Bleeding in association with surgery is a common problem encountered by surgeons and of referral to haematologists. An understanding of the mechanism of blood coagulation is important in order to understand the basis of disorders of haemostasis and the common laboratory tests.

The fundamental step in blood coagulation is the formation of insoluble fibrin strands. The cleavage of small polypeptide chains from the soluble parent fibrinogen molecule is sufficient to achieve this transformation. However, this is only the last step in a series of enzymatic reactions that take place during coagulation (Fig. 1) 77. The coagulation cascade is initiated in two ways. The extrinsic arm is activated when tissue factor forms a complex with factor VII. The resultant complex activates factor X directly, which has a central role in both pathways. Factor X may also be activated through the intrinsic pathway, when negatively-charged subendothelial collagen is exposed and activates factors XII and XI. Factor X forms a complex with calcium and factor V. This complex cleaves prothrombin to produce thrombin, which in turn cleaves fibrinopeptides A and B from soluble fibrinogen to yield insoluble fibrin strands.

The basic tests included in a clotting screen are the prothrombin time and the kaolin cephalin clotting time: the former tests the extrinsic pathway, and the latter tests the intrinsic pathway (see below).

There are naturally occurring anticoagulants. The most important of these are antithrombin III, protein C, and protein S. Deficiency of these factors may result in a thrombotic tendency (see below).

Platelets are essential for normal haemostasis. They are particularly important in the formation of the primary response to vascular injury, when activated platelets coalesce to form a platelet plug in the transected vessel.

Coagulation tests are carried out on blood anticoagulated with sodium citrate. In contrast to other laboratory samples, it is very important that the correct volume of blood is collected in the appropriate tube. Tests must be carried out on fresh samples as coagulation factors are labile, and if there is undue delay in sending a specimen to the laboratory there will be spurious prolongation of the clotting times. Contamination with heparin (e.g. from indwelling cannulae) will also result in spurious results. It is not possible to state normal ranges which are universally applicable for most clotting tests, as individual laboratories are likely to use slightly different techniques and reagents. Individual laboratories will issue their own normal ranges.

Brain extract (rich in tissue factor) and calcium are added to the test plasma, and the time taken for a clot to develop is measured. The prothrombin time is a test of the function of the extrinsic pathway, and is sensitive to isolated or combined deficiencies of factors II, V, VII, X, and fibrinogen.

The results of a prothrombin time are usually expressed as a ratio, comparing the result obtained on the patient test plasma to that obtained on a normal plasma sample Equation 20

The normal prothrombin time ratio should be close to 1.0. The International Normalized Ratio (INR) is for practical purposes the same as the prothrombin time ratio.

Kaolin, phospholipid, and calcium are added to the test plasma, and the time taken for a clot to appear is measured. Kaolin activates factor XII, just like collagen. The kaolin cephalin clotting time thus tests function of the intrinsic pathway, and is sensitive to isolated or combined deficiencies of factors XII, XI, X, IX, VIII, V, II, or fibrinogen.

Thrombin is added to patient plasma, and the time taken for a clot to develop is measured. The thrombin time is prolonged when the fibrinogen level is low, or in the presence of inhibitors of thrombin (e.g. heparin).

Normal range is 150 to 400 × 10&sup9;/l. Significant bleeding may occur when the count falls below 80 × 10&sup9;/l. Abnormal bleeding may occasionally occur if platelet function is abnormal, even when the platelet count is normal or even increased (e.g. myelodysplasia, polycythaemia).

The bleeding time is a simple test of platelet function. A sphygmomanometer cuff is inflated to 40 mmHg around the upper arm. A 5-mm incision is made with a special blade. The incision is wiped with blotting paper every 30 s, and the bleeding time is taken as the time when blood stops oozing. The normal bleeding time is less than 9 min. The bleeding time will be prolonged in association with thrombocytopenia and disorders associated with defective platelet function (e.g. von Willebrand's disease, some cases of myeloproliferative disease).

Acquired disorders of haemostasis are encountered much more frequently than congenital defects. In order to exclude the possibility of a congenital disorder of coagulation (for which specific therapy may be available) it is important to try to establish from a personal history whether there have been spontaneous haemorrhagic problems (e.g. epistaxis, haemarthrosis, gastrointestinal bleeding) in the past or bleeding after previous surgery (e.g. tonsillectomy, appendicectomy) or dental extractions. The family history should also be elicited. Symptoms of congenital disorders of haemostasis usually appear early in life. However, it should be borne in mind that mild forms of congenital disorders such as haemophilia may only become evident after surgery or major trauma. Acquired disorders of haemostasis may present in a number of ways, ranging from sudden life-threatening bleeding after surgery or childbirth at one end of the spectrum to minor purpura or an increased bruising tendency at the other.

