stomach after a process of purification. Although it was noted by Aristotle that the asplenic state is compatible with life, the complexity of its function and anatomy are still a major topic of medical and scientific study. Thus the care of patients with splenic disorders often requires the care of a multidisciplinary team of haematologists, pathologists, and surgeons.

At 5 weeks of gestation, the part of the gut tube destined to become the stomach appears as a fusiform swelling in the foregut. It is attached to the dorsal body wall by a peritoneal fold, the dorsal mesogastrium. During the following weeks the stomach rotates in both its longitudinal and anteroposterior axes. As a result, the pyloric end moves to the right and the cardiac to the left, the left side faces anteriorly and the right posteriorly. These rotations result in the dorsal mesogastrium being pulled out and pouched to the left of the median plane. The pouch developed in this way lies behind the stomach and forms the omental bursa. At the same time, the fetal splenic tissue develops from condensations of mesoderm in the dorsal mesogastrium. This condensation has the effect of dividing the mesogastrium into two parts, that between the stomach and the fetal splenic tissue forming the gastrosplenic ligament and that between it and the kidney becoming the lienorenal ligament. The mesenchymal condensations then fuse to form the spleen.

Simple splenic agenesis is a rare isolated abnormality but is found in 4.0 per cent of children with congenital cardiac disease. As the spleen develops from independent collections of mesoderm which then fuse, both accessory spleens (splenunculi) and polysplenia can occur. Splenunculi are found in approximately 10 to 30 per cent of the population, and are therefore one of the most frequently found congenital anomalies. They usually occur in the gastrosplenic ligament and the greater omentum. The spleen may retain its fetal lobulated form or show deep notches on its diaphragmatic surface. In the rare condition of polysplenia, two to nine distinct parts of the spleen are found due to failure of splenic fusion.

The spleen is a lymphatic organ situated in the left hypochondrium between the gastric fundus and the left hemidiaphragm. Its long axis is in the line of the tenth rib and the hilum is in the angle between the stomach and the left kidney and makes contact with the tail of the pancreas. It is invested by visceral peritoneum. The diaphragmatic surface is moulded into a reciprocal convexity and the visceral surface has impressions from the stomach, left kidney, pancreas, and splenic flexure. Its size and weight vary depending on age and in different conditions, but a normal adult spleen weighs approximately 150 g (range 80–300 g), and measures 12 × 7 × 3 cm.

The long axis of the spleen lies along the tenth rib. The lower pole does not normally project beyond the midaxillary line. When the spleen enlarges the long axis extends down along the tenth rib and the anterior border approaches the costal margin. The spleen must at least treble in size before becoming palpable, which it does by passing in front of the splenic flexure. During clinical examination the spleen is palpated as a left upper quadrant mass which is ballotable, and is identified by the splenic notch and the lack of resonance to percussion over the mass ( Fig. 1 2233 and see also Fig. 4 2236).

The arterial supply to the spleen passes through the splenic artery, a tortuous branch of the coeliac axis (usually from a common stem with a hepatic artery), running along the superior surface of the body and tail of the pancreas. The short gastric and left gastroepiploic branches of the splenic artery pass between the layers of the gastrosplenic ligament. The splenic artery divides at the splenic hilum into superior and inferior branches which then further divide into four or five segmental branches, each serving one segment of spleen. Radio-opaque dye injection and corrosion casting techniques demonstrate each segment to be distinct, with a separate arterial supply and venous drainage. Such studies have clarified the existence of pyramidal segments in the splenic polar regions and wedge shaped segments in the central part of the spleen. Further subsegmentation has also been demonstrated, being similar in appearance and blood supply to accessory spleens or splenunculi. Vascular anastomoses between segments are uncommon. This knowledge concerning the arrangement of the segmental vasculature is important when considering splenic salvage surgery (see below).

The splenic vein forms from five or more tributaries leaving the splenic hilum. It then runs behind the pancreas to join the superior mesenteric vein, thus forming the portal vein. Its tributaries correspond to the arterial branches. The lymphatic drainage of the spleen comprises extensive efferent vessels in the white pulp which run with the arterioles and emerge from nodes at the hilum. These nodes drain through the retropancreatic nodes to those at the coeliac axis. There are no afferent vessels. Sympathetic nerve fibres run from the coeliac plexus and mainly innervate branches of the splenic artery.

The spleen is surrounded by serosa and a collagenous capsule which contains smooth muscle fibres penetrating the parenchyma. These trabeculae, with the reticular framework, support the cells of the spleen and surround the vessels in the splenic pulp. The trabeculae are dense connective tissue fibres which are rich in collagen and elastic tissue. In many mammals they contain non-striated myocytes but in man these are few in number. Between the trabeculae there is a network of reticular fibres supporting the splenic parenchyma, which consists of the red and white pulp.

On macroscopic examination of the cut surface of the spleen, the white pulp appears as discrete nodules embedded in a red matrix, the red pulp. The red pulp consists of the venous sinuses and the splenic pulp cords of Billroth. The sinuses are lined by unusual elongated spindle shaped endothelial cells and are large, complex, blood containing cavities embedded in macrophage rich tissue attached to the splenic reticulum forming the cords. The cords contain dendritic and fibroblastic reticular cells. In contrast, the white pulp comprises the periarteriolar lymphoid sheaths (see below) and the adjoining follicles (malpighian bodies) containing a germinal centre.

