CLINICAL RESEARCH www.jasn.org

Pretransplant Kinetics of Anti-HLA in Patients on the Waiting List for Kidney Transplantation

Matteo Togninalli,1,2 Daisuke Yoneoka ,1,2 Antonios G.A. Kolios,3 Karsten Borgwardt,1,2 and Jakob Nilsson 3

1Machine Learning and Computational Biology Laboratory, Department of Biosystems Science and Engineering, ETH Zurich, Zurich, Switzerland; 2SIB Swiss Institute of Bioinformatics, Lausanne, Switzerland; and 3Department of , University Hospital Zurich, Zurich, Switzerland

ABSTRACT Background Patients on organ transplant waiting lists are evaluated for preexisting alloimmunity to min- imize episodes of acute and chronic rejection by regularly monitoring for changes in alloimmune status. There are few studies on how alloimmunity changes over time in patients on kidney allograft waiting lists, and an apparent lack of research-based evidence supporting currently used monitoring intervals. Methods To investigate the dynamics of alloimmune responses directed at HLA , we retrospec- tively evaluated data on anti-HLA antibodies measured by the single- bead assay from 627 wai- tlisted patients who subsequently received a kidney transplant at University Hospital Zurich, Switzerland, between 2008 and 2017. Our analysis focused on a filtered dataset comprising 467 patients who had at least two assay measurements. Results Within the filtered dataset, we analyzed potential changes in mean fluorescence intensity values (reflecting bound anti-HLA antibodies) between consecutive measurements for individual patients in re- lation to the time interval between measurements. Using multiple approaches, we found no correlation between these two factors. However, when we stratified the dataset on the basis of documented previous immunizing events (transplant, pregnancy, or transfusion), we found significant differences in the magni- tude of change in alloimmune status, especially among patients with a previous transplant versus patients without such a history. Further efforts to cluster patients according to statistical properties related to alloimmune status kinetics were unsuccessful, indicating considerable complexity in individual variability. Conclusions Alloimmune kinetics in patients on a kidney transplant waiting list do not appear to be related to the interval between measurements, but are instead associated with alloimmunization history. This suggests that an individualized strategy for alloimmune status monitoring may be preferable to currently used intervals.

JASN 30: ccc–ccc, 2019. doi: https://doi.org/10.1681/ASN.2019060594

Tominimize episodes of acute and chronic rejection associated with substantially increased incidence of of a transplanted organ, patients on the organ trans- both these rejection types in previously published plant waiting list are regularly monitored for signs of preexisting alloimmunity. The information on alloimmunity is then incorporated into the organ Received June 12, 2019. Accepted August 19, 2019. allocation algorithm used, so that organs will not Published online ahead of print. Publication date available at be offered to potential recipients with preexisting www.jasn.org. fi alloimmunity to the speci c donor organ. Opti- Correspondence: Dr. Jakob Nilsson, Department of Immunol- mally, this will lead to reduced incidences of acute ogy, University Hospital Zurich, Gloriastrasse 23, CH-8091 Zurich, and chronic graft rejection because evidence of Switzerland. Email: [email protected] preexisting donor specific alloimmunity has been Copyright © 2019 by the American Society of Nephrology