In most circumstances, initiation of coagulation is a local phenomenon and this is an appropriate reaction to local vascular injury which has resulted in bleeding, e.g. at the site of a surgical incision. Disseminated intravascular coagulation (DIC) is a consequence of explosive activation of the coagulation cascade throughout the vascular tree. Paradoxically, this results in a bleeding tendency and symptoms related to vascular occlusion are relatively rare. This is because the initial thrombus formed in response to the triggering of disseminated intravascular coagulation undergoes very rapid lysis. If the initial trigger persists, further cycles of coagulation and instantaneous lysis rapidly result in the depletion of coagulation factors, including fibrinogen, and consumption of platelets.

Most cases of disseminated intravascular coagulation are triggered by septicaemia. Gram-negative organisms are often implicated. Malignant disease is also an important cause, particularly when there are multiple metastases. Carcinoma of the lung, pancreas, stomach, and prostate are particularly associated with this complication. Promyelocytic leukaemia is also frequently complicated by disseminated intravascular coagulation.

In overt cases, there is widespread bruising with extensive purpura. There may be persistent oozing of blood from surgical wounds and venepuncture sites. Bleeding from mucosal surfaces (e.g. epistaxis) is common. Occasionally, there are signs of vascular occlusion in the distal limbs.

In fulminant cases of disseminated intravascular coagulation the blood is incoagulable. Both the kaolin cephalin clotting time and prothrombin time are markedly prolonged, but will be corrected by the addition of normal plasma to the patient's plasma. The thrombin time is also prolonged, reflecting depletion of fibrinogen. The fibrinogen level is usually below 1.0 g/l (normal range 2–4 g/l). Levels of fibrin degradation tests in the blood will be very high, reflecting hyperfibrinolysis. The platelet count is usually reduced, and may be less than 50 × 10&sup9;/l in severe cases. Examination of the blood film typically reveals the presence of fragmented erythrocytes.

Fresh frozen plasma should be infused, as this is a good source of coagulation factors, including fibrinogen. As a rough guideline, three to four packs will be required initially. Cryoprecipitate is a very good source of fibrinogen, and has the advantage of being more concentrated so that volume overload may be avoided. If available, cryoprecipitate should be given as well as fresh frozen plasma. Platelet concentrates should also be transfused: 10 to 12 packs should suffice as initial therapy. There is no convincing evidence that administration of antithrombin III concentrates is beneficial. It is important not to overlook the fact that patients often need blood in addition to plasma products. Maintenance of circulating blood volume and an adequate haemoglobin level are important objectives, as tissue hypoxia will only exacerbate disseminated intravascular coagulation.

Contrary to what might be imagined, administration of inhibitors of fibrinolysis (e.g. tranexamic acid) are of no value in the treatment of disseminated intravascular coagulation. The use of such agents may precipitate overt thrombosis. In the past low doses of heparin were administered in cases of disseminated intravascular coagulation, in an attempt to break the vicious cycle of initial thrombosis and subsequent lysis. Such therapy is no longer widely advocated, and it is usually possible to control the activation process by judicious use of blood products and treatment of the underlying cause.

The liver is the principal site of synthesis of coagulation factors. Both acute and chronic liver diseases are thus frequently associated with haemostatic abnormalities. Thrombocytopenia of moderate severity (50–100 × 10&sup9;/l) is a frequent finding in patients with chronic liver disease. Another factor which contributes to the bleeding tendency of chronic liver disease is increased fibrinolytic activity, associated with a decreased plasma level of naturally occurring &agr;&sub2;-antiplasmin.

The most frequent haemorrhagic problems are oesophageal and gastrointestinal haemorrhage, as well as bleeding from biopsy sites and during and after surgery. Bleeding into soft tissues is only rarely encountered.

The most common laboratory findings are a marked reduction in the plasma levels of all coagulation factors except factor VIII. Both the prothrombin time and kaolin cephalin clotting time are prolonged.

Vitamin K should be administered when it is suspected that the haemorrhagic disorder is due, at least in part, to deficiency of the vitamin (see below). Fresh frozen plasma contains all of the coagulation factors and inhibitors present in blood and is suitable for the correction of the multiple abnormalities associated with liver disease. Usually two to three bags of fresh frozen plasma should suffice to correct the haemostatic defect.

Platelet concentrates are usually of little use in patients with liver disease and thrombocytopenia, as the infused platelets are rapidly sequestered in the liver and spleen. Desmopressin (DDAVP) at a dose of 0.3 &mgr;g/kg can be used to shorten the bleeding time before invasive procedures such as biopsy or laparoscopy.

Patients with chronic renal failure often have a bleeding tendency, due in part to poor platelet function. This may be corrected by dialysis. Those with nephrotic syndrome, by contrast, may develop thrombotic complications such as renal vein thrombosis and deep venous thrombosis. The most frequent haemorrhagic manifestations observed in uraemic patients, whether on chronic haemodialysis or not, are usually from mucosal surfaces (gastrointestinal bleeding, epistaxis, menorrhagia). Retroperitoneal haemorrhage, bleeding into the pericardial and pleural spaces, and intracranial haemorrhage may develop occasionally. Patients do not usually bleed after surgical procedures, but renal biopsies are sometimes complicated by the formation of an intrarenal haematoma.