The distribution and organization of the red and white pulp is on a complex vascular arrangement (Fig. 2) 2234. Blood flow from the splenic artery is to the trabecular arteries from which central arteries supply the white pulp. The adventitia of the central arteries is then replaced by a periarteriolar lymphoid sheath and at various points these are enlarged to form the splenic follicles. When stimulated by antigen the follicles form germinal centres, which are similar in structure to lymphoid follicles. The central arterioles lie to one side of the germinal centres. They then lose their periarteriolar lymphoid sheath and continue into the red pulp as several very straight branches or penicilli. The arterioles become smaller and some gain a sheath of phagocytic cells. The fate of the terminal capillaries is controversial: they either enter a cord (open circulation), or a sinus (closed circulation). Present evidence supports either theory and it seems likely that the two circulatory routes coexist. In the normal spleen it is probable that most of the flow is via the closed route, allowing a rapid transit time of blood through the organ (1–2 min). In contrast, in the diseased spleen, flow may well be predominantly via the slower ‘open’ system, when transit times of 30 to 60 min have been recorded, with blood pooling in the splenic cords. The pulp venules collect the blood from both the sinuses and the cords and carry it to the trabecular veins and hence to the hilum and the splenic veins.

The spleen removes effete and damaged red cells from the circulation as well as particulate foreign matter, microbes, antigens, and cellular debris. This process occurs in the sinuses and the splenic cords by the action of the endothelial macrophages. In addition, intact cells are held up and siderotic granules, as well as Howell-Jolly and Heinz bodies (nuclear remnants and precipitated haemoglobin or globin subunits respectively) are removed before the red cells are returned to the circulation.

The spleen comprises the largest single accumulation of lymphoid cells in the body, containing 25 per cent of the total T lymphocyte population and 10 to 15 per cent of the B lymphocyte population. Blood-borne cells and antigen are emptied into the marginal zone of the white pulp, where antigen is taken up by the follicular dendritic cells and presented to immunocompetent cells leading to antibody production by plasma cells and an increase in size of the germinal centres in white pulp.

This function is less marked in man compared with other species but the spleen does contain a large volume of blood (approximately 8 per cent of the red cell mass) either in the venous sinuses or in the reticular meshes of the splenic cords. During emergencies such as anoxia large volumes of blood can be discharged into the circulation. Enlarged spleens may contain a much larger proportion of the blood volume (up to 40 per cent).

The red pulp contains groups of myelocytes, erythroblasts, and megakaryocytes. From the fourth month of intrauterine life some degree of haemopoiesis occurs in the human fetus within the spleen, although this is a minor site compared with the liver. Although a large bulk of lymphoid tissue, the mature spleen is not a major site of lymphopoiesis. However, stimulation in the white pulp by antigenic challenge does result in the proliferation of germinal centres which contributes to the circulating pool of competent T and B cells and macrophages. In disease states, especially in myeloproliferative disorders, thalassaemias, and chronic haemolytic anaemias, this may also occur.

Damage to the spleen may result from accidental or iatrogenic trauma, or may present as a ‘delayed’ or ‘spontaneous’ rupture. It has been estimated that more than 20 per cent of all splenectomies are performed for damage to the spleen occurring during the course of another abdominal operation. The most common type of non-iatrogenic injury to result in splenic trauma is non-penetrating blunt trauma to the upper abdomen, often associated with fractures of the left lower ribs. Delayed rupture is said to occur when the clinical signs develop after a delay of at least 48 h from the initial injury. It is generally believed to result from tearing of the capsule by the expansion of a subcapsular haematoma. The clinical management and presentation are the same as for immediate rupture. ‘Spontaneous’ rupture of a diseased spleen usually results from trivial trauma, often forgotten by the patient. It occurs most commonly as a complication of malaria, but is the most common cause of death in infectious mononucleosis.

If the trauma is localized to the spleen, in the absence of other intraperitoneal damage, physical signs depend on the degree of blood loss and vary from minimal left upper quadrant tenderness in an otherwise well patient, to signs of peritonitis accompanied by evidence of shock. Peritoneal lavage can aid the diagnosis of intra-abdominal bleeding, especially in cases which are otherwise difficult to assess, such as an obtunded patient and those with multiple potential sites of blood loss. After clinical assessment of the degree of blood loss and appropriate resuscitation with colloid or blood, a management plan as suggested by the algorithm in Fig. 3 2235 can be followed. The stable patient should be investigated with abdominal ultrasound, CT scanning, or splenic scintiscanning.

Laparotomy is indicated for splenic trauma if there is obvious evidence of continuing blood loss despite adequate resuscitation, or if there is clinical suspicion of additional trauma to other organs such as the liver, pancreas, or bowel. In several series of patients with abdominal trauma in adults, the incidence of major associated intra-abdominal injuries with blunt splenic trauma is approximately 30 to 60 per cent. If there is any doubt as to the presence of an associated intra-abdominal injury, laparotomy should not be delayed. The surgical treatment of these injuries is dictated by the operative findings. At laparotomy the decision to perform splenectomy or splenic repair (splenorrhaphy), the techniques of which are discussed later, will depend largely on the degree of trauma and the expertise available. A simple grading system with suggested management is shown in Table 2 585.

