Clinical Review

INTRODUCTION

Transfusion therapy is an essential part of hematology practice, allowing for curative therapy of diseases such as leukemia, aplastic anemia, and aggressive lymphomas. Nonetheless, transfusions are associated with significant risks, including transmission of infections and transfusion-related reactions. Controversy remains about key issues in transfusion therapy, such as triggers for red cell transfusions. This article reviews the available blood products (Table 1) and indications for transfusion along with the associated risks, and also discusses specific clinical situations, such as massive transfusion.

BLOOD PRODUCTS

WHOLE BLOOD

Whole blood is the product of 1 unit of donated blood plus anticoagulant/preservative, and by definition contains 1 unit of plasma and red cells. Whole blood can be stored for 5 weeks. Although it was the standard product in the past, whole blood is rarely used since 1 unit of donated blood can now be fractionated into 1 unit of red blood cells (RBC), 1 unit of platelets, and 1 unit of fresh frozen plasma (FFP). Thus, the use of whole blood for just a single transfusion represents a waste of resources. There are 2 exceptions. One is autologous blood donations, which are whole blood units. Second, whole blood is increasingly being used in massive transfusions for trauma patients, with the rationale being that all essential blood components are being transfused at once.¹

PACKED RED CELLS

The remaining red cell mass after most of the plasma is removed is called the “packed” red cell unit (hematocrit = 70%–80%), and so red cells are often called “packed” red cells, or PRBC. A preservative is added to improve the flow of blood and to provide “nutrients” for the red cells, and this reduces the hematocrit to approximately 60%. The volume of a red cell unit is approximately 340 mL. In the average adult, 1 unit of RBC raises the hematocrit by 3%. The indications for transfusion of red cells are to increase red cell mass, and thus oxygen delivery, in patients who are compromised by their anemia.

Several randomized trials have helped define the indications for red cell transfusions and justify lower hematocrit thresholds for initiating transfusion. The TRICC (Transfusion Requirements in Critical Care Investigators for the Canadian Critical Care Trials Group) trial showed that in critical care patients (30-day mortality, 18.7%–23.3%), a conservative transfusion strategy of waiting until the hematocrit was below 21% had the same outcomes as transfusing at a threshold of 24%.² The TRACS (Transfusion Requirements After Cardiac Surgery) trial showed that a hematocrit target of 24% had the same benefit as a target of 30% in patients who had undergone cardiac bypass surgery.³ For patients with acute myocardial infarction, the outcomes were worse with aggressive transfusion at a hematocrit of 30% compared to 24%.⁴ In patients with upper gastrointestinal bleeding, a hemoglobin transfusion trigger of 7 g/dL was associated with a lower mortality than a trigger of 9 g/dL (5% versus 9%).⁵ Finally, the FOCUS (Transfusion Trigger Trial for Functional Outcomes in Cardiovascular Patients Undergoing Surgical Hip Fracture Repair) trial showed that in older patients (average age 80 years) who had undergone hip fracture surgery, transfusions based on symptoms and not a fixed trigger of 30% had the same outcomes but considerable savings in blood products.⁶ Based on these trials, decisions regarding when to transfuse patients should be based on symptoms and not “numbers.” Young patients, especially those with reversible anemias, can tolerate low blood counts and should not be transfused based on an arbitrary number.

PLATELETS

Several types of platelet products exist. One unit of platelet concentrate is derived from 1 unit of donor blood. Plateletpheresis from volunteer donors is also used to harvest platelets, with the resulting product referred to as plateletpheresis platelets. One unit of single-donor (pheresis) platelets is equivalent to 6 platelet concentrates. Finally, HLA-matched platelets are single-donor pheresis units that are obtained from an HLA-matched donor. This product should be ordered only if there is evidence of HLA antibodies (see Platelet Alloimmunization section).

The dose of platelets for the average patient is 6 units of platelet concentrate or 1 pheresis unit. In theory, 1 unit of platelet concentrate can raise the count by 5 to 7 × 10³/µL, but often this response is blunted by concurrent illness or bleeding. In patients who appear to have a poor response, the platelet count can be checked 15 minutes after platelet infusion. No rise or a minimal rise (< 2 × 10³/µL) in the platelet count is suggestive of platelet refractoriness, while a good 15-minute response but poor 24-hour count is more suggestive of consumption—fever, sepsis, drug, or splenomegaly—and not refractoriness.

The indication for platelet transfusion depends on the clinical situation. For patients with immune thrombocytopenia, platelets should not be transfused unless there is life-threatening bleeding. For stable patients with marrow aplasia from chemotherapy, a cut-off of a morning platelet count of less than 10 × 10³/µL has been shown to be as safe as higher levels for prophylactic transfusions.⁷ For patients with active bleeding, the platelet count should be kept above 50 × 10³/µL. Patients with acquired or inherited platelet dysfunction may benefit from transfusion no matter the platelet count.

Patients exposed to transfused white cells with different HLA antigens can develop antibodies to these antigens.⁸ Anti-HLA antibodies are common in patients who previously have received transfused blood that is not leukodepleted and in patients who have been pregnant. Since platelets carry class I HLA antigens, they will be rapidly destroyed by anti-HLA antibodies when transfused into these patients. In patients transfused for aplastic anemia or myelodysplasia, as many as 90% will become HLA-immunized. The incidence is lower in patients receiving chemotherapy but still can be as high as 60% to 90%.^9,10 Patients who have developed anti-HLA antibodies can respond to transfused platelets matched for HLA antigens. Unfortunately, some patients will either be a rare HLA type or be so heavily immunized that they will not respond to any platelet transfusion.

The significance of alloimmunization centers on 2 concepts: recognition and avoidance. Patients with HLA antibodies will fail to have an increment of their platelet counts with transfusions. Accordingly, patients who do not experience an increase in their count 15 minutes after the transfusion may have HLA antibodies. One can test for the presence of anti-HLA antibodies, although some patients instead have specific antiplatelet antibodies that will not respond to HLA-matched platelets. In patients who have been pregnant or previously transfused and are scheduled to undergo transplant or aggressive chemotherapy, it is wise to test for anti-HLA antibodies in order to plan their transfusion needs. The evidence suggests that transfused white cells are responsible for initiating the anti-HLA response. Trials have shown that giving leukodepleted blood products may reduce the incidence of alloimmunization, so patients who are not HLA-alloimmunized should receive only leukodepleted products.¹¹A difficult problem is bleeding in patients who are refractory to platelet transfusion.¹² Patients who test positive for the presence of anti-HLA antibodies can receive transfusions of HLA-matched platelets.¹³ Unfortunately, matched platelet transfusions are not effective in 20% to 70% of these patients. Also, since some loci are difficult to match, effective products may be unavailable. Finally, as many as 25% of patients have antiplatelet antibodies in which HLA-matched products will be ineffective. Platelet cross-matching can be performed to find compatible units for these patients, but this may not always be successful. In the patient who is totally refractory to platelet transfusion, consider drugs as an etiology of antiplatelet antibodies (especially vancomycin).¹⁴ Use of antifibrinolytic agents such as epsilon-aminocaproic acid or tranexamic acid may decrease the incidence of minor bleeding, but these are ineffective for major bleeding. “Platelet drips”—infusing either a platelet concentrate per hour or 1 plateletpheresis unit every 6 hours—may be given as a continuous infusion, but there is no evidence that this is helpful.¹⁵

FRESH FROZEN PLASMA

FFP is made from 1 unit of donated whole blood, with an average volume of 225 mL per unit. One unit of FFP can increase coagulation factor levels by 5% and fibrinogen by 10 mg/dL in the average stable patient. FFP can take 20 to 30 minutes to thaw before use, so in situations where FFP is needed quickly, the blood bank must be informed to “keep ahead” some units. Units of FFP that have been thawed but not used can be stored refrigerated for 5 days to prevent wasting blood products.

The indications for FFP are limited to several situations. These include a documented coagulation defect that can be corrected by a reasonable amount of FFP, such as factor V deficiency and factor XI deficiency, disseminated intravascular coagulation (DIC), reversal of warfarin, and massive transfusions. FFP is also used for the therapy of thrombotic thrombocytopenic purpura.

There is little justification for FFP transfusion in many of the clinical settings in which it is commonly used. For example, FFP is given for minor elevations of the INR in patients with liver disease, despite literature showing not only that the INR rise is not reflective of coagulation defects, but also that patients with liver disease may even be thrombophilic.^16,17 Reviews of FFP use found limited evidence-based indications for its use.^18,19 Also, several studies have shown that transfusion of FFP is not effective at reversing minor elevations of the INR (1.3–1.8).²⁰ In a meta-analysis, FFP was associated with increased risk for lung injury and a trend toward increased mortality.¹⁸

Cryoprecipitate is produced from 1 unit of FFP that is thawed at 4°C. The precipitate is resuspended with 10 mL of saline or FFP and refrozen for storage. One unit contains at least 150 mg of fibrinogen and 80 units of factor VIII, along with von Willebrand factor. Thawing time for cryoprecipitate is approximately 20 minutes.

Cryoprecipitate is used to raise the fibrinogen level in patients with DIC or massive transfusion with hemodilution. It is third-line therapy in the treatment of type 1 von Willebrand disease and is second-line therapy in the treatment of patients with other types of von Willebrand disease. Currently, von Willebrand factor concentrates are the preferred replacement product for von Willebrand disease. Cryoprecipitate can be used as a source of factor VIII for hemophiliacs, but the preferred product for these patients is the super pure factor VIII concentrates or recombinant products. Cryoprecipitate can also be used to shorten the bleeding time of uremic patients, but its effectiveness for this is controversial.

GRANULOCYTES

Granulocytes are harvested by leukopheresis of normal donors, with a target yield of 1010 granulocytes from each donor. To reach this target, the donors are often “stimulated” with neutrophil growth factors. The harvesting procedure can take 3 hours and is associated with some risks to the donor (eg, citrate toxicity). The current indications for granulocytes are very limited since the advent of neutrophil growth factors and improved antimicrobials.²¹ They can be useful in the neutropenic patient with a documented bacterial infection in whom the white blood cell count is not expected to recover in the near future. Given the difficulty of keeping the count up, these transfusions have been mainly used in treating small children.

SPECIAL BLOOD PRODUCTS

IRRADIATED BLOOD PRODUCTS

Irradiation of blood is performed for only one reason: to prevent transfusion-related graft-versus-host disease (TGVHD) (Table 2).²² The irradiation can be performed at the blood center or in the transfusion service of larger hospitals. The units are not radioactive and can be transfused safely to other patients. There is increased leakage of potassium in irradiated units of blood, so the units need to be transfused within 14 days; in patients potentially sensitive to potassium (eg, neonates), the units must be transfused within 24 hours. Patients undergoing stem cell transplant, those receiving either interuterine transfusions or products from relatives, any patient with Hodgkin disease or receiving purine analogs or alemtuzumab, and patients with severe congenital immune deficiencies should receive irradiated blood. Most would also advocate that patients with hematologic malignances receiving chemotherapy receive irradiated products, but this is more controversial.

LEUKPDEPLETED BLOOD

Contamination of blood products by white blood cells is increasingly being recognized as a possible cause of adverse effects in transfused patients, including febrile transfusion reactions, inducing HLA alloimmunization, immunosuppression, disease transmission, and TGVHD. Reducing white cells can reduce the incidence of all of these complications except TGVHD. Currently, white cells are removed by infusion through filters that trap the cells. This can be done either at the bedside, in the blood bank, or at the donor center. The majority of red cells provided by blood centers in many areas of the country are already leukoreduced, eliminating the need for labor-intensive filtration at the transfusion center or bedside. Platelets collected by plateletpheresis methods can also be made leukocyte-poor. The current indications for leukodepleted productions are:

1. Prevention of febrile transfusion reactions in patients with previous documented reactions
2. Prevention of HLA alloimmunization (ineffective if patient has received 1 or more blood products not leukodepleted or is already HLA immunized)
3. Prevention of cytomegalovirus (CMV) infection

CMV can be transmitted through any cellular blood product—red cells and platelets. For patients who are CMV-negative and receiving transplants, especially stem cell transplants, a new CMV infection can be devastating.²¹ For years only blood from CMV-negative donors was used to transfuse CMV-negative patients. This policy is effective in preventing CMV infection, but because 50% of the population is positive for CMV antibodies, it may potentially lead to shortages of products that could be transfused to the patient. Currently, leukoreduced blood products are used since leukofiltration of the blood is just as effective as transfusion of CMV-negative blood in preventing infections and allows greater use of all blood products.²³

COMPLICATIONS OF TRANSFUSIONS

HEMOLYTIC TRANSFUSION REACTION

There are 2 forms of hemolytic reactions—immediate and delayed.²⁴ The immediate reaction is associated with fevers, hypotension, back pain, and oliguria. In severe cases, DIC and renal failure may occur. The immediate reaction is due to transfusion of blood that reacts with the recipient’s preformed high-titer blood antigen antibodies, most often to ABO. This is fatal 2% of the time and occurs almost always as a result of errors in correct identification of the patient. Reactions are due to recipient antibodies attacking donated RBCs, resulting in release of hemoglobin and red cell membrane–antigen complexes. These complexes are believed to lead to the hypotension, fevers, chills, and renal damage associated with the hemolytic reaction. Treatment consists of immediately stopping the transfusion, notifying the blood bank, vigorous intravenous hydration to keep the urine output over 100 mL/hr, and supportive therapy.

The delayed reaction can range in severity from an abrupt drop in the hematocrit to normal response to transfusion but the patient developing a positive Coombs’ test. The delayed response is due to an anamnestic response to blood-group antigens. When the patient is exposed to the same antigen, there is a rise in antibody titer leading to the reaction. Some alloantibodies can lead to a brisk reaction, most often anti-Kidd. The frequency with which delayed transfusion reactions occur is underestimated because mild reactions often do not get worked up or even discovered.

ALLERGIC REACTIONS

Allergic reactions are common (1%–3% of transfusions) and occur in patients having antibodies to proteins in donor blood, which can lead to hives and itching with transfusions. Most of the time these allergic reactions are mild and can be treated with antihistamines. Prophylaxis with antihistamines is not indicated for future transfusions unless the reactions are frequent. Rarely these reactions can be associated with shock and hypotension. Patients who are immunoglobulin (Ig) A–deficient can develop anaphylactic reactions to IgA-containing blood products. Patients with severe allergic reactions need to have their IgA measured and, if deficient, receive only washed units or plasma from IgA-deficient donors to prevent future severe reactions.

FEBRILE REACTIONS

The most common transfusion reaction is a febrile reaction that occurs after the transfusion starts and that sometimes can be complicated by chills. This reaction often occurs due to the presence of leukocyte debris and cytokines in the donated blood. Therapy is supportive and involves stopping the transfusion and administering acetaminophen, but since hemolytic transfusion reactions can present with fever all patients need to be thoroughly evaluated. The incidence of reactions can be decreased by using leukodepleted blood and plateletpheresis platelets. Most patients do not benefit from receiving prophylactic acetaminophen for future transfusion unless they have multiple reactions.

Once thought a rare complication, transfusion-related acute lung injury (TRALI) is increasingly being recognized, with an incidence of approximately 1:5000 patients; it is now the most frequent cause of transfusion-related death.^24,25 TRALI is noncardiac pulmonary edema and typically manifests clinically with hypoxemia, fever, bilateral infiltrates, and hypotension 2 to 6 hours after blood is given. Ventilatory support is often required. Recovery is usually rapid (24–48 hours) and complete. The etiology is complex. In many cases, transfused anti-HLA antibodies react with the recipient’s white cells leading to pulmonary damage. Another theory is that transfusion of preformed cytokines leads to pulmonary damage. Because plasma products from multiparous women are most often associated with anti-HLA antibodies, the restricted use of blood products from women has decreased the incidence of TRALI over the past few years.²⁶

TRANSFUSION-ASSOCIATED CIRCULATORY OVERLOAD

Increasingly it being recognized that volume overload resulting from transfusions can lead to significant morbidity.²⁷ Patients with heart or renal disease or patients who already have compromised fluid status are at risk for transfusion-associated circulatory overload (TACO). Another risk factor is transfusion of multiple blood products. Patients with TACO develop dyspnea within 6 hours of transfusion, but do not have fever or rash with the dyspnea. The diagnosis is made by demonstrating circulatory overload (eg, high venous pressure, B-type natriuretic peptide). Treatment is aggressive diuresis. Strategies to prevent TACO include judicious use of blood products, especially in patients at risk for TACO, and the use of prophylactic diuretics, especially with red cell or plasma transfusions.²⁸

TRANSFUSION-RELATED GRAFT-VERSUS-HOST DISEASE

TGVHD is a rare reaction, but one that is most often fatal.²⁹ TGVHD occurs when donor lymphocytes attack the blood recipient’s organs—skin, liver, intestines, and marrow. This is very rare in the normal blood recipient unless the donor and recipient share some HLA haplotypes.³⁰ In immunosuppressed patients, TGVHD can occur with lesser degrees of HLA similarity, with cases reported in blood recipients who are mainly patients with Hodgkin disease or acute leukemia undergoing chemotherapy, and in patients receiving purine analogs. TGVHD had not been reported in AIDS patients despite profound immunosuppression, perhaps because the milieu of the patient does not allow lymphocyte expansion. Symptoms of TGVHD are an erythematous rash that may progress to epidermal toxic necrolysis, liver dysfunction, diarrhea, and pancytopenia. TGVHD is prevented by irradiating blood products given to at-risk patients with 2500 to 3500 rads. Directed blood donation from all blood relatives should also be irradiated. TGVHD cannot be prevented by leukopoor blood because the minute amount of lymphocytes that are not filtered still can lead to these complications.

POST-TRANSFUSION PURPURA

Patients with post-transfusion purpura (PTP) develop severe thrombocytopenia (< 10 × 10³/µL) with often severe bleeding 1 to 2 weeks after receiving any type of blood product.³¹ Patients who develop PTP most often lack platelet antigen PLA1 or other platelet antigens. For unknown reasons, exposure to the antigens from the transfusion leads to rapid destruction of the patient’s own platelets. The diagnostic clue is thrombocytopenia in a patient, typically female, who has received a red cell or platelet blood product in the past 7 to 10 days. Treatment consists of intravenous immunoglobulin³² and plasmapheresis to remove the offending antibody. If patients with a history of PTP require further transfusions, only PLA1-negative platelets should be given.

IRON OVERLOAD

Every transfusion of red cells delivers approximately 250 mg of iron to the recipient. Since there is no natural way of ridding the body of iron, heavily transfused patients are at risk of iron overload. This is most often seen in children heavily transfused for thalassemia. Starting in the second decade of life, these individuals will develop endocrinopathies due to iron overload, liver problems, and often fatal cardiomyopathies. Studies have shown that chelation of iron with deferoxamine can be effective in preventing this fatal complication.³³ Oral iron chelators such as deferasirox and deferiprone are also effective. The risk of iron overload in heavily transfused patients with myelodysplasia or other transfusion-dependent anemias is unclear, and uncertainty exists about the need for chelation.³⁴

Young patients who face years of transfusions should be started on iron chelation to avoid iron overload. For older patients with transfusion-dependent anemia, iron chelation therapy should be considered if their life expectancy is long (years to decades) or special studies such as T2-weighted cardiac magnetic resonance imaging showing iron overloading.³⁵

INFECTIOUS COMPLICATIONS

Concern over transmission of HIV infection via blood products in the late 1980s led to both a reduction in blood product use and a greater awareness of infectious complications of transfusion and their prevention. However, no blood product can ever be assumed to be safe for 2 reasons. One is that blood products can transmit infections during a “window period”—the time before a contaminated product can be detected by testing. The second is that blood is not screened for all potential infections (eg, babesiosis or new infections such as West Nile virus at the start of the outbreak). Risk of infection is reduced in 2 ways: deferral of potential infectious donors and blood product testing.

As part of the donation process, potential blood donors are asked a series of questions to see if they have risk factors for infections (eg, recent travel to malarious areas, recent tattoos), and if they answer positive are deferred from donating blood. Blood products are then tested for infectious agents by a combination of methods including detection of viral antigen, antibody response to infections, and more recently polymerase chain reaction (PCR).³⁶ Current screening includes syphilis testing; testing for antibodies to HIV, HTLV (human T-lymphotropic virus), hepatitis C virus, hepatitis B core antigen (HBcAg), hepatitis B surface antigen, and PCR for HIV, hepatitis B virus, HCV, and West Nile virus. Some centers also test for Trypanosoma cruzi, the cause of Chagas disease.

In the past, the numerically most common transfusion-related disease was hepatitis, first B and then C.^37,38 The first step in eliminating these infections was to stop paying donors for blood products. With the introduction of effective testing for hepatitis B and then C, the incidence of transfusion-related hepatitis has plummeted.³⁶ For example, with the introduction of a diagnostic test for hepatitis C, the estimated risk has fallen from 5% to less than 1 per million. Currently, the risk of transmission of hepatitis B and C, HIV, and HTLV is less than 1 in a million.³⁸

Despite this testing, blood transfusions can transmit a variety of infections, including malaria and babesiosis.³⁹ Any new blood-borne infection introduced into the population can get into the blood supply as well. For example, at the start of the West Nile virus epidemic, there was a cluster of transfusion-transmitted cases that resulted in severe and sometimes fatal illness in immunosuppressed patients, but this issue has been addressed with the development of a PCR assay for screening blood.⁴⁰ The rate of transfusion-related babesiosis has been increasing and screening for the causative parasite is being considered.

Acutely bleeding patients can require large amounts of transfusion products. Early data showed high mortality rates with transfusion of more than 20 units of blood,⁴¹ but with modern blood banking techniques and improved laboratory testing, this rate has decreased dramatically, with survival rates of 43% to 70% in patients transfused with more than 50 units of blood.⁴²

The basic approach to massive transfusions is to first transfuse the patient to maintain hemodynamic stability while specific blood tests are being obtained, and then to use the results of these early tests to guide the rest of the resuscitation. An important component is the ability to rapidly deliver standard packages of red cells, usually 6 to 10 units at a time, to the bleeding patient. To avoid delay while the patient’s blood is being typed, the first products delivered are blood group O Rh-positive units. Given the shortage of Rh-negative blood, this should be reserved for only empiric therapy of women of child-bearing age. Once the blood type is known, the patient can be switched over to type-specific blood.