Prolongation of the bleeding time, associated with a normal or only slightly reduced platelet count and normal coagulation tests are the usual findings in uraemia. There is an inverse relationship between the haematocrit and the bleeding time in renal failure. Haemodialysis or transfusion of platelet concentrates may produce transient shortening of the bleeding time. The infusion of eight to ten bags of cryoprecipitate in uraemic patients is usually followed by shortening or even return to normal of the bleeding time. DDAVP (desmopressin) at a dose of 0.3–0.4 &mgr;g/kg usually restores the bleeding time to normal in uraemic subjects 1 h after intravenous infusion. Transfusion of red cell concentrates in order to correct anaemia and maintain the haematocrit above 0.30 also shortens the bleeding time. Administration of erythropoietin has a similar effect.

Vitamin K is necessary for the synthesis of coagulation factors II (prothrombin), VII, IX, and X. Vitamin K is fat soluble and is absorbed effectively only in the presence of bile salts. Some vitamin K is also synthesized by colonic bacteria. Little vitamin K is stored in the body and in certain conditions symptoms of deficiency may become evident within a few weeks.

Deficiency of vitamin K is associated with prolongation of both the prothrombin time and the partial thromboplastin time. The thrombin time and plasma fibrinogen concentration are normal, which helps in the exclusion of disseminated intravascular coagulation, and the platelet count is normal. Typical haemorrhagic manifestations are easy bruising, and bleeding from sites of injury or from the gums or gastrointestinal tract.

Debilitated patients undergoing surgery are particularly vulnerable, as dietary deficiency may be compounded by the administration of broad-spectrum antibiotics which kill off gut bacteria that synthesize the vitamin. Vitamin K will not be effectively absorbed from the gastrointestinal tract when there is obstruction of the bile duct. Malabsorption of vitamin K may also occur in a number of other conditions, e.g. coeliac disease, intestinal fistulae. Vitamin K deficiency as a consequence of partial biliary tree obstruction may contribute to the development of impaired haemostasis in chronic hepatic disorders such as cirrhosis. An injection of vitamin K may shorten an abnormally long prothrombin time in such cases.

Obviously, the consumption of anticoagulants will result in a bleeding tendency! Where unexpected bleeding occurs during or after surgery, consideration should be given to the possibility that the patient is on anticoagulant therapy. This problem may arise when urgent surgery is carried out, without full medical details of the patient being available (e.g. the patient is unconscious).

Where the possibility is suspected, further details may be sought from others involved in the medical care of the patient. Patients taking oral anticoagulants usually carry a medical advisory card. The diagnosis may be confirmed by the finding of a significantly prolonged prothrombin time. The kaolin cephalin clotting time is often only slightly prolonged, and the thrombin time is normal. The platelet count will also be normal.

Bleeding after cardiopulmonary bypass may be due to the presence of heparin (see below). Contamination with heparin of blood samples drawn from venous cannulae for clotting studies is a very common cause of spurious laboratory results. This may lead to further time-consuming tests and delays in surgery before it can be confirmed that the patient does not have a disorder of haemostasis. Blood for clotting studies should be drawn from a peripheral vein if venous cannulae are flushed with heparin. Inadvertent full heparinization of patients prior to surgery has been reported. In these cases, full-strength heparin was used to flush indwelling venous cannulae, rather than the specific dilute preparations. The use of low-dose subcutaneous heparin does not significantly alter laboratory tests of haemostasis and does not result in a generalized bleeding tendency during surgery.

Administration of streptokinase as thrombolytic therapy (e.g. for myocardial infarction) within 10 days or so after surgery or other invasive procedure such as biopsy may be hazardous as serious bleeding may ensue.

Blood collected in citrate phosphate dextrose with added adenine (CPDA1) has a shelf life of 35 days at 4°C. However, levels of all the coagulation factors decline during storage. Platelets in stored blood also rapidly lose their viability.

For these reasons, haemorrhagic problems may develop when a patient's blood is replaced by large quantities of stored blood within a short period of time. Microvascular bleeding is a typical manifestation of the impaired haemostasis. Examples include bleeding from mucous membranes, oozing from catheter sites which persists after application of pressure, continuous oozing from surgical wounds, and generalized petechiae. Impaired haemostasis is, of course, only one of several important problems encountered in patients receiving a massive blood transfusion. These include hypocalcaemia due to citrate overload, hyperkalaemia, and hypothermia.

When there is no underlying medical complication, replacement of up to one blood volume (8–10 units of blood in an adult) is not likely to be associated with significant haemostatic problems. Laboratory tests of coagulation may help to identify patients who need additional blood components to improve haemostasis when larger volumes of blood are transfused. A platelet count of less than 50 × 10&sup9;/l, prothrombin time ratio of 1.8 or more, and a plasma fibrinogen level of 0.5 g/l or less are strongly associated with microvascular bleeding. Platelet support may be required once the patient has received 15 or more units of blood. Fresh frozen plasma is a source of coagulation factors, including fibrinogen.

