Journal Logo

Clinical and Translational Research

Antibody Removal Before ABO-Incompatible Renal Transplantation: How Much Plasma Exchange Is Therapeutic?

Lawrence, Christopher1,3; Galliford, Jack W.1; Willicombe, Michelle K.1; McLean, Adam G.1; Lesabe, Mary1; Rowan, Fiona1; Papalois, Vassilios1; Regan, Fiona1,2; Taube, David1

Author Information
doi: 10.1097/TP.0b013e31823360cf


Renal transplantation is the optimal therapy for patients with end-stage renal disease (ESRD), but the waiting list for deceased donor transplantation continues to grow (1). The relative shortage of organs for transplantation from deceased donors has led to an increase in the number of live donor transplants being performed (2). As a result of anti-human leukocyte antigen (anti-HLA) sensitization or blood group (ABO) incompatibility, not all donor-recipient pairs are suitable for simple live donor transplantation. Data from the United States suggest that 30% of patients are sensitized with anti-HLA antibody, and that between two individuals, there is a 35% chance of ABO antibody incompatibility (3). Montgomery (4) calculates that there are 6000 prevalent patients on the waiting list in the United States and 3500 incident patients each year who have an otherwise willing and able but antibody-incompatible donor. Assuming similar donor-recipient characteristics, then approximately 1/10th of these numbers would be affected in the UK (in 2009, there were 6920 patients on the kidney alone waiting list in the UK [2] compared with 76,089 patients in the United States [5]).

The growth of the waiting list and the compelling evidence of the survival benefit of transplanted patients compared with wait-listed patients (6) has led a drive to increase the donor pool by consideration of more unconventional or high-risk techniques. These strategies include transplantation from HLA-incompatible (7) donors, which would normally be considered a barrier to transplantation (i.e., those against whom the recipient has a positive complement-dependent cytotoxic or flow crossmatch); ABO-incompatible (8) (ABOi) donors or novel strategies such as the paired and pooled matching scheme (9).

The largest worldwide experience of ABOi transplantation is in Japan (10) where sociocultural reasons limit the supply of deceased donor organs. The technique has gained in popularity in North America (11, 12), Australia (13), Europe (14), and the United Kingdom (15). Most ABOi transplantation is preceded by the start of an immunosuppressive regimen, usually incorporating a leukocyte depleting monoclonal antibody, and removal of preformed anti-ABO antibody, to a level considered safe, before transplantation. Many published regimens for antibody removal exist. The main methods used for physically removing circulating antibody are centrifugal plasma exchange (therapeutic plasma exchange [TPE]) (11, 12, 15), double filtration plasmapheresis (13), and use of immunoadsorption (IA) columns (14). The three techniques have never been studied in a head-to-head trial, and the choice of technique is likely to be strongly influenced by institutional preference and practical considerations (including cost).

Despite the increasing frequency of ABOi transplantation several questions remain unanswered, including which antibody removal regimen is superior; whether a cut-off antibody titer exists beyond which ABOi transplantation should not be attempted; at what antibody titer. it is safe to proceed to transplantation, and how much plasma exchange is required to achieve a safe antibody titer.

The Johns Hopkins group has reported that it is more difficult to remove antibody from patients with higher pre-TPE anti-ABO titers (16), and set out a schedule for the number of planned pre- and posttransplant TPE treatments based on pre-TPE IgG titer. However, it is not clear whether this schedule can be applied uniformly to other centers where immunosuppressive regimens or titer measurement methods may vary. Recently, Flint et al. (17) have reported the relationship they observed between plasma exchange and the starting anti-ABO titer. However, in their study, the cut-off titer for proceeding to transplantation was less than 1:32 (by tube method), and it is not clear whether a linear relationship persists if antibody removal is pursued to lower anti-ABO titers.