In the past decade, there has been a shift toward increasing the amount of plasma given to patients receiving massive transfusions. This shift has occurred for 2 reasons. One is that modeling of coagulation changes in massive bleeding suggests the need for larger amounts of plasma to correct defects than have previously been recommended.⁴³ The other reason is based on analysis of resuscitation protocols used in military and civilian trauma centers showing that giving red cells and plasma units in a 1:1 ratio appears to be associated with improved outcomes in massive transfusion. Several studies have extended this concept to platelets, again suggesting improved survival with 1 unit of random donor platelets given 1:1 with red cells and plasma units. The PROPPR (Prospective Observational Multicenter Major Trauma Transfusion) study compared a 1:1:1 to 1:1:2 ratio in patients with severe trauma and major bleeding and found less exsanguination and faster achievement of hemostasis in the first 24 hours.⁴⁴ This has led to the widespread adoption of the 1:1 ratio by most trauma centers, and by default to other massive transfusion situations despite the lack of clinical trial data.⁴⁵

One barrier to increased use is that plasma is kept frozen and requires 20 minutes to thaw. Many institutions are now keeping inventories of thawed plasma available for immediate use, ranging from 2 to 4 units of group AB plasma to keeping their entire inventory as liquid plasma.⁴⁶ Plasma that is thawed but not used can be relabeled as “thawed plasma” and kept for up to 5 days. Also, many centers now use group A plasma for massive transfusions as this rarely leads to transfusion reactions and is much more available.⁴⁷ Research is currently under way on lyophilized plasma, which can be stored at room temperature and can be rapidly reconstituted for emergency use.

The standard approach for laboratory testing is obtaining 5 tests: hematocrit, platelet count, INR/prothrombin time, activated partial thromboplastin time (aPTT), and fibrinogen.⁴⁸ Product selection is guided by these tests, and they are repeated at regular intervals during the massive transfusion. A typical protocol is shown in Table 3. It is important as part of any protocol to have a flow chart that records laboratory results and products given that any member of the team can easily view.

The transfusion threshold for a low hematocrit depends on the stability of the patient. If the hematocrit is below 30% and the patient is bleeding or hemodynamically unstable, one should transfuse packed red cells. Stable patients can tolerate lower hematocrits, and an aggressive transfusion policy may even be detrimental.^2,49 If the patient is bleeding, has florid DIC, or has received platelet aggregation inhibitors, then keeping the platelet count above 50 × 10³/µL is reasonable. There are data regarding massive transfusion showing that keeping the platelet count above 50 × 10³/µL resulted in less microvascular bleeding.⁵⁰ The conventional dose of platelets is 6 to 8 platelet concentrates or 1 plateletpheresis unit.

While in the past fibrinogen targets of 50 to 100 mg/dL were recommended, recent data indicate that a target of 150 mg/dL or higher may be more appropriate.^51–53 Severe fibrinolysis may occur in certain clinical situations such as brain injuries, hepatic trauma, or ischemic limb reperfusion, and the use of large amounts of cryoprecipitate can be anticipated. In patients with an INR greater than 2 and an abnormal aPTT, one can give 2 to 4 units of FFP. For an aPTT greater than 1.5 times normal, 2 to 4 units of plasma should be given. Elevation of the aPTT above 1.8 times normal control is associated with microvascular bleeding in trauma patients.⁵⁴ Patients with marked abnormalities (eg, anaPTT more than 2 times normal) may require aggressive therapy with at least 15 to 30 mL/kg (4–8 units for an average adult) of plasma.⁵⁵

Recently there has been increasing interest in the use of thromboelastography (TEG) in massive transfusion.⁵⁶ This is a point-of-care assay performed on fresh whole blood that can assess multiple facets of hemostasis, including coagulation, platelet function, and fibrinolysis.^57,58 TEG is performed by placing a 0.35-mL sample of whole blood into an oscillating container with a sensor pin that measures the force of thrombus formation. TEG measures 5 parameters:

• r time: time from starting TEG until clot formation
• K time: time needed for tracing to go from 2 mm to 20 mm
• alpha angle: slope of tracing between r and K time
• MA: greatest amplitude of TEG tracing
• Whole blood lysis index: amplitude of tracing 60 minutes after MA.

Several centers have incorporated TEG into resuscitation protocols that include standardized strategies for responding to abnormalities. Data suggest that use of TEG may decrease the use of blood products, especially in cardiac surgery, but this has not been prospectively studied in massive transfusions.^56,59

COMPLICATIONS OF MASSIVE TRANSFUSIONS

Electrolyte abnormalities are unusual even in patients who receive massive transfusions.⁶⁰ Platelet concentrates and plasma contain citrate that can chelate calcium. However, the citrate is rapidly metabolized, and it is rare to see clinically significant hypocalcemia. Although empiric calcium replacement is often recommended, one study suggests that this is associated with a worse outcome and should not be done.⁶¹ If hypocalcemia is a clinical concern, then levels should be drawn to guide therapy. Stored blood is acidic, with a pH of 6.5 to 6.9. However, acidosis attributed solely to transfused blood is rare and most often is a reflection of the patient’s stability. Empirical bicarbonate replacement has been associated with severe alkalosis and is not recom mended.^62,63 Although potassium leaks out of stored red cells, even older units of blood contain only 8 mEq/L of potassium, so hyperkalemia is usually not a concern.

PATIENTS WITH AUTOIMMUNE HEMOLYTIC ANEMIA

Patients with autoimmune hemolytic anemia can be difficult to transfuse,⁶⁴ because the autoantibody can interfere with several aspects of the transfusion services evaluation. In some patients the autoantibody can be so strong that the patient’s blood type cannot be determined. In most patients, the final step of the cross-match—mixing the donor blood with recipient plasma—will show noncompatibility due to the autoantibodies reacting with any red cells.

The first step when transfusing a patient with autoimmune hemolytic anemia is to draw several tubes of blood for the transfusion service before any potential transfusions. This allows the transfusion service to remove the autoantibodies so they can screen for underlying alloantibodies. Second, if the patient requires immediate transfusion, then type-specific or O-negative blood should be given. If the patient has not been recently (months) transfused, the incidence of a severe transfusion reaction is low. The first unit should be infused slowly with close observation of the patient. For patients who have been multiply transfused, the use of an “in-vivo” cross-match may be helpful. This is where the patient is slowly transfused 10 to 15 mL of blood over 15 minutes. The the plasma and urine are then assessed for signs of hemolysis and, if negative, the remaining product is given.

REFUSAL OF BLOOD PRODUCTS

The initial step in managing patients who refuse blood products is to find out why they are refusing them. Many patients have an exaggerated fear of HIV and other infectious agents, so discussing the very low risk for infection transmission can often resolve the situation. The most common reason for refusal of blood products is religious belief. Jehovah’s Witness patients will refuse blood products due to their interpretation of the Bible.⁶⁵ All members will refuse red cells, plasma, and platelets, while decisions about “derived” blood products—products made by manipulation of the original donated units—are a matter of conscience. These include cryoprecipitate, intravenous gammaglobulin, and albumin.

In an elective situation, the first step is to discuss with the patient those products that are a matter of conscience and clearly document this. The patient’s blood count and iron stores should be assessed to identify any correctible causes of anemia or low iron stores before surgery. The use of erythropoietin to correct blood counts before surgery is controversial, as this may increase thrombosis risk and is contraindicated in patients with curable tumors.

For patients with acute blood loss, use of intravenous iron combined with high-dose erythropoietin is the most common approach to raise the blood count.⁶⁵ A recommended erythropoietin dose is 300 units/kg 3 times a week, dropping to 100 units/kg 3 times weekly until the goal hematocrit is reached. Another often overlooked step is to consolidate and minimize laboratory testing. The most important step is to be respectful of the patient and their beliefs. Many larger cities have liaisons that can help with interactions between Jehovah’s Witness patients and the health care system.

DESMOPRESSIN

Desmopressin (DDAVP) is a synthetic analog of antidiuretic hormone that raises the levels of both factor VIII and von Willebrand protein severalfold.⁶⁶ Desmopressin is effective in supporting hemostasis in patients with a wide variety of congenital and acquired bleeding disorders. However, desmopressin does not reduce blood loss before routine surgery in a healthy patient and should not be used for this purpose.⁶⁷

TRANEXAMIC ACID

Tranexamic acid is an antifibrinolytic agent that blocks the binding of plasmin to fibrin.⁶⁸ This agent was first shown to be useful in disorders that involve excessive fibrinolysis^69–73 or as adjunctive therapy for oral or dental procedures in patients with a bleeding diathesis. In patients with severe thrombocytopenia, the use of antifibrinolytic agents may reduce bleeding. Increasing data shows that tranexamic acid can prevent blood loss in a variety of surgeries including heart bypass, liver transplantation, and orthopedic surgery.⁷⁴ Patients across these settings have decreased blood loss and need for transfusion with no increased risk of thrombosis. The CRASH-2 study showed that the use of tranexamic acid significantly reduced mortality in trauma patients.⁷⁵ The WOMEN trial demonstrated that 1 g of tranexamic acid given to women with blood loss of more than 500 mL after vaginal delivery or 1000 mL after cesarean section has a risk reduction of death of 0.81 with no increased risk of thrombosis.⁷⁶ Given this abundant data, it is clear tranexamic acid needs to be part of any massive transfusion protocol.⁷⁷

RECOMBINANT FACTOR VIIa

Recombinant factor VIIa (rVIIa) was originally developed as a “bypass” agent to support hemostasis in hemophiliacs.⁷⁸ However, the use of rVIIa for a wide array of bleeding disorders, including patients with factor VII and XI deficiency and Glanzmann thrombasthenia, has been reported.⁷⁹ Increasingly, rVIIa is being used as a “universal hemostatic agent” for patients with uncontrolled bleeding from any mechanism.⁸⁰ Multiple case reports have described the use of rVIIa for bleeding in cardiac surgery patients, obstetrical bleeding, reversal of anticoagulation, and trauma.⁸¹ Unfortunately, little formal trial data exists to put these anecdotes into perspective, and formal review of clinical trial results has shown no benefit.^82,83 However, when used in older patients, especially those with vascular risk factors, the risk of arterial thrombosis appears to increase.⁸⁴ In the trials for intracranial hemorrhage, the thrombosis rate was 5% to 9%, and rates up to 10% for arterial events were seen in older patients in a review of all trials.^85–87 Given the lack of data but the evidence of risk, rVIIa use should be restricted to patients with documented bleeding disorders that have been shown to benefit by its use.

INTRODUCTION

Transfusion therapy is an essential part of hematology practice, allowing for curative therapy of diseases such as leukemia, aplastic anemia, and aggressive lymphomas. Nonetheless, transfusions are associated with significant risks, including transmission of infections and transfusion-related reactions. Controversy remains about key issues in transfusion therapy, such as triggers for red cell transfusions. This article reviews the available blood products (Table 1) and indications for transfusion along with the associated risks, and also discusses specific clinical situations, such as massive transfusion.

BLOOD PRODUCTS

WHOLE BLOOD

Whole blood is the product of 1 unit of donated blood plus anticoagulant/preservative, and by definition contains 1 unit of plasma and red cells. Whole blood can be stored for 5 weeks. Although it was the standard product in the past, whole blood is rarely used since 1 unit of donated blood can now be fractionated into 1 unit of red blood cells (RBC), 1 unit of platelets, and 1 unit of fresh frozen plasma (FFP). Thus, the use of whole blood for just a single transfusion represents a waste of resources. There are 2 exceptions. One is autologous blood donations, which are whole blood units. Second, whole blood is increasingly being used in massive transfusions for trauma patients, with the rationale being that all essential blood components are being transfused at once.¹

PACKED RED CELLS

The remaining red cell mass after most of the plasma is removed is called the “packed” red cell unit (hematocrit = 70%–80%), and so red cells are often called “packed” red cells, or PRBC. A preservative is added to improve the flow of blood and to provide “nutrients” for the red cells, and this reduces the hematocrit to approximately 60%. The volume of a red cell unit is approximately 340 mL. In the average adult, 1 unit of RBC raises the hematocrit by 3%. The indications for transfusion of red cells are to increase red cell mass, and thus oxygen delivery, in patients who are compromised by their anemia.

Several randomized trials have helped define the indications for red cell transfusions and justify lower hematocrit thresholds for initiating transfusion. The TRICC (Transfusion Requirements in Critical Care Investigators for the Canadian Critical Care Trials Group) trial showed that in critical care patients (30-day mortality, 18.7%–23.3%), a conservative transfusion strategy of waiting until the hematocrit was below 21% had the same outcomes as transfusing at a threshold of 24%.² The TRACS (Transfusion Requirements After Cardiac Surgery) trial showed that a hematocrit target of 24% had the same benefit as a target of 30% in patients who had undergone cardiac bypass surgery.³ For patients with acute myocardial infarction, the outcomes were worse with aggressive transfusion at a hematocrit of 30% compared to 24%.⁴ In patients with upper gastrointestinal bleeding, a hemoglobin transfusion trigger of 7 g/dL was associated with a lower mortality than a trigger of 9 g/dL (5% versus 9%).⁵ Finally, the FOCUS (Transfusion Trigger Trial for Functional Outcomes in Cardiovascular Patients Undergoing Surgical Hip Fracture Repair) trial showed that in older patients (average age 80 years) who had undergone hip fracture surgery, transfusions based on symptoms and not a fixed trigger of 30% had the same outcomes but considerable savings in blood products.⁶ Based on these trials, decisions regarding when to transfuse patients should be based on symptoms and not “numbers.” Young patients, especially those with reversible anemias, can tolerate low blood counts and should not be transfused based on an arbitrary number.

PLATELETS

Several types of platelet products exist. One unit of platelet concentrate is derived from 1 unit of donor blood. Plateletpheresis from volunteer donors is also used to harvest platelets, with the resulting product referred to as plateletpheresis platelets. One unit of single-donor (pheresis) platelets is equivalent to 6 platelet concentrates. Finally, HLA-matched platelets are single-donor pheresis units that are obtained from an HLA-matched donor. This product should be ordered only if there is evidence of HLA antibodies (see Platelet Alloimmunization section).

The dose of platelets for the average patient is 6 units of platelet concentrate or 1 pheresis unit. In theory, 1 unit of platelet concentrate can raise the count by 5 to 7 × 10³/µL, but often this response is blunted by concurrent illness or bleeding. In patients who appear to have a poor response, the platelet count can be checked 15 minutes after platelet infusion. No rise or a minimal rise (< 2 × 10³/µL) in the platelet count is suggestive of platelet refractoriness, while a good 15-minute response but poor 24-hour count is more suggestive of consumption—fever, sepsis, drug, or splenomegaly—and not refractoriness.

The indication for platelet transfusion depends on the clinical situation. For patients with immune thrombocytopenia, platelets should not be transfused unless there is life-threatening bleeding. For stable patients with marrow aplasia from chemotherapy, a cut-off of a morning platelet count of less than 10 × 10³/µL has been shown to be as safe as higher levels for prophylactic transfusions.⁷ For patients with active bleeding, the platelet count should be kept above 50 × 10³/µL. Patients with acquired or inherited platelet dysfunction may benefit from transfusion no matter the platelet count.

Patients exposed to transfused white cells with different HLA antigens can develop antibodies to these antigens.⁸ Anti-HLA antibodies are common in patients who previously have received transfused blood that is not leukodepleted and in patients who have been pregnant. Since platelets carry class I HLA antigens, they will be rapidly destroyed by anti-HLA antibodies when transfused into these patients. In patients transfused for aplastic anemia or myelodysplasia, as many as 90% will become HLA-immunized. The incidence is lower in patients receiving chemotherapy but still can be as high as 60% to 90%.^9,10 Patients who have developed anti-HLA antibodies can respond to transfused platelets matched for HLA antigens. Unfortunately, some patients will either be a rare HLA type or be so heavily immunized that they will not respond to any platelet transfusion.

The significance of alloimmunization centers on 2 concepts: recognition and avoidance. Patients with HLA antibodies will fail to have an increment of their platelet counts with transfusions. Accordingly, patients who do not experience an increase in their count 15 minutes after the transfusion may have HLA antibodies. One can test for the presence of anti-HLA antibodies, although some patients instead have specific antiplatelet antibodies that will not respond to HLA-matched platelets. In patients who have been pregnant or previously transfused and are scheduled to undergo transplant or aggressive chemotherapy, it is wise to test for anti-HLA antibodies in order to plan their transfusion needs. The evidence suggests that transfused white cells are responsible for initiating the anti-HLA response. Trials have shown that giving leukodepleted blood products may reduce the incidence of alloimmunization, so patients who are not HLA-alloimmunized should receive only leukodepleted products.¹¹A difficult problem is bleeding in patients who are refractory to platelet transfusion.¹² Patients who test positive for the presence of anti-HLA antibodies can receive transfusions of HLA-matched platelets.¹³ Unfortunately, matched platelet transfusions are not effective in 20% to 70% of these patients. Also, since some loci are difficult to match, effective products may be unavailable. Finally, as many as 25% of patients have antiplatelet antibodies in which HLA-matched products will be ineffective. Platelet cross-matching can be performed to find compatible units for these patients, but this may not always be successful. In the patient who is totally refractory to platelet transfusion, consider drugs as an etiology of antiplatelet antibodies (especially vancomycin).¹⁴ Use of antifibrinolytic agents such as epsilon-aminocaproic acid or tranexamic acid may decrease the incidence of minor bleeding, but these are ineffective for major bleeding. “Platelet drips”—infusing either a platelet concentrate per hour or 1 plateletpheresis unit every 6 hours—may be given as a continuous infusion, but there is no evidence that this is helpful.¹⁵

FRESH FROZEN PLASMA

FFP is made from 1 unit of donated whole blood, with an average volume of 225 mL per unit. One unit of FFP can increase coagulation factor levels by 5% and fibrinogen by 10 mg/dL in the average stable patient. FFP can take 20 to 30 minutes to thaw before use, so in situations where FFP is needed quickly, the blood bank must be informed to “keep ahead” some units. Units of FFP that have been thawed but not used can be stored refrigerated for 5 days to prevent wasting blood products.

The indications for FFP are limited to several situations. These include a documented coagulation defect that can be corrected by a reasonable amount of FFP, such as factor V deficiency and factor XI deficiency, disseminated intravascular coagulation (DIC), reversal of warfarin, and massive transfusions. FFP is also used for the therapy of thrombotic thrombocytopenic purpura.

There is little justification for FFP transfusion in many of the clinical settings in which it is commonly used. For example, FFP is given for minor elevations of the INR in patients with liver disease, despite literature showing not only that the INR rise is not reflective of coagulation defects, but also that patients with liver disease may even be thrombophilic.^16,17 Reviews of FFP use found limited evidence-based indications for its use.^18,19 Also, several studies have shown that transfusion of FFP is not effective at reversing minor elevations of the INR (1.3–1.8).²⁰ In a meta-analysis, FFP was associated with increased risk for lung injury and a trend toward increased mortality.¹⁸

Cryoprecipitate is produced from 1 unit of FFP that is thawed at 4°C. The precipitate is resuspended with 10 mL of saline or FFP and refrozen for storage. One unit contains at least 150 mg of fibrinogen and 80 units of factor VIII, along with von Willebrand factor. Thawing time for cryoprecipitate is approximately 20 minutes.

Cryoprecipitate is used to raise the fibrinogen level in patients with DIC or massive transfusion with hemodilution. It is third-line therapy in the treatment of type 1 von Willebrand disease and is second-line therapy in the treatment of patients with other types of von Willebrand disease. Currently, von Willebrand factor concentrates are the preferred replacement product for von Willebrand disease. Cryoprecipitate can be used as a source of factor VIII for hemophiliacs, but the preferred product for these patients is the super pure factor VIII concentrates or recombinant products. Cryoprecipitate can also be used to shorten the bleeding time of uremic patients, but its effectiveness for this is controversial.

GRANULOCYTES

Granulocytes are harvested by leukopheresis of normal donors, with a target yield of 1010 granulocytes from each donor. To reach this target, the donors are often “stimulated” with neutrophil growth factors. The harvesting procedure can take 3 hours and is associated with some risks to the donor (eg, citrate toxicity). The current indications for granulocytes are very limited since the advent of neutrophil growth factors and improved antimicrobials.²¹ They can be useful in the neutropenic patient with a documented bacterial infection in whom the white blood cell count is not expected to recover in the near future. Given the difficulty of keeping the count up, these transfusions have been mainly used in treating small children.

SPECIAL BLOOD PRODUCTS

IRRADIATED BLOOD PRODUCTS

Irradiation of blood is performed for only one reason: to prevent transfusion-related graft-versus-host disease (TGVHD) (Table 2).²² The irradiation can be performed at the blood center or in the transfusion service of larger hospitals. The units are not radioactive and can be transfused safely to other patients. There is increased leakage of potassium in irradiated units of blood, so the units need to be transfused within 14 days; in patients potentially sensitive to potassium (eg, neonates), the units must be transfused within 24 hours. Patients undergoing stem cell transplant, those receiving either interuterine transfusions or products from relatives, any patient with Hodgkin disease or receiving purine analogs or alemtuzumab, and patients with severe congenital immune deficiencies should receive irradiated blood. Most would also advocate that patients with hematologic malignances receiving chemotherapy receive irradiated products, but this is more controversial.