JASN 30: ccc–ccc,2019 ISSN : 1046-6673/3011-ccc 1 CLINICAL RESEARCH www.jasn.org studies.1,2 As the majority of alloimmune responses in human Significance Statement organ transplantation are directed against the polymorphic HLA proteins, the immunologic monitoring is focused on Because the presence of pretransplant donor-specific anti-HLA assessing preexisting toward nonself HLAs. Assays antibodies is associated with increased organ rejection risk, patients to evaluate preexisting alloimmunity have been difficult on transplant waiting lists are regularly monitored for changes in their alloimmune status. In this retrospective analysis, the authors to develop, and thus the immunologic monitoring for alloim- investigated the dynamics of anti-HLA antibodies over time in pa- munity has focused on evaluating antibodies directed against tients on a kidney transplant waiting list. Their findings show that the nonself HLA proteins.3 This is assessed, with high sensitivity, kinetics of alloimmunity are highly individualized and do not appear by use of single-antigen bead (SAB) technology, where differ- to correlate with the interval between measurements. However, the fi ent HLA protein variants are immobilized on fluorescent magnitude of alloimmune status change increased signi cantly in patients with a previous transplant versus those without such a beads, so that one individual bead will only hold a single history. This suggests that an individualized strategy for alloimmune HLA antigen.4 reactivity to a specific SAB is assessed status monitoring of patients on organ transplant waiting lists on the by evaluating the mean fluorescence intensity (MFI) of the basis of their alloimmunization history might be preferable tocurrent bound anti-HLA antibodies. HLA typing of organ donors be- recommendations for regular monitoring. fore transplantation, in combination with pretransplant SAB analysis in the recipient, facilitates the assessment of donor- undergoing kidney transplantation at the University Hospital fi speci c antibodies (DSA), so that transplant pairs with pres- Zurich. ence of DSA can be avoided. Different clinical pretransplant SAB MFI cut-offs are used at different kidney transplant cen- ters, and studies suggest that an optimal cut-off for identifying METHODS patients with increased risk of rejection could be somewhere between 1000 and 2000 MFI.5 Patient Population Alloimmunity is a dynamic process and, as such, the alloim- All patients transplanted with a kidney between 2008 and 2017 mune status of an individual patient may change over time. at the University Hospital Zurich for whom a SAB analysis had Recognized alloimmunization events include blood transfu- been performed were included in the study. Ethical approval of sions, pregnancies, and organ transplantations, but other im- the study was given by the regional ethical review board. Clin- munologic events, such as vaccinations and changes in ongoing ical data were retrospectively collected and data on immuni- immunosuppressive therapies, may also have an influence on a zation events were available through the mandatory reporting patient’s alloimmune status.6–9 As the current alloimmune system for patients listed for organ transplantation within the status of the transplant recipient is central to the pretransplant Swiss Organ Allocation System. The data on previous immu- individualized immunologic risk stratification, patients on the nization included immunizing events occurring both before organ transplant waiting list are usually monitored for changes patients being listed and during the waiting time. in their alloimmune status at regular intervals, and after new alloimmunization events, using anti-HLA antibody detection Laboratory Analysis assays. Before the introduction of recombinant erythropoie- Presence of circulating anti–HLA-A,-B,-Cw,-DR,-DQ, tin, many patients with ESRD would regularly receive blood and -DP antibodies was analyzed in native serum using SAB transfusions and this likely influenced the praxis of alloim- assays (One Lambda, Inc., Canoga Park, CA) on a Luminex mune status monitoring, with many centers testing serum platform. Only serum samples collected before transplanta- from patients on the organ transplant waiting list every 3–4 tion, while the patients were actively listed on the kidney trans- months. A recommendation for alloimmune status monitor- plant waiting list, were analyzed. Upon being listed for kidney ing every 3 months is also included in the current guidelines transplantation, the patients were initially screened using the from the European Federation of Immunogenetics, whereas Luminex Mix Assay (One Lambda, Inc.) or investigated with the American Society for leaves it up to the the SAB assay, depending on previous history regulations within different transplant networks to decide and time period. For SAB- and Mix-negative patients, subse- upon the appropriate interval of alloimmune status monitor- quent monitoring was made using the Luminex Mix Assay. ing. There are few studies on how alloimmune responses For SAB-positive patients, subsequent monitoring was per- change over time in patients on the transplant waiting list, formed with SAB. For patients that had a positive Luminex and to our knowledge, there are no published studies support- Mix Assay on subsequent screenings, a new SAB analysis was ing the 3-month recommendation. The paucity of knowledge performed and if it was positive, all subsequent analyses were within this field was also highlighted in a recent publication made with SAB. Depending on the patients’ accrued waiting from the Sensitizing in Transplantation: Assessment of Risk time, the immunization status of the patient, and the presence (STAR) working group.10 of immunizing events, samples were analyzed with SAB at In this study, we aim to investigate the dynamics of different time intervals in an individualized manner. Data alloimmune responses, as assessed with the SAB assay, by on the occurrence of new immunizing events was obtained retrospectively examining values collected from patients continuously from the mandatory reporting system for