Frequently, centers report results of antibody removal programs only in patients who reach transplantation and exclude the outcomes of those that do not. Some patients enter antibody removal programs, and consequently receive augmented, and potentially, long-lasting immunosuppression without the guarantee of successful transplantation. Most ABOi programs use potent induction immunosuppression such as antithymocyte globulin (18, 19) or rituximab (20, 21) The combination of augmented immunosuppression, including biologic agents, and extensive TPE results in expensive treatment protocols.

Montgomery (4, 16) has argued that paired and pooled schemes should not be seen as being in competition with antibody removal programs but rather as complimentary systems to maximize successful donor-recipient pairs. It is important when selecting patients for antibody removal to be aware of the likelihood of progressing to transplantation so that the potential recipient and donor are able to give fully informed consent. In Montgomery's proposed scheme patients with an anti-ABO IgG titer more than 1:128 would not progress straight to antibody removal and ABOi transplantation but instead enter the paired and pooled scheme only proceeding to ABOi transplantation if no match is forthcoming from the paired matching run.

The purpose of this study was to review our ABOi live donor transplant program and assess the efficacy of our antibody removal regimen in achieving the cut-off titer for transplantation and to determine how much plasma exchange is required to achieve transplantation. Importantly we wished to define those patients in whom antibody removal was not sufficient to facilitate transplantation and to use this information to establish a cut-off titer above which patients should consider, in the first instance, an alternative route to transplantation such as preemptive referral to the paired and pooled scheme. The aim of this approach is to avoid unnecessary disappointment, exposure to potentially harmful, and long-lasting immunosuppression (22) and to reduce healthcare costs.


In fifty-one (91%) patients (21f:30m, mean age 49.3±10.9 years) the post-TPE IgG titer reached less than or equal to 1:4, and the transplant proceeded. In five patients, despite extensive TPE, the cut off was not reached, and transplantation did not go ahead. Outcomes are compared with 331 consecutive ABO compatible (ABOc) live donor kidney transplants performed during a similar period.

Patient and Graft Survival

Mean follow-up was 30.3±19 months (range 6–80 months). Patient survival in the ABOi cohort was 97.7% at 1 year, and 94.2% at 3 years compared with 98.9% at 1 year and 97.9% at 3 years in the ABOc cohort (log-rank P=ns). Two patients in the ABOi cohort have died, one after 10 months because of sepsis (graft failed) and one after 25 months because of pneumonitis (good allograft function). Allograft survival was 93.9% at 1 year, 90.9% at 2 and 3 years in the ABOi cohort (compare with 96.4% 1 year and 90.3% 3-year graft survival in the ABOc cohort, log-rank P=ns). Five grafts have been lost form the ABOi cohort, one as a result of technical failure (hemorrhage), one as a result of early and irreversible anti-HLA antibody-mediated rejection (AMR; <1 month), one related to immunosuppression withdrawal for overwhelming sepsis (8 months) in the patient who subsequently died, one as a result of chronic mixed cellular and humoral rejection (22 months), and one as a result of chronic transplant glomerulopathy (48 months).


Biopsy proven rejection-free survival (All Banff grades) in the ABOi cohort was 82.2% at 1 year and 65.5% at 3 years (compare with 82.1% at 1 year and 75.2% at 3 years in the ABOc cohort, log-rank P=ns). AMR-free survival in the ABOi cohort was 90.1% at 1 year and 82.1% at 2 and 3 years. There was a nonsignificant trend for patients with a preformed donor-specific antibody (DSA) to experience more AMR than those without a preformed anti-HLA DSA (log-rank P=0.19).

Allograft Function

Allograft function, as measured by serum creatinine, was 130.6±30.0 μmol/L at 1 year and 120.4±28 μmol/L at 3 years. Allograft function was not significantly different from contemporary ABOc transplant recipients (128.5±51.0 μmol/L at 1 year and 132.6±54.5 μmol/L at 3 years, both p=ns). There was no statistically significant difference in patient survival, allograft survival, rejection-free survival, or graft function between rituximab/daclizumab- and Campath-treated patients.