LEUKPDEPLETED BLOOD

Contamination of blood products by white blood cells is increasingly being recognized as a possible cause of adverse effects in transfused patients, including febrile transfusion reactions, inducing HLA alloimmunization, immunosuppression, disease transmission, and TGVHD. Reducing white cells can reduce the incidence of all of these complications except TGVHD. Currently, white cells are removed by infusion through filters that trap the cells. This can be done either at the bedside, in the blood bank, or at the donor center. The majority of red cells provided by blood centers in many areas of the country are already leukoreduced, eliminating the need for labor-intensive filtration at the transfusion center or bedside. Platelets collected by plateletpheresis methods can also be made leukocyte-poor. The current indications for leukodepleted productions are:

1. Prevention of febrile transfusion reactions in patients with previous documented reactions
2. Prevention of HLA alloimmunization (ineffective if patient has received 1 or more blood products not leukodepleted or is already HLA immunized)
3. Prevention of cytomegalovirus (CMV) infection

CMV can be transmitted through any cellular blood product—red cells and platelets. For patients who are CMV-negative and receiving transplants, especially stem cell transplants, a new CMV infection can be devastating.²¹ For years only blood from CMV-negative donors was used to transfuse CMV-negative patients. This policy is effective in preventing CMV infection, but because 50% of the population is positive for CMV antibodies, it may potentially lead to shortages of products that could be transfused to the patient. Currently, leukoreduced blood products are used since leukofiltration of the blood is just as effective as transfusion of CMV-negative blood in preventing infections and allows greater use of all blood products.²³

COMPLICATIONS OF TRANSFUSIONS

HEMOLYTIC TRANSFUSION REACTION

There are 2 forms of hemolytic reactions—immediate and delayed.²⁴ The immediate reaction is associated with fevers, hypotension, back pain, and oliguria. In severe cases, DIC and renal failure may occur. The immediate reaction is due to transfusion of blood that reacts with the recipient’s preformed high-titer blood antigen antibodies, most often to ABO. This is fatal 2% of the time and occurs almost always as a result of errors in correct identification of the patient. Reactions are due to recipient antibodies attacking donated RBCs, resulting in release of hemoglobin and red cell membrane–antigen complexes. These complexes are believed to lead to the hypotension, fevers, chills, and renal damage associated with the hemolytic reaction. Treatment consists of immediately stopping the transfusion, notifying the blood bank, vigorous intravenous hydration to keep the urine output over 100 mL/hr, and supportive therapy.

The delayed reaction can range in severity from an abrupt drop in the hematocrit to normal response to transfusion but the patient developing a positive Coombs’ test. The delayed response is due to an anamnestic response to blood-group antigens. When the patient is exposed to the same antigen, there is a rise in antibody titer leading to the reaction. Some alloantibodies can lead to a brisk reaction, most often anti-Kidd. The frequency with which delayed transfusion reactions occur is underestimated because mild reactions often do not get worked up or even discovered.

ALLERGIC REACTIONS

Allergic reactions are common (1%–3% of transfusions) and occur in patients having antibodies to proteins in donor blood, which can lead to hives and itching with transfusions. Most of the time these allergic reactions are mild and can be treated with antihistamines. Prophylaxis with antihistamines is not indicated for future transfusions unless the reactions are frequent. Rarely these reactions can be associated with shock and hypotension. Patients who are immunoglobulin (Ig) A–deficient can develop anaphylactic reactions to IgA-containing blood products. Patients with severe allergic reactions need to have their IgA measured and, if deficient, receive only washed units or plasma from IgA-deficient donors to prevent future severe reactions.

FEBRILE REACTIONS

The most common transfusion reaction is a febrile reaction that occurs after the transfusion starts and that sometimes can be complicated by chills. This reaction often occurs due to the presence of leukocyte debris and cytokines in the donated blood. Therapy is supportive and involves stopping the transfusion and administering acetaminophen, but since hemolytic transfusion reactions can present with fever all patients need to be thoroughly evaluated. The incidence of reactions can be decreased by using leukodepleted blood and plateletpheresis platelets. Most patients do not benefit from receiving prophylactic acetaminophen for future transfusion unless they have multiple reactions.

Once thought a rare complication, transfusion-related acute lung injury (TRALI) is increasingly being recognized, with an incidence of approximately 1:5000 patients; it is now the most frequent cause of transfusion-related death.^24,25 TRALI is noncardiac pulmonary edema and typically manifests clinically with hypoxemia, fever, bilateral infiltrates, and hypotension 2 to 6 hours after blood is given. Ventilatory support is often required. Recovery is usually rapid (24–48 hours) and complete. The etiology is complex. In many cases, transfused anti-HLA antibodies react with the recipient’s white cells leading to pulmonary damage. Another theory is that transfusion of preformed cytokines leads to pulmonary damage. Because plasma products from multiparous women are most often associated with anti-HLA antibodies, the restricted use of blood products from women has decreased the incidence of TRALI over the past few years.²⁶

TRANSFUSION-ASSOCIATED CIRCULATORY OVERLOAD

Increasingly it being recognized that volume overload resulting from transfusions can lead to significant morbidity.²⁷ Patients with heart or renal disease or patients who already have compromised fluid status are at risk for transfusion-associated circulatory overload (TACO). Another risk factor is transfusion of multiple blood products. Patients with TACO develop dyspnea within 6 hours of transfusion, but do not have fever or rash with the dyspnea. The diagnosis is made by demonstrating circulatory overload (eg, high venous pressure, B-type natriuretic peptide). Treatment is aggressive diuresis. Strategies to prevent TACO include judicious use of blood products, especially in patients at risk for TACO, and the use of prophylactic diuretics, especially with red cell or plasma transfusions.²⁸

TRANSFUSION-RELATED GRAFT-VERSUS-HOST DISEASE

TGVHD is a rare reaction, but one that is most often fatal.²⁹ TGVHD occurs when donor lymphocytes attack the blood recipient’s organs—skin, liver, intestines, and marrow. This is very rare in the normal blood recipient unless the donor and recipient share some HLA haplotypes.³⁰ In immunosuppressed patients, TGVHD can occur with lesser degrees of HLA similarity, with cases reported in blood recipients who are mainly patients with Hodgkin disease or acute leukemia undergoing chemotherapy, and in patients receiving purine analogs. TGVHD had not been reported in AIDS patients despite profound immunosuppression, perhaps because the milieu of the patient does not allow lymphocyte expansion. Symptoms of TGVHD are an erythematous rash that may progress to epidermal toxic necrolysis, liver dysfunction, diarrhea, and pancytopenia. TGVHD is prevented by irradiating blood products given to at-risk patients with 2500 to 3500 rads. Directed blood donation from all blood relatives should also be irradiated. TGVHD cannot be prevented by leukopoor blood because the minute amount of lymphocytes that are not filtered still can lead to these complications.

POST-TRANSFUSION PURPURA

Patients with post-transfusion purpura (PTP) develop severe thrombocytopenia (< 10 × 10³/µL) with often severe bleeding 1 to 2 weeks after receiving any type of blood product.³¹ Patients who develop PTP most often lack platelet antigen PLA1 or other platelet antigens. For unknown reasons, exposure to the antigens from the transfusion leads to rapid destruction of the patient’s own platelets. The diagnostic clue is thrombocytopenia in a patient, typically female, who has received a red cell or platelet blood product in the past 7 to 10 days. Treatment consists of intravenous immunoglobulin³² and plasmapheresis to remove the offending antibody. If patients with a history of PTP require further transfusions, only PLA1-negative platelets should be given.

IRON OVERLOAD

Every transfusion of red cells delivers approximately 250 mg of iron to the recipient. Since there is no natural way of ridding the body of iron, heavily transfused patients are at risk of iron overload. This is most often seen in children heavily transfused for thalassemia. Starting in the second decade of life, these individuals will develop endocrinopathies due to iron overload, liver problems, and often fatal cardiomyopathies. Studies have shown that chelation of iron with deferoxamine can be effective in preventing this fatal complication.³³ Oral iron chelators such as deferasirox and deferiprone are also effective. The risk of iron overload in heavily transfused patients with myelodysplasia or other transfusion-dependent anemias is unclear, and uncertainty exists about the need for chelation.³⁴

Young patients who face years of transfusions should be started on iron chelation to avoid iron overload. For older patients with transfusion-dependent anemia, iron chelation therapy should be considered if their life expectancy is long (years to decades) or special studies such as T2-weighted cardiac magnetic resonance imaging showing iron overloading.³⁵

INFECTIOUS COMPLICATIONS

Concern over transmission of HIV infection via blood products in the late 1980s led to both a reduction in blood product use and a greater awareness of infectious complications of transfusion and their prevention. However, no blood product can ever be assumed to be safe for 2 reasons. One is that blood products can transmit infections during a “window period”—the time before a contaminated product can be detected by testing. The second is that blood is not screened for all potential infections (eg, babesiosis or new infections such as West Nile virus at the start of the outbreak). Risk of infection is reduced in 2 ways: deferral of potential infectious donors and blood product testing.

As part of the donation process, potential blood donors are asked a series of questions to see if they have risk factors for infections (eg, recent travel to malarious areas, recent tattoos), and if they answer positive are deferred from donating blood. Blood products are then tested for infectious agents by a combination of methods including detection of viral antigen, antibody response to infections, and more recently polymerase chain reaction (PCR).³⁶ Current screening includes syphilis testing; testing for antibodies to HIV, HTLV (human T-lymphotropic virus), hepatitis C virus, hepatitis B core antigen (HBcAg), hepatitis B surface antigen, and PCR for HIV, hepatitis B virus, HCV, and West Nile virus. Some centers also test for Trypanosoma cruzi, the cause of Chagas disease.

In the past, the numerically most common transfusion-related disease was hepatitis, first B and then C.^37,38 The first step in eliminating these infections was to stop paying donors for blood products. With the introduction of effective testing for hepatitis B and then C, the incidence of transfusion-related hepatitis has plummeted.³⁶ For example, with the introduction of a diagnostic test for hepatitis C, the estimated risk has fallen from 5% to less than 1 per million. Currently, the risk of transmission of hepatitis B and C, HIV, and HTLV is less than 1 in a million.³⁸

Despite this testing, blood transfusions can transmit a variety of infections, including malaria and babesiosis.³⁹ Any new blood-borne infection introduced into the population can get into the blood supply as well. For example, at the start of the West Nile virus epidemic, there was a cluster of transfusion-transmitted cases that resulted in severe and sometimes fatal illness in immunosuppressed patients, but this issue has been addressed with the development of a PCR assay for screening blood.⁴⁰ The rate of transfusion-related babesiosis has been increasing and screening for the causative parasite is being considered.

Acutely bleeding patients can require large amounts of transfusion products. Early data showed high mortality rates with transfusion of more than 20 units of blood,⁴¹ but with modern blood banking techniques and improved laboratory testing, this rate has decreased dramatically, with survival rates of 43% to 70% in patients transfused with more than 50 units of blood.⁴²

The basic approach to massive transfusions is to first transfuse the patient to maintain hemodynamic stability while specific blood tests are being obtained, and then to use the results of these early tests to guide the rest of the resuscitation. An important component is the ability to rapidly deliver standard packages of red cells, usually 6 to 10 units at a time, to the bleeding patient. To avoid delay while the patient’s blood is being typed, the first products delivered are blood group O Rh-positive units. Given the shortage of Rh-negative blood, this should be reserved for only empiric therapy of women of child-bearing age. Once the blood type is known, the patient can be switched over to type-specific blood.

In the past decade, there has been a shift toward increasing the amount of plasma given to patients receiving massive transfusions. This shift has occurred for 2 reasons. One is that modeling of coagulation changes in massive bleeding suggests the need for larger amounts of plasma to correct defects than have previously been recommended.⁴³ The other reason is based on analysis of resuscitation protocols used in military and civilian trauma centers showing that giving red cells and plasma units in a 1:1 ratio appears to be associated with improved outcomes in massive transfusion. Several studies have extended this concept to platelets, again suggesting improved survival with 1 unit of random donor platelets given 1:1 with red cells and plasma units. The PROPPR (Prospective Observational Multicenter Major Trauma Transfusion) study compared a 1:1:1 to 1:1:2 ratio in patients with severe trauma and major bleeding and found less exsanguination and faster achievement of hemostasis in the first 24 hours.⁴⁴ This has led to the widespread adoption of the 1:1 ratio by most trauma centers, and by default to other massive transfusion situations despite the lack of clinical trial data.⁴⁵

One barrier to increased use is that plasma is kept frozen and requires 20 minutes to thaw. Many institutions are now keeping inventories of thawed plasma available for immediate use, ranging from 2 to 4 units of group AB plasma to keeping their entire inventory as liquid plasma.⁴⁶ Plasma that is thawed but not used can be relabeled as “thawed plasma” and kept for up to 5 days. Also, many centers now use group A plasma for massive transfusions as this rarely leads to transfusion reactions and is much more available.⁴⁷ Research is currently under way on lyophilized plasma, which can be stored at room temperature and can be rapidly reconstituted for emergency use.

The standard approach for laboratory testing is obtaining 5 tests: hematocrit, platelet count, INR/prothrombin time, activated partial thromboplastin time (aPTT), and fibrinogen.⁴⁸ Product selection is guided by these tests, and they are repeated at regular intervals during the massive transfusion. A typical protocol is shown in Table 3. It is important as part of any protocol to have a flow chart that records laboratory results and products given that any member of the team can easily view.

The transfusion threshold for a low hematocrit depends on the stability of the patient. If the hematocrit is below 30% and the patient is bleeding or hemodynamically unstable, one should transfuse packed red cells. Stable patients can tolerate lower hematocrits, and an aggressive transfusion policy may even be detrimental.^2,49 If the patient is bleeding, has florid DIC, or has received platelet aggregation inhibitors, then keeping the platelet count above 50 × 10³/µL is reasonable. There are data regarding massive transfusion showing that keeping the platelet count above 50 × 10³/µL resulted in less microvascular bleeding.⁵⁰ The conventional dose of platelets is 6 to 8 platelet concentrates or 1 plateletpheresis unit.

While in the past fibrinogen targets of 50 to 100 mg/dL were recommended, recent data indicate that a target of 150 mg/dL or higher may be more appropriate.^51–53 Severe fibrinolysis may occur in certain clinical situations such as brain injuries, hepatic trauma, or ischemic limb reperfusion, and the use of large amounts of cryoprecipitate can be anticipated. In patients with an INR greater than 2 and an abnormal aPTT, one can give 2 to 4 units of FFP. For an aPTT greater than 1.5 times normal, 2 to 4 units of plasma should be given. Elevation of the aPTT above 1.8 times normal control is associated with microvascular bleeding in trauma patients.⁵⁴ Patients with marked abnormalities (eg, anaPTT more than 2 times normal) may require aggressive therapy with at least 15 to 30 mL/kg (4–8 units for an average adult) of plasma.⁵⁵

Recently there has been increasing interest in the use of thromboelastography (TEG) in massive transfusion.⁵⁶ This is a point-of-care assay performed on fresh whole blood that can assess multiple facets of hemostasis, including coagulation, platelet function, and fibrinolysis.^57,58 TEG is performed by placing a 0.35-mL sample of whole blood into an oscillating container with a sensor pin that measures the force of thrombus formation. TEG measures 5 parameters:

• r time: time from starting TEG until clot formation
• K time: time needed for tracing to go from 2 mm to 20 mm
• alpha angle: slope of tracing between r and K time
• MA: greatest amplitude of TEG tracing
• Whole blood lysis index: amplitude of tracing 60 minutes after MA.

Several centers have incorporated TEG into resuscitation protocols that include standardized strategies for responding to abnormalities. Data suggest that use of TEG may decrease the use of blood products, especially in cardiac surgery, but this has not been prospectively studied in massive transfusions.^56,59

COMPLICATIONS OF MASSIVE TRANSFUSIONS

Electrolyte abnormalities are unusual even in patients who receive massive transfusions.⁶⁰ Platelet concentrates and plasma contain citrate that can chelate calcium. However, the citrate is rapidly metabolized, and it is rare to see clinically significant hypocalcemia. Although empiric calcium replacement is often recommended, one study suggests that this is associated with a worse outcome and should not be done.⁶¹ If hypocalcemia is a clinical concern, then levels should be drawn to guide therapy. Stored blood is acidic, with a pH of 6.5 to 6.9. However, acidosis attributed solely to transfused blood is rare and most often is a reflection of the patient’s stability. Empirical bicarbonate replacement has been associated with severe alkalosis and is not recom mended.^62,63 Although potassium leaks out of stored red cells, even older units of blood contain only 8 mEq/L of potassium, so hyperkalemia is usually not a concern.

PATIENTS WITH AUTOIMMUNE HEMOLYTIC ANEMIA

Patients with autoimmune hemolytic anemia can be difficult to transfuse,⁶⁴ because the autoantibody can interfere with several aspects of the transfusion services evaluation. In some patients the autoantibody can be so strong that the patient’s blood type cannot be determined. In most patients, the final step of the cross-match—mixing the donor blood with recipient plasma—will show noncompatibility due to the autoantibodies reacting with any red cells.

The first step when transfusing a patient with autoimmune hemolytic anemia is to draw several tubes of blood for the transfusion service before any potential transfusions. This allows the transfusion service to remove the autoantibodies so they can screen for underlying alloantibodies. Second, if the patient requires immediate transfusion, then type-specific or O-negative blood should be given. If the patient has not been recently (months) transfused, the incidence of a severe transfusion reaction is low. The first unit should be infused slowly with close observation of the patient. For patients who have been multiply transfused, the use of an “in-vivo” cross-match may be helpful. This is where the patient is slowly transfused 10 to 15 mL of blood over 15 minutes. The the plasma and urine are then assessed for signs of hemolysis and, if negative, the remaining product is given.

REFUSAL OF BLOOD PRODUCTS

The initial step in managing patients who refuse blood products is to find out why they are refusing them. Many patients have an exaggerated fear of HIV and other infectious agents, so discussing the very low risk for infection transmission can often resolve the situation. The most common reason for refusal of blood products is religious belief. Jehovah’s Witness patients will refuse blood products due to their interpretation of the Bible.⁶⁵ All members will refuse red cells, plasma, and platelets, while decisions about “derived” blood products—products made by manipulation of the original donated units—are a matter of conscience. These include cryoprecipitate, intravenous gammaglobulin, and albumin.

In an elective situation, the first step is to discuss with the patient those products that are a matter of conscience and clearly document this. The patient’s blood count and iron stores should be assessed to identify any correctible causes of anemia or low iron stores before surgery. The use of erythropoietin to correct blood counts before surgery is controversial, as this may increase thrombosis risk and is contraindicated in patients with curable tumors.

For patients with acute blood loss, use of intravenous iron combined with high-dose erythropoietin is the most common approach to raise the blood count.⁶⁵ A recommended erythropoietin dose is 300 units/kg 3 times a week, dropping to 100 units/kg 3 times weekly until the goal hematocrit is reached. Another often overlooked step is to consolidate and minimize laboratory testing. The most important step is to be respectful of the patient and their beliefs. Many larger cities have liaisons that can help with interactions between Jehovah’s Witness patients and the health care system.

DESMOPRESSIN

Desmopressin (DDAVP) is a synthetic analog of antidiuretic hormone that raises the levels of both factor VIII and von Willebrand protein severalfold.⁶⁶ Desmopressin is effective in supporting hemostasis in patients with a wide variety of congenital and acquired bleeding disorders. However, desmopressin does not reduce blood loss before routine surgery in a healthy patient and should not be used for this purpose.⁶⁷

TRANEXAMIC ACID

Tranexamic acid is an antifibrinolytic agent that blocks the binding of plasmin to fibrin.⁶⁸ This agent was first shown to be useful in disorders that involve excessive fibrinolysis^69–73 or as adjunctive therapy for oral or dental procedures in patients with a bleeding diathesis. In patients with severe thrombocytopenia, the use of antifibrinolytic agents may reduce bleeding. Increasing data shows that tranexamic acid can prevent blood loss in a variety of surgeries including heart bypass, liver transplantation, and orthopedic surgery.⁷⁴ Patients across these settings have decreased blood loss and need for transfusion with no increased risk of thrombosis. The CRASH-2 study showed that the use of tranexamic acid significantly reduced mortality in trauma patients.⁷⁵ The WOMEN trial demonstrated that 1 g of tranexamic acid given to women with blood loss of more than 500 mL after vaginal delivery or 1000 mL after cesarean section has a risk reduction of death of 0.81 with no increased risk of thrombosis.⁷⁶ Given this abundant data, it is clear tranexamic acid needs to be part of any massive transfusion protocol.⁷⁷

RECOMBINANT FACTOR VIIa

Recombinant factor VIIa (rVIIa) was originally developed as a “bypass” agent to support hemostasis in hemophiliacs.⁷⁸ However, the use of rVIIa for a wide array of bleeding disorders, including patients with factor VII and XI deficiency and Glanzmann thrombasthenia, has been reported.⁷⁹ Increasingly, rVIIa is being used as a “universal hemostatic agent” for patients with uncontrolled bleeding from any mechanism.⁸⁰ Multiple case reports have described the use of rVIIa for bleeding in cardiac surgery patients, obstetrical bleeding, reversal of anticoagulation, and trauma.⁸¹ Unfortunately, little formal trial data exists to put these anecdotes into perspective, and formal review of clinical trial results has shown no benefit.^82,83 However, when used in older patients, especially those with vascular risk factors, the risk of arterial thrombosis appears to increase.⁸⁴ In the trials for intracranial hemorrhage, the thrombosis rate was 5% to 9%, and rates up to 10% for arterial events were seen in older patients in a review of all trials.^85–87 Given the lack of data but the evidence of risk, rVIIa use should be restricted to patients with documented bleeding disorders that have been shown to benefit by its use.

INTRODUCTION

Transfusion therapy is an essential part of hematology practice, allowing for curative therapy of diseases such as leukemia, aplastic anemia, and aggressive lymphomas. Nonetheless, transfusions are associated with significant risks, including transmission of infections and transfusion-related reactions. Controversy remains about key issues in transfusion therapy, such as triggers for red cell transfusions. This article reviews the available blood products (Table 1) and indications for transfusion along with the associated risks, and also discusses specific clinical situations, such as massive transfusion.

BLOOD PRODUCTS

WHOLE BLOOD

Whole blood is the product of 1 unit of donated blood plus anticoagulant/preservative, and by definition contains 1 unit of plasma and red cells. Whole blood can be stored for 5 weeks. Although it was the standard product in the past, whole blood is rarely used since 1 unit of donated blood can now be fractionated into 1 unit of red blood cells (RBC), 1 unit of platelets, and 1 unit of fresh frozen plasma (FFP). Thus, the use of whole blood for just a single transfusion represents a waste of resources. There are 2 exceptions. One is autologous blood donations, which are whole blood units. Second, whole blood is increasingly being used in massive transfusions for trauma patients, with the rationale being that all essential blood components are being transfused at once.¹

PACKED RED CELLS

The remaining red cell mass after most of the plasma is removed is called the “packed” red cell unit (hematocrit = 70%–80%), and so red cells are often called “packed” red cells, or PRBC. A preservative is added to improve the flow of blood and to provide “nutrients” for the red cells, and this reduces the hematocrit to approximately 60%. The volume of a red cell unit is approximately 340 mL. In the average adult, 1 unit of RBC raises the hematocrit by 3%. The indications for transfusion of red cells are to increase red cell mass, and thus oxygen delivery, in patients who are compromised by their anemia.