The proportion of splenic injuries suitable for splenic repair will vary with the type of trauma seen in an individual centre and ranges from 30 to 90 per cent. The presence of extensive hilar injury makes it unlikely that the spleen will be suitable for repair, and avulsion or extensive splenic fragmentation requires splenectomy, as should failure to achieve haemostasis following attempted splenorrhaphy. If splenic repair is undertaken in conjunction with resection of splenic tissue, consideration should be given to the amount of spleen which will remain. The ‘critical mass’ required to confer substantial protection from bacterial infection in experimental models seems to be approximately one-quarter to one-third of the spleen. Splenic repair in the presence of significant peritoneal contamination is controversial, as is the repair of a diseased ruptured spleen, but is probably inadvisable.

Following splenic salvage surgery, careful postoperative monitoring is necessary to alert the surgeon to rebleeding. If the clinical condition deteriorates and there is evidence of continuing blood loss, re-exploration of the abdomen should not be delayed. This assessment may be difficult, or indeed impossible in the presence of multiple other injuries, and this may make splenectomy preferable to attempted repair in such circumstances.

The complications of splenic repair are few, but reported series are small. True ‘failures’ seem rare and the majority represent inadequate assessment of the splenic damage at the time of laparotomy, missed injuries, or over-ambitious attempts to repair severely damaged spleens.

A suspected diagnosis of splenic trauma in a patient whose cardiovascular system is stable can be investigated using ultrasound, CT scanning, or splenic scintiscanning. If the radiological investigation confirms the presence of a splenic injury, a decision should be taken as to whether it is appropriate to manage the patient in a conservative manner. The proportion of cases judged to be suitable for non-operative management varies from 20 to 45 per cent. Patient care is ideally carried out in a high dependency or intensive care unit where careful cardiovascular and haematological monitoring with repeated clinical examination is performed. Bed rest is imposed and, assuming the patient remains stable, discharge to the ward can occur after 24 to 48 h of observation, following which activity can be increased to discharge. It is prudent to restrict activity for some 4 to 6 weeks and to avoid contact sports for up to 6 months. Regular rescanning (with either ultrasound or isotope scintigraphy) will monitor resolution of the trauma, which is complete at 3 months in 90 per cent of cases. Deterioration, or evidence of bleeding not controlled by transfusion, is an indication for surgery.

Up to nearly one-third of cases managed in this way will ‘fail’ and require surgery. Variations between series (failure rates of zero to 30 per cent) reflect the differences in clinical judgement as to which patients are treated in an non-operative manner. In those instances in which conservative management fails and a laparotomy is needed, most result in the patient losing the spleen. Indeed, the protagonists of splenic repair would argue that more spleens would be preserved overall if early laparotomy were to replace attempts at conservative management.

Conservative (non-operative) management of splenic trauma has been most widely practised in children, because the increased incidence of overwhelming postsplenectomy infection in the child behoves the surgeons to attempt to conserve the spleen if possible. Fortunately, there are various reasons why conservative management of splenic injury in children is more likely to be successful than in adults. The type of injury suffered by children is less likely to be of the penetrating type (as associated with crimes of violence), and multiple injuries are also less common. Both criteria make non-operative management safer in the child. Furthermore, children have a better capacity for haemostasis, an increased resilience of the cardiovascular system to hypovolaemia, and an increased compliance of the capsule and septa. All these factors improve the chances of a successful outcome of conservative management.

Idiopathic thrombocytopenic purpura is a condition affecting females between the ages of 15 and 50. However, it can be associated with other conditions, e.g. SLE, chronic lymphatic leukaemia, Hodgkin's disease. It presents with bruising, usually after trauma or pressure, and examination reveals variable numbers of petechial haemorrhages in the skin. The clinical course is often intermittent and chronic. The platelets of affected patients become sensitized by antiplatelet IgG autoantibodies and are then removed from the circulation. Many patients need no treatment if the platelet count remains over 50 × 10&sup9;/l and no spontaneous bleeding occurs. If the count falls below 20 × 10&sup9;/l the risk of bleeding increases. Spontaneous recovery occurs in under 10 per cent of patients and treatment is initially directed at decreasing the levels of the autoantibodies and decreasing the rate of destruction of the platelets. Eighty per cent of patients go into remission on steroid therapy (60 mg prednisolone/day). If this does not result in remission, or continuing high doses are needed, or the platelet count remains low, splenectomy is indicated, the spleen being the initial and major site of antibody production and also a major site of platelet destruction. Good results are usually obtained and in Oxford in a series of 45 consecutive patients with this condition, 90 per cent of the patients had a satisfactory outcome, no longer requiring steroids to maintain an adequate platelet level. However, in more heavily sensitized patients platelets can continue to be destroyed elsewhere in the reticuloendothelial system, including the liver. A good response is most likely in patients under the age of 45 years, in those in whom the thrombocytopenia is less severe, and in those who have shown at least an initial response to steroid therapy. At operation, the spleen is usually of normal size and no special technique is necessary for its removal, although a careful search must be made to ensure that no splenunculi remain. Although the preoperative platelet count is low, platelet transfusion is not commenced until the splenic vessels have been ligated, when 6 to 10 units of platelets may be transfused. In addition, some patients may benefit from high doses of intravenous gammaglobulin preoperatively to raise the platelet count to an acceptable level (greater than 20 × 10&sup9;/l).

