Br. J. Surg. 1992, Vol. 79, June, 513-51 6
L. S. Jensen, A. J. Andersen*, P. M . Christiansent, P. Hokland$, C. 0. Juhl, G. Madsen”, J. Mortensen*, C. Meller-Nielsent, F. Hanberg-Serensen and M. Hoklandg University Departments of Surgical Gastroenterology, Aarhus County Hospital and *Aarhus Municipal Hospital, ?Department of Surgery, Randers County Hospital, $University Department of Medicine and Haematology, Aarhus County Hospital and $Department of Microbiology, Aarhus University, Denmark Correspondence to: Dr L. S. Jensen, University Department of Surgical Gastroenterology L, Aarhus
Countv HosDital. . . DK-8000 Aarhus. Denmark
Postoperative infection and natural killer cell function following blood transfusion in patients undergoing elective colorectal surgery Thefrequency of infection in 197patients undergoing elective colorectal surgery and having either no blood transfusion, transfusion with whole blood, or filtered blood free from leucocytes and platelets was investigated in aprospective randomized trial. Natural killer cellfunction was measured before operation and 3 , 7 and 30 days after surgery in 60 consecutive patients. Of the patients 104 required blood transfusion; 48 received filtered blood and 56 underwent whole blood transfusion. Postoperative infections developed in 13 patients transfused with whole blood (23 per cent, 9.5 per cent confidence interval 13-32 per cent), in one patient transfused with blood free f r o m leucocytes and platelets ( 2 per cent, 9.5 per cent confidence interval 0.05-11 per cent) and in two non-transfusedpatients ( 2per cent, 95 per cent conjidence interval 0.3-8per cent) ( P < 0.01). Natural killer cellfunction was signiJicantly ( P < 0.001) impaired up to 30 days after surgery in patients transfused with whole blood. These data provide a strong case against the use of whole blood transfusion in patients undergoing elective colorectal surgery.
Blood transfusion is associated with certain risks and during the past three or four decades attention has been drawn to its possible effects on the immune system. Blood transfusion produces immunosuppression and this may have beneficial clinical consequences, e.g. in renal transplantation, where transfusion-induced changes are associated with a decreased rate of allograft rejection’-3. Significant decreases in the rate of recurrence of Crohn’s disease have also been described after transfusion4. In other conditions the immunosuppressive action of blood transfusion may be dzleterious. Perioperative transfusion increases the recurrence rate and decreases survival rates in patients with c ~ l o r e c t a land ~ ~ ~breast’ cancer. These studies have been retrospective and it has not been proven that blood transfusion is the direct cause of these effects. Blood transfusion may be a significant factor affecting postoperative immune system function, leading to an increased susceptibility to i n f e ~ t i o n *and ~ ~ multiple organ failure”. Leucocytes may mediate the effects of cancer recurrence and allograft rejection” ;whole blood has a better effect than packed red cells on allograft survival”. The underlying mechanism for this is not clear but it has been demonstrated that whole blood transfusion causes reduced natural killer (NK ) cell function and a decreased T he1per:T suppressor lymphocyte ratio’ 3 . Whole blood amplifies the impairment in delayed hypersensitivity after operationI4. Blood free from white cells may be the safest preparation for preventing postoperative infection when transfusion is necessary. A prospective study was designed to compare the frequency of infection in patients undergoing elective colorectal surgery and either having no transfusion, or being randomized to transfusion with whole blood or blood free from white cells. Immune function was monitored in all patients from one of the three participating departments of surgery.