Several randomized trials have helped define the indications for red cell transfusions and justify lower hematocrit thresholds for initiating transfusion. The TRICC (Transfusion Requirements in Critical Care Investigators for the Canadian Critical Care Trials Group) trial showed that in critical care patients (30-day mortality, 18.7%–23.3%), a conservative transfusion strategy of waiting until the hematocrit was below 21% had the same outcomes as transfusing at a threshold of 24%.² The TRACS (Transfusion Requirements After Cardiac Surgery) trial showed that a hematocrit target of 24% had the same benefit as a target of 30% in patients who had undergone cardiac bypass surgery.³ For patients with acute myocardial infarction, the outcomes were worse with aggressive transfusion at a hematocrit of 30% compared to 24%.⁴ In patients with upper gastrointestinal bleeding, a hemoglobin transfusion trigger of 7 g/dL was associated with a lower mortality than a trigger of 9 g/dL (5% versus 9%).⁵ Finally, the FOCUS (Transfusion Trigger Trial for Functional Outcomes in Cardiovascular Patients Undergoing Surgical Hip Fracture Repair) trial showed that in older patients (average age 80 years) who had undergone hip fracture surgery, transfusions based on symptoms and not a fixed trigger of 30% had the same outcomes but considerable savings in blood products.⁶ Based on these trials, decisions regarding when to transfuse patients should be based on symptoms and not “numbers.” Young patients, especially those with reversible anemias, can tolerate low blood counts and should not be transfused based on an arbitrary number.

PLATELETS

Several types of platelet products exist. One unit of platelet concentrate is derived from 1 unit of donor blood. Plateletpheresis from volunteer donors is also used to harvest platelets, with the resulting product referred to as plateletpheresis platelets. One unit of single-donor (pheresis) platelets is equivalent to 6 platelet concentrates. Finally, HLA-matched platelets are single-donor pheresis units that are obtained from an HLA-matched donor. This product should be ordered only if there is evidence of HLA antibodies (see Platelet Alloimmunization section).

The dose of platelets for the average patient is 6 units of platelet concentrate or 1 pheresis unit. In theory, 1 unit of platelet concentrate can raise the count by 5 to 7 × 10³/µL, but often this response is blunted by concurrent illness or bleeding. In patients who appear to have a poor response, the platelet count can be checked 15 minutes after platelet infusion. No rise or a minimal rise (< 2 × 10³/µL) in the platelet count is suggestive of platelet refractoriness, while a good 15-minute response but poor 24-hour count is more suggestive of consumption—fever, sepsis, drug, or splenomegaly—and not refractoriness.

The indication for platelet transfusion depends on the clinical situation. For patients with immune thrombocytopenia, platelets should not be transfused unless there is life-threatening bleeding. For stable patients with marrow aplasia from chemotherapy, a cut-off of a morning platelet count of less than 10 × 10³/µL has been shown to be as safe as higher levels for prophylactic transfusions.⁷ For patients with active bleeding, the platelet count should be kept above 50 × 10³/µL. Patients with acquired or inherited platelet dysfunction may benefit from transfusion no matter the platelet count.

Patients exposed to transfused white cells with different HLA antigens can develop antibodies to these antigens.⁸ Anti-HLA antibodies are common in patients who previously have received transfused blood that is not leukodepleted and in patients who have been pregnant. Since platelets carry class I HLA antigens, they will be rapidly destroyed by anti-HLA antibodies when transfused into these patients. In patients transfused for aplastic anemia or myelodysplasia, as many as 90% will become HLA-immunized. The incidence is lower in patients receiving chemotherapy but still can be as high as 60% to 90%.^9,10 Patients who have developed anti-HLA antibodies can respond to transfused platelets matched for HLA antigens. Unfortunately, some patients will either be a rare HLA type or be so heavily immunized that they will not respond to any platelet transfusion.

The significance of alloimmunization centers on 2 concepts: recognition and avoidance. Patients with HLA antibodies will fail to have an increment of their platelet counts with transfusions. Accordingly, patients who do not experience an increase in their count 15 minutes after the transfusion may have HLA antibodies. One can test for the presence of anti-HLA antibodies, although some patients instead have specific antiplatelet antibodies that will not respond to HLA-matched platelets. In patients who have been pregnant or previously transfused and are scheduled to undergo transplant or aggressive chemotherapy, it is wise to test for anti-HLA antibodies in order to plan their transfusion needs. The evidence suggests that transfused white cells are responsible for initiating the anti-HLA response. Trials have shown that giving leukodepleted blood products may reduce the incidence of alloimmunization, so patients who are not HLA-alloimmunized should receive only leukodepleted products.¹¹A difficult problem is bleeding in patients who are refractory to platelet transfusion.¹² Patients who test positive for the presence of anti-HLA antibodies can receive transfusions of HLA-matched platelets.¹³ Unfortunately, matched platelet transfusions are not effective in 20% to 70% of these patients. Also, since some loci are difficult to match, effective products may be unavailable. Finally, as many as 25% of patients have antiplatelet antibodies in which HLA-matched products will be ineffective. Platelet cross-matching can be performed to find compatible units for these patients, but this may not always be successful. In the patient who is totally refractory to platelet transfusion, consider drugs as an etiology of antiplatelet antibodies (especially vancomycin).¹⁴ Use of antifibrinolytic agents such as epsilon-aminocaproic acid or tranexamic acid may decrease the incidence of minor bleeding, but these are ineffective for major bleeding. “Platelet drips”—infusing either a platelet concentrate per hour or 1 plateletpheresis unit every 6 hours—may be given as a continuous infusion, but there is no evidence that this is helpful.¹⁵

FRESH FROZEN PLASMA

FFP is made from 1 unit of donated whole blood, with an average volume of 225 mL per unit. One unit of FFP can increase coagulation factor levels by 5% and fibrinogen by 10 mg/dL in the average stable patient. FFP can take 20 to 30 minutes to thaw before use, so in situations where FFP is needed quickly, the blood bank must be informed to “keep ahead” some units. Units of FFP that have been thawed but not used can be stored refrigerated for 5 days to prevent wasting blood products.

The indications for FFP are limited to several situations. These include a documented coagulation defect that can be corrected by a reasonable amount of FFP, such as factor V deficiency and factor XI deficiency, disseminated intravascular coagulation (DIC), reversal of warfarin, and massive transfusions. FFP is also used for the therapy of thrombotic thrombocytopenic purpura.

There is little justification for FFP transfusion in many of the clinical settings in which it is commonly used. For example, FFP is given for minor elevations of the INR in patients with liver disease, despite literature showing not only that the INR rise is not reflective of coagulation defects, but also that patients with liver disease may even be thrombophilic.^16,17 Reviews of FFP use found limited evidence-based indications for its use.^18,19 Also, several studies have shown that transfusion of FFP is not effective at reversing minor elevations of the INR (1.3–1.8).²⁰ In a meta-analysis, FFP was associated with increased risk for lung injury and a trend toward increased mortality.¹⁸

Cryoprecipitate is produced from 1 unit of FFP that is thawed at 4°C. The precipitate is resuspended with 10 mL of saline or FFP and refrozen for storage. One unit contains at least 150 mg of fibrinogen and 80 units of factor VIII, along with von Willebrand factor. Thawing time for cryoprecipitate is approximately 20 minutes.

Cryoprecipitate is used to raise the fibrinogen level in patients with DIC or massive transfusion with hemodilution. It is third-line therapy in the treatment of type 1 von Willebrand disease and is second-line therapy in the treatment of patients with other types of von Willebrand disease. Currently, von Willebrand factor concentrates are the preferred replacement product for von Willebrand disease. Cryoprecipitate can be used as a source of factor VIII for hemophiliacs, but the preferred product for these patients is the super pure factor VIII concentrates or recombinant products. Cryoprecipitate can also be used to shorten the bleeding time of uremic patients, but its effectiveness for this is controversial.

GRANULOCYTES

Granulocytes are harvested by leukopheresis of normal donors, with a target yield of 1010 granulocytes from each donor. To reach this target, the donors are often “stimulated” with neutrophil growth factors. The harvesting procedure can take 3 hours and is associated with some risks to the donor (eg, citrate toxicity). The current indications for granulocytes are very limited since the advent of neutrophil growth factors and improved antimicrobials.²¹ They can be useful in the neutropenic patient with a documented bacterial infection in whom the white blood cell count is not expected to recover in the near future. Given the difficulty of keeping the count up, these transfusions have been mainly used in treating small children.

SPECIAL BLOOD PRODUCTS

IRRADIATED BLOOD PRODUCTS

Irradiation of blood is performed for only one reason: to prevent transfusion-related graft-versus-host disease (TGVHD) (Table 2).²² The irradiation can be performed at the blood center or in the transfusion service of larger hospitals. The units are not radioactive and can be transfused safely to other patients. There is increased leakage of potassium in irradiated units of blood, so the units need to be transfused within 14 days; in patients potentially sensitive to potassium (eg, neonates), the units must be transfused within 24 hours. Patients undergoing stem cell transplant, those receiving either interuterine transfusions or products from relatives, any patient with Hodgkin disease or receiving purine analogs or alemtuzumab, and patients with severe congenital immune deficiencies should receive irradiated blood. Most would also advocate that patients with hematologic malignances receiving chemotherapy receive irradiated products, but this is more controversial.

LEUKPDEPLETED BLOOD

Contamination of blood products by white blood cells is increasingly being recognized as a possible cause of adverse effects in transfused patients, including febrile transfusion reactions, inducing HLA alloimmunization, immunosuppression, disease transmission, and TGVHD. Reducing white cells can reduce the incidence of all of these complications except TGVHD. Currently, white cells are removed by infusion through filters that trap the cells. This can be done either at the bedside, in the blood bank, or at the donor center. The majority of red cells provided by blood centers in many areas of the country are already leukoreduced, eliminating the need for labor-intensive filtration at the transfusion center or bedside. Platelets collected by plateletpheresis methods can also be made leukocyte-poor. The current indications for leukodepleted productions are:

1. Prevention of febrile transfusion reactions in patients with previous documented reactions
2. Prevention of HLA alloimmunization (ineffective if patient has received 1 or more blood products not leukodepleted or is already HLA immunized)
3. Prevention of cytomegalovirus (CMV) infection

CMV can be transmitted through any cellular blood product—red cells and platelets. For patients who are CMV-negative and receiving transplants, especially stem cell transplants, a new CMV infection can be devastating.²¹ For years only blood from CMV-negative donors was used to transfuse CMV-negative patients. This policy is effective in preventing CMV infection, but because 50% of the population is positive for CMV antibodies, it may potentially lead to shortages of products that could be transfused to the patient. Currently, leukoreduced blood products are used since leukofiltration of the blood is just as effective as transfusion of CMV-negative blood in preventing infections and allows greater use of all blood products.²³

COMPLICATIONS OF TRANSFUSIONS

HEMOLYTIC TRANSFUSION REACTION

There are 2 forms of hemolytic reactions—immediate and delayed.²⁴ The immediate reaction is associated with fevers, hypotension, back pain, and oliguria. In severe cases, DIC and renal failure may occur. The immediate reaction is due to transfusion of blood that reacts with the recipient’s preformed high-titer blood antigen antibodies, most often to ABO. This is fatal 2% of the time and occurs almost always as a result of errors in correct identification of the patient. Reactions are due to recipient antibodies attacking donated RBCs, resulting in release of hemoglobin and red cell membrane–antigen complexes. These complexes are believed to lead to the hypotension, fevers, chills, and renal damage associated with the hemolytic reaction. Treatment consists of immediately stopping the transfusion, notifying the blood bank, vigorous intravenous hydration to keep the urine output over 100 mL/hr, and supportive therapy.

The delayed reaction can range in severity from an abrupt drop in the hematocrit to normal response to transfusion but the patient developing a positive Coombs’ test. The delayed response is due to an anamnestic response to blood-group antigens. When the patient is exposed to the same antigen, there is a rise in antibody titer leading to the reaction. Some alloantibodies can lead to a brisk reaction, most often anti-Kidd. The frequency with which delayed transfusion reactions occur is underestimated because mild reactions often do not get worked up or even discovered.

ALLERGIC REACTIONS

Allergic reactions are common (1%–3% of transfusions) and occur in patients having antibodies to proteins in donor blood, which can lead to hives and itching with transfusions. Most of the time these allergic reactions are mild and can be treated with antihistamines. Prophylaxis with antihistamines is not indicated for future transfusions unless the reactions are frequent. Rarely these reactions can be associated with shock and hypotension. Patients who are immunoglobulin (Ig) A–deficient can develop anaphylactic reactions to IgA-containing blood products. Patients with severe allergic reactions need to have their IgA measured and, if deficient, receive only washed units or plasma from IgA-deficient donors to prevent future severe reactions.

FEBRILE REACTIONS

The most common transfusion reaction is a febrile reaction that occurs after the transfusion starts and that sometimes can be complicated by chills. This reaction often occurs due to the presence of leukocyte debris and cytokines in the donated blood. Therapy is supportive and involves stopping the transfusion and administering acetaminophen, but since hemolytic transfusion reactions can present with fever all patients need to be thoroughly evaluated. The incidence of reactions can be decreased by using leukodepleted blood and plateletpheresis platelets. Most patients do not benefit from receiving prophylactic acetaminophen for future transfusion unless they have multiple reactions.

Once thought a rare complication, transfusion-related acute lung injury (TRALI) is increasingly being recognized, with an incidence of approximately 1:5000 patients; it is now the most frequent cause of transfusion-related death.^24,25 TRALI is noncardiac pulmonary edema and typically manifests clinically with hypoxemia, fever, bilateral infiltrates, and hypotension 2 to 6 hours after blood is given. Ventilatory support is often required. Recovery is usually rapid (24–48 hours) and complete. The etiology is complex. In many cases, transfused anti-HLA antibodies react with the recipient’s white cells leading to pulmonary damage. Another theory is that transfusion of preformed cytokines leads to pulmonary damage. Because plasma products from multiparous women are most often associated with anti-HLA antibodies, the restricted use of blood products from women has decreased the incidence of TRALI over the past few years.²⁶

TRANSFUSION-ASSOCIATED CIRCULATORY OVERLOAD

Increasingly it being recognized that volume overload resulting from transfusions can lead to significant morbidity.²⁷ Patients with heart or renal disease or patients who already have compromised fluid status are at risk for transfusion-associated circulatory overload (TACO). Another risk factor is transfusion of multiple blood products. Patients with TACO develop dyspnea within 6 hours of transfusion, but do not have fever or rash with the dyspnea. The diagnosis is made by demonstrating circulatory overload (eg, high venous pressure, B-type natriuretic peptide). Treatment is aggressive diuresis. Strategies to prevent TACO include judicious use of blood products, especially in patients at risk for TACO, and the use of prophylactic diuretics, especially with red cell or plasma transfusions.²⁸

TRANSFUSION-RELATED GRAFT-VERSUS-HOST DISEASE

TGVHD is a rare reaction, but one that is most often fatal.²⁹ TGVHD occurs when donor lymphocytes attack the blood recipient’s organs—skin, liver, intestines, and marrow. This is very rare in the normal blood recipient unless the donor and recipient share some HLA haplotypes.³⁰ In immunosuppressed patients, TGVHD can occur with lesser degrees of HLA similarity, with cases reported in blood recipients who are mainly patients with Hodgkin disease or acute leukemia undergoing chemotherapy, and in patients receiving purine analogs. TGVHD had not been reported in AIDS patients despite profound immunosuppression, perhaps because the milieu of the patient does not allow lymphocyte expansion. Symptoms of TGVHD are an erythematous rash that may progress to epidermal toxic necrolysis, liver dysfunction, diarrhea, and pancytopenia. TGVHD is prevented by irradiating blood products given to at-risk patients with 2500 to 3500 rads. Directed blood donation from all blood relatives should also be irradiated. TGVHD cannot be prevented by leukopoor blood because the minute amount of lymphocytes that are not filtered still can lead to these complications.

POST-TRANSFUSION PURPURA

Patients with post-transfusion purpura (PTP) develop severe thrombocytopenia (< 10 × 10³/µL) with often severe bleeding 1 to 2 weeks after receiving any type of blood product.³¹ Patients who develop PTP most often lack platelet antigen PLA1 or other platelet antigens. For unknown reasons, exposure to the antigens from the transfusion leads to rapid destruction of the patient’s own platelets. The diagnostic clue is thrombocytopenia in a patient, typically female, who has received a red cell or platelet blood product in the past 7 to 10 days. Treatment consists of intravenous immunoglobulin³² and plasmapheresis to remove the offending antibody. If patients with a history of PTP require further transfusions, only PLA1-negative platelets should be given.

IRON OVERLOAD

Every transfusion of red cells delivers approximately 250 mg of iron to the recipient. Since there is no natural way of ridding the body of iron, heavily transfused patients are at risk of iron overload. This is most often seen in children heavily transfused for thalassemia. Starting in the second decade of life, these individuals will develop endocrinopathies due to iron overload, liver problems, and often fatal cardiomyopathies. Studies have shown that chelation of iron with deferoxamine can be effective in preventing this fatal complication.³³ Oral iron chelators such as deferasirox and deferiprone are also effective. The risk of iron overload in heavily transfused patients with myelodysplasia or other transfusion-dependent anemias is unclear, and uncertainty exists about the need for chelation.³⁴

Young patients who face years of transfusions should be started on iron chelation to avoid iron overload. For older patients with transfusion-dependent anemia, iron chelation therapy should be considered if their life expectancy is long (years to decades) or special studies such as T2-weighted cardiac magnetic resonance imaging showing iron overloading.³⁵

INFECTIOUS COMPLICATIONS

Concern over transmission of HIV infection via blood products in the late 1980s led to both a reduction in blood product use and a greater awareness of infectious complications of transfusion and their prevention. However, no blood product can ever be assumed to be safe for 2 reasons. One is that blood products can transmit infections during a “window period”—the time before a contaminated product can be detected by testing. The second is that blood is not screened for all potential infections (eg, babesiosis or new infections such as West Nile virus at the start of the outbreak). Risk of infection is reduced in 2 ways: deferral of potential infectious donors and blood product testing.

As part of the donation process, potential blood donors are asked a series of questions to see if they have risk factors for infections (eg, recent travel to malarious areas, recent tattoos), and if they answer positive are deferred from donating blood. Blood products are then tested for infectious agents by a combination of methods including detection of viral antigen, antibody response to infections, and more recently polymerase chain reaction (PCR).³⁶ Current screening includes syphilis testing; testing for antibodies to HIV, HTLV (human T-lymphotropic virus), hepatitis C virus, hepatitis B core antigen (HBcAg), hepatitis B surface antigen, and PCR for HIV, hepatitis B virus, HCV, and West Nile virus. Some centers also test for Trypanosoma cruzi, the cause of Chagas disease.

In the past, the numerically most common transfusion-related disease was hepatitis, first B and then C.^37,38 The first step in eliminating these infections was to stop paying donors for blood products. With the introduction of effective testing for hepatitis B and then C, the incidence of transfusion-related hepatitis has plummeted.³⁶ For example, with the introduction of a diagnostic test for hepatitis C, the estimated risk has fallen from 5% to less than 1 per million. Currently, the risk of transmission of hepatitis B and C, HIV, and HTLV is less than 1 in a million.³⁸

Despite this testing, blood transfusions can transmit a variety of infections, including malaria and babesiosis.³⁹ Any new blood-borne infection introduced into the population can get into the blood supply as well. For example, at the start of the West Nile virus epidemic, there was a cluster of transfusion-transmitted cases that resulted in severe and sometimes fatal illness in immunosuppressed patients, but this issue has been addressed with the development of a PCR assay for screening blood.⁴⁰ The rate of transfusion-related babesiosis has been increasing and screening for the causative parasite is being considered.

Acutely bleeding patients can require large amounts of transfusion products. Early data showed high mortality rates with transfusion of more than 20 units of blood,⁴¹ but with modern blood banking techniques and improved laboratory testing, this rate has decreased dramatically, with survival rates of 43% to 70% in patients transfused with more than 50 units of blood.⁴²

The basic approach to massive transfusions is to first transfuse the patient to maintain hemodynamic stability while specific blood tests are being obtained, and then to use the results of these early tests to guide the rest of the resuscitation. An important component is the ability to rapidly deliver standard packages of red cells, usually 6 to 10 units at a time, to the bleeding patient. To avoid delay while the patient’s blood is being typed, the first products delivered are blood group O Rh-positive units. Given the shortage of Rh-negative blood, this should be reserved for only empiric therapy of women of child-bearing age. Once the blood type is known, the patient can be switched over to type-specific blood.

In the past decade, there has been a shift toward increasing the amount of plasma given to patients receiving massive transfusions. This shift has occurred for 2 reasons. One is that modeling of coagulation changes in massive bleeding suggests the need for larger amounts of plasma to correct defects than have previously been recommended.⁴³ The other reason is based on analysis of resuscitation protocols used in military and civilian trauma centers showing that giving red cells and plasma units in a 1:1 ratio appears to be associated with improved outcomes in massive transfusion. Several studies have extended this concept to platelets, again suggesting improved survival with 1 unit of random donor platelets given 1:1 with red cells and plasma units. The PROPPR (Prospective Observational Multicenter Major Trauma Transfusion) study compared a 1:1:1 to 1:1:2 ratio in patients with severe trauma and major bleeding and found less exsanguination and faster achievement of hemostasis in the first 24 hours.⁴⁴ This has led to the widespread adoption of the 1:1 ratio by most trauma centers, and by default to other massive transfusion situations despite the lack of clinical trial data.⁴⁵

One barrier to increased use is that plasma is kept frozen and requires 20 minutes to thaw. Many institutions are now keeping inventories of thawed plasma available for immediate use, ranging from 2 to 4 units of group AB plasma to keeping their entire inventory as liquid plasma.⁴⁶ Plasma that is thawed but not used can be relabeled as “thawed plasma” and kept for up to 5 days. Also, many centers now use group A plasma for massive transfusions as this rarely leads to transfusion reactions and is much more available.⁴⁷ Research is currently under way on lyophilized plasma, which can be stored at room temperature and can be rapidly reconstituted for emergency use.