This is an autosomal dominant hereditary disorder characterized by small spherocytic red cells. Abnormalities in the red cell membrane transform the red cells to a spherical shape which results in a reduced ability to deform and therefore to entrapment in the spleen. The clinical presentation is commonly in childhood but may be delayed until later in life. Mild intermittent jaundice occurs associated with mild anaemia, splenomegaly, and gallstones. The diagnosis is made on examination of a peripheral blood film where reticulocytes and spherocytes are seen. Serum bilirubin levels are commonly elevated.

The anaemia and shortened red cell survival can be corrected by splenectomy which is normally delayed until over the age of 6 years to minimize the risk of postsplenectomy infection. The red cells retain their spherical appearance but in the absence of the spleen they have a near normal survival. The long-term results of the operation are excellent, relapses being attributed to missed splenunculi, but with no very convincing evidence.

This condition is characterized by oval shaped red cells caused by an abnormality in the red cell membrane. There are three types, two inherited as an autosomal dominant and one as a recessive trait. Most heterozygous carriers require no treatment, but if the anaemia is symptomatic, partial relief can be obtained from splenectomy.

This is a condition in which marked splenomegaly can occur. The patient may become transfusion dependent because of the development of hypersplenism and in these selected cases there may be a benefit from splenectomy.

In this condition splenic infarction is not uncommon, and only in rare cases is removal of the spleen indicated. On occasion benefit will be obtained from the operation in the presence of hypersplenism and ‘splenic sequestration syndrome’.

Hypersplenism is a clinical syndrome with many causes. It is characterized by splenic enlargement, haematological cytopenia (reduction in one or more cellular components of the blood), maturation arrest in the marrow, and the premature release of immature cells into the circulation. Red cell and/or platelet survival may be decreased. Anaemia, neutropenia, and thrombocytopenia can all occur, either alone or in any combination, the relative severity of the three depending on the nature of the underlying disease. Usually the decision as to whether the patient will benefit from splenectomy is made on the basis of knowledge about the particular disease process, and the severity of the haematological cytopenia. In difficult cases it may be necessary to use special tests of splenic function (summarized in Table 4 587). The combination of tests used will be dictated by the individual clinical problem under investigation, but even these do not always allow the haematological outcome of splenectomy to be predicted.

The most common indications for splenectomy for hypersplenism are lymphoma, chronic lymphatic leukaemia, and hairy cell leukaemia. After splenectomy for lymphoma, and chronic lymphatic leukaemia 80 to 90 per cent of cases will achieve haematological correction. This can be performed even in quite ill patients with widespread disease, where correction of the haematological cytopenia by splenectomy allows chemotherapy to be commenced and remission, either partial or complete, to be obtained as a result (Fig. 6) 2238. In hairy cell leukaemia, which is a condition of middle aged men that usually presents with hypersplenism, 90 per cent of patients have splenomegaly. Premature destruction of blood cells occurs during their passage through the spleen and splenectomy is the treatment of choice, with a resultant 5-year survival of 50 per cent. &agr;-Interferon is currently used for patients with little or no splenomegaly or in those in whom splenectomy may be contraindicated.

Splenectomy may be required to achieve a diagnosis (in the absence of palpable nodes for biopsy), to allow staging to occur, or to correct a haematological cytopenia of hypersplenism (see above). On occasion, splenectomy may be required to relieve the symptoms of gross splenomegaly (such as satiety and left upper quadrant discomfort).

In any large series of splenectomies, 10 to 20 per cent of operations are diagnostic laparotomies for enlarged spleens in the absence of other nodal tissue for biopsy. Of these, approximately three-quarters of the operations will allow a histological diagnosis of lymphoma to be made. Indeed, unless the spleen histology is quite normal, patients with undiagnosed splenomegaly are at high risk of subsequently developing a malignant lymphoma.

In Hodgkin's disease, staging laparotomy has been a cornerstone in the management of patients and has been crucial in deciding the stage of the disease and the treatment regimen. In general terms, stages I and IIA are treated with local nodal irradiation whereas stage III (and IV) is managed with systemic chemotherapy. Until the mid–1970s, all patients with clinical and radiological stages I, II, and III disease were subjected to a staging laparotomy unless there were specific medical contraindications. The operation comprises a splenectomy, liver biopsy, and nodal sampling from the coeliac axis, para-aortic and mesenteric nodes. In the last decade, the advent of CT scanning and a wider use of chemotherapy in the earlier stages of the disease have made staging laparotomy less common, although many series have highlighted the pitfalls of relying on radiological, as opposed to surgical staging.