Thrombocytopenia is often present during cardiopulmonary bypass. There is also impairment of platelet function, associated with prolongation of the bleeding time. Platelet dysfunction is related to contact with the synthetic surface of the oxygenator, and probably the induced hypothermia. Many patients with coronary artery disease may be taking aspirin. Patients taking aspirin before cardiopulmonary bypass surgery are at risk of excessive blood loss during the procedure. It is therefore recommended that aspirin be discontinued at least 5 days before cardiopulmonary bypass. Despite abnormalities in platelet number and function, there is no evidence that routine perioperative transfusion of platelet concentrates is necessary.

Cardiopulmonary bypass is associated with a drop in the plasma levels of most coagulation factors, which is primarily attributable to haemodilution. As with platelet concentrates, it is not necessary routinely to transfuse fresh frozen plasma during cardiopulmonary bypass.

Heparin is routinely administered in order to prevent extracorporeal clotting in the oxygenator. Protamine sulphate is administered at the end of surgery in order to neutralize remaining heparin. In gross excess protamine itself acts as an anticoagulant. Following initial adequate heparin neutralization, the reappearance of active heparin in the bloodstream may occur 2 to 6 h later. This rebound effect is caused by the delayed return of sequestered extravascular heparin which occurs when peripheral perfusion improves. Thrombocytopenia may complicate heparin therapy in about 5 per cent of patients receiving the drug. The onset of thrombocytopenia is typically between 6 and 12 days after exposure to the drug. This is due to the development of an antibody which induces platelet activation. The possibility of heparin-induced thrombocytopenia should always be considered in the differential diagnosis when thrombocytopenia develops after cardiopulmonary bypass. Platelet concentrates should only be transfused if there are bleeding complications as arterial thrombosis has been reported after transfusion of platelets.

The administration of aprotinin, an inhibitor of plasmin, significantly reduces intraoperative and postoperative blood loss associated with cardiopulmonary bypass. The requirement for blood transfusion is reduced, and the actual operating time may be shortened.

Haemophilia is the most common congenital disorder of coagulation and affects 1 in 10 000 males. Haemophilia A is caused by a deficiency of factor VIII in the circulating blood. In its severe form (less than 2 per cent of the normal factor VIII) it is characterized by recurrent joint bleeds, intramuscular bleeding, and excessive bruising after trauma. Recurrent bleeds may result in crippling arthritis. Haemophilia B (Christmas disease) is a clinically identical disorder caused by deficiency of factor IX. Both conditions show X-linked inheritance, but in one-third of cases of haemophilia there is no preceding family history as the condition results from a new mutation. Female carriers invariably have a high enough factor level to protect them from spontaneous haemorrhagic complications, although treatment with coagulation factor concentrate may be required before surgery.

Some 6 per cent of patients with haemophilia A and 1 per cent of those with haemophilia B develop inhibitory IgG antibodies directed against the infused factor. Management of patients with high titre antibodies can be difficult. A satisfactory clinical response may often be obtained by increasing the dose of human factor VIII administered, although the antibody titre may be subsequently boosted. Porcine factor VIII may be effective where the antibody titre against human factor VIII is very high.

The patient's plasma must be screened for the presence of inhibitory antibodies prior to surgery. Elective, non-urgent surgery is not advisable if a patient is known to have inhibitory antibodies.

Chronic liver disease is a problem in many haemophiliacs. This is usually due to persistent infection with hepatitis B or C. All patients who are likely to receive coagulation factor concentrates should be vaccinated against hepatitis B. Documentation of abnormal hepatic function is of practical importance to the surgeon for a number of reasons.

1.In established liver disease, the synthesis of a number of coagulation factors is disturbed. Fresh plasma may be required in addition to factor concentrate to ensure haemostasis.

2.The dosage of certain drugs may need to be modified.

3.There is a potential risk to staff of infection through needlestick injuries.

Liver transplantation has been carried out in a small number of haemophiliacs with serious liver disease, and this has the added bonus of curing haemophilia as both factor VIII and factor IX are synthesized by hepatocytes. Of course, this radical treatment cannot be routinely applied to haemophiliacs!

Approximately one-third of haemophiliacs in the United Kingdom were infected with the human immunodeficiency virus (HIV) between the years 1979 and 1985. The extent to which haemophiliacs in various countries were infected with HIV is very variable. Infection with HIV is associated with a progressive decline in the number of CD4+ T lymphocytes in the blood. Cellular immunity is compromised and patients may eventually develop opportunistic infections. Thrombocytopenia is seen in approximately 10 per cent of patients infected with HIV. Thrombocytopenia is clearly particularly dangerous in haemophiliacs who already have a bleeding tendency. Zidovudine therapy usually raises the platelet count considerably.

There is no evidence that surgery accelerates progression to AIDS by promoting the decline in the number of CD4+ T lymphocytes. Many surgeons are concerned about the possibility of transmission of HIV (see also Chapter 2.2 1). However, the risk of infection to health care workers through body fluids or inoculation is much less than that associated with hepatitis B. A prospective study of needlestick injuries among health care staff and involving patients known to be infected with HIV has demonstrated a risk of infection of around 0.3 per cent. The vast majority of such injuries are associated with the resheathing of needles: this should be avoided, and all used needles should be discarded in a suitable container immediately after use. Needlestick injuries amongst surgeons due to suture needles appear to be less dangerous than injuries associated with hollow needles used for venepuncture. The prophylactic administration of zidovudine after a needlestick injury does not guarantee protection against infection, and the drug has a number of toxic effects. The risk of transmission through blood or other bodily fluids coming into contact with intact skin is negligible.