Patients and methods All patients admitted for elective colorectal surgery between August 1989 and September 1990 were considered for inclusion in the study,
~7-1323/92/060513-04 0 1992 Butterworth-Heinemann Ltd
which had been approved by the local ethical committee. The exclusion criteria were: age less than 18 years; acute colorectal surgery; and immunosuppressive treatment. After giving informed consent, patients in whom transfusion was indicated were randomly allocated using sealed envelopes to receive either whole blood or blood free from white cells. The latter was obtained by passing whole blood through a filtering device (Pall-filter; MEDA, Copenhagen, Denmark) thereby reducing the leucocyte and platelet content by 99.98 per cent. Red cells and plasma passed through the filter, which was mounted at the bedside and placed between the blood bag and the transfusion tube. All blood used was less than 3 weeks old. The indications for transfusion were a haematocrit before operation of less than 0.35 or a haemoglobin level below 7.2 mmoll- I . Intraoperative or postoperative transfusions were given according to the haemodynamic status of the patient. All patients received cefuroxime 3 g and metronidazole 1.5 g intravenously as one peroperative dose of antibiotic prophylaxis. Oral antibiotics were not given. The bowel was prepared with oral laxative and saline enema or Golytely (AAKH, Aarhus, Denmark). After surgery, abdominal wounds were closed with absorbable sutures for the peritoneum and fascia and non-absorbable sutures to the skin. Antibiotics were not used in the wounds; neither were subcutaneous drains. Follow-up examination included daily wound inspection by an independent assessor from the third day after operation until suture removal or patient discharge. All patients were followed in the outpatient clinic for 30 days after operation. Abdominal wound infection was defined as accumulation of pus with spontaneous discharge or requiring surgical drainage. Wounds with purulent discharge after the patient returned home were included in the infected group. Deep surgical infection consisted of: ( 1 ) intra-abdominal abscess defined as an intraperitoneal or pelvic collection of pus diagnosed by ultrasonography, laparotomy or spontaneous discharge; and ( 2) septicaemia defined as the presence of spiking fever associated with positive blood culture. The frequency of anastomotic leakage was diagnosed by fever and faecal fistula or peritonitis. Pneumonia was diagnosed by fever and infiltrate on chest radiography. Other surgical complications such as abdominal wound dehiscence, bleeding and ileus were considered separately. Blood samples for measurement of haematocrit, haemoglobin, serum albumin, immunoglobulin A (IgA), IgG, IgM and NK cell function were taken before operation and on days 3, 7 and 30 after operation. Cytomegalovirus (CMV) titre (IgG and IgM), Epstein-Barr
513
Infection following transfusion for colorectal surgery: L. S. Jensen et al.
virus (EBV) titre (IgG and IgM) and herpes virus titre were measured before operation and 30 days afterwards. NK cell function was evaluated in 60 consecutive patients from one of the participating departments of surgery by a modified whole blood assay as previously described15. A volume of 100 pl heparinized blood was added to varying numbers of 51Cr-radiolabelled K562 target cells. After a 4-h incubation at 37°C supernatants were counted for isotope release and cytotoxicity expressed as the percentage specific "Cr release according to the formula:
(E-b) x Percentage "Cr release =
(
V , - V, x H
V,
)-(S-b) x 100
( M x 0.8) - S
where V, is the total assay volume, V, is the volume of blood assayed, H is the percentage haematocrit/100, E is the total test count, b is the background c.p.m., S the spontaneous and M the maximal 51Crrelease. For each patient sample assayed, blood from a normal control person with known cell activity was tested in parallel as a control. Patients in whom the protocol was violated during operation were withdrawn from evaluation. Differences between groups and variables of patients with and without infection were tested using the x2 test or Student's t test. Results were considered significant at P
L. S. Jensen, A. J. Andersen*, P. M . Christiansent, P. Hokland$, C. 0. Juhl, G. Madsen”, J. Mortensen*, C. Meller-Nielsent, F. Hanberg-Serensen and M. Hoklandg University Departments of Surgical Gastroenterology, Aarhus County Hospital and *Aarhus Municipal Hospital, ?Department of Surgery, Randers County Hospital, $University Department of Medicine and Haematology, Aarhus County Hospital and $Department of Microbiology, Aarhus University, Denmark Correspondence to: Dr L. S. Jensen, University Department of Surgical Gastroenterology L, Aarhus
Countv HosDital. . . DK-8000 Aarhus. Denmark
Postoperative infection and natural killer cell function following blood transfusion in patients undergoing elective colorectal surgery Thefrequency of infection in 197patients undergoing elective colorectal surgery and having either no blood transfusion, transfusion with whole blood, or filtered blood free from leucocytes and platelets was investigated in aprospective randomized trial. Natural killer cellfunction was measured before operation and 3 , 7 and 30 days after surgery in 60 consecutive patients. Of the patients 104 required blood transfusion; 48 received filtered blood and 56 underwent whole blood transfusion. Postoperative infections developed in 13 patients transfused with whole blood (23 per cent, 9.5 per cent confidence interval 13-32 per cent), in one patient transfused with blood free f r o m leucocytes and platelets ( 2 per cent, 9.5 per cent confidence interval 0.05-11 per cent) and in two non-transfusedpatients ( 2per cent, 95 per cent conjidence interval 0.3-8per cent) ( P < 0.01). Natural killer cellfunction was signiJicantly ( P < 0.001) impaired up to 30 days after surgery in patients transfused with whole blood. These data provide a strong case against the use of whole blood transfusion in patients undergoing elective colorectal surgery.