Although CT scanning can diagnose nodal disease with greater reliability than other radiological methods, it cannot detect small deposits in normal sized nodes. Furthermore, no imaging technique can reliably detect micronodular disease in the spleen (Fig. 7) 2239. Currently, although practices will vary somewhat from centre to centre, staging laparotomy in Hodgkin's disease is often restricted to those patients in whom a definite histological diagnosis of intra-abdominal disease will affect management. In early disease, (stages I and II), laparotomy will change the stage (usually by increasing it), in up to one-third of patients, about 30 per cent having occult involvement of the spleen and 3 per cent occult involvement of the liver. A higher proportion of these patients have a positive laparotomy in presence of B symptoms, adverse histological subtypes (lymphocyte depleted or mixed cellularity subtypes), or multiple nodal sites. Conversely, a lower risk of intra-abdominal disease is seen in those patients with nodular sclerosing and lymphocyte predominant disease. If all patients are subjected to the operation, 70 per cent will have had an unnecessary laparotomy, with its attendant morbidity. Furthermore, although treatment of stage III disease does vary from that of stages I and IIA, it has been shown that those patients who receive local treatment only and then relapse with disease below the diaphragm can then be treated and the outcome of patients thus salvaged seems no worse than those patients in whom stage III disease was diagnosed at the outset. Thus staging laparotomy is now performed infrequently.

1.Abscess. Abscess is a relatively rare condition, general requiring a splenectomy.

2.Hydatid disease (see Section 41.8 173)

3.Part of a cancer operation. Because of the local extension of malignant tumours, the spleen may need to be resected as a part of operations to remove tumours of the greater curve of the stomach, the splenic flexure, or the tail of the pancreas. Routine splenectomy as part of radical gastric surgery for malignancy is no longer routinely practised.

4.Immunosuppression. Splenectomy will prolong renal allograft survival in experimental models in the rat, and has been shown to produce improved survival of human renal allografts in association with conventional immunosuppressive drugs. However, as time progresses there is an increased death rate from overwhelming infection in immunosuppressed renal transplant patients which negates the earlier improvement in graft survival. As a result splenectomy is no longer practised as part of an immunosuppressive protocol in clinical renal transplantation.

The observation that the asplenic state was compatible with life was made by Aristotle and was confirmed by experiments in the 17th and 18th centuries by Wren and Morgagni. It is unknown when the first splenectomy was performed but splenectomy as treatment for a disease process was possibly first performed in 1549 by Adriana Zaccarello (although there is dispute as to whether the removed organ was in fact an ovarian mass!). The first splenectomy for trauma was performed in 1678 by Nicholas Matthias. In 1928 William Mayo reported a series of 500 splenectomies with a mortality rate of 10 per cent. In the absence of a clear understanding of splenic function, subsequent reports of healthy survivors led to the concept that there were no ill effects from splenectomy. Despite the fact that in 1919 Morris and Bullock had reported that asplenic rats were susceptible to infection and had a shorter life-span than other rats, the caution advised by these authors went unheeded for over 30 years. In 1953, a report from King and Schumacker demonstrated an increased susceptibility to infection and death from sepsis in infants who had undergone splenectomy for congenital spherocytosis. It had been noted by Billroth that haemostasis could occur in a traumatized spleen without removal and that non-operative management of splenic trauma was therefore feasible. Early in this century, initial attempts at non-operative management and splenorrhaphy resulted in poor results and splenectomy was the rule for trauma. In the latter half of this century, with the recognition of the risk of overwhelming postsplenectomy infection, especially in children, non-operative management has become an option once more and current management of trauma is usually dictated by the age of the patient, the experience of the institution, the individual surgeon, and the type of trauma.

Under general anaesthesia with endotracheal intubation and muscle relaxation, the patient is positioned supine on the operating table. After the intravenous administration of a prophylactic antibiotic (a penicillin or second generation cephalosporin), the abdomen is opened through an upper midline, a left paramedian, a subcostal, or, rarely, a thoracoabdominal approach. Having opened the peritoneal cavity, a laparotomy is performed appropriate to the indication for surgery. Before mobilization of the spleen it is advisable to divide, if necessary after diathermy or ligation, any adhesions between the lower pole of the spleen and the greater omentum or splenic flexure. The colon can then be retracted down with the aid of a moistened abdominal pack. The right-handed surgeon then retracts the spleen medially with the left hand in order to facilitate division of the posterior layer of the lienorenal ligament behind the spleen by sharp dissection. This is usually avascular unless splenomegaly has resulted in vascular adhesions at this site. The gastrosplenic ligament is then divided between ligatures on the short gastric vessels passing to the upper pole of the spleen, care being taken not to include any gastric tissue in the ligatures. Having divided these two ligaments the spleen can be delivered medially into the wound and the splenic pedicle identified. The splenic artery and vein are then double ligated separately, without damaging the tail of the pancreas in the splenic hilum. After checking for haemostasis the abdomen is closed and a drain is not usually needed. In the presence of oozing, a suction drain can be left in the left upper quadrant for 24 to 48 h. If there has been suspected or definite damage to the tail of the pancreas a tube drain should be inserted to drain a potential pancreatic fistula.

When splenectomy is performed for trauma, an upper midline incision is recommended in order that a full laparotomy may be performed as necessary. In such cases the lienorenal ligament may have been avulsed and the spleen may deliver into the wound with little mobilization. However, if salvage surgery is planned (see below), then mobilization must be performed with extra care.