Wherever possible, elective surgery should be carried out in a centre experienced in the management of haemophilia. All registered patients should carry a medical card which states the precise diagnosis, factor level, inhibitor status and blood group. The HIV status of a patient will not be stated on the card. It must be remembered that female carriers of haemophilia A or B may themselves also have low levels of factor VIII or IX respectively, and may require treatment prior to surgery or dental work.

Patients scheduled for elective surgery should be admitted 1 or 2 days beforehand for blood tests (Table 3) 112. Patients with severe haemophilia are likely to need infusions of concentrate for at least 7 to 10 days after any type of surgery. They will often need, therefore, to stay in hospital for a longer period than other patients and this must be planned for accordingly. It is all too common to see patients scheduled for discharge 2 days after ‘minor’ surgery. There is no such thing as ‘minor’ surgery for haemophiliacs.

All patients receiving coagulation factor concentrates (even high-purity ones) should be vaccinated against hepatitis B. If some time has elapsed since completion of the course of vaccination, immune status should be checked and the antibody level (anti-HBs) checked. Surgeons dealing with haemophiliacs should certainly be vaccinated against hepatitis B. No vaccine is as yet available against hepatitis C.

In some countries where concentrate is not readily available, some units have achieved considerable savings by good co-ordination of activities. If a patient is scheduled to have surgery and will be in hospital and having concentrate for some days thereafter, other procedures can be carried out at the same time (e.g. dental work, excision of skin lesions, vasectomy etc.).

On the day of surgery, concentrate should be infused within 1 h before surgery: a dose of 50 i.u./kg of factor VIII or 75 i.u./kg of factor IX should suffice to boost the plasma level to well within the normal range. Pre- and postinfusion plasma levels of factor should be checked in the laboratory to confirm that a suitable level has been achieved for surgery. After surgery, twice-daily bolus infusions of factor VIII will be required for at least 5 days. Further daily doses may be required for a further 5 days or so, until soft tissue healing is achieved. Factor IX has a longer half-life, and only one infusion a day is necessary. Analgesics must not be given by intramuscular injection, and aspirin is contraindicated.

To minimize the risk of bleeding from wounds, it is sensible to change dressings after infusion of a dose of concentrate. Similarly, physiotherapy is best scheduled immediately after infusion. Intramuscular injections (e.g. premedication prior to surgery, postoperative analgesia) must not be given, as a large intramuscular haematoma is likely to develop.

Deep venous thrombosis is not a problem in patients with haemophilia A, and prophylaxis (e.g. heparin, low-dose warfarin) is not necessary, even in orthopaedic surgery. Such treatment may also interfere with laboratory monitoring of plasma factor levels after infusion of concentrate. Thrombosis has been reported in patients receiving prothrombin-complex concentrates, and high-purity factor IX concentrates are to be recommended for patients with haemophilia B undergoing surgery.

Very occasionally, patients with isolated congenital deficiencies of other coagulation factors may be encountered. Deficiencies of fibrinogen, factor V, factor X, factor VII, factor XI, or XIII may be associated with a serious bleeding tendency. Similar practical guidelines to those set out above apply. Specific plasma-derived concentrates of most of these coagulation factors are available commercially.

It is not necessary to perform tests of haemostasis on all patients scheduled for elective surgery. However, if a history of a possible bleeding tendency emerges during outpatient investigation, patients should certainly be referred for investigation prior to surgery.

Basic tests of coagulation include the prothrombin time, kaolin cephalin clotting time, and the platelet count. Where the platelet count is normal (or even elevated) a bleeding time should be measured to exclude the possibility of congenital or acquired defects of platelet function. It is important to try to distinguish between congenital disorders on the basis of the personal and family history. Table 4 113 is intended only as a guideline to how the screening test may help in the diagnosis of specific defects.

Where bleeding is encountered during surgery and there is no obvious source of haemorrhage, a prothrombin time, kaolin cephalin clotting time, and platelet count should be requested in the first instance. The administration of two to three units of fresh frozen plasma is a useful practical measure until the cause is identified. Platelet concentrates should be given if the platelet count is below 50 × 10&sup9;/l.

The term thrombophilia is used to describe familial or acquired disorders of the haemostatic mechanism which predispose to thrombosis. Inherited deficiencies of antithrombin III, protein C, and protein S are important causes of a thrombotic tendency. Antithrombin III inactivates factors IX, X, XI, and thrombin. Protein C and its cofactor protein S are vitamin K-dependent proteins which inactivate factors V and VIII. Such patients are particularly vulnerable to the development of deep venous thrombosis and/or pulmonary embolism after surgery if adequate precautions are not taken.