The standard approach for laboratory testing is obtaining 5 tests: hematocrit, platelet count, INR/prothrombin time, activated partial thromboplastin time (aPTT), and fibrinogen.⁴⁸ Product selection is guided by these tests, and they are repeated at regular intervals during the massive transfusion. A typical protocol is shown in Table 3. It is important as part of any protocol to have a flow chart that records laboratory results and products given that any member of the team can easily view.

The transfusion threshold for a low hematocrit depends on the stability of the patient. If the hematocrit is below 30% and the patient is bleeding or hemodynamically unstable, one should transfuse packed red cells. Stable patients can tolerate lower hematocrits, and an aggressive transfusion policy may even be detrimental.^2,49 If the patient is bleeding, has florid DIC, or has received platelet aggregation inhibitors, then keeping the platelet count above 50 × 10³/µL is reasonable. There are data regarding massive transfusion showing that keeping the platelet count above 50 × 10³/µL resulted in less microvascular bleeding.⁵⁰ The conventional dose of platelets is 6 to 8 platelet concentrates or 1 plateletpheresis unit.

While in the past fibrinogen targets of 50 to 100 mg/dL were recommended, recent data indicate that a target of 150 mg/dL or higher may be more appropriate.^51–53 Severe fibrinolysis may occur in certain clinical situations such as brain injuries, hepatic trauma, or ischemic limb reperfusion, and the use of large amounts of cryoprecipitate can be anticipated. In patients with an INR greater than 2 and an abnormal aPTT, one can give 2 to 4 units of FFP. For an aPTT greater than 1.5 times normal, 2 to 4 units of plasma should be given. Elevation of the aPTT above 1.8 times normal control is associated with microvascular bleeding in trauma patients.⁵⁴ Patients with marked abnormalities (eg, anaPTT more than 2 times normal) may require aggressive therapy with at least 15 to 30 mL/kg (4–8 units for an average adult) of plasma.⁵⁵

Recently there has been increasing interest in the use of thromboelastography (TEG) in massive transfusion.⁵⁶ This is a point-of-care assay performed on fresh whole blood that can assess multiple facets of hemostasis, including coagulation, platelet function, and fibrinolysis.^57,58 TEG is performed by placing a 0.35-mL sample of whole blood into an oscillating container with a sensor pin that measures the force of thrombus formation. TEG measures 5 parameters:

• r time: time from starting TEG until clot formation
• K time: time needed for tracing to go from 2 mm to 20 mm
• alpha angle: slope of tracing between r and K time
• MA: greatest amplitude of TEG tracing
• Whole blood lysis index: amplitude of tracing 60 minutes after MA.

Several centers have incorporated TEG into resuscitation protocols that include standardized strategies for responding to abnormalities. Data suggest that use of TEG may decrease the use of blood products, especially in cardiac surgery, but this has not been prospectively studied in massive transfusions.^56,59

COMPLICATIONS OF MASSIVE TRANSFUSIONS

Electrolyte abnormalities are unusual even in patients who receive massive transfusions.⁶⁰ Platelet concentrates and plasma contain citrate that can chelate calcium. However, the citrate is rapidly metabolized, and it is rare to see clinically significant hypocalcemia. Although empiric calcium replacement is often recommended, one study suggests that this is associated with a worse outcome and should not be done.⁶¹ If hypocalcemia is a clinical concern, then levels should be drawn to guide therapy. Stored blood is acidic, with a pH of 6.5 to 6.9. However, acidosis attributed solely to transfused blood is rare and most often is a reflection of the patient’s stability. Empirical bicarbonate replacement has been associated with severe alkalosis and is not recom mended.^62,63 Although potassium leaks out of stored red cells, even older units of blood contain only 8 mEq/L of potassium, so hyperkalemia is usually not a concern.

PATIENTS WITH AUTOIMMUNE HEMOLYTIC ANEMIA

Patients with autoimmune hemolytic anemia can be difficult to transfuse,⁶⁴ because the autoantibody can interfere with several aspects of the transfusion services evaluation. In some patients the autoantibody can be so strong that the patient’s blood type cannot be determined. In most patients, the final step of the cross-match—mixing the donor blood with recipient plasma—will show noncompatibility due to the autoantibodies reacting with any red cells.

The first step when transfusing a patient with autoimmune hemolytic anemia is to draw several tubes of blood for the transfusion service before any potential transfusions. This allows the transfusion service to remove the autoantibodies so they can screen for underlying alloantibodies. Second, if the patient requires immediate transfusion, then type-specific or O-negative blood should be given. If the patient has not been recently (months) transfused, the incidence of a severe transfusion reaction is low. The first unit should be infused slowly with close observation of the patient. For patients who have been multiply transfused, the use of an “in-vivo” cross-match may be helpful. This is where the patient is slowly transfused 10 to 15 mL of blood over 15 minutes. The the plasma and urine are then assessed for signs of hemolysis and, if negative, the remaining product is given.

REFUSAL OF BLOOD PRODUCTS

The initial step in managing patients who refuse blood products is to find out why they are refusing them. Many patients have an exaggerated fear of HIV and other infectious agents, so discussing the very low risk for infection transmission can often resolve the situation. The most common reason for refusal of blood products is religious belief. Jehovah’s Witness patients will refuse blood products due to their interpretation of the Bible.⁶⁵ All members will refuse red cells, plasma, and platelets, while decisions about “derived” blood products—products made by manipulation of the original donated units—are a matter of conscience. These include cryoprecipitate, intravenous gammaglobulin, and albumin.

In an elective situation, the first step is to discuss with the patient those products that are a matter of conscience and clearly document this. The patient’s blood count and iron stores should be assessed to identify any correctible causes of anemia or low iron stores before surgery. The use of erythropoietin to correct blood counts before surgery is controversial, as this may increase thrombosis risk and is contraindicated in patients with curable tumors.

For patients with acute blood loss, use of intravenous iron combined with high-dose erythropoietin is the most common approach to raise the blood count.⁶⁵ A recommended erythropoietin dose is 300 units/kg 3 times a week, dropping to 100 units/kg 3 times weekly until the goal hematocrit is reached. Another often overlooked step is to consolidate and minimize laboratory testing. The most important step is to be respectful of the patient and their beliefs. Many larger cities have liaisons that can help with interactions between Jehovah’s Witness patients and the health care system.

DESMOPRESSIN

Desmopressin (DDAVP) is a synthetic analog of antidiuretic hormone that raises the levels of both factor VIII and von Willebrand protein severalfold.⁶⁶ Desmopressin is effective in supporting hemostasis in patients with a wide variety of congenital and acquired bleeding disorders. However, desmopressin does not reduce blood loss before routine surgery in a healthy patient and should not be used for this purpose.⁶⁷

TRANEXAMIC ACID

Tranexamic acid is an antifibrinolytic agent that blocks the binding of plasmin to fibrin.⁶⁸ This agent was first shown to be useful in disorders that involve excessive fibrinolysis^69–73 or as adjunctive therapy for oral or dental procedures in patients with a bleeding diathesis. In patients with severe thrombocytopenia, the use of antifibrinolytic agents may reduce bleeding. Increasing data shows that tranexamic acid can prevent blood loss in a variety of surgeries including heart bypass, liver transplantation, and orthopedic surgery.⁷⁴ Patients across these settings have decreased blood loss and need for transfusion with no increased risk of thrombosis. The CRASH-2 study showed that the use of tranexamic acid significantly reduced mortality in trauma patients.⁷⁵ The WOMEN trial demonstrated that 1 g of tranexamic acid given to women with blood loss of more than 500 mL after vaginal delivery or 1000 mL after cesarean section has a risk reduction of death of 0.81 with no increased risk of thrombosis.⁷⁶ Given this abundant data, it is clear tranexamic acid needs to be part of any massive transfusion protocol.⁷⁷

RECOMBINANT FACTOR VIIa

Recombinant factor VIIa (rVIIa) was originally developed as a “bypass” agent to support hemostasis in hemophiliacs.⁷⁸ However, the use of rVIIa for a wide array of bleeding disorders, including patients with factor VII and XI deficiency and Glanzmann thrombasthenia, has been reported.⁷⁹ Increasingly, rVIIa is being used as a “universal hemostatic agent” for patients with uncontrolled bleeding from any mechanism.⁸⁰ Multiple case reports have described the use of rVIIa for bleeding in cardiac surgery patients, obstetrical bleeding, reversal of anticoagulation, and trauma.⁸¹ Unfortunately, little formal trial data exists to put these anecdotes into perspective, and formal review of clinical trial results has shown no benefit.^82,83 However, when used in older patients, especially those with vascular risk factors, the risk of arterial thrombosis appears to increase.⁸⁴ In the trials for intracranial hemorrhage, the thrombosis rate was 5% to 9%, and rates up to 10% for arterial events were seen in older patients in a review of all trials.^85–87 Given the lack of data but the evidence of risk, rVIIa use should be restricted to patients with documented bleeding disorders that have been shown to benefit by its use.

Reverse total shoulder arthroplasty (RTSA) is a common treatment for rotator cuff tear arthropathy. We performed a systematic review of all the RTSA literature to answer if we are treating the same patients with RTSA, across the world.

A systematic review was registered with PROSPERO, the international prospective register of systematic reviews, and performed with Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines using 3 publicly available free databases. Therapeutic clinical outcome investigations reporting RTSA outcomes with levels of evidence I to IV were eligible for inclusion. All study, subject, and surgical technique demographics were analyzed and compared between continents. Statistical comparisons were conducted using linear regression, analysis of variance (ANOVA), Fisher's exact test, and Pearson's chi-square test.

There were 103 studies included in the analysis (8973 patients; 62% female; mean age, 70.9 ± 6.7 years; mean length of follow-up, 34.3 ± 19.3 months) that had a low Modified Coleman Methodology Score (MCMS) (mean, 36.9 ± 8.7: poor). Most patients (60.8%) underwent RTSA for a diagnosis of rotator cuff arthropathy, whereas 1% underwent RTSA for fracture; indications varied by continent. There were no consistent reports of preopeartive or postoperative scores from studies in any region. Studies from North America reported significantly higher postoperative external rotation (34.1° ± 13.3° vs 19.3° ± 8.9°) (P < .001) and a greater change in flexion (69.0° ± 24.5° vs 56.3° ± 11.3°) (P = .004) compared with studies from Europe. North America had the greatest total number of publications followed by Europe. The total yearly number of publications increased each year (P < .001), whereas the MCMS decreased each year (P = .037).

The quantity, but not the quality of RTSA studies is increasing. Indications for RTSA varied by continent, although most patients underwent RTSA for rotator cuff arthropathy. The majority of patients undergoing RTSA are female over the age of 60 years for a diagnosis of rotator cuff arthropathy with pseudoparalysis.

Continue to: Reverse total shoulder arthroplasty...

Reverse total shoulder arthroplasty (RTSA) is a common procedure with indications including rotator cuff tear arthropathy, proximal humerus fractures, and others.^1,2 Studies have shown excellent, reliable, short- and mid-term outcomes in patients treated with RTSA for various indications.^3-5 Al-Hadithy and colleagues⁶ reviewed 41 patients who underwent RTSA for pseudoparalysis secondary to rotator cuff tear arthropathy and, at a mean follow-up of 5 years, found significant improvements in range of motion (ROM) as well as age-adjusted Constant and Oxford Outcome scores. Similarly, Ross and colleagues⁷ evaluated outcomes of RTSA in 28 patients in whom RTSA was performed for 3- or 4-part proximal humerus fractures, and found both good clinical and radiographic outcomes with no revision surgeries at a mean follow-up of 54.9 months. RTSA is performed across the world, with specific implant designs, specifically humeral head inclination, but is more common in some areas when compared with others.^3,8,9

The number of RTSAs performed has steadily increased over the past 20 years, with recent estimates of approximately 20,000 RTSAs performed in the United States in 2011.^10,11 However, there is little information about the similarities and differences between those patients undergoing RTSA in various parts of the world regarding surgical indications, patient demographics, and outcomes. The purpose of this study is to perform a systematic review and meta-analysis of the RTSA body of literature to both identify and compare characteristics of studies published (level of evidence, whether a conflict of interest existed), patients analyzed (age, gender), and surgical indications performed across both continents and countries. Essentially, the study aims to answer the question, "Across the world, are we treating the same patients?" The authors hypothesized that there would be no significant differences in RTSA publications, subjects, and indications based on both the continent and country of publication.

METHODS

A systematic review was conducted according to PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) guidelines using a PRISMA checklist.¹² A systematic review registration was performed using PROSPERO, the international prospective register of systematic reviews (registration number CRD42014010578).¹³Two reviewers independently conducted the search on March 25, 2014, using the following databases: Medline, Cochrane Central Register of Controlled Trials, SportDiscus, and CINAHL. The electronic search citation algorithm utilized was: (((((reverse[Title/Abstract]) AND shoulder[Title/Abstract]) AND arthroplasty[Title/Abstract]) NOT arthroscopic[Title/Abstract]) NOT cadaver[Title/Abstract]) NOT biomechanical[Title/Abstract]. English language Level I to IV evidence (2011 update by the Oxford Centre for Evidence-Based Medicine¹⁴) clinical studies were eligible. Medical conference abstracts were ineligible for inclusion. All references within included studies were cross-referenced for inclusion if missed by the initial search with any additionally located studies screened for inclusion. Duplicate subject publications within separate unique studies were not reported twice, but rather the study with longer duration follow-up or, if follow-up was equal, the study with the greater number of patients was included. Level V evidence reviews, letters to the editor, basic science, biomechanical and cadaver studies, total shoulder arthroplasty (TSA) papers, arthroscopic shoulder surgery papers, imaging, surgical techniques, and classification studies were excluded.

A total of 255 studies were identified, and, after implementation of the exclusion criteria, 103 studies were included in the final analysis (Figure 1). Subjects of interest in this systematic review underwent RTSA for one of many indications including rotator cuff tear arthropathy, osteoarthritis, rheumatoid arthritis, posttraumatic arthritis, instability, revision from a previous RTSA for instability, infection, acute proximal humerus fracture, revision from a prior proximal humerus fracture, revision from a prior hemiarthroplasty, revision from a prior TSA, osteonecrosis, pseudoparalysis, tumor, and a locked shoulder dislocation. There was no minimum follow-up or rehabilitation requirement. Study and subject demographic parameters analyzed included year of publication, years of subject enrollment, presence of study financial conflict of interest, number of subjects and shoulders, gender, age, body mass index, diagnoses treated, and surgical positioning. Clinical outcome scores sought were the DASH (Disability of the Arm, Shoulder, and Hand), SPADI (Shoulder Pain And Disability Index), Absolute Constant, ASES (American Shoulder and Elbow Score), KSS (Korean Shoulder Score), SST-12 (Simple Shoulder Test), SF-12 (12-item Short Form), SF-36 (36-item Short Form), SSV (Subjective Shoulder Value), EQ-5D (EuroQol-5 Dimension), SANE (Single Assessment Numeric Evaluation), Rowe Score for Instability, Oxford Instability Score, UCLA (University of California, Los Angeles) activity score, Penn Shoulder Score, and VAS (visual analog scale). In addition, ROM (forward elevation, abduction, external rotation, internal rotation) was analyzed. Radiographs and magnetic resonance imaging data were extracted when available. The methodological quality of the study was evaluated using the MCMS (Modified Coleman Methodology Score).¹⁵

STATISTICAL ANALYSIS

First, the number of publications per year, level of evidence, and Modified Coleman Methodology Score were tested for association with the calendar year using linear regression. Second, demographic data were tested for association with the continent using Pearson’s chi-square test or ANOVA. Third, indications were tested for association with the continent using Fisher’s exact test. Finally, clinical outcome scores and ROM were tested for association with the continent using ANOVA. Statistical significance was extracted from studies when available. Statistical significance was defined as P < .05.

Continue to: RESULTS...

 

 

RESULTS

There were 103 studies included in the analysis (Figure 1). A total of 8973 patients were included, 62% of whom were female with a mean age of 70.9 ± 6.7 years (Table 1). The average follow-up was 34.3 ± 19.3 months. North America had the overall greatest total number of publications on RTSA, followed by Europe (Figure 2). The total yearly number of publications increased by a mean of 1.95 publications each year (P < .001). There was no association between the mean level of evidence with the year of publication (P = .296) (Figure 3). Overall, the rating of studies was poor for the MCMS (mean 36.9 ± 8.7). The MCMS decreased each year by a mean of 0.76 points (P = .037) (Figure 4).

Table 1. Demographic Data by Continent

	North America	Europe	Asia	Australia	Total	P-value
Number of studies	52	43	4	4	103	-
Number of subjects	6158	2609	51	155	8973	-
Level of evidence						0.693
II	5 (10%)	3 (7%)	0 (0%)	0 (0%)	8 (8%)
III	10 (19%)	4 (9%)	0 (0%)	1 (25%)	15 (15%)
IV	37 (71%)	36 (84%)	4 (100%)	3 (75%)	80 (78%)
Mean MCMS	34.6 ± 8.4	40.2 ± 8.0	32.5 12.4	34.5 ± 6.6	36.9 ± 8.7	0.010
Institutional collaboration						1.000
Multi-center	7 (14%)	6 (14%)	0 (0%)	0 (0%)	13 (13%)
Single-center	45 (86%)	37 (86%)	4 (100%)	4 (100%)	90 (87%)
Financial conflict of interest						0.005
Present	28 (54%)	15 (35%)	0 (0%)	0 (0%)	43 (42%)
Not present	19 (37%)	16 (37%)	4 (100%)	4 (100%)	43 (42%)
Not reported	5 (10%)	12 (28%)	0 (0%)	0 (0%)	17 (17%)
Sex						N/A
Male	2157 (38%)	1026 (39%)	13 (25%)	61 (39%)	3257 (38%)
Female	3520 (62%)	1622 (61%)	38 (75%)	94 (61%)	5274 (62%)
Mean age (years)	71.3 ± 5.6	70.1 ± 7.9	68.1 ± 5.3	76.9 ± 3.0	70.9 ± 6.7	0.191
Minimum age (mean across studies)	56.9 ± 12.8	52.8 ± 15.7	62.8 ± 6.2	68.0 ± 12.1	55.6 ± 14.3	0.160
Maximum age (mean across studies)	82.1 ± 8.6	83.0 ± 5.5	73.0 ± 9.4	85.0 ± 7.9	82.2 ± 7.6	0.079
Mean length of follow-up (months)	26.5 ± 13.7	43.1 ± 21.7	29.4 ± 7.9	34.2 ± 16.6	34.3 ± 19.3	<0.001
Prosthesis type						N/A
Cemented	988 (89%)	969 (72%)	0 (0%)	8 (16%)	1965 (78%)
Press fit	120 (11%)	379 (28%)	0 (0%)	41 (84%)	540 (22%)

Abbreviations: MCMS, Modified Coleman Methodology Score; N/A, not available.

 

In studies that reported press-fit vs cemented prostheses, the highest percentage of press-fit prostheses compared with cemented prostheses was seen in Australia (84% press-fit), whereas the highest percentage of cemented prostheses was seen in North America (89% cemented). A higher percentage of studies from North America had a financial conflict of interest (COI) than did those from other countries (54% had a COI).

Continue to: Rotator cuff tear arthropathy...

 

 

Rotator cuff tear arthropathy was the most common indication for RTSA overall in 5459 patients, followed by pseudoparalysis in 1352 patients (Tables 2 and 3). While studies in North America reported rotator cuff tear arthropathy as the indication for RTSA in 4418 (75.8%) patients, and pseudoparalysis as the next most common indication in 535 (9.2%) patients, studies from Europe reported rotator cuff tear arthropathy as the indication in 895 (33.5%) patients, and pseudoparalysis as the indication in 795 (29.7%) patients. Studies from Asia also had a relatively even split between rotator cuff tear arthropathy and pseudoparalysis (45.3% vs 37.8%), whereas those from Australia were mostly rotator cuff tear arthropathy (77.7%).

Table 2. Number (Percent) of Studies With Each Indication by Continent

	North America	Europe	Asia	Australia	Total	P-value
Rotator cuff arthropathy	29 (56%)	19 (44%)	3 (75%)	3 (75%)	54 (52%)	0.390
Osteoarthritis	4 (8%)	10 (23%)	1 (25%)	1 (25%)	16 (16%)	0.072
Rheumatoid arthritis	9 (17%)	10 (23%)	0 (0%)	2 (50%)	21 (20%)	0.278
Post-traumatic arthritis	3 (6%)	5 (12%)	0 (0%)	1 (25%)	9 (9%)	0.358
Instability	6 (12%)	3 (7%)	0 (0%)	1 (25%)	10 (10%)	0.450
Revision of previous RTSA for instability	5 (10%)	1 (2%)	0 (0%)	1 (25%)	7 (7%)	0.192
Infection	4 (8%)	1 (2%)	1 (25%)	0 (0%)	6 (6%)	0.207
Unclassified acute proximal humerus fracture	9 (17%)	5 (12%)	1 (25%)	1 (25%)	16  (16%)	0.443
Acute 2-part proximal humerus fracture	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	N/A
Acute 3-part proximal humerus fracture	2 (4%)	0 (0%)	0 (0%)	0 (0%)	2 (2%)	0.574
Acute 4-part proximal humerus fracture	5 (10%)	0 (0%)	0 (0%)	0 (0%)	5 (5%)	0.183
Acute 3- or 4-part proximal humerus fracture	6 (12%)	2 (5%)	0 (0%)	0 (0%)	8 (8%)	0.635
Revised from previous nonop proximal humerus fracture	7 (13%)	3 (7%)	0 (0%)	0 (0%)	10 (10%)	0.787
Revised from ORIF	1 (2%)	1 (2%)	0 (0%)	0 (0%)	2 (2%)	1.000
Revised from CRPP	0 (0%)	1 (2%)	0 (0%)	0 (0%)	1 (1%)	0.495
Revised from hemi	8 (15%)	4 (9%)	0 (0%)	1 (25%)	13 (13%)	0.528
Revised from TSA	15 (29%)	11 (26%)	0 (0%)	2 (50%)	28 (27%)	0.492
Osteonecrosis	4 (8%)	2 (5%)	1 (25%)	0 (0%)	7 (7%)	0.401
Pseudoparalysis irreparable tear without arthritis	20 (38%)	18 (42%)	2 (50%)	1 (25%)	41 (40%)	0.919
Bone tumors	0 (0%)	4 (9.3%)	0 (0%)	0 (0%)	4 (4%)	0.120
Locked shoulder dislocation	0 (0%)	0 (0%)	1 (25%)	0 (0%)	1 (1%)	0.078

Abbreviations: CRPP, closed reduction and percutaneous pinning; ORIF, open reduction internal fixation; RTSA, reverse total shoulder arthroplasty; TSA, total shoulder arthroplasty.