Preoperative TPE

The mean number of TPE actually delivered to achieve transplantation was 8.3±5. The number of TPE delivered before transplantation correlated closely with both the most recent pretreatment anti-ABO IgG titer (r2=0.51, P<0.01) and the highest ever IgG titer (r2=0.47, P<0.01). The correlation between the number of TPE required and IgG titer was not improved by correcting the number of plasma exchanges for body mass, nor for estimated plasma volume (using the Kaplan equation (23) to estimate the patient's plasma volume).

Twenty-one (41%) patients required TPE on the day of surgery because of antibody rebound. These patients tended to have higher starting anti-ABO titers (median 1:128).

We were able to retrospectively assess the minimum number of TPEs that could have been provided to obtain a post-TPE IgG titer less than or equal to 1:4 (calculated as the number of TPE performed when the anti-ABO titer first fell to less than or equal to 1:4). The minimum number of TPE required was 4±4.5, significantly fewer than the number of TPE actually administered (TTest, P<0.001). The best correlation observed was between the pretreatment IgG titer and the minimum number of TPE required (r2=0.85). There seems to be an exponential relationship between IgG titer and number of TPE required (Fig. 1). This graph allows us to predict the likely number of TPE required for any given IgG titer although there is increasing variation at high titers.

Exponential relationship between pre-treatment anti-ABO titer and the minimum number of TPE required to achieve a pre-operative titer of less than or equal to 1:4 (r2=0.85).

Patients given rituximab and daclizumab induction therapy required less preoperative TPE than patients receiving Campath 1H (median of 7 vs. 9 TPE, Wilcoxon p=ns) However, rituximab/daclizumab-treated patients had lower peak titers (median 64 vs. 16) and pre-TPE titers (median 64 vs. 16) than Campath-1H-treated patients.

Failure to Reach Transplantation

Five patients (8.9%) did not reach the cut-off IgG titer of less than or equal to 1:4 (rituximab/daclizumab 3, Campath-1H 2) despite receiving extensive TPE (15.2±6 vs. 8.3±5, TTest P<0.05) and treatment was abandoned. All five patients were ABO O with ABO A (four A1, one A2) potential donors (Fisher's exact test, P<0.05). The pretreatment IgG titers (Fig. 2) were 1:1024 in one patient, 1:512 in three patients, and 1:128 in one patient.

Distribution of pre-treatment anti-ABO titers of patients accepted for antibody removal in the ABOi program.

The likelihood of achieving successful antibody removal is therefore 49/50 (95.6%), if the IgG titer is less than or equal to 1:256 but only 2/6 (33.3%), if the IgG titer is more than or equal to 1:512 (χ2 27.6, P<0.0001). It is noteworthy that both patients successfully transplanted from titers of 1:512 had A2 donors.

Of the patients who did not achieve ABOi transplantation one has subsequently been transplanted through the paired and pooled scheme, two have been transplanted from well matched deceased donors and two await deceased donor transplantation. All three who were subsequently transplanted currently have excellent graft function and have not experienced a high incidence of complications, including immunosuppression related infections.


This is the first study using the relationship between the starting anti-ABO IgG titer and the number of TPE required to achieve ABOi renal transplantation to demonstrate a cut-off titer above which patients should not enter the ABOi program. Our model suggests that there is an exponential relationship between IgG titer and the number of plasma exchanges required to reach our target titer. This has allowed us to predict with a reasonable degree of accuracy, from the starting titer, how much TPE is likely to be required. This has the twin effects of saving resources by administering fewer TPEs than historically (we have reduced the minimum number of TPEs delivered from five to three) and of improving logistics as transplants are less likely to be delayed by underestimating the amount of TPE that may be required.

Conversely, we can predict which patients should not enter the ABOi program. As transplantation is only achieved in 33.3% of patients with anti-ABO titers more than or equal to 1:512, but 95.6% of patients with titers less than or equal to 1:256 we will now, ordinarily, only accept patients for ABOi transplantation if the IgG titer is less than or equal to 1:256. This has the effect of avoiding unnecessary potent immunosuppression in patients unlikely to reach transplantation and allows for more efficient use of resources.