At present, the great majority of thrombotic episodes remain unexplained in that no underlying disorder may be identified. Occasionally, acquired blood disorders associated with a definite thrombotic tendency may be diagnosed after an episode of thrombosis. These include thrombocytosis secondary to myeloproliferative disease, polycythaemia, and the lupus anticoagulant. The development of a lupus anticoagulant is associated with prolongation of the kaolin cephalin clotting time and, in about a quarter of cases, prolongation of the prothrombin time. The former is not corrected by the addition of normal plasma to the patient's own plasma. There may be thrombocytopenia. Despite these findings, the patient is paradoxically at risk of thrombosis and there is no undue risk of haemorrhage (even in association with surgery). Although originally described in association with systemic lupus erythematosus, it is important to appreciate that the great majority of cases arise in people who have no evidence of the disorder and who are perfectly well.

Consideration should be given to investigating patients encountered in clinical practice who have thrombosis at an unusually early age, or in an unusual site (e.g. mesenteric vein thrombosis), or where there is a suggestion of a familial tendency to thrombosis. Laboratory testing for thrombophilia is difficult once a patient is anticoagulated, as heparin reduces the plasma level of antithrombin III whilst warfarin boosts the plasma level of antithrombin III but lowers the level of proteins C and S. Investigation is best deferred until the patient has been off anticoagulants for at least 2 months.

By no means all patients with documented thrombophilia will experience spontaneous thrombosis during their lifetime. Studies suggest that approximately half of such patients will experience spontaneous thrombosis. Where deficiency of antithrombin III, protein C, protein S, or lupus anticoagulant are identified in patients without a history of thrombosis, it is not usual practice to initiate long-term prophylactic warfarin therapy as this treatment itself is not without risks. However, such patients should certainly receive some form of prophylaxis to cover surgery and pregnancy. Standard low-dose subcutaneous heparin therapy is perfectly satisfactory for general surgery. Higher doses of heparin or low-intensity anticoagulation with warfarin are advisable to cover major orthopaedic procedures. Affected women should not take oestrogen-containing oral contraceptives or hormone replacement therapy after the menopause. Plasma-derived concentrates of antithrombin III are available and may be given in addition to subcutaneous heparin to cover surgical procedures. Concentrates of protein C and S are not yet available.

An episode of thrombosis in a person with documented thrombophilia should be treated with heparin and warfarin in the standard fashion. Patients with antithrombin III deficiency may require more heparin than usual in order to achieve an adequately prolonged kaolin cephalin clotting time. Consideration should be given to long-term oral anticoagulation following an episode of thrombosis, if there are no contraindications.

Subjects with sickle haemoglobin (HbS) are found throughout tropical Africa, scattered in countries bordering the Mediterranean, in the Middle East and India. With the ease of world travel there can be no surgeon who can assume that a patient with sickle-cell disease requiring surgery will not come under his or her care.

Operating on a patient with sickle-cell disease significantly increases the risk to the patient of sickle-related complications. These may vary in severity from a painful limb crisis to life threatening neurological or respiratory crises but the risk may be minimized by appropriate perioperative care.

Dehydration, anoxia, hypothermia, and hypotension must be avoided. Local or regional anaesthesia may be preferable to general anaesthesia but limb tourniquets must not be used in order to avoid local tissue anoxia. For very minor surgery of short duration the above precautions may be adequate. For more major surgery the level of HbS should be reduced by blood transfusion. The most effective, albeit complicated, method is to exchange transfuse the patient preoperatively with the aim being to reduce the percentage of HbS from a predicted starting point of approximately 90 per cent to below 20 per cent. Once this is achieved the HbS percentage is maintained below 20 per cent by regular transfusions of red cells until the patient is mobilized and active postoperatively.

If emergency surgery is necessary in an unprepared patient an exchange transfusion to lower the HbS as far as possible is the best measure if time and the patient's condition allow. Otherwise, adherence to the general measures described above should be performed. Wound and pulmonary infections in the postoperative period should be rigorously sought and treated.

Finally, patients with sickle-cell disease may have minor cardiac and renal impairment but this will be rarely severe enough, at least in younger patients, to occasion any alarm peroperatively.

The risk for patients with sickle-cell trait undergoing surgery is not much greater than for normal individuals. However, the simple precautions of avoiding dehydration, anoxia etc. should be followed. Preoperative blood transfusion is not needed.

Patients with thalassaemia trait can undergo surgery without any special precautions whatsoever. In contrast, patients with &bgr;-thalassaemia major or anaemic patients with thalassaemia intermedia (a hotch-potch of thalassaemic disorders of a severity somewhere between thalassaemia major and thalassaemia trait) should be transfused up to a normal haemoglobin preoperatively. There is no need for exchange transfusion in this group.

The optimal routine treatment for patients with &bgr;-thalassaemia is regular blood transfusion to maintain a normal haemoglobin level. Those patients who have not simultaneously undergone a rigorous iron chelating programme may become iron overloaded and the clinical consequences of this, especially the risk of impaired cardiac function, will be of importance to the surgeon and anaesthetist in a patient needing surgery. There have recently been some encouraging reports of improvement in cardiac function in iron overloaded patients after aggressive high-dose therapy with desferrioxamine. Therapeutic benefit may take many months to achieve but is worthwhile if time allows.