2 JASN JASN 30: ccc–ccc,2019 www.jasn.org CLINICAL RESEARCH patients listed for organ transplantation within the Swiss Or- 200daysafterthetransfusionwascalculatedforclass1 gan Allocation System. The analyses were also made in relation and class 2 separately. to the local transplant regulation in Switzerland that stipulates a 4-month interval for sampling of patients that Data Exploration, Statistical Analyses, and Clustering are active on the kidney transplant waiting list. Serums were Analysis analyzed without the addition of EDTA during the whole The data preprocessing and analysis was conducted using py- study; if a prozone effect was suspected (highly immunized thon and the following libraries: pandas, numpy, and scikit- patient), the serum was heat inactivated and analyzed again. learn. The figures were generated using the matplotlib and If a prozone effect was detected, all subsequent serums from seaborn libraries. The Welch t test was used to determine sig- the same patient were heat inactivated. HLA antibodies toward nificance for all tests to account both for differences in sample class 1 and class 2 HLA antigens were sometimes separately sizes and in variance of the distributions at hand. This choice analyzed at different time points, depending on the immuni- was motivated by the large sample sizes of the group at hand. zation status of the patient. For each SAB analysis, the maximum For the differences between time percentiles, a Bonferroni cor- MFI for an HLA antigen was defined as the highest-ranked rection was applied to account for multiple hypotheses testing. bead carrying the designated HLA antigen. During the study The scikit-learn library was used for the implementation of period, the protocol for SAB measurements in our laboratory clustering (Kmeans, DBSCAN [R1], and hierarchical cluster- remained the same, and two different lot numbers from the ing) and dimensionality reduction algorithms (multidimen- SAB kits were used (One Lambda, Inc.). We were unable to find sional scaling, t-SNE [R3], principal component analysis). any statistically significant change between serums analyzed with the different lots. Our SAB analysis is subject to annual Code Availability external and internal proficiency testing; for the external test- The code used to generate the figures is available at https:// ing, single-bead MFIs from ten serums are compared between github.com/BorgwardtLab/alloimmunity-kinetics. all six Swiss transplant laboratories. Assay agreement be- tween laboratories on MFI cut-offs (1000 MFI) is typically be- tween 90% and 99% and the coefficient of variation is typically RESULTS around 15%. For internal evaluation, MFI values from single beads are compared as the same serum is analyzed by different Dataset Structure laboratory staff at different time points, with intralaboratory Anti-HLA antibody specificities and corresponding MFIvalues coefficients of variation typically being around 7%. were consecutively collected for all patients who had under- gone at least one SAB analysis and were transplanted with a Calculation of MFI Change kidney during the follow-up period, at the University Hospital Changes in MFI over time were computed as the total abso- Zurich, Switzerland between 2008 and 2017. In total 627 pa- lute delta of the MFI values between two consecutive mea- tients fulfilled these criteria and were included into the study. surements, where all changes in MFI towards individual Additionally, because we focused on the temporal evolution of HLA antigens were compounded. All negative HLA antigens the MFI values over time, we excluded all patients where only a were given an arbitrary value of 499 MFI to allow for DMFI cal- single SAB measurement had been performed before trans- culations. To capture all dynamical changes in MFI values plantation. The filtered dataset, comprising patients who had at contained within the dataset, only measurements of MFI least two SAB analyses, contains 1881 time points; SAB mea- change between consecutive measurements in individual pa- surements were performed on a total of 467 patients, which tients were included. Mean absolute MFI change was calcu- equals an average of 4.03 measurement dates per patient. Ad- lated by dividing the total absolute MFI change on the number ditional clinical characteristics of included patients are presen- of positive HLA antigens detected within the dataset (80 for ted in Table 1. Analyses on both class 1 and class 2 anti-HLA class 1 and 43 for class 2). For analysis of MFI changes that antibodies were conducted, but as these measurements were crossed common clinical MFI cut-offs, each anti-HLA antigen sometimes individually performed, there are slightly more MFI change over a chosen cut-off (both in the positive and class 1 analyses in the dataset according to the results of the negative direction) between two consecutive measurements anti-HLA antibody screening (Table 1). As repeated SAB- was awarded a point. The points were then added from con- based monitoring was almost exclusively performed for pa- secutive measurements within the same time percentile and tients with detectable anti-HLA antibodies (.1000 MFI), the differences in points were analyzed between the different filtered dataset consists of a group of patients where 92.3% percentiles. Stratified analyses were performed by grouping had a peak anti-HLA antibody .1000 MFI (Table 1). This MFI changes crossing clinical cut-offs by the clinical cova- means that a large group of patients without evidence of pre- riates of the patients at hand. For analyses on the effect of existing alloimmunity (SAB negative) are not represented in transfusions on MFI change, the mean absolute MFI change the filtered dataset. In total, positive MFI values were found for occurring between an SAB measurement performed before SABs with 80 different HLA class 1 antigens and 43 different transfusion and an SAB measurement performed up to HLA class 2 antigens within the dataset.