 

Table 3. Number of Patients With Each Indication as Reported by Individual Studies by Continent

	North America	Europe	Asia	Australia	Total
Rotator cuff arthropathy	4418	895	24	122	5459
Osteoarthritis	90	251	1	14	356
Rheumatoid arthritis	59	87	0	2	148
Post-traumatic arthritis	62	136	0	1	199
Instability	23	15	0	1	39
Revision of previous RTSA for instability	29	2	0	1	32
Infection	28	11	2	0	41
Unclassified acute proximal humerus fracture	42	30	4	8	84
Acute 3-part proximal humerus fracture	60	0	0	0	6
Acute 4-part proximal humerus fracture	42	0	0	0	42
Acute 3- or 4-part proximal humerus fracture	92	46	0	0	138
Revised from previous nonop proximal humerus fracture	43	53	0	0	96
Revised from ORIF	3	9	0	0	12
Revised from CRPP	0	3	0	0	3
Revised from hemi	105	51	0	1	157
Revised from TSA	192	246	0	5	443
Osteonecrosis	9	6	1	0	16
Pseudoparalysis irreparable tear without arthritis	535	795	20	2	1352
Bone tumors	0	38	0	0	38
Locked shoulder dislocation	0	0	1	0	1

Abbreviations: CRPP, closed reduction and percutaneous pinning; ORIF, open reduction internal fixation; RTSA, reverse total shoulder arthroplasty; TSA, total shoulder arthroplasty.

 

The ASES, SST-12, and VAS scores were the most frequently reported outcome scores in studies from North America, whereas the Absolute Constant score was the most common score reported in studies from Europe (Table 4). Studies from North America reported significantly higher postoperative external rotation (34.1° ± 13.3° vs 19.3° ± 8.9°) (P < .001) and a greater change in flexion (69.0° ± 24.5° vs 56.3° +/- 11.3°) (P = .004) compared with studies from Europe (Table 5).

Table 4. Outcomes by Continent

Metric (number of studies)	North America	Europe	Asia	Australia	P-value
DASH	1	2	0	0
Preoperative	54.0	62.0 ± 8.5	-	-	0.582
Postoperative	24.0	32.0 ± 2.8	-	-	0.260
Change	-30.0	-30.0 ± 11.3	-	-	1.000
SPADI	2	0	0	0
Preoperative	80.0 ± 4.2	-	-	-	N/A
Postoperative	34.8 ± 1.1	-	-	-	N/A
Change	-45.3 ± 3.2	-	-	-	N/A
Absolute constant	2	27	0	1
Preopeartive	33.0 ± 0.0	28.2 ± 7.1	-	20.0	0.329
Postoperative	54.5 ± 7.8	62.9 ± 9.0	-	65.0	0.432
Change	+21.5 ± 7.8	+34.7 ± 8.0	-	+45.0	0.044
ASES	13	0	2	0
Preoperative	33.2 ± 5.4	-	32.5 ± 3.5	-	0.867
Postoperative	73.9 ± 6.8	-	75.7 ± 10.8	-	0.752
Change	+40.7 ± 6.5	-	+43.2 ± 14.4	-	0.670
UCLA	3	2	1	0
Preoperative	10.1 ± 3.4	11.2 ± 5.7	12.0	-	0.925
Postoperative	24.5 ± 3.1	24.3 ± 3.7	24.0	-	0.991
Change	+14.4 ± 1.6	+13.1 ± 2.0	+12.0	-	0.524
KSS	0	0	2	0
Preopeartive	-	-	38.2 ± 1.1	-	N/A
Postoperative	-	-	72.3 ± 6.0	-	N/A
Change	-	-	+34.1 ± 7.1	-	N/A
SST-12	12	1	0	0
Preoperative	1.9 ± 0.8	1.2	-	-	N/A
Postoperative	7.1 ± 1.5	5.6	-	-	N/A
Change	+5.3 ± 1.2	+4.4	-	-	N/A
SF-12	1	0	0	0
Preoperative	34.5	-	-	-	N/A
Postoperative	38.5	-	-	-	N/A
Change	+4.0	-	-	-	N/A
SSV	0	5	0	0
Preopeartive	-	22.0 ± 7.4	-	-	N/A
Postoperative	-	63.4 ± 7.9	-	-	N/A
Change	-	+41.4 ± 2.1	-	-	N/A
EQ-5D	0	2	0	0
Preoperative	-	0.5 ± 0.2	-	-	N/A
Postoperative	-	0.8 ± 0.1	-	-	N/A
Change	-	+0.3 ± 0.1	-	-	N/A
OOS	1	0	0	0
Preoperative	24.7	-	-	-	N/A
Postoperative	14.9	-	-	-	N/A
Change	-9.9	-	-	-	N/A
Rowe	0	1	0	0
Preoperative	-	50.2	-	-	N/A
Postoperative	-	82.1	-	-	N/A
Change	-	31.9	-	-	N/A
Oxford	0	2	0	0
Preoperative	-	119.9 ± 138.8	-	-	N/A
Postoperative	-	39.9 ± 3.3	-	-	N/A
Change	-	-80.6 ± 142.2	-	-	N/A
Penn	1	0	0	0
Preoperative	24.9	-	-	-	N/A
Postoperative	66.4	-	-	-	N/A
Change	+41.5	-	-	-	N/A
VAS	10	1	1	1
Preoperative	6.6 ± 0.8	7.0	8.4	7.0	N/A
Postoperative	2.0 ± 0.7	1.0	0.8	0.8	N/A
Change	-4.6 ± 0.8	-6.0	-7.6	-6.2	N/A
SF-36 physical	2	0	0	0
Preoperative	32.7 ± 1.2	-	-	-	N/A
Postoperative	39.6 ± 4.0	-	-	-	N/A
Change	+7.0 ± 2.8	-	-	-	N/A
SF-36 mental	2	0	0	0
Preoperative	43.6 ± 2.8	-	-	-	N/A
Postoperative	48.1 ± 1.0	-	-	-	N/A
Change	+4.5 ± 1.8	-	-	-	N/A

Abbreviations: ASES, American Shoulder and Elbow Surgeon score; DASH, Disability of the Arm, Shoulder, and Hand; EQ-5D, EuroQol-5 Dimension; KSS, Korean Shoulder Scoring system; N/A, not available; OOS, Orthopaedic Outcome Score; SF, short form; SPADI, Shoulder Pain and Disability Index; SST, Simple Shoulder Test; SSV, Subjective Shoulder Value; UCLA, University of California, Los Angeles; VAS, visual analog scale.

 

Table 5. Shoulder Range of Motion, by Continent

Metric (number of studies)	North America	Europe	Asia	Australia	P-value
Flexion	18	22	1	1
Preoperative	57.6 ± 17.9	65.5 ± 17.2	91.0	30.0	0.060
Postoperative	126.6 ± 14.4	121.8 ± 19.0	133.0	150.0	0.360
Change	+69.0 ± 24.5	+56.3 ± 11.3	+42.0	120.0	0.004
Abduction	11	12	1	0
Preoperative	53.7 ± 25.0	52.0 ± 19.0	88.0	-	0.311
Postoperative	109.3 ± 15.1	105.4 ± 19.8	131.0	-	0.386
Change	55.5 ± 25.5	53.3 ± 8.3	43.0	-	0.804
External rotation	17	19	0	0
Preoperative	19.4 ± 9.9	11.2 ± 6.1	-	-	0.005
Postoperative	34.1 ± 13.3	19.3 ± 8.9	-	-	<0.001
Change	+14.7 ± 13.2	+8.1 ± 8.5	-	-	0.079

Continue to: DISCUSSION...

 

 

DISCUSSION

RTSA is a common procedure performed in many different areas of the world for a variety of indications. The study hypotheses were partially confirmed, as there were no significant differences seen in the characteristics of the studies published and patients analyzed; although, the majority of studies from North America reported rotator cuff tear arthropathy as the primary indication for RTSA, whereas studies from Europe were split between rotator cuff tear arthropathy and pseudoparalysis as the primary indication. Hence, based on the current literature the study proved that we are treating the same patients. Despite this finding, we may be treating them for different reasons with an RTSA.

RTSA has become a standard procedure in the United States, with >20,000 RTSAs performed in 2011.¹⁰ This number will continue to increase as it has over the past 20 years given the aging population in the United States, as well as the expanding indications for RTSA.¹¹ Indications of RTSA have become broad, although the main indication remains as rotator cuff tear arthropathy (>60% of all patients included in this study), and pseudoparalysis (>15% of all patients included in this study). Results for RTSA for rotator cuff tear arthropathy and pseudoparalysis have been encouraging.^16,17 Frankle and colleagues¹⁶ evaluated 60 patients who underwent RTSA for rotator cuff tear arthropathy at a minimum of 2 years follow-up (average, 33 months). The authors found significant improvements in all measured clinical outcome variables (P < .0001) (ASES, mean function score, mean pain score, and VAS) as well as ROM, specifically forward flexion increased from 55° to 105.1°, and abduction increased from 41.4° to 101.8°. Similarly, Werner and colleagues¹⁷ evaluated 58 consecutive patients who underwent RTSA for pseudoparalysis secondary to irreparable rotator cuff dysfunction at a mean follow-up of 38 months. Overall, significant improvements (P < .0001) were seen in the SSV score, relative Constant score, and Constant score for pain, active anterior elevation (42° to 100° following RTSA), and active abduction (43° to 90° following RTSA).

It is essential to understand the similarities and differences between patients undergoing RTSA in different parts of the world so the literature from various countries can be compared between regions, and conclusions extrapolated to the correct patients. For example, an interesting finding in this study is that the majority of patients in North America have their prosthesis cemented whereas the majority of patients in Australia have their prosthesis press-fit. While the patients each continent is treating are not significantly different (mostly older women), the difference in surgical technique could have implications in long- or short-term functional outcomes. Prior studies have shown no difference in axial micromotion between cemented and press-fit humeral components, but the clinical implications surrounding this are not well defined.¹⁸ Small series comparing cementless to cemented humeral prosthesis in RTSA have found no significant differences in clinical outcomes or postoperative ROM, but larger series are necessary to validate these outcomes.¹⁹ However, studies have shown lower rates of postoperative infections in patients who receive antibiotic-loaded cement compared with those who receive plain bone cement following RTSA.²⁰

Similarly, as the vast majority of patients in North America had an RTSA for rotator cuff arthropathy (75.8%) whereas those from Europe had RTSA almost equally for rotator cuff arthropathy (33.5%) and pseudoparalysis (29.7%), one must ensure similar patient populations before attempting to extrapolate results of a study from a different country to patients in other areas. Fortunately, the clinical results following RTSA for either indication have been good.^6,21,22

One final point to consider is the cost effectiveness of the implant. Recent evidence has shown that RTSA is associated with a higher risk for in-hospital death, multiple perioperative complications, prolonged hospital stay, and increased hospital cost when compared with TSA.²³ This data may be biased as the patient selection for RTSA varies from that of TSA, but it is a point that must be considered. Other studies have shown that an RTSA is a cost-effective treatment option for treating patients with rotator cuff tear arthropathy, and is a more cost-effective option in treating rotator cuff tear arthropathy than hemiarthroplasty.^24,25 Similarly, RTSA offers a more cost-effective treatment option with better outcomes for patients with acute proximal humerus fractures when compared with open reduction internal fixation and hemiarthroplasty.²⁶ However, TSA is a more cost-effective treatment option than RTSA for patients with glenohumeral osteoarthritis.²⁷ With changing reimbursement in healthcare, surgeons must scrutinize not only anticipated outcomes with specific implants but the cost effectiveness of these implants as well. Further cost analysis studies are necessary to determine the ideal candidate for an RTSA.

LIMITATIONS

Despite its extensive review of the literature, this study had several limitations. While 2 independent authors searched for studies, it is possible that some studies were missed during the search process, introducing possible selection bias. No abstracts or unpublished works were included which could have introduced publication bias. Several studies did not report all variables the authors examined, and this could have skewed some of the results since the reporting of additional variables could have altered the data to show significant differences in some measured variables. As outcome measures for various pathologies were not compared, conclusions cannot be drawn on the best treatment option for various indications. As case reports were included, this could have lowered both the MCMS as well as the average in studies reporting outcomes. Furthermore, given the overall poor quality of the underlying data available for this study, the validity/generalizability of the results could be limited as the level of evidence of this systematic review is only as high as the studies it includes. There are subtle differences between rotator cuff arthropathy and pseudoparalysis, and some studies may have classified patients differently than others, causing differences in indications. Finally, as the primary goal of this study was to report on demographics, no evaluation of concomitant pathology at the time of surgery or rehabilitation protocols was performed.

CONCLUSION

The quantity, but not the quality of RTSA studies is increasing. Indications for RTSA varied by continent although most patients underwent RTSA for rotator cuff arthropathy. The majority of patients undergoing RTSA are female over the age of 60 years for a diagnosis of rotator cuff arthropathy with pseudoparalysis.

This paper will be judged for the Resident Writer’s Award.

Reverse total shoulder arthroplasty (RTSA) is a common treatment for rotator cuff tear arthropathy. We performed a systematic review of all the RTSA literature to answer if we are treating the same patients with RTSA, across the world.

A systematic review was registered with PROSPERO, the international prospective register of systematic reviews, and performed with Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines using 3 publicly available free databases. Therapeutic clinical outcome investigations reporting RTSA outcomes with levels of evidence I to IV were eligible for inclusion. All study, subject, and surgical technique demographics were analyzed and compared between continents. Statistical comparisons were conducted using linear regression, analysis of variance (ANOVA), Fisher's exact test, and Pearson's chi-square test.

There were 103 studies included in the analysis (8973 patients; 62% female; mean age, 70.9 ± 6.7 years; mean length of follow-up, 34.3 ± 19.3 months) that had a low Modified Coleman Methodology Score (MCMS) (mean, 36.9 ± 8.7: poor). Most patients (60.8%) underwent RTSA for a diagnosis of rotator cuff arthropathy, whereas 1% underwent RTSA for fracture; indications varied by continent. There were no consistent reports of preopeartive or postoperative scores from studies in any region. Studies from North America reported significantly higher postoperative external rotation (34.1° ± 13.3° vs 19.3° ± 8.9°) (P < .001) and a greater change in flexion (69.0° ± 24.5° vs 56.3° ± 11.3°) (P = .004) compared with studies from Europe. North America had the greatest total number of publications followed by Europe. The total yearly number of publications increased each year (P < .001), whereas the MCMS decreased each year (P = .037).

The quantity, but not the quality of RTSA studies is increasing. Indications for RTSA varied by continent, although most patients underwent RTSA for rotator cuff arthropathy. The majority of patients undergoing RTSA are female over the age of 60 years for a diagnosis of rotator cuff arthropathy with pseudoparalysis.

Continue to: Reverse total shoulder arthroplasty...

Reverse total shoulder arthroplasty (RTSA) is a common procedure with indications including rotator cuff tear arthropathy, proximal humerus fractures, and others.^1,2 Studies have shown excellent, reliable, short- and mid-term outcomes in patients treated with RTSA for various indications.^3-5 Al-Hadithy and colleagues⁶ reviewed 41 patients who underwent RTSA for pseudoparalysis secondary to rotator cuff tear arthropathy and, at a mean follow-up of 5 years, found significant improvements in range of motion (ROM) as well as age-adjusted Constant and Oxford Outcome scores. Similarly, Ross and colleagues⁷ evaluated outcomes of RTSA in 28 patients in whom RTSA was performed for 3- or 4-part proximal humerus fractures, and found both good clinical and radiographic outcomes with no revision surgeries at a mean follow-up of 54.9 months. RTSA is performed across the world, with specific implant designs, specifically humeral head inclination, but is more common in some areas when compared with others.^3,8,9

The number of RTSAs performed has steadily increased over the past 20 years, with recent estimates of approximately 20,000 RTSAs performed in the United States in 2011.^10,11 However, there is little information about the similarities and differences between those patients undergoing RTSA in various parts of the world regarding surgical indications, patient demographics, and outcomes. The purpose of this study is to perform a systematic review and meta-analysis of the RTSA body of literature to both identify and compare characteristics of studies published (level of evidence, whether a conflict of interest existed), patients analyzed (age, gender), and surgical indications performed across both continents and countries. Essentially, the study aims to answer the question, "Across the world, are we treating the same patients?" The authors hypothesized that there would be no significant differences in RTSA publications, subjects, and indications based on both the continent and country of publication.

METHODS

A systematic review was conducted according to PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) guidelines using a PRISMA checklist.¹² A systematic review registration was performed using PROSPERO, the international prospective register of systematic reviews (registration number CRD42014010578).¹³Two reviewers independently conducted the search on March 25, 2014, using the following databases: Medline, Cochrane Central Register of Controlled Trials, SportDiscus, and CINAHL. The electronic search citation algorithm utilized was: (((((reverse[Title/Abstract]) AND shoulder[Title/Abstract]) AND arthroplasty[Title/Abstract]) NOT arthroscopic[Title/Abstract]) NOT cadaver[Title/Abstract]) NOT biomechanical[Title/Abstract]. English language Level I to IV evidence (2011 update by the Oxford Centre for Evidence-Based Medicine¹⁴) clinical studies were eligible. Medical conference abstracts were ineligible for inclusion. All references within included studies were cross-referenced for inclusion if missed by the initial search with any additionally located studies screened for inclusion. Duplicate subject publications within separate unique studies were not reported twice, but rather the study with longer duration follow-up or, if follow-up was equal, the study with the greater number of patients was included. Level V evidence reviews, letters to the editor, basic science, biomechanical and cadaver studies, total shoulder arthroplasty (TSA) papers, arthroscopic shoulder surgery papers, imaging, surgical techniques, and classification studies were excluded.

A total of 255 studies were identified, and, after implementation of the exclusion criteria, 103 studies were included in the final analysis (Figure 1). Subjects of interest in this systematic review underwent RTSA for one of many indications including rotator cuff tear arthropathy, osteoarthritis, rheumatoid arthritis, posttraumatic arthritis, instability, revision from a previous RTSA for instability, infection, acute proximal humerus fracture, revision from a prior proximal humerus fracture, revision from a prior hemiarthroplasty, revision from a prior TSA, osteonecrosis, pseudoparalysis, tumor, and a locked shoulder dislocation. There was no minimum follow-up or rehabilitation requirement. Study and subject demographic parameters analyzed included year of publication, years of subject enrollment, presence of study financial conflict of interest, number of subjects and shoulders, gender, age, body mass index, diagnoses treated, and surgical positioning. Clinical outcome scores sought were the DASH (Disability of the Arm, Shoulder, and Hand), SPADI (Shoulder Pain And Disability Index), Absolute Constant, ASES (American Shoulder and Elbow Score), KSS (Korean Shoulder Score), SST-12 (Simple Shoulder Test), SF-12 (12-item Short Form), SF-36 (36-item Short Form), SSV (Subjective Shoulder Value), EQ-5D (EuroQol-5 Dimension), SANE (Single Assessment Numeric Evaluation), Rowe Score for Instability, Oxford Instability Score, UCLA (University of California, Los Angeles) activity score, Penn Shoulder Score, and VAS (visual analog scale). In addition, ROM (forward elevation, abduction, external rotation, internal rotation) was analyzed. Radiographs and magnetic resonance imaging data were extracted when available. The methodological quality of the study was evaluated using the MCMS (Modified Coleman Methodology Score).¹⁵

STATISTICAL ANALYSIS

First, the number of publications per year, level of evidence, and Modified Coleman Methodology Score were tested for association with the calendar year using linear regression. Second, demographic data were tested for association with the continent using Pearson’s chi-square test or ANOVA. Third, indications were tested for association with the continent using Fisher’s exact test. Finally, clinical outcome scores and ROM were tested for association with the continent using ANOVA. Statistical significance was extracted from studies when available. Statistical significance was defined as P < .05.

Continue to: RESULTS...

 

 

RESULTS

There were 103 studies included in the analysis (Figure 1). A total of 8973 patients were included, 62% of whom were female with a mean age of 70.9 ± 6.7 years (Table 1). The average follow-up was 34.3 ± 19.3 months. North America had the overall greatest total number of publications on RTSA, followed by Europe (Figure 2). The total yearly number of publications increased by a mean of 1.95 publications each year (P < .001). There was no association between the mean level of evidence with the year of publication (P = .296) (Figure 3). Overall, the rating of studies was poor for the MCMS (mean 36.9 ± 8.7). The MCMS decreased each year by a mean of 0.76 points (P = .037) (Figure 4).

Table 1. Demographic Data by Continent

	North America	Europe	Asia	Australia	Total	P-value
Number of studies	52	43	4	4	103	-
Number of subjects	6158	2609	51	155	8973	-
Level of evidence						0.693
II	5 (10%)	3 (7%)	0 (0%)	0 (0%)	8 (8%)
III	10 (19%)	4 (9%)	0 (0%)	1 (25%)	15 (15%)
IV	37 (71%)	36 (84%)	4 (100%)	3 (75%)	80 (78%)
Mean MCMS	34.6 ± 8.4	40.2 ± 8.0	32.5 12.4	34.5 ± 6.6	36.9 ± 8.7	0.010
Institutional collaboration						1.000
Multi-center	7 (14%)	6 (14%)	0 (0%)	0 (0%)	13 (13%)
Single-center	45 (86%)	37 (86%)	4 (100%)	4 (100%)	90 (87%)
Financial conflict of interest						0.005
Present	28 (54%)	15 (35%)	0 (0%)	0 (0%)	43 (42%)
Not present	19 (37%)	16 (37%)	4 (100%)	4 (100%)	43 (42%)
Not reported	5 (10%)	12 (28%)	0 (0%)	0 (0%)	17 (17%)
Sex						N/A
Male	2157 (38%)	1026 (39%)	13 (25%)	61 (39%)	3257 (38%)
Female	3520 (62%)	1622 (61%)	38 (75%)	94 (61%)	5274 (62%)
Mean age (years)	71.3 ± 5.6	70.1 ± 7.9	68.1 ± 5.3	76.9 ± 3.0	70.9 ± 6.7	0.191
Minimum age (mean across studies)	56.9 ± 12.8	52.8 ± 15.7	62.8 ± 6.2	68.0 ± 12.1	55.6 ± 14.3	0.160
Maximum age (mean across studies)	82.1 ± 8.6	83.0 ± 5.5	73.0 ± 9.4	85.0 ± 7.9	82.2 ± 7.6	0.079
Mean length of follow-up (months)	26.5 ± 13.7	43.1 ± 21.7	29.4 ± 7.9	34.2 ± 16.6	34.3 ± 19.3	<0.001
Prosthesis type						N/A
Cemented	988 (89%)	969 (72%)	0 (0%)	8 (16%)	1965 (78%)
Press fit	120 (11%)	379 (28%)	0 (0%)	41 (84%)	540 (22%)

Abbreviations: MCMS, Modified Coleman Methodology Score; N/A, not available.