Splenectomy is a common operation in thalassaemic patients. The increased risk of infection after splenectomy is well described. Briefly, patients should receive Pneumovax preoperatively and then penicillin by mouth for a minimum of 2 years postoperatively. In addition, an interesting thrombotic problem has been identified in some patients after splenectomy. Persistent thrombocytosis after splenectomy occurs predictably in patients where anaemia persists. The thrombocytosis may be associated with life threatening thromboembolic complications and this phenomenon has been well documented in patients with HbH disease. HbH disease occurs where there is a failure of three of the four &agr; globin genes to form &agr; globin chains with the clinical phenotype that of thalassaemia intermedia. A number of reports cite persistent thrombocytosis in these patients after splenectomy and the subsequent development of thrombophlebitis and thromboembolism.

In patients with a persistent thrombocytosis (>500 × 10&sup9;/l) after splenectomy it is wise to use life-long aspirin postoperatively and treat with warfarin those patients who develop signs of superficial thrombophlebitis or other thrombotic events despite the aspirin.

Among these disorders are included chronic myeloid leukaemia, primary proliferative polycythaemia (formerly known as polycythaemia rubra vera), essential thrombocythaemia, and myelofibrosis. In each of these disorders the platelet count may be significantly raised (often greater than 1000 × 10&sup9;/l) and in the perioperative period this may greatly enhance a thrombotic tendency. In addition, patients with polycythaemia will have a high haemoglobin associated with an increase in red cell mass which will further enhance the risk of thrombosis. The situation may be further complicated by an increased haemorrhagic tendency due to abnormal platelet function in each of the above disorders giving rise to the situation where a single patient is simultaneously at increased risk of thrombosis and bleeding.

Wherever possible surgery in such patients should be planned and appropriate treatment given to the patient to render the blood counts normal or near normal preoperatively. Once this is achieved it is good practice to give antithrombotic prophylaxis peroperatively. Preoperative coagulation tests should include, in addition to the platelet count, a bleeding time, prothrombin time, and activated partial thromboplastin time. Patients with a prolonged bleeding time, with the implication of impaired platelet function, should have platelets available to be transfused in the operative and postoperative period as necessary.

Patients with uncontrolled myeloproliferative disorders who are admitted with acute surgical problems requiring emergency surgery are at significant risk of thrombotic and haemorrhagic complications. Patients with polycythaemia may be venesected down to a normal haemoglobin relatively quickly (isovolumetric replacement of red cells by colloid is recommended) but a high platelet count cannot be quickly corrected. Perioperative anticoagulation is essential to reduce the risk of thrombosis. Bleeding perioperatively can be managed as above with platelet transfusions and fresh frozen plasma if there is prolongation of the prothrombin or activated partial thromboplastin time.

Perhaps the single most commonly performed operation in myeloproliferative disorders is splenectomy. The indications for this operation are beyond the scope of this section (see Chapter 37 21). A number of reports of splenectomy in the management of myeloproliferative disorders have been published. Mortality rates vary from 4 to 28 per cent and complication rates from 47 to 56 per cent. The major reported causes of morbidity were haemorrhage despite apparently appropriate precautions, followed by infection and thrombosis. Infections in one series occurred in 21 of 96 patients after splenectomy with the respiratory system, wound, and subphrenic space being the affected sites. The high mortality and morbidity rates for splenectomy in patients with myeloproliferative disorders compare unfavourably with the relatively trouble free course seen in patients undergoing splenectomy for autoimmune disorders such as idiopathic thrombocytopenic purpura, and autoimmune haemolytic anaemia (see later).

The indications for splenectomy in patients with lymphoproliferative disorders are lessening. Fewer staging laparotomies and splenectomies in patients with Hodgkin's disease are performed nowadays compared with 10 to 20 years ago and splenomegaly with hypersplenism due to tumour infiltration in patients with lymphoma can often be successfully treated by chemotherapy.

Nevertheless there remains a steady trickle of patients, often elderly and with massively enlarged spleens, who require surgery either for local symptoms induced by the splenomegaly, hypersplenism, or both (see also Chapter 37 21). Thrombocytopenia is the most common haemostatic abnormality and platelet transfusions may be needed perioperatively. Infection and bleeding are the most common complications but the postoperative risks are substantially less in patients with lymphoproliferative disorders compared with myeloproliferative disorders.

Myelodysplasia is a clonal disorder of bone marrow resulting in the production of morphologically and functionally abnormal red cells, white cells, and platelets. Infection and bleeding are common complications. The median survival is between 2 and 3 years from diagnosis and approximately one-third of patients will develop acute myeloid leukaemia in the terminal stages of their illness. In the early stages a typical patient may be noted to be mildly anaemic (often with a raised mean cell volume), leucopenic, and thrombocytopenic, or any combination of these.