When performing splenectomy for massive splenomegaly, a subcostal incision can be used if the costal margin is wide and if the spleen does not extend below the umbilicus, in which event a midline incision is preferable. In the absence of adhesions between the diaphragm and the spleen, the size of the spleen stretches the ligaments and mobilization is not usually difficult. However, in the presence of dense adhesions, these should ideally be divided under direct vision, if necessary after ligation of the splenic artery.

As the spleen is a delicate, easily traumatized organ, splenic trauma can result in haemorrhage. For the reason discussed below, salvage of the spleen is preferable to splenectomy in certain situations. Various techniques have been described.

For trivial trauma, pressure can be combined with topical agents to arrest bleeding. Fibrin glue is a highly concentrated form of human fibrinogen and clotting factors. It is used by spraying a thin layer directly on to the injury or injecting it into or over a fractured surface. For knife or bullet tracts it can be injected deep into the base of the injury and slowly withdrawn. This distends the tract slightly to allow for both a haemostatic and tamponade effect. The glue can also be used in association with sutures or to seal the splenic surface after a partial resection. It has advantage over cyanoacrylate adhesive in that it is less histiotoxic. Microfibrillar collagen acts by trapping and then activating platelets. It forms a firm, adherent coagulum with an affinity for moist surfaces. It has been shown to be hypoallogenic, exciting little in the way of tissue reaction and is absorbed in 3 to 6 weeks. It should be applied with dry instruments and in sufficient amounts to cover the bleeding surface to a depth of several millimetres. Following application pressure should then be applied for about 5 min. Linear or stellate cracks can be packed and it is most successful if the surface blood flow can be reduced to a mild to moderate ooze. Other topical agents include gelatin foam, thrombin, and bovine collagen. These techniques are applicable to minor superficial lacerations or capsular tears without parenchymal damage.

The main splenic artery can be ligated in continuity in an attempt to reduce blood loss from a traumatized spleen and this can be performed without inevitable splenic infarction and loss of function provided the short gastric vessels are intact. Arterial ligation may also be used in addition to the other techniques described below. The splenic artery branches can be dissected at the hilum. The most constant is the superior polar artery which is the first branch from the main artery before entering the hilum. The lower polar vessels may not be obvious outside the spleen and it may be necessary to ligate the main vessel distal to the superior polar branch. The branches to bleeding segments may be ligated individually, either with or without partial splenectomy. Ligation of the short gastric vessels may be indicated for bleeding from the upper pole. Following ligation of any vessels, a careful examination of the spleen should be made so that any devitalized splenic tissue can be resected by sharp dissection, diathermy, or blunt finger dissection. During this technique, as vessels are encountered they may be clipped, tied, or under-run. When under-running with a stitch the parenchyma of the spleen may cut through, but the vessel wall will hold and residual bleeding can be controlled with the aid of haemostatic agents as described above. Stitches can be tied over pledgets of absorbable gelatin sponge (Gelfoam), oxidized regenerated cellulose (Surgicell), or compressed microfibrillar collagen.

Before any repair is undertaken the spleen must be fully mobilized to assess the extent of damage. During this procedure, extra care must be taken not to strip the posterior capsule, as consequent blood loss may preclude a successful repair. Capsular avulsion is most likely to occur if the incision in the posterior peritoneum is made too close to the spleen. Any peritoneal folds to the lower pole of the spleen from the greater omentum must be carefully divided. Following this incision, blunt dissection posteriorly is continued to elevate the spleen and the tail of the pancreas. If time permits, isolation of the splenic artery with a vascular sling is ideal to allow occlusion with a vascular clamp if necessary. The veins are fragile and no attempt should be made to isolate them.

Splenorrhaphy should aim both to achieve control of bleeding and to avoid causing or leaving behind infarcted splenic tissue. Simple suture of the torn spleen often results in further bleeding, but the use of a buttress technique with collagen, omentum, or Teflon pledgets can be effective. Haemostatic agents can be used in addition to the sutures (see above). The spleen may be wrapped in the greater omentum to assist haemostasis, and Dexon and Vicryl mesh can be used to achieve a similar effect.

In view of the segmental nature of the vascular supply, if one pole or segment of the spleen has been extensively damaged, partial splenectomy can be successfully performed. Haemostatic agents or an omental buttress can then be applied to the raw splenic surface following resection.

In general, it is appropriate to drain the left upper quadrant following a repair procedure. Drainage, in addition to clinical assessment, can be used to determine the success of the procedure. These procedures are often technically more demanding than a simple splenectomy, and it must be constantly remembered that the short- and long-term morbidity following such a procedure must be less than that resulting from splenectomy in order for it to be an acceptable part of the treatment.

There are certain complications which are specific to the operation of splenectomy, as opposed to those seen after any abdominal operation. Left lower lobe atelectasis is common, and all patients should receive active physiotherapy from the first postoperative day. Gastric ileus is usually short-lived and a nasogastric tube is not needed as a routine practice. Postsplenectomy fever is described and cannot always be attributed to atelectasis or subphrenic haematoma. In the absence of a definite cause it is a self limiting feature.