It seems on initial inspection that patients treated with rituximab and daclizumab required less pretransplant TPE than those treated with Campath-1H. However, this must be interpreted with great caution, predominantly because the median ABO antibody titer is two dilutions higher in the Campath-treated group, who one would therefore expect to require more TPE. The two induction regimens are probably equally effective in facilitating antibody removal although Campath-1H is less costly than the combination of rituximab and daclizumab.

It is interesting that all five patients who did not achieve transplantation were ABO O with potential ABO A donors and although treatment success was not significantly different from those with ABO B donors, it may be that anti-A antibody is more difficult to remove than ABO B, possibly because ABO A glycoprotein is more antigenic than group B glycoprotein (24).

Many controversies relating to the provision of ABOi renal transplantation remain. It is not known, for example, at what titer it is prudent to proceed to transplantation. Centers performing ABOi transplantation vary in their cut-off titer for transplantation ranging between 1:4 and 1:16 and one unit (25) has reported using a fixed TPE protocol than having an antibody titer target (it has been suggested that antibody titer goal-directed TPE can unnecessarily complicate the timing of surgery).

ABO-incompatible renal transplantation in the UK is in its infancy and our conservative pretransplant IgG titer cut off, established at the commencement of this program in 2003, of less than or equal to 1:4 reflects this, but also is based on safety evidence (26)and demonstrates an understanding that the intermeasurement variability of antibody titer may vary by more than one dilution (27, 28). By using this cut off we have never experienced hyperacute rejection.

It is possible that the use of Glycorex-ABO IA columns (Glycorex Transplantation AB, Lund, Sweden) may have facilitated transplantation in the patients with high starting titers who did not reach a pretransplant titer of less than or equal to 1:4 with conventional TPE because of more complete antibody removal with each treatment (26). However, a recent article from Genberg et al. (29) demonstrates that IA columns are not universally successful in patients with high inter treatment antibody rebound or because of inadequate adsorption of core-chain dependent antibodies (the removal of which required conventional TPE).

This review of our results to date has enabled us to make safety and cost improvements to our ABOi program. The risk of exposing patients to a significant amount of long-lasting immunosuppression without the guarantee of successful transplantation is not undertaken lightly and patients with titers higher than 1:256 may be better served by another route to transplantation such as the paired and pooled scheme.


Patient Selection

The first ABOi transplant in our center was performed in May 2003 and up until February 2010 fifty-six patients have entered the program with 51 successfully achieving transplantation. All patients were assessed in the live donor clinic with selection criteria being the presence of dialysis dependent ESRD or a glomerular filtration rate less than 20 mL/min (and deteriorating) and the absence of a suitable ABOc live donor.

Recipient blood samples were sent to the NHS Blood & Transplant, Colindale, for assessment of anti-ABO antibody titers. Titers were performed using standard techniques using dithiothreitol-treated sera tested against A1 or B, non donor, red cells by the indirect antiglobulin technique, initially by tube technique and subsequently, after a validation exercise showing comparable results, using modified Diamed gel cards (28). As is customary, we report IgG anti-ABO antibody titers (11).

Patient Demographics

The mean age of patients was 50.3±11 years (range 24.7–72.1 years). Most patients were male (m 34: f 22). The cause of ESRD was adult polycystic kidney disease 15, diabetes mellitus 10, IgA nephropathy 6, other 18, and unknown 7.

The ABO antibody barrier to be overcome is shown in Table 1. The median antibody titer (both highest historical measurement and immediately before plasma exchange) was 1:64. The distribution of ABO antibody titer is shown in Figure 2. Ten patients (19.6%) also had a preformed DSA (class I: 1, class II: 6, and class I+II: 3). All patients had a negative T- and B-cell cytotoxic-dependent crossmatch and a negative T-cell flow crossmatch both immediately before transplant and historically. B-cell flow crossmatch was not performed because of low sensitivity and specificity.