JASN 30: ccc–ccc,2019 Pretransplant Anti-HLA Antibody Kinetics 3 CLINICAL RESEARCH www.jasn.org

Table 1. Clinical characteristics and anti-HLA antibody status No Previous Patients in the Age at Mean Previously Previous Previous Characteristic Immunizing Filtered Dataset Transplant Waiting Time Transplanted Pregnancy Transfusion Events Men 265 (56.7%) 49.2 yr 2.3 yr 70 — 60 135 Women 202 (43.3%) 50.0 yr 2.3 yr 45 98 21 38 Total 467 49.6 yr 2.3 yr 115 (24.6%) 98 (21.0%) 81 (17.3%) 173 (37.0%) HLA Class 1 HLA Class 2 Total A B Cw DR DQ DP Patients with SAB analysis 455 440 467 Total dates 1720 1651 3371 Total analyzed beads 166,840 156,845 323,685 Proportion of patients 48.6% 56.1% 30.4% 45.8% 40.9% 33.4% 92.3% with peak MFI.1000 Proportion of patients 19.3% 23.3% 7.7% 12.2% 15.2% 4.9% 38.1% with peak MFI.5000 Proportion of patients 10.1% 12.0% 3.6% 7.3% 10.1% 1.1% 22.3% with peak MFI.10,000

The interval between measurements (IBM) is a critically time for transplantation, but previous detection of anti-HLA important component in understanding how anti-HLA anti- antibodies and the occurrence of new immunizing events bodies evolve over time in individual patients, and is therefore a during the waiting time also influenced monitoring intervals. central part of this study. The dataset is very diverse in this Figure 1A shows the distribution of the individual SAB mea- regard because SAB analyses were performed at disparate time surements over time for all patients in the filtered dataset. We intervals, mostly depending on estimated additional waiting also analyzed peak MFI values for all detected anti-HLA

A B Locus ABCwDR DQ DP 25000 20000 15000

10000

5000 scale) g Patients Patients Peak Value (lo

1000

500

2008 2010 2012 2014 2016 2018 A B Cw DR DQ DP Year Locus

Figure 1. Characteristics of the investigated dataset. (A) Sampling timeline of the 467 patients in the filtered dataset, every line represents an individual patient and every dot represents a SAB measurement. The lines are colored according to the date of the first SAB measurement. (B) The peak MFI values for anti-HLA antibodies directed at the different HLA antigens, for every patient, are displayed on a single line and colored according to MFI value. The measurements are grouped by locus and patients are sorted by the sum of their total peak MFI value.

4 JASN JASN 30: ccc–ccc,2019 www.jasn.org CLINICAL RESEARCH

A B Time between measurements <98 days <98 days 98 days - 175 days 176 days - 225 days 226 days - 361 days 98 days - 175 days > 361 days

176 days - 225 days

226 days - 361 days Time between measurements

> 361 days

0 500 1000 1500 2000 2500 3000 3500 4000 0 500 1000 1500 2000 2500 3000 3500 4000 Mean absolute MFI change between measurements Mean absolute MFI change between measurements C D Time between measurements <97 days <97 days 97 days - 175 days 176 days - 225 days 226 days - 362 days 97 days - 175 days > 362 days

176 days - 225 days

226 days - 362 days Time between measurements

> 362 days

0 500 1000 1500 2000 2500 3000 3500 4000 0 500 1000 1500 2000 2500 3000 3500 4000 Mean absolute MFI change between measurements Mean absolute MFI change between measurements