Blood transfusion is associated with certain risks and during the past three or four decades attention has been drawn to its possible effects on the immune system. Blood transfusion produces immunosuppression and this may have beneficial clinical consequences, e.g. in renal transplantation, where transfusion-induced changes are associated with a decreased rate of allograft rejection’-3. Significant decreases in the rate of recurrence of Crohn’s disease have also been described after transfusion4. In other conditions the immunosuppressive action of blood transfusion may be dzleterious. Perioperative transfusion increases the recurrence rate and decreases survival rates in patients with c ~ l o r e c t a land ~ ~ ~breast’ cancer. These studies have been retrospective and it has not been proven that blood transfusion is the direct cause of these effects. Blood transfusion may be a significant factor affecting postoperative immune system function, leading to an increased susceptibility to i n f e ~ t i o n *and ~ ~ multiple organ failure”. Leucocytes may mediate the effects of cancer recurrence and allograft rejection” ;whole blood has a better effect than packed red cells on allograft survival”. The underlying mechanism for this is not clear but it has been demonstrated that whole blood transfusion causes reduced natural killer (NK ) cell function and a decreased T he1per:T suppressor lymphocyte ratio’ 3 . Whole blood amplifies the impairment in delayed hypersensitivity after operationI4. Blood free from white cells may be the safest preparation for preventing postoperative infection when transfusion is necessary. A prospective study was designed to compare the frequency of infection in patients undergoing elective colorectal surgery and either having no transfusion, or being randomized to transfusion with whole blood or blood free from white cells. Immune function was monitored in all patients from one of the three participating departments of surgery.
Patients and methods All patients admitted for elective colorectal surgery between August 1989 and September 1990 were considered for inclusion in the study,
~7-1323/92/060513-04 0 1992 Butterworth-Heinemann Ltd
which had been approved by the local ethical committee. The exclusion criteria were: age less than 18 years; acute colorectal surgery; and immunosuppressive treatment. After giving informed consent, patients in whom transfusion was indicated were randomly allocated using sealed envelopes to receive either whole blood or blood free from white cells. The latter was obtained by passing whole blood through a filtering device (Pall-filter; MEDA, Copenhagen, Denmark) thereby reducing the leucocyte and platelet content by 99.98 per cent. Red cells and plasma passed through the filter, which was mounted at the bedside and placed between the blood bag and the transfusion tube. All blood used was less than 3 weeks old. The indications for transfusion were a haematocrit before operation of less than 0.35 or a haemoglobin level below 7.2 mmoll- I . Intraoperative or postoperative transfusions were given according to the haemodynamic status of the patient. All patients received cefuroxime 3 g and metronidazole 1.5 g intravenously as one peroperative dose of antibiotic prophylaxis. Oral antibiotics were not given. The bowel was prepared with oral laxative and saline enema or Golytely (AAKH, Aarhus, Denmark). After surgery, abdominal wounds were closed with absorbable sutures for the peritoneum and fascia and non-absorbable sutures to the skin. Antibiotics were not used in the wounds; neither were subcutaneous drains. Follow-up examination included daily wound inspection by an independent assessor from the third day after operation until suture removal or patient discharge. All patients were followed in the outpatient clinic for 30 days after operation. Abdominal wound infection was defined as accumulation of pus with spontaneous discharge or requiring surgical drainage. Wounds with purulent discharge after the patient returned home were included in the infected group. Deep surgical infection consisted of: ( 1 ) intra-abdominal abscess defined as an intraperitoneal or pelvic collection of pus diagnosed by ultrasonography, laparotomy or spontaneous discharge; and ( 2) septicaemia defined as the presence of spiking fever associated with positive blood culture. The frequency of anastomotic leakage was diagnosed by fever and faecal fistula or peritonitis. Pneumonia was diagnosed by fever and infiltrate on chest radiography. Other surgical complications such as abdominal wound dehiscence, bleeding and ileus were considered separately. Blood samples for measurement of haematocrit, haemoglobin, serum albumin, immunoglobulin A (IgA), IgG, IgM and NK cell function were taken before operation and on days 3, 7 and 30 after operation. Cytomegalovirus (CMV) titre (IgG and IgM), Epstein-Barr
513
Infection following transfusion for colorectal surgery: L. S. Jensen et al.
virus (EBV) titre (IgG and IgM) and herpes virus titre were measured before operation and 30 days afterwards. NK cell function was evaluated in 60 consecutive patients from one of the participating departments of surgery by a modified whole blood assay as previously described15. A volume of 100 pl heparinized blood was added to varying numbers of 51Cr-radiolabelled K562 target cells. After a 4-h incubation at 37°C supernatants were counted for isotope release and cytotoxicity expressed as the percentage specific "Cr release according to the formula:
(E-b) x Percentage "Cr release =
(
V , - V, x H
V,
)-(S-b) x 100
( M x 0.8) - S
where V, is the total assay volume, V, is the volume of blood assayed, H is the percentage haematocrit/100, E is the total test count, b is the background c.p.m., S the spontaneous and M the maximal 51Crrelease. For each patient sample assayed, blood from a normal control person with known cell activity was tested in parallel as a control. Patients in whom the protocol was violated during operation were withdrawn from evaluation. Differences between groups and variables of patients with and without infection were tested using the x2 test or Student's t test. Results were considered significant at P