Distribution of recipient-donor antibody combinations in transplanted patients


As previously reported (15), the first 25 patients received an immunosuppressive regimen consisting of rituximab (1 g on d-14 and 1 g on d0), daclizumab (2 mg/Kg on d0 and d14), tacrolimus (0.15 mg/kg/day titrated to 12-hr trough tacrolimus levels of 8–12 ng/mL), mycophenolate mofetil (750 mg two times per day [b.i.d.], titrated to 12-hr trough mycophenolic acid levels of 1.2–2.4 mg/L) with short course steroids (methylprednisolone 500 mg on d0 followed by prednisone 30 mg b.i.d. d0–3 and 30 mg one time per day [o.d.] d4–7 then stopped). The subsequent 31 patients received Campath 1H (30 mg on d-14 and 20 mg on d0), tacrolimus (0.15 mg/kg/day titrated to 12-hr trough tacrolimus levels of 8–12 ng/mL) with short course steroids (methylprednisolone 500 mg on d-14 and d0 followed by prednisone 30 mg b.i.d. d0–3 and 30 mg o.d. d4–7 then stopped).

Antibody Removal

All patients received plasma exchange before transplantation. TPE was performed with the COBE Spectra centrifugal blood separator and consisted of a standard 3 L exchange using acid citrate dextrose-A as anticoagulant. Plasma volume was replaced with 5% human albumin, or in the event of the fibrinogen falling to less than 1 g/L fresh frozen plasma (FFP) of the kidney donor's, ABO group was substituted. Patients received 100 mg/Kg replacement intravenous immunoglobulin (IVIg) after each TPE.

Blood samples for anti-ABO antibody titers were sent to the National Blood Service pre- and post-TPE and transplantation went ahead if the post-TPE IgG anti-ABO antibody titer was less than or equal to 1:4. If the target was not met then surgery was postponed, and TPE continued until the desired titer was achieved.

Although a post-TPE anti-ABO titer of less than or equal to 1:4 may be rapidly achieved the significance of antibody rebound (after TPE) is unclear. Many patients who achieve a titer of 1:4 or less after TPE on the day before surgery may have a much higher titer on the morning of surgery, it is our practice in the event of excessive antibody rebound to perform a final TPE (using FFP) on the morning of the surgery. There is much interpatient variability in the extent of antibody rebound, and it is clear that it is not entirely ameliorated by the administration of IVIg post-TPE.

The procedure was abandoned if the post-TPE IgG titer did not achieve less than or equal to 1:4 despite extensive TPE. After transplantation patients received two protocol TPE using FFP during postoperative d1–3, with replacement IVIg as above.


The authors thank the staff of the AUCHI dialysis unit who perform the TPE and the staff of the Red Cell Immunology Department of the NHS Blood & Transplant, North London Centre.