From a surgical point of view the major risks are of haemorrhage and infection. Even if the platelet count is normal it is likely that functionally the platelets will be abnormal and a preoperative bleeding time should be checked. As with the myeloproliferative disorders platelets for transfusion should be available perioperatively and close attention given to prompt treatment of postoperative infection.

Splenectomy is a standard form of treatment in patients with idiopathic thrombocytopenic purpura who have failed to achieve a sustained response to primary drug therapy. In childhood idiopathic thrombocytopenic purpura up to 80 per cent of patients will remit spontaneously or after treatment with prednisolone and/or intravenous immunoglobulin (IV IgG). In contrast, first-line therapy results in sustained complete remission in only around 25 per cent of adults with this disorder.

Up to 90 per cent of patients with idiopathic thrombocytopenic purpura who have failed to respond to prednisolone, or who relapse after the withdrawal of prednisolone, will obtain a rapid rise in platelet count after treatment with intravenous immunoglobulin; unfortunately the response is usually shortlived. However, maintenance therapy with intravenous immunoglobulin may achieve long-term remission in 40 to 50 per cent of patients.

The expense of maintenance intravenous immunoglobulin means that splenectomy is the most common treatment option in patients who have failed to achieve a sustained response with prednisolone.

Improvement in the platelet count preoperatively can be achieved with prednisolone in those patients who have previously shown a response to this drug, or with intravenous immunoglobulin which improves the platelet count in up to 90 per cent of patients with idiopathic thrombocytopenic purpura.

The risk of adrenal suppression in patients treated with prednisolone for weeks or months preoperatively would necessitate increasing doses of steroid perioperatively. Also patients treated with prednisolone have an increased risk of perioperative infection. Intravenous immunoglobulin does not carry these risks and its use preoperatively may enhance the response to splenectomy. In preparing a patient for surgery intravenous immunoglobulin is the treatment of choice where the means and facilities allow it to be given. A typical dose is 0.4 g/kg body weight/day for 5 days. Platelet transfusion perioperatively should rarely be required nowadays.

Splenectomy will return the platelet count to normal, or substantially improve it, in around 60 to 70 per cent of patients. Only a small number (about 10 per cent) will be left with continuing haemorrhagic problems.

Splenectomy is the treatment of choice for the patients with autoimmune haemolytic anaemia who fail to respond to treatment with prednisolone. Splenectomy is effective in more than 50 per cent of cases. In contrast to patients with idiopathic thrombocytopenic purpura intravenous immunoglobulin has little or no place in the management of patients with autoimmune haemolyic anaemia. There are no specific precautions to be taken preoperatively. Severe anaemia can be corrected by blood transfusion although the half-life of the transfused red cells will be markedly shortened by immune destruction.

Postoperatively a transient increase in platelet count (often to >1000 × 10&sup9;/l) is common and perioperative antithrombotic prophylaxis should be given.

Bickerstaff KI, Morris PJ. Splenectomy for massive splenomegaly. Br J Surg 1987; 74: 346–49.

Blood Transfusion Task Force, British Committee for Standardization in Haematology. Guidelines for transfusion for massive blood loss. Clin Lab Haematol 1988; 10: 265–73.

Bloom AL, Forbes D, Thomas DP, Tuddenham EGD, eds. Thrombosis and haemostasis. 3rd edn. Edinburgh: Churchill Livingstone, 1993.

British Committee for Standards in Haematology. Guidelines for the use of fresh frozen plasma Transfusion Med 1992; 2: 57–63.

Cohen AR, Mizanin J, Schwartz E. Rapid removal of excessive iron with daily dose intravenous chelation therapy. J Pediatr 1989; 115: 151–5.

Dacie JV. The Haemolytic Anaemias, Vol. 2. Edinburgh: Churchill Livingstone—Longman Group UK Ltd, 1988.

Danforth DN, Fraker DL. Splenectomy for the massively enlarged spleen. American Surgeon 1991; 52: 108–13.

Duthie RB, Rizza CRC, Giangrande PLF, Dodd C. Management of musculoskeletal problems in haemophilia. Oxford: University Press, 1993.

Gill PG, Souter RG, Morris PJ. Splenectomy for hypersplenism in malignant lymphomas Br J Surg 1981; 1: 29–33.

Hirsch J, Dacie JV. Persistent post-splenectomy thrombocytosis and thrombo-embolism. A consequence of continuing anaemia. Br J Haematol 1966; 12: 44–53.

Jacobs P, Wood L, Dent DM. Splenectomy in the chronic myeloproliferative syndromes. S Afr Med J 1992; 81: 499–503.

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

Morris PJ, Cooper IA, Madigan JP. Splenectomy for hypersplenism in malignant lymphoma. Lancet 1975, ii: 250–2.

Poller L, Thomson JM, eds. Thrombosis and its management. Edinburgh: Churchill Livingstone, 1993.

Rizza CRC, Lowe GDO, eds. Haemophilia and its management. London: Balliere Tindall, 1994.

Woodman RC, Harker LA. Bleeding complications associated with cardiopulmonary bypass. Blood 1990; 76: 1680–97.