The platelet count often increases to between 600 and 1000 × 10&sup9;/l postoperatively, usually peaking between days 7 to 12. This rise is usually transitory but may last up to 3 months. If a level of 1000 × 10&sup9;/l is reached it is a sensible precaution to administer aspirin (150 mg/day) to prevent deep venous thrombosis.

Splenectomy is the traditional treatment of the traumatized spleen, but it is now well recognized that there are disadvantages to the asplenic state, the most important of these being the potential for the development of overwhelming postsplenectomy infection, a term coined by Diamond in 1969. The clinical course, consists of a fulminant bacteraemia, a frequent absence of a septic focus, coma, shock, consumptive coagulopathy, and adrenal haemorrhage. The bacteraemia is most commonly associated with the encapsulated organism Streptococcus pneumoniae, and less commonly with Neisseria meningitidis, Escherichia coli, and Haemophilus influenzae.

Removal of the spleen renders the patient more susceptible to infection for two reasons. Firstly, the absence of the splenic red pulp results in impairment of phagocytosis and clearance of exogenous organisms. Secondly, antibody production and subsequent bacterial opsonization are reduced by the removal of the white pulp, and poorly opsonized bacteria are cleared less effectively by the liver.

Estimates of the incidence of overwhelming postsplenectomy infection and its associated mortality vary, but large overviews found an overall incidence of 4.2, 5, and 3.8 per cent respectively, with all series showing the incidence in the paediatric age group to be consistently higher than in the adult population. The incidence of overwhelming postsplenectomy infection varies according to the indication for which the splenectomy is performed. Some 1.5 to 2.5 per cent of children develop overwhelming postsplenectomy infection when the spleen is removed for trauma, but the greatest risk is seen in patients undergoing splenectomy for congenital anaemias, portal hypertension, or lymphoreticular tumours. The time of development of overwhelming postsplenectomy infection is variable in relation to the splenectomy. Although it has been described up to 30 years after loss of the spleen, it usually occurs within 3 years after the operation. It is of particular importance as this infection is associated with a very high mortality rate (25–75 per cent reported in many series). It is primarily to avoid this complication that a renewed interest in splenic preservation has arisen, especially in children. Attempts to preserve the spleen can be achieved by conservative, non-operative management, or by one or other of the splenic salvage operations described in the previous section.

The effect of prophylactic antibiotics on overwhelming postsplenectomy infection is controversial. Traditionally, oral penicillin has been given postoperatively for periods which have varied from 6 weeks to 5 years, or until the onset of puberty if splenectomy is performed in a child. As the time of greatest risk is during the first 2 or 3 years after splenectomy it would not be unreasonable to give all patients some form of antibiotic prophylaxis at least during that time. However, many septic episodes in asplenic patients are secondary to organisms not sensitive to penicillin and patient compliance is a problem. As a compromise, it seems appropriate to treat patients with prophylactic antibiotics only if they are at high risk, such as children and immunosuppressed patients. For the prevention of infection, antibiotics may be given in reduced dosage to avoid undesirable side-effects. Penicillin V (250 mg daily) is a recommended prophylactic dose. It is likely that antibiotics are of more use when given promptly and appropriately when infection develops, and clinicians should ensure that asplenic patients are supplied with amoxycillin to take at the first sign of a respiratory illness or fever. Amoxycillin is recommended firstly because of the higher blood levels achieved with oral administration, and secondly because of its activity against H. influenzae.

Vaccines are available to the pneumococci, Haemophilus influenzae, and meningococci. The pneumococcal vaccine consists of purified, capsular polysaccharide antigens of the 23 most prevalent serotypes of S. pneumoniae. Most healthy adults show at least a twofold rise in antibody titre within 2 weeks. Antibody titres of 256 ng immunoglobulin nitrogen/ml or more are indicative of immunity. Raised antibody levels have been demonstrated 5 years after vaccination but the exact duration of protection is now known. Splenectomy results in a reduced antibody responsiveness and this is most marked in the young and those with malignancy or on immunosuppressive drugs. Although vaccination failures have been reported, in the case of elective surgery, vaccination is recommended and should be given as soon as a decision to operate has been taken. In a trauma victim, vaccination can be given in the postoperative period and the resulting antibody levels will be protective in the majority of cases, although the antibody levels are less than 50 per cent of those achieved if vaccination is given in the presence of an intact spleen. Following vaccination, antibody titres remain high for at least a year before slowly declining, but this rate is increased in immunosuppressed patients and those with lymphoma. Although the incidence of pneumococcal infection has been shown to be reduced following vaccination, protection is not guaranteed. The reasons for this include the fact that about half the organisms causing post-splenectomy sepsis are not pneumococcal, and further more, 20 per cent of subtypes are not covered. Antigenic types also vary and therefore protection from a particular vaccine may be incomplete. Serious side-effects of the vaccine are rare, but ‘booster’ doses are not recommended as severe reactions are common.