 

In studies that reported press-fit vs cemented prostheses, the highest percentage of press-fit prostheses compared with cemented prostheses was seen in Australia (84% press-fit), whereas the highest percentage of cemented prostheses was seen in North America (89% cemented). A higher percentage of studies from North America had a financial conflict of interest (COI) than did those from other countries (54% had a COI).

Continue to: Rotator cuff tear arthropathy...

 

 

Rotator cuff tear arthropathy was the most common indication for RTSA overall in 5459 patients, followed by pseudoparalysis in 1352 patients (Tables 2 and 3). While studies in North America reported rotator cuff tear arthropathy as the indication for RTSA in 4418 (75.8%) patients, and pseudoparalysis as the next most common indication in 535 (9.2%) patients, studies from Europe reported rotator cuff tear arthropathy as the indication in 895 (33.5%) patients, and pseudoparalysis as the indication in 795 (29.7%) patients. Studies from Asia also had a relatively even split between rotator cuff tear arthropathy and pseudoparalysis (45.3% vs 37.8%), whereas those from Australia were mostly rotator cuff tear arthropathy (77.7%).

Table 2. Number (Percent) of Studies With Each Indication by Continent

	North America	Europe	Asia	Australia	Total	P-value
Rotator cuff arthropathy	29 (56%)	19 (44%)	3 (75%)	3 (75%)	54 (52%)	0.390
Osteoarthritis	4 (8%)	10 (23%)	1 (25%)	1 (25%)	16 (16%)	0.072
Rheumatoid arthritis	9 (17%)	10 (23%)	0 (0%)	2 (50%)	21 (20%)	0.278
Post-traumatic arthritis	3 (6%)	5 (12%)	0 (0%)	1 (25%)	9 (9%)	0.358
Instability	6 (12%)	3 (7%)	0 (0%)	1 (25%)	10 (10%)	0.450
Revision of previous RTSA for instability	5 (10%)	1 (2%)	0 (0%)	1 (25%)	7 (7%)	0.192
Infection	4 (8%)	1 (2%)	1 (25%)	0 (0%)	6 (6%)	0.207
Unclassified acute proximal humerus fracture	9 (17%)	5 (12%)	1 (25%)	1 (25%)	16  (16%)	0.443
Acute 2-part proximal humerus fracture	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	N/A
Acute 3-part proximal humerus fracture	2 (4%)	0 (0%)	0 (0%)	0 (0%)	2 (2%)	0.574
Acute 4-part proximal humerus fracture	5 (10%)	0 (0%)	0 (0%)	0 (0%)	5 (5%)	0.183
Acute 3- or 4-part proximal humerus fracture	6 (12%)	2 (5%)	0 (0%)	0 (0%)	8 (8%)	0.635
Revised from previous nonop proximal humerus fracture	7 (13%)	3 (7%)	0 (0%)	0 (0%)	10 (10%)	0.787
Revised from ORIF	1 (2%)	1 (2%)	0 (0%)	0 (0%)	2 (2%)	1.000
Revised from CRPP	0 (0%)	1 (2%)	0 (0%)	0 (0%)	1 (1%)	0.495
Revised from hemi	8 (15%)	4 (9%)	0 (0%)	1 (25%)	13 (13%)	0.528
Revised from TSA	15 (29%)	11 (26%)	0 (0%)	2 (50%)	28 (27%)	0.492
Osteonecrosis	4 (8%)	2 (5%)	1 (25%)	0 (0%)	7 (7%)	0.401
Pseudoparalysis irreparable tear without arthritis	20 (38%)	18 (42%)	2 (50%)	1 (25%)	41 (40%)	0.919
Bone tumors	0 (0%)	4 (9.3%)	0 (0%)	0 (0%)	4 (4%)	0.120
Locked shoulder dislocation	0 (0%)	0 (0%)	1 (25%)	0 (0%)	1 (1%)	0.078

Abbreviations: CRPP, closed reduction and percutaneous pinning; ORIF, open reduction internal fixation; RTSA, reverse total shoulder arthroplasty; TSA, total shoulder arthroplasty.

 

Table 3. Number of Patients With Each Indication as Reported by Individual Studies by Continent

	North America	Europe	Asia	Australia	Total
Rotator cuff arthropathy	4418	895	24	122	5459
Osteoarthritis	90	251	1	14	356
Rheumatoid arthritis	59	87	0	2	148
Post-traumatic arthritis	62	136	0	1	199
Instability	23	15	0	1	39
Revision of previous RTSA for instability	29	2	0	1	32
Infection	28	11	2	0	41
Unclassified acute proximal humerus fracture	42	30	4	8	84
Acute 3-part proximal humerus fracture	60	0	0	0	6
Acute 4-part proximal humerus fracture	42	0	0	0	42
Acute 3- or 4-part proximal humerus fracture	92	46	0	0	138
Revised from previous nonop proximal humerus fracture	43	53	0	0	96
Revised from ORIF	3	9	0	0	12
Revised from CRPP	0	3	0	0	3
Revised from hemi	105	51	0	1	157
Revised from TSA	192	246	0	5	443
Osteonecrosis	9	6	1	0	16
Pseudoparalysis irreparable tear without arthritis	535	795	20	2	1352
Bone tumors	0	38	0	0	38
Locked shoulder dislocation	0	0	1	0	1

Abbreviations: CRPP, closed reduction and percutaneous pinning; ORIF, open reduction internal fixation; RTSA, reverse total shoulder arthroplasty; TSA, total shoulder arthroplasty.

 

The ASES, SST-12, and VAS scores were the most frequently reported outcome scores in studies from North America, whereas the Absolute Constant score was the most common score reported in studies from Europe (Table 4). Studies from North America reported significantly higher postoperative external rotation (34.1° ± 13.3° vs 19.3° ± 8.9°) (P < .001) and a greater change in flexion (69.0° ± 24.5° vs 56.3° +/- 11.3°) (P = .004) compared with studies from Europe (Table 5).

Table 4. Outcomes by Continent

Metric (number of studies)	North America	Europe	Asia	Australia	P-value
DASH	1	2	0	0
Preoperative	54.0	62.0 ± 8.5	-	-	0.582
Postoperative	24.0	32.0 ± 2.8	-	-	0.260
Change	-30.0	-30.0 ± 11.3	-	-	1.000
SPADI	2	0	0	0
Preoperative	80.0 ± 4.2	-	-	-	N/A
Postoperative	34.8 ± 1.1	-	-	-	N/A
Change	-45.3 ± 3.2	-	-	-	N/A
Absolute constant	2	27	0	1
Preopeartive	33.0 ± 0.0	28.2 ± 7.1	-	20.0	0.329
Postoperative	54.5 ± 7.8	62.9 ± 9.0	-	65.0	0.432
Change	+21.5 ± 7.8	+34.7 ± 8.0	-	+45.0	0.044
ASES	13	0	2	0
Preoperative	33.2 ± 5.4	-	32.5 ± 3.5	-	0.867
Postoperative	73.9 ± 6.8	-	75.7 ± 10.8	-	0.752
Change	+40.7 ± 6.5	-	+43.2 ± 14.4	-	0.670
UCLA	3	2	1	0
Preoperative	10.1 ± 3.4	11.2 ± 5.7	12.0	-	0.925
Postoperative	24.5 ± 3.1	24.3 ± 3.7	24.0	-	0.991
Change	+14.4 ± 1.6	+13.1 ± 2.0	+12.0	-	0.524
KSS	0	0	2	0
Preopeartive	-	-	38.2 ± 1.1	-	N/A
Postoperative	-	-	72.3 ± 6.0	-	N/A
Change	-	-	+34.1 ± 7.1	-	N/A
SST-12	12	1	0	0
Preoperative	1.9 ± 0.8	1.2	-	-	N/A
Postoperative	7.1 ± 1.5	5.6	-	-	N/A
Change	+5.3 ± 1.2	+4.4	-	-	N/A
SF-12	1	0	0	0
Preoperative	34.5	-	-	-	N/A
Postoperative	38.5	-	-	-	N/A
Change	+4.0	-	-	-	N/A
SSV	0	5	0	0
Preopeartive	-	22.0 ± 7.4	-	-	N/A
Postoperative	-	63.4 ± 7.9	-	-	N/A
Change	-	+41.4 ± 2.1	-	-	N/A
EQ-5D	0	2	0	0
Preoperative	-	0.5 ± 0.2	-	-	N/A
Postoperative	-	0.8 ± 0.1	-	-	N/A
Change	-	+0.3 ± 0.1	-	-	N/A
OOS	1	0	0	0
Preoperative	24.7	-	-	-	N/A
Postoperative	14.9	-	-	-	N/A
Change	-9.9	-	-	-	N/A
Rowe	0	1	0	0
Preoperative	-	50.2	-	-	N/A
Postoperative	-	82.1	-	-	N/A
Change	-	31.9	-	-	N/A
Oxford	0	2	0	0
Preoperative	-	119.9 ± 138.8	-	-	N/A
Postoperative	-	39.9 ± 3.3	-	-	N/A
Change	-	-80.6 ± 142.2	-	-	N/A
Penn	1	0	0	0
Preoperative	24.9	-	-	-	N/A
Postoperative	66.4	-	-	-	N/A
Change	+41.5	-	-	-	N/A
VAS	10	1	1	1
Preoperative	6.6 ± 0.8	7.0	8.4	7.0	N/A
Postoperative	2.0 ± 0.7	1.0	0.8	0.8	N/A
Change	-4.6 ± 0.8	-6.0	-7.6	-6.2	N/A
SF-36 physical	2	0	0	0
Preoperative	32.7 ± 1.2	-	-	-	N/A
Postoperative	39.6 ± 4.0	-	-	-	N/A
Change	+7.0 ± 2.8	-	-	-	N/A
SF-36 mental	2	0	0	0
Preoperative	43.6 ± 2.8	-	-	-	N/A
Postoperative	48.1 ± 1.0	-	-	-	N/A
Change	+4.5 ± 1.8	-	-	-	N/A

Abbreviations: ASES, American Shoulder and Elbow Surgeon score; DASH, Disability of the Arm, Shoulder, and Hand; EQ-5D, EuroQol-5 Dimension; KSS, Korean Shoulder Scoring system; N/A, not available; OOS, Orthopaedic Outcome Score; SF, short form; SPADI, Shoulder Pain and Disability Index; SST, Simple Shoulder Test; SSV, Subjective Shoulder Value; UCLA, University of California, Los Angeles; VAS, visual analog scale.

 

Table 5. Shoulder Range of Motion, by Continent

Metric (number of studies)	North America	Europe	Asia	Australia	P-value
Flexion	18	22	1	1
Preoperative	57.6 ± 17.9	65.5 ± 17.2	91.0	30.0	0.060
Postoperative	126.6 ± 14.4	121.8 ± 19.0	133.0	150.0	0.360
Change	+69.0 ± 24.5	+56.3 ± 11.3	+42.0	120.0	0.004
Abduction	11	12	1	0
Preoperative	53.7 ± 25.0	52.0 ± 19.0	88.0	-	0.311
Postoperative	109.3 ± 15.1	105.4 ± 19.8	131.0	-	0.386
Change	55.5 ± 25.5	53.3 ± 8.3	43.0	-	0.804
External rotation	17	19	0	0
Preoperative	19.4 ± 9.9	11.2 ± 6.1	-	-	0.005
Postoperative	34.1 ± 13.3	19.3 ± 8.9	-	-	<0.001
Change	+14.7 ± 13.2	+8.1 ± 8.5	-	-	0.079

Continue to: DISCUSSION...

 

 

DISCUSSION

RTSA is a common procedure performed in many different areas of the world for a variety of indications. The study hypotheses were partially confirmed, as there were no significant differences seen in the characteristics of the studies published and patients analyzed; although, the majority of studies from North America reported rotator cuff tear arthropathy as the primary indication for RTSA, whereas studies from Europe were split between rotator cuff tear arthropathy and pseudoparalysis as the primary indication. Hence, based on the current literature the study proved that we are treating the same patients. Despite this finding, we may be treating them for different reasons with an RTSA.

RTSA has become a standard procedure in the United States, with >20,000 RTSAs performed in 2011.¹⁰ This number will continue to increase as it has over the past 20 years given the aging population in the United States, as well as the expanding indications for RTSA.¹¹ Indications of RTSA have become broad, although the main indication remains as rotator cuff tear arthropathy (>60% of all patients included in this study), and pseudoparalysis (>15% of all patients included in this study). Results for RTSA for rotator cuff tear arthropathy and pseudoparalysis have been encouraging.^16,17 Frankle and colleagues¹⁶ evaluated 60 patients who underwent RTSA for rotator cuff tear arthropathy at a minimum of 2 years follow-up (average, 33 months). The authors found significant improvements in all measured clinical outcome variables (P < .0001) (ASES, mean function score, mean pain score, and VAS) as well as ROM, specifically forward flexion increased from 55° to 105.1°, and abduction increased from 41.4° to 101.8°. Similarly, Werner and colleagues¹⁷ evaluated 58 consecutive patients who underwent RTSA for pseudoparalysis secondary to irreparable rotator cuff dysfunction at a mean follow-up of 38 months. Overall, significant improvements (P < .0001) were seen in the SSV score, relative Constant score, and Constant score for pain, active anterior elevation (42° to 100° following RTSA), and active abduction (43° to 90° following RTSA).

It is essential to understand the similarities and differences between patients undergoing RTSA in different parts of the world so the literature from various countries can be compared between regions, and conclusions extrapolated to the correct patients. For example, an interesting finding in this study is that the majority of patients in North America have their prosthesis cemented whereas the majority of patients in Australia have their prosthesis press-fit. While the patients each continent is treating are not significantly different (mostly older women), the difference in surgical technique could have implications in long- or short-term functional outcomes. Prior studies have shown no difference in axial micromotion between cemented and press-fit humeral components, but the clinical implications surrounding this are not well defined.¹⁸ Small series comparing cementless to cemented humeral prosthesis in RTSA have found no significant differences in clinical outcomes or postoperative ROM, but larger series are necessary to validate these outcomes.¹⁹ However, studies have shown lower rates of postoperative infections in patients who receive antibiotic-loaded cement compared with those who receive plain bone cement following RTSA.²⁰

Similarly, as the vast majority of patients in North America had an RTSA for rotator cuff arthropathy (75.8%) whereas those from Europe had RTSA almost equally for rotator cuff arthropathy (33.5%) and pseudoparalysis (29.7%), one must ensure similar patient populations before attempting to extrapolate results of a study from a different country to patients in other areas. Fortunately, the clinical results following RTSA for either indication have been good.^6,21,22

One final point to consider is the cost effectiveness of the implant. Recent evidence has shown that RTSA is associated with a higher risk for in-hospital death, multiple perioperative complications, prolonged hospital stay, and increased hospital cost when compared with TSA.²³ This data may be biased as the patient selection for RTSA varies from that of TSA, but it is a point that must be considered. Other studies have shown that an RTSA is a cost-effective treatment option for treating patients with rotator cuff tear arthropathy, and is a more cost-effective option in treating rotator cuff tear arthropathy than hemiarthroplasty.^24,25 Similarly, RTSA offers a more cost-effective treatment option with better outcomes for patients with acute proximal humerus fractures when compared with open reduction internal fixation and hemiarthroplasty.²⁶ However, TSA is a more cost-effective treatment option than RTSA for patients with glenohumeral osteoarthritis.²⁷ With changing reimbursement in healthcare, surgeons must scrutinize not only anticipated outcomes with specific implants but the cost effectiveness of these implants as well. Further cost analysis studies are necessary to determine the ideal candidate for an RTSA.

LIMITATIONS

Despite its extensive review of the literature, this study had several limitations. While 2 independent authors searched for studies, it is possible that some studies were missed during the search process, introducing possible selection bias. No abstracts or unpublished works were included which could have introduced publication bias. Several studies did not report all variables the authors examined, and this could have skewed some of the results since the reporting of additional variables could have altered the data to show significant differences in some measured variables. As outcome measures for various pathologies were not compared, conclusions cannot be drawn on the best treatment option for various indications. As case reports were included, this could have lowered both the MCMS as well as the average in studies reporting outcomes. Furthermore, given the overall poor quality of the underlying data available for this study, the validity/generalizability of the results could be limited as the level of evidence of this systematic review is only as high as the studies it includes. There are subtle differences between rotator cuff arthropathy and pseudoparalysis, and some studies may have classified patients differently than others, causing differences in indications. Finally, as the primary goal of this study was to report on demographics, no evaluation of concomitant pathology at the time of surgery or rehabilitation protocols was performed.

CONCLUSION

The quantity, but not the quality of RTSA studies is increasing. Indications for RTSA varied by continent although most patients underwent RTSA for rotator cuff arthropathy. The majority of patients undergoing RTSA are female over the age of 60 years for a diagnosis of rotator cuff arthropathy with pseudoparalysis.

This paper will be judged for the Resident Writer’s Award.

Reverse total shoulder arthroplasty (RTSA) is a common treatment for rotator cuff tear arthropathy. We performed a systematic review of all the RTSA literature to answer if we are treating the same patients with RTSA, across the world.

A systematic review was registered with PROSPERO, the international prospective register of systematic reviews, and performed with Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines using 3 publicly available free databases. Therapeutic clinical outcome investigations reporting RTSA outcomes with levels of evidence I to IV were eligible for inclusion. All study, subject, and surgical technique demographics were analyzed and compared between continents. Statistical comparisons were conducted using linear regression, analysis of variance (ANOVA), Fisher's exact test, and Pearson's chi-square test.

There were 103 studies included in the analysis (8973 patients; 62% female; mean age, 70.9 ± 6.7 years; mean length of follow-up, 34.3 ± 19.3 months) that had a low Modified Coleman Methodology Score (MCMS) (mean, 36.9 ± 8.7: poor). Most patients (60.8%) underwent RTSA for a diagnosis of rotator cuff arthropathy, whereas 1% underwent RTSA for fracture; indications varied by continent. There were no consistent reports of preopeartive or postoperative scores from studies in any region. Studies from North America reported significantly higher postoperative external rotation (34.1° ± 13.3° vs 19.3° ± 8.9°) (P < .001) and a greater change in flexion (69.0° ± 24.5° vs 56.3° ± 11.3°) (P = .004) compared with studies from Europe. North America had the greatest total number of publications followed by Europe. The total yearly number of publications increased each year (P < .001), whereas the MCMS decreased each year (P = .037).

The quantity, but not the quality of RTSA studies is increasing. Indications for RTSA varied by continent, although most patients underwent RTSA for rotator cuff arthropathy. The majority of patients undergoing RTSA are female over the age of 60 years for a diagnosis of rotator cuff arthropathy with pseudoparalysis.

Continue to: Reverse total shoulder arthroplasty...

Reverse total shoulder arthroplasty (RTSA) is a common procedure with indications including rotator cuff tear arthropathy, proximal humerus fractures, and others.^1,2 Studies have shown excellent, reliable, short- and mid-term outcomes in patients treated with RTSA for various indications.^3-5 Al-Hadithy and colleagues⁶ reviewed 41 patients who underwent RTSA for pseudoparalysis secondary to rotator cuff tear arthropathy and, at a mean follow-up of 5 years, found significant improvements in range of motion (ROM) as well as age-adjusted Constant and Oxford Outcome scores. Similarly, Ross and colleagues⁷ evaluated outcomes of RTSA in 28 patients in whom RTSA was performed for 3- or 4-part proximal humerus fractures, and found both good clinical and radiographic outcomes with no revision surgeries at a mean follow-up of 54.9 months. RTSA is performed across the world, with specific implant designs, specifically humeral head inclination, but is more common in some areas when compared with others.^3,8,9

The number of RTSAs performed has steadily increased over the past 20 years, with recent estimates of approximately 20,000 RTSAs performed in the United States in 2011.^10,11 However, there is little information about the similarities and differences between those patients undergoing RTSA in various parts of the world regarding surgical indications, patient demographics, and outcomes. The purpose of this study is to perform a systematic review and meta-analysis of the RTSA body of literature to both identify and compare characteristics of studies published (level of evidence, whether a conflict of interest existed), patients analyzed (age, gender), and surgical indications performed across both continents and countries. Essentially, the study aims to answer the question, "Across the world, are we treating the same patients?" The authors hypothesized that there would be no significant differences in RTSA publications, subjects, and indications based on both the continent and country of publication.

METHODS

A systematic review was conducted according to PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) guidelines using a PRISMA checklist.¹² A systematic review registration was performed using PROSPERO, the international prospective register of systematic reviews (registration number CRD42014010578).¹³Two reviewers independently conducted the search on March 25, 2014, using the following databases: Medline, Cochrane Central Register of Controlled Trials, SportDiscus, and CINAHL. The electronic search citation algorithm utilized was: (((((reverse[Title/Abstract]) AND shoulder[Title/Abstract]) AND arthroplasty[Title/Abstract]) NOT arthroscopic[Title/Abstract]) NOT cadaver[Title/Abstract]) NOT biomechanical[Title/Abstract]. English language Level I to IV evidence (2011 update by the Oxford Centre for Evidence-Based Medicine¹⁴) clinical studies were eligible. Medical conference abstracts were ineligible for inclusion. All references within included studies were cross-referenced for inclusion if missed by the initial search with any additionally located studies screened for inclusion. Duplicate subject publications within separate unique studies were not reported twice, but rather the study with longer duration follow-up or, if follow-up was equal, the study with the greater number of patients was included. Level V evidence reviews, letters to the editor, basic science, biomechanical and cadaver studies, total shoulder arthroplasty (TSA) papers, arthroscopic shoulder surgery papers, imaging, surgical techniques, and classification studies were excluded.