1. Schold J, Srinivas TR, Sehgal AR, et al. Half of kidney transplant candidates who are older than 60 years now placed on the waiting list will die before receiving a deceased-donor transplant. Clin J Am Soc Nephrol 2009; 4: 1239.
2. UK Transplant. Transplant activity in the United Kingdom 2009–2009, p 7.
3. Segev DL, Gentry DL, Warren DS, et al. Kidney paired donation and optimizing the use of live donor organs. JAMA 2005; 293: 1883.
4. Montgomery RA. Renal transplantation across HLA and ABO antibody barriers: Integrating paired donation into desensitization protocols. Am J Trans 2010; 10: 449.
5. HHS/HRSA/HSB/DOT. 2009 OPTN/SRTR Annual report 1999–2008. N2009. Rockville, MD.
6. Merion RM, Ashby VB, Wolfe RA, et al. Deceased donor characteristics and the survival benefit of kidney transplantation. JAMA 2005; 294: 2726.
7. Palmer A, Taube D, Welsh K, et al. Removal of anti-HLA antibodies by extracorporeal immunoadsorption to enable renal transplantation. Lancet 1989; 1: 10.
8. Nelson PW, Helling TS, Pierce GE, et al. Successful transplantation of blood group A2 kidneys in to non-A recipients. Transplantation 1988; 45: 316.
9. Johnson R, Allen JE, Fuggle SV, et al. UK Transplant NHSBT. Early experience of paired living donation in the United Kingdom. Transplantation 2008; 86: 1672.
10. Toma H, Tanabe K, Tokumoto K. Long term outcome of ABO- incompatible renal transplantation. Urol Clin North Am 2001; 28: 769.
11. Tobian AA, Shirey RS, Montgomery RA, et al. The critical role of plasmapheresis in ABO-incompatible renal transplantation. Transfusion 2008; 48: 2453.
12. Winters JL, Gloor JM, Pireda AA, et al. Plasma exchange conditioning for ABO-incompatible renal transplantation. J Clin Apheresis 2004; 19: 79.
13. Cohney SJ, Walker RG, Haeusler MN, et al. Blood group incompatibility in kidney transplantation: Definitely time to re-examine! Med J Aust 2007; 187: 306.
14. Tyden G. The European experience. Transplantation 2007; 84: S2.
15. Galliford J, Charif R, Chan KK, et al. ABO incompatible living renal transplantation with a steroid sparing protocol. Transplantation 2008; 86: 901.
16. Montgomery RA, Locke JE, King KE, et al. ABO Incompatible renal transplantation: A paradigm ready for broad implementation. Transplantation 2009; 87: 1246.
17. Flint SM, Walker RG, Hogan C, et al. Successful ABO-incompatible kidney transplantation with antibody removal and standard immunosuppression. Am J Transpl 2011; 11: 1016.
18. Thibaudin D, Alamartine E, Mariat C, et al. Long-term kinetics of T-lymphocyte subsets in kidney transplant recipients: Influence of anti-T cell antibodies and association with post transplant malignancies. Transplantation 2005; 80: 1514.
19. Sollinger H, Kaplan B, Pescovitz MD, et al. Basiliximab versus antithymocyte globulin for prevention of acute renal allograft rejection. Transplantation 2001; 72: 1915.
20. Onrust SV, Lamb HM, Balfour JA. Rituximab. Drugs 1999; 58: 78.
21. Habicht A, Broker V, Blume C, et al. Increase in infectious complications in ABO incompatible kidney transplant. Nephrol Dial Transplant [Epub ahead of print; doi: 10.1093/ndt/gfr215].
22. Scarsi M, Bossini N, Malacarne F, et al. The number of circulating recent thymic emigrants is severely reduced 1 year after a single dose of alemtuzumab in renal transplant recipients. Transpl Int 2010; 23: 786.
23. Kaplan AA. A simple and accurate method for prescribing plasma exchange. ASAIO Trans 1990; 36: M597.
24. Makidono R. Antigenicity of erythrocytes as analyzed in terms of their cross-reactivity. Microbiol Immunol 1979; 23: 791.
25. Sivakumaran P, Vo AA, Villicana R, et al. Therapeutic plasma exchange for desensitization prior to transplantation in ABO incompatible renal allografts. J Clin Apher 2009; 24: 155.
26. Rydberg L, Bengtsson A, Samuelson O, et al. In vitro assessment of a new ABO immunosorbent with synthetic carbohydrates attached to sepharose. Transpl Int 2005; 17: 666.
27. Cohney SJ, Hogan C, Haeusler M, et al. Variability of anti-blood group titres according to methodology and clinical relevance in ABO-incompatible renal transplantation. Am J Trans 2007; 7: S2 [A42].
28. Kumlien G, Wilpert J, Safwenberg J, et al. Comparing the tube and gel techniques for ABO antibody titration as performed in three European centres. Transplantation 2007; 84(12 Suppl): S17.
29. Genberg H, Kumlien G, Wennberg L, et al. The efficacy of antigen-specific immunoadsorption and rebound of anti-A/B antibodies in ABO-incompatible kidney transplantation. Nephrol Dial Transplant 2011;26:2394–2400.

ABO incompatible; Kidney transplantation; Plasma exchange

© 2011 Lippincott Williams & Wilkins, Inc.