Figure 2. MFI variability is, on average, not affected by the IBM. (A and C) Swarm plots depicting the mean of the absolute MFI change, between consecutive SAB measurements, for five different IBM categories (in days), for (A) class 1 and (C) class 2 anti-HLA antibodies. The IBM category is denoted by differently colored dots. (B and D) Mean and SD of the mean absolute MFI change for the five IBM categories. Each group contains the same number of measurements: (B) n=253 for class 1 and (D) n=242 for class 2. After correcting for multiple hypothesis testing, no pair of time differences showed a statistically significant mean absolute MFI change. antibodies from every single patient. We then plotted the the relation between IBM and changes in anti-HLA antibody peak MFI value signatures, according to locus, for every in- MFI in consecutive measurements, in individual patients. The cluded patient, ordering them by their total peak MFI value IBM in days was compared with the mean of absolute DMFI (Figure 1B). Visual inspection allows one to distinguish clus- values between measurements. The IBM categories were cho- ters of patients with a stronger signal for each of the different sen so that they contain the same number of measurements loci. There are two major clusters of patients, one with a (i.e., percentiles). In total, 1265 individual patient IBM were broader alloimmunization status showing positive MFI values analyzed for class 1 measurements and 1210 IBM for class 2. across many of the assayed HLA antigens, and one cluster of No clear pattern of time dependency emerged in the visual patients with negative peak MFI values across most loci. Ad- presentation of the data for either class 1 or class 2 (Figure 2, ditionally, antibodies toward HLA-A, -B, -DR, and -DQ anti- A and C). From looking at the data, it is clear that changes in gens are the most common within our dataset, as can also be total mean MFI occur within all IBM categories in a subgroup seen in Table 1. of patients, whereas the majority of measurements show very little mean absolute MFI change regardless of the IBM (Figure Time Dependency 2, B and D). For statistical validation, we ran two-sample To investigate whether changes in alloimmunization status are Welch tests on the five different percentiles to test the hypoth- influenced by the time between measurements, we investigated esis that the means of the total absolute MFI change are not

JASN 30: ccc–ccc,2019 Pretransplant Anti-HLA Antibody Kinetics 5 CLINICAL RESEARCH www.jasn.org

Class I Class II 4000

3500

3000

2500

2000

1500

between measurements 1000 Mean absolute MFI change 500

0

3 46 3 46 Time interval (months) Time interval (months)

Figure 3. MFI variability is not significantly different across three relevant clinical time intervals for class 1 and class 2 SAB mea- surements. The mean absolute MFI change between consecutive SAB measurements, in individual patients, performed at an interval shorter (in green) and longer (in blue) than the denoted time interval in months was compared. There were no statistically significant differences between measurements obtained with a shorter or longer time interval, for any of the investigated time intervals, either for class 1 (left) or class 2 (right) (P.0.06; see Supplemental Table 3 for details). different from each other in the five temporal categories. No Time Dependency of Changes Over Clinical Cut-Offs pair of time differences showed a significant total MFI change Decisions to accept or reject kidney offers for patients on the after a Bonferroni multiple hypothesis correction (a=0.005; transplant waiting list are influenced by the presence of DSA P$0.010 for class 1 and P$0.06 for class 2; see Supplemental in the recipient. This makes changes in MFI that cross com- Tables 1 and 2 for detailed results). To test whether the non- monly used clinical MFI cut-offs (in the positive or negative significant total MFI change was a product of our division of direction) important. To investigate whether the number of the data into five percentiles, we tested if there were significant crossings of clinical cut-offs was associated with IBM, we in- differences in total MFI change between measurements ob- vestigated crossings over the clinical cut-offs 1000 and 5000 tained before and after commonly used monitoring intervals. MFI. As can be seen in Figure 4, the pattern of crossings over In line with our previous findings, we could not find any sta- the 1000 MFI cut-off is comparable with the previous analysis tistically significant differences in total MFI change when we of total MFI change, with the number of crossings being sim- used 3, 4, and 6 months as time intervals (Figure 3, Supple- ilarly distributed over the different percentiles. In addition, no mentalTable3).Toinvestigateifthelackofsignificant pair of time difference showed a statistically significant change total MFI change between our investigated time points was in crossings after multiple hypothesis correction (a=0.005; influenced by individualized alloimmunity monitoring of P$0.007 for class 1 and P$0.009 for class 2; Supplemental immunized patients, we filtered our dataset by excluding all Tables 6 and 7). An analog analysis of crossings over the 5000 measurements from patients who encountered at least one MFI cut-off did not show any significant differences between documented immunizing event or who had measurements the percentiles (P$0.04; Figure 5, Supplemental Tables 8 and performed ,30 days apart. This resulted in a reduced dataset 9). Thus, changes in MFI over commonly used clinical MFI containing 82 patients with multiple class 1 measurements and cut-offs are not associated with the IBM in our dataset. 75 patients with multiple class 2 measurements. Mean absolute MFI change between measurements were naturally much Stratified Analyses smaller in terms of absolute values in these patients without a To investigate if we could find groups of patients with similar documented immunization event (Supplemental Figure 1). In- anti-HLA antibody kinetics we looked more closely at sub- terestingly, here as well, no pair of time differences showed a groups of patients according to clinical characteristics. We significant total MFI change after multiple hypothesis correc- used our previous approach of quantifying the alloimmune tion (a=0.005; P$0.08 for class 1 and P$0.45 for class 2; see status change over time, by analyzing the number of MFI Supplemental Tables 4 and 5). In conclusion, across all patients changes over the 1000 MFI cut-off between measurements. and in a subgroup without a documented immunization event, Stratifying the dataset by sex or age did not show a difference we were unable to find any significant correlation between IBM in the pattern of alloimmune status change (Supplemental and mean changes in anti-HLA antibody MFIs. Figure 2). However, when the measurements were stratified