A total of 255 studies were identified, and, after implementation of the exclusion criteria, 103 studies were included in the final analysis (Figure 1). Subjects of interest in this systematic review underwent RTSA for one of many indications including rotator cuff tear arthropathy, osteoarthritis, rheumatoid arthritis, posttraumatic arthritis, instability, revision from a previous RTSA for instability, infection, acute proximal humerus fracture, revision from a prior proximal humerus fracture, revision from a prior hemiarthroplasty, revision from a prior TSA, osteonecrosis, pseudoparalysis, tumor, and a locked shoulder dislocation. There was no minimum follow-up or rehabilitation requirement. Study and subject demographic parameters analyzed included year of publication, years of subject enrollment, presence of study financial conflict of interest, number of subjects and shoulders, gender, age, body mass index, diagnoses treated, and surgical positioning. Clinical outcome scores sought were the DASH (Disability of the Arm, Shoulder, and Hand), SPADI (Shoulder Pain And Disability Index), Absolute Constant, ASES (American Shoulder and Elbow Score), KSS (Korean Shoulder Score), SST-12 (Simple Shoulder Test), SF-12 (12-item Short Form), SF-36 (36-item Short Form), SSV (Subjective Shoulder Value), EQ-5D (EuroQol-5 Dimension), SANE (Single Assessment Numeric Evaluation), Rowe Score for Instability, Oxford Instability Score, UCLA (University of California, Los Angeles) activity score, Penn Shoulder Score, and VAS (visual analog scale). In addition, ROM (forward elevation, abduction, external rotation, internal rotation) was analyzed. Radiographs and magnetic resonance imaging data were extracted when available. The methodological quality of the study was evaluated using the MCMS (Modified Coleman Methodology Score).¹⁵

STATISTICAL ANALYSIS

First, the number of publications per year, level of evidence, and Modified Coleman Methodology Score were tested for association with the calendar year using linear regression. Second, demographic data were tested for association with the continent using Pearson’s chi-square test or ANOVA. Third, indications were tested for association with the continent using Fisher’s exact test. Finally, clinical outcome scores and ROM were tested for association with the continent using ANOVA. Statistical significance was extracted from studies when available. Statistical significance was defined as P < .05.

Continue to: RESULTS...

 

 

RESULTS

There were 103 studies included in the analysis (Figure 1). A total of 8973 patients were included, 62% of whom were female with a mean age of 70.9 ± 6.7 years (Table 1). The average follow-up was 34.3 ± 19.3 months. North America had the overall greatest total number of publications on RTSA, followed by Europe (Figure 2). The total yearly number of publications increased by a mean of 1.95 publications each year (P < .001). There was no association between the mean level of evidence with the year of publication (P = .296) (Figure 3). Overall, the rating of studies was poor for the MCMS (mean 36.9 ± 8.7). The MCMS decreased each year by a mean of 0.76 points (P = .037) (Figure 4).

Table 1. Demographic Data by Continent

	North America	Europe	Asia	Australia	Total	P-value
Number of studies	52	43	4	4	103	-
Number of subjects	6158	2609	51	155	8973	-
Level of evidence						0.693
II	5 (10%)	3 (7%)	0 (0%)	0 (0%)	8 (8%)
III	10 (19%)	4 (9%)	0 (0%)	1 (25%)	15 (15%)
IV	37 (71%)	36 (84%)	4 (100%)	3 (75%)	80 (78%)
Mean MCMS	34.6 ± 8.4	40.2 ± 8.0	32.5 12.4	34.5 ± 6.6	36.9 ± 8.7	0.010
Institutional collaboration						1.000
Multi-center	7 (14%)	6 (14%)	0 (0%)	0 (0%)	13 (13%)
Single-center	45 (86%)	37 (86%)	4 (100%)	4 (100%)	90 (87%)
Financial conflict of interest						0.005
Present	28 (54%)	15 (35%)	0 (0%)	0 (0%)	43 (42%)
Not present	19 (37%)	16 (37%)	4 (100%)	4 (100%)	43 (42%)
Not reported	5 (10%)	12 (28%)	0 (0%)	0 (0%)	17 (17%)
Sex						N/A
Male	2157 (38%)	1026 (39%)	13 (25%)	61 (39%)	3257 (38%)
Female	3520 (62%)	1622 (61%)	38 (75%)	94 (61%)	5274 (62%)
Mean age (years)	71.3 ± 5.6	70.1 ± 7.9	68.1 ± 5.3	76.9 ± 3.0	70.9 ± 6.7	0.191
Minimum age (mean across studies)	56.9 ± 12.8	52.8 ± 15.7	62.8 ± 6.2	68.0 ± 12.1	55.6 ± 14.3	0.160
Maximum age (mean across studies)	82.1 ± 8.6	83.0 ± 5.5	73.0 ± 9.4	85.0 ± 7.9	82.2 ± 7.6	0.079
Mean length of follow-up (months)	26.5 ± 13.7	43.1 ± 21.7	29.4 ± 7.9	34.2 ± 16.6	34.3 ± 19.3	<0.001
Prosthesis type						N/A
Cemented	988 (89%)	969 (72%)	0 (0%)	8 (16%)	1965 (78%)
Press fit	120 (11%)	379 (28%)	0 (0%)	41 (84%)	540 (22%)

Abbreviations: MCMS, Modified Coleman Methodology Score; N/A, not available.

 

In studies that reported press-fit vs cemented prostheses, the highest percentage of press-fit prostheses compared with cemented prostheses was seen in Australia (84% press-fit), whereas the highest percentage of cemented prostheses was seen in North America (89% cemented). A higher percentage of studies from North America had a financial conflict of interest (COI) than did those from other countries (54% had a COI).

Continue to: Rotator cuff tear arthropathy...

 

 

Rotator cuff tear arthropathy was the most common indication for RTSA overall in 5459 patients, followed by pseudoparalysis in 1352 patients (Tables 2 and 3). While studies in North America reported rotator cuff tear arthropathy as the indication for RTSA in 4418 (75.8%) patients, and pseudoparalysis as the next most common indication in 535 (9.2%) patients, studies from Europe reported rotator cuff tear arthropathy as the indication in 895 (33.5%) patients, and pseudoparalysis as the indication in 795 (29.7%) patients. Studies from Asia also had a relatively even split between rotator cuff tear arthropathy and pseudoparalysis (45.3% vs 37.8%), whereas those from Australia were mostly rotator cuff tear arthropathy (77.7%).

Table 2. Number (Percent) of Studies With Each Indication by Continent

	North America	Europe	Asia	Australia	Total	P-value
Rotator cuff arthropathy	29 (56%)	19 (44%)	3 (75%)	3 (75%)	54 (52%)	0.390
Osteoarthritis	4 (8%)	10 (23%)	1 (25%)	1 (25%)	16 (16%)	0.072
Rheumatoid arthritis	9 (17%)	10 (23%)	0 (0%)	2 (50%)	21 (20%)	0.278
Post-traumatic arthritis	3 (6%)	5 (12%)	0 (0%)	1 (25%)	9 (9%)	0.358
Instability	6 (12%)	3 (7%)	0 (0%)	1 (25%)	10 (10%)	0.450
Revision of previous RTSA for instability	5 (10%)	1 (2%)	0 (0%)	1 (25%)	7 (7%)	0.192
Infection	4 (8%)	1 (2%)	1 (25%)	0 (0%)	6 (6%)	0.207
Unclassified acute proximal humerus fracture	9 (17%)	5 (12%)	1 (25%)	1 (25%)	16  (16%)	0.443
Acute 2-part proximal humerus fracture	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	N/A
Acute 3-part proximal humerus fracture	2 (4%)	0 (0%)	0 (0%)	0 (0%)	2 (2%)	0.574
Acute 4-part proximal humerus fracture	5 (10%)	0 (0%)	0 (0%)	0 (0%)	5 (5%)	0.183
Acute 3- or 4-part proximal humerus fracture	6 (12%)	2 (5%)	0 (0%)	0 (0%)	8 (8%)	0.635
Revised from previous nonop proximal humerus fracture	7 (13%)	3 (7%)	0 (0%)	0 (0%)	10 (10%)	0.787
Revised from ORIF	1 (2%)	1 (2%)	0 (0%)	0 (0%)	2 (2%)	1.000
Revised from CRPP	0 (0%)	1 (2%)	0 (0%)	0 (0%)	1 (1%)	0.495
Revised from hemi	8 (15%)	4 (9%)	0 (0%)	1 (25%)	13 (13%)	0.528
Revised from TSA	15 (29%)	11 (26%)	0 (0%)	2 (50%)	28 (27%)	0.492
Osteonecrosis	4 (8%)	2 (5%)	1 (25%)	0 (0%)	7 (7%)	0.401
Pseudoparalysis irreparable tear without arthritis	20 (38%)	18 (42%)	2 (50%)	1 (25%)	41 (40%)	0.919
Bone tumors	0 (0%)	4 (9.3%)	0 (0%)	0 (0%)	4 (4%)	0.120
Locked shoulder dislocation	0 (0%)	0 (0%)	1 (25%)	0 (0%)	1 (1%)	0.078

Abbreviations: CRPP, closed reduction and percutaneous pinning; ORIF, open reduction internal fixation; RTSA, reverse total shoulder arthroplasty; TSA, total shoulder arthroplasty.

 

Table 3. Number of Patients With Each Indication as Reported by Individual Studies by Continent

	North America	Europe	Asia	Australia	Total
Rotator cuff arthropathy	4418	895	24	122	5459
Osteoarthritis	90	251	1	14	356
Rheumatoid arthritis	59	87	0	2	148
Post-traumatic arthritis	62	136	0	1	199
Instability	23	15	0	1	39
Revision of previous RTSA for instability	29	2	0	1	32
Infection	28	11	2	0	41
Unclassified acute proximal humerus fracture	42	30	4	8	84
Acute 3-part proximal humerus fracture	60	0	0	0	6
Acute 4-part proximal humerus fracture	42	0	0	0	42
Acute 3- or 4-part proximal humerus fracture	92	46	0	0	138
Revised from previous nonop proximal humerus fracture	43	53	0	0	96
Revised from ORIF	3	9	0	0	12
Revised from CRPP	0	3	0	0	3
Revised from hemi	105	51	0	1	157
Revised from TSA	192	246	0	5	443
Osteonecrosis	9	6	1	0	16
Pseudoparalysis irreparable tear without arthritis	535	795	20	2	1352
Bone tumors	0	38	0	0	38
Locked shoulder dislocation	0	0	1	0	1

Abbreviations: CRPP, closed reduction and percutaneous pinning; ORIF, open reduction internal fixation; RTSA, reverse total shoulder arthroplasty; TSA, total shoulder arthroplasty.

 

The ASES, SST-12, and VAS scores were the most frequently reported outcome scores in studies from North America, whereas the Absolute Constant score was the most common score reported in studies from Europe (Table 4). Studies from North America reported significantly higher postoperative external rotation (34.1° ± 13.3° vs 19.3° ± 8.9°) (P < .001) and a greater change in flexion (69.0° ± 24.5° vs 56.3° +/- 11.3°) (P = .004) compared with studies from Europe (Table 5).

Table 4. Outcomes by Continent

Metric (number of studies)	North America	Europe	Asia	Australia	P-value
DASH	1	2	0	0
Preoperative	54.0	62.0 ± 8.5	-	-	0.582
Postoperative	24.0	32.0 ± 2.8	-	-	0.260
Change	-30.0	-30.0 ± 11.3	-	-	1.000
SPADI	2	0	0	0
Preoperative	80.0 ± 4.2	-	-	-	N/A
Postoperative	34.8 ± 1.1	-	-	-	N/A
Change	-45.3 ± 3.2	-	-	-	N/A
Absolute constant	2	27	0	1
Preopeartive	33.0 ± 0.0	28.2 ± 7.1	-	20.0	0.329
Postoperative	54.5 ± 7.8	62.9 ± 9.0	-	65.0	0.432
Change	+21.5 ± 7.8	+34.7 ± 8.0	-	+45.0	0.044
ASES	13	0	2	0
Preoperative	33.2 ± 5.4	-	32.5 ± 3.5	-	0.867
Postoperative	73.9 ± 6.8	-	75.7 ± 10.8	-	0.752
Change	+40.7 ± 6.5	-	+43.2 ± 14.4	-	0.670
UCLA	3	2	1	0
Preoperative	10.1 ± 3.4	11.2 ± 5.7	12.0	-	0.925
Postoperative	24.5 ± 3.1	24.3 ± 3.7	24.0	-	0.991
Change	+14.4 ± 1.6	+13.1 ± 2.0	+12.0	-	0.524
KSS	0	0	2	0
Preopeartive	-	-	38.2 ± 1.1	-	N/A
Postoperative	-	-	72.3 ± 6.0	-	N/A
Change	-	-	+34.1 ± 7.1	-	N/A
SST-12	12	1	0	0
Preoperative	1.9 ± 0.8	1.2	-	-	N/A
Postoperative	7.1 ± 1.5	5.6	-	-	N/A
Change	+5.3 ± 1.2	+4.4	-	-	N/A
SF-12	1	0	0	0
Preoperative	34.5	-	-	-	N/A
Postoperative	38.5	-	-	-	N/A
Change	+4.0	-	-	-	N/A
SSV	0	5	0	0
Preopeartive	-	22.0 ± 7.4	-	-	N/A
Postoperative	-	63.4 ± 7.9	-	-	N/A
Change	-	+41.4 ± 2.1	-	-	N/A
EQ-5D	0	2	0	0
Preoperative	-	0.5 ± 0.2	-	-	N/A
Postoperative	-	0.8 ± 0.1	-	-	N/A
Change	-	+0.3 ± 0.1	-	-	N/A
OOS	1	0	0	0
Preoperative	24.7	-	-	-	N/A
Postoperative	14.9	-	-	-	N/A
Change	-9.9	-	-	-	N/A
Rowe	0	1	0	0
Preoperative	-	50.2	-	-	N/A
Postoperative	-	82.1	-	-	N/A
Change	-	31.9	-	-	N/A
Oxford	0	2	0	0
Preoperative	-	119.9 ± 138.8	-	-	N/A
Postoperative	-	39.9 ± 3.3	-	-	N/A
Change	-	-80.6 ± 142.2	-	-	N/A
Penn	1	0	0	0
Preoperative	24.9	-	-	-	N/A
Postoperative	66.4	-	-	-	N/A
Change	+41.5	-	-	-	N/A
VAS	10	1	1	1
Preoperative	6.6 ± 0.8	7.0	8.4	7.0	N/A
Postoperative	2.0 ± 0.7	1.0	0.8	0.8	N/A
Change	-4.6 ± 0.8	-6.0	-7.6	-6.2	N/A
SF-36 physical	2	0	0	0
Preoperative	32.7 ± 1.2	-	-	-	N/A
Postoperative	39.6 ± 4.0	-	-	-	N/A
Change	+7.0 ± 2.8	-	-	-	N/A
SF-36 mental	2	0	0	0
Preoperative	43.6 ± 2.8	-	-	-	N/A
Postoperative	48.1 ± 1.0	-	-	-	N/A
Change	+4.5 ± 1.8	-	-	-	N/A

Abbreviations: ASES, American Shoulder and Elbow Surgeon score; DASH, Disability of the Arm, Shoulder, and Hand; EQ-5D, EuroQol-5 Dimension; KSS, Korean Shoulder Scoring system; N/A, not available; OOS, Orthopaedic Outcome Score; SF, short form; SPADI, Shoulder Pain and Disability Index; SST, Simple Shoulder Test; SSV, Subjective Shoulder Value; UCLA, University of California, Los Angeles; VAS, visual analog scale.

 

Table 5. Shoulder Range of Motion, by Continent

Metric (number of studies)	North America	Europe	Asia	Australia	P-value
Flexion	18	22	1	1
Preoperative	57.6 ± 17.9	65.5 ± 17.2	91.0	30.0	0.060
Postoperative	126.6 ± 14.4	121.8 ± 19.0	133.0	150.0	0.360
Change	+69.0 ± 24.5	+56.3 ± 11.3	+42.0	120.0	0.004
Abduction	11	12	1	0
Preoperative	53.7 ± 25.0	52.0 ± 19.0	88.0	-	0.311
Postoperative	109.3 ± 15.1	105.4 ± 19.8	131.0	-	0.386
Change	55.5 ± 25.5	53.3 ± 8.3	43.0	-	0.804
External rotation	17	19	0	0
Preoperative	19.4 ± 9.9	11.2 ± 6.1	-	-	0.005
Postoperative	34.1 ± 13.3	19.3 ± 8.9	-	-	<0.001
Change	+14.7 ± 13.2	+8.1 ± 8.5	-	-	0.079

Continue to: DISCUSSION...

 

 

DISCUSSION

RTSA is a common procedure performed in many different areas of the world for a variety of indications. The study hypotheses were partially confirmed, as there were no significant differences seen in the characteristics of the studies published and patients analyzed; although, the majority of studies from North America reported rotator cuff tear arthropathy as the primary indication for RTSA, whereas studies from Europe were split between rotator cuff tear arthropathy and pseudoparalysis as the primary indication. Hence, based on the current literature the study proved that we are treating the same patients. Despite this finding, we may be treating them for different reasons with an RTSA.

RTSA has become a standard procedure in the United States, with >20,000 RTSAs performed in 2011.¹⁰ This number will continue to increase as it has over the past 20 years given the aging population in the United States, as well as the expanding indications for RTSA.¹¹ Indications of RTSA have become broad, although the main indication remains as rotator cuff tear arthropathy (>60% of all patients included in this study), and pseudoparalysis (>15% of all patients included in this study). Results for RTSA for rotator cuff tear arthropathy and pseudoparalysis have been encouraging.^16,17 Frankle and colleagues¹⁶ evaluated 60 patients who underwent RTSA for rotator cuff tear arthropathy at a minimum of 2 years follow-up (average, 33 months). The authors found significant improvements in all measured clinical outcome variables (P < .0001) (ASES, mean function score, mean pain score, and VAS) as well as ROM, specifically forward flexion increased from 55° to 105.1°, and abduction increased from 41.4° to 101.8°. Similarly, Werner and colleagues¹⁷ evaluated 58 consecutive patients who underwent RTSA for pseudoparalysis secondary to irreparable rotator cuff dysfunction at a mean follow-up of 38 months. Overall, significant improvements (P < .0001) were seen in the SSV score, relative Constant score, and Constant score for pain, active anterior elevation (42° to 100° following RTSA), and active abduction (43° to 90° following RTSA).

It is essential to understand the similarities and differences between patients undergoing RTSA in different parts of the world so the literature from various countries can be compared between regions, and conclusions extrapolated to the correct patients. For example, an interesting finding in this study is that the majority of patients in North America have their prosthesis cemented whereas the majority of patients in Australia have their prosthesis press-fit. While the patients each continent is treating are not significantly different (mostly older women), the difference in surgical technique could have implications in long- or short-term functional outcomes. Prior studies have shown no difference in axial micromotion between cemented and press-fit humeral components, but the clinical implications surrounding this are not well defined.¹⁸ Small series comparing cementless to cemented humeral prosthesis in RTSA have found no significant differences in clinical outcomes or postoperative ROM, but larger series are necessary to validate these outcomes.¹⁹ However, studies have shown lower rates of postoperative infections in patients who receive antibiotic-loaded cement compared with those who receive plain bone cement following RTSA.²⁰

Similarly, as the vast majority of patients in North America had an RTSA for rotator cuff arthropathy (75.8%) whereas those from Europe had RTSA almost equally for rotator cuff arthropathy (33.5%) and pseudoparalysis (29.7%), one must ensure similar patient populations before attempting to extrapolate results of a study from a different country to patients in other areas. Fortunately, the clinical results following RTSA for either indication have been good.^6,21,22

One final point to consider is the cost effectiveness of the implant. Recent evidence has shown that RTSA is associated with a higher risk for in-hospital death, multiple perioperative complications, prolonged hospital stay, and increased hospital cost when compared with TSA.²³ This data may be biased as the patient selection for RTSA varies from that of TSA, but it is a point that must be considered. Other studies have shown that an RTSA is a cost-effective treatment option for treating patients with rotator cuff tear arthropathy, and is a more cost-effective option in treating rotator cuff tear arthropathy than hemiarthroplasty.^24,25 Similarly, RTSA offers a more cost-effective treatment option with better outcomes for patients with acute proximal humerus fractures when compared with open reduction internal fixation and hemiarthroplasty.²⁶ However, TSA is a more cost-effective treatment option than RTSA for patients with glenohumeral osteoarthritis.²⁷ With changing reimbursement in healthcare, surgeons must scrutinize not only anticipated outcomes with specific implants but the cost effectiveness of these implants as well. Further cost analysis studies are necessary to determine the ideal candidate for an RTSA.

LIMITATIONS

Despite its extensive review of the literature, this study had several limitations. While 2 independent authors searched for studies, it is possible that some studies were missed during the search process, introducing possible selection bias. No abstracts or unpublished works were included which could have introduced publication bias. Several studies did not report all variables the authors examined, and this could have skewed some of the results since the reporting of additional variables could have altered the data to show significant differences in some measured variables. As outcome measures for various pathologies were not compared, conclusions cannot be drawn on the best treatment option for various indications. As case reports were included, this could have lowered both the MCMS as well as the average in studies reporting outcomes. Furthermore, given the overall poor quality of the underlying data available for this study, the validity/generalizability of the results could be limited as the level of evidence of this systematic review is only as high as the studies it includes. There are subtle differences between rotator cuff arthropathy and pseudoparalysis, and some studies may have classified patients differently than others, causing differences in indications. Finally, as the primary goal of this study was to report on demographics, no evaluation of concomitant pathology at the time of surgery or rehabilitation protocols was performed.

CONCLUSION

The quantity, but not the quality of RTSA studies is increasing. Indications for RTSA varied by continent although most patients underwent RTSA for rotator cuff arthropathy. The majority of patients undergoing RTSA are female over the age of 60 years for a diagnosis of rotator cuff arthropathy with pseudoparalysis.

This paper will be judged for the Resident Writer’s Award.