Splenosis, or spontaneous regrowth of splenic tissue following splenectomy for trauma, possibly accounts for the lower incidence of overwhelming postsplenectomy infection in patients who have had splenectomy for trauma rather than for other reasons. Detection of splenosis by scantiscanning is complicated by the presence of splenunculi. The precise incidence of splenosis is unknown but may be as high as 50 per cent following trauma. Splenic function of the tissue can be assessed using red cell morphology, tests of phagocytic function, the presence of Howell-Jolly bodies, IgM, IgG, and IgA levels, and scanning. It is clear that the presence of splenosis does not eliminate the risk of developing overwhelming postsplenectomy infection, as there have been documented reports of deaths from overwhelming postsplenectomy infection occurring in patients in whom splenosis has been demonstrated at postmortem examination. It has been estimated that 25 to 30 g of tissue is needed for protection against overwhelming postsplenectomy infection.

In an attempt to conserve functioning splenic tissue in cases in which splenunculi are not present and where splenorrhaphy is not possible, splenic tissue can be autotransplanted at the time of surgery by dicing the spleen into small pieces and implanting them into the abdominal wall or an omental pouch. Both techniques can be shown to result in functioning splenic tissue and it is clear from reports that such transplants can be shown to survive. Experimental work has shown a variable potential for decreasing the incidence of infection in asplenic animal models, but as an overwhelming postsplenectomy infection affects only 2.5 per cent of patients following splenectomy for trauma, clinical reports of ‘successful’ transplants are anecdotal, and it is unclear as to whether it should be routine practice. There is experimental evidence that suggests that the mere presence of functioning splenic tissue in itself is insufficient to protect against overwhelming postsplenectomy infection (whether from splenosis or as a result of autotransplantation), and it seems likely that the normal splenic vasculature is crucial for maximum protection. Overall it is a procedure which is not associated with morbidity, and there seems little to be lost by performing it.

Bickerstaff KI, Morris PJ. Splenectomy for massive splenomegaly. Br J Surg, 1987; 74: 346–9.

Cooper MJ, Williamson RCN. Splenectomy: indications, hazards and alternatives. Br J Surg, 1984; 71: 173–80.

Di Padova F, Durig M, Wadstrom, Harder F. Role of the spleen in immune response to polyvalent pneumococcal vaccine. Br Med J, 1983; 287: 1829–32.

Ellison EC, Fabri PJ. Complications of splenectomy. Etiology, prevention and management. Surg Clin N Am, 1983; 63: 1313–30.

Franke EL, Neu HC. Post splenectomy infection. Surg Clin N Am, 1981; 61: 135–55.

Ghosh S, Symes JM, Walsh TH. Splenic repair for trauma. Br J Surg, 1988; 75: 1139–40.

Gill PG, Souter RG, Morris PJ. Results of surgical staging of Hodgkin's disease. Br J Surg, 1980; 67: 478–81.

Gill PG, Souter RG, Morris PJ. Splenectomy for hypersplenism in malignant lymphomas. Br J Surg, 1981; 68: 29–33.

Greco RS, Alvarez FE. Protection against pneumococcal bacteremia by partial splenectomy. Surg Gynecol Obstet, 1981; 152: 67–9.

Holdsworth RJ, Irving AD, Cushchieri A. Postsplenectomy sepsis and its mortality rate: actual versus perceived risks. Br J Surg, 1991; 78: 1031–8.

King H, Schumaacker HB. Splenic studies. I. Susceptibility to infection after splenectomy performed in infancy. Ann Surg, 1952; 136: 239–42.

Kram HB, del Junco T, Clark SR, Ocampo HP, Shoemaker WC. Techniques of splenic preservation using fibrin glue. J Trauma, 1990; 30: 97–101.

Lewis SM. The spleen—mysteries resolved and unresolved. Clinics in Haematol 1983; 363–73.

Malangoni MA, Levine AW, Droege EA, Aprahamian C, Condon RE. Management of injury to the spleen in adults. Ann Surg, 1984; 200: 702–5.

Mayo W. A review of 500 splenectomies with special reference to mortality and results. Ann Surg, 1928; 88: 409–15.

McIndoe AH. Delayed haemorrhage following traumatic rupture of the spleen. Br J Surg 1932; 20: 249–68.

Mitchell A, Morris PJ. Surgery of the spleen. Clinics in Haematol, 1983; 12: 565–90.

Morgenstern L, Shapiro SJ. Techniques of splenic conservation. Arch Surg, 1979; 114: 449–54.

Morris DH, Bullock FD. The importance of the spleen in resistance to infection. Ann Surg, 1919; 70: 513–21.

Morris PJ, Cooper IA, Madigan JP. Splenectomy for haematological cytopenias in patients with malignant lymphomas. Lancet, 1975; ii: 250–3.

Mucha P, Daly RC, Farnell MB. Selective management of blunt splenic trauma. J Trauma, 1986; 26: 970–9.

Redmond HP, Redmond JM, Rooney BP, Duignan JP, Bouchier-Hayes DJ. Surgical anatomy of the human spleen. Br J Surg, 1989; 76: 198–201.

Schwartz PE, Sterioff S, Mucha P, Melton LJ, Offord KP. Post splenectomy sepsis and mortality in adults. JAMA, 1982; 248: 2279–83.

Shackford SR, Sise MJ, Virgilio RW, Peters RM. Evaluation of splenorrhaphy: a grading system for splenic trauma. J Trauma, 1981; 21: 538–42.

The splenic injury study group. (1987) Splenic injury: a prospective multicentre study on non-operative and operative treatment. Br J Surg, 1987; 74: 310–13.