6 JASN JASN 30: ccc–ccc,2019 www.jasn.org CLINICAL RESEARCH

A B Time between measurements <98 days <98 days 98 days - 175 days 176 days - 225 days 226 days - 361 days 98 days - 175 days > 361 days

176 days - 225 days

226 days - 361 days Time between measurements > 361 days

0 5 101520253035 0 5 101520253035 Number of threshold crossings Number of threshold crossings C D Time between measurements <97 days <97 days 97 days - 175 days 176 days - 225 days 226 days - 362 days 97 days - 175 days > 362 days

176 days - 225 days

226 days - 362 days Time between measurements > 362 days

0 5 101520253035 0 5 101520253035 Number of threshold crossings Number of threshold crossings

Figure 4. The IBM does not significantly affect the number of SAB specificities that cross the clinical cut-off of MFI 1000 between consecutive measurements. (A and C) Swarm plots of the number of SAB specificities, in individual patients, with MFI values that cross the 1000 MFI threshold between consecutive measurements for five different IBM categories (in days), for (A) class 1 and (C) class 2 HLA antigens. The IBM category is denoted by differently colored dots. (B and D) Individual swarm plots of the number of 1000 MFI threshold crossings for the five IBM categories. Each group contains the same number of measurements: (B) n=253 for class 1 and (D) n=242 for class 2. After correcting for multiple hypothesis testing, no pair of time differences showed a significantly different mean number of threshold crossings. on documented immunizing events, we found large differ- The quantity of alloimmune status change within the group ences in alloimmune status changes between these subgroups of patients without a documented alloimmunization event (Figure 6). As expected, patients with previous transplanta- was moderate for changes over the 1000 MFI cut-off, where tions showed the largest extent of alloimmune status change, a mean of 0.9 crossings per measurement were seen, compared but the change interestingly appeared similar over the differ- with a mean of 5.0 crossings in the previously transplanted ent investigated IBM (Figure 6). This was evident both in com- patients. Furthermore, 41.8% of the patients in the group parison with other documented immunizing events, such as without a documented immunizing event had at least one previous pregnancies or blood transfusions, and with patients measured crossing over the 1000 MFI for class 1 (38.8% without previous documented immunization events. The class 2) during their waiting time, as compared with 74.1% amount of alloimmune status change was also significantly in the group of previously transplanted patients (63.4% class increased in the two groups of patients with previous preg- 2) (Supplemental Tables 10 and 11). As large, positive MFI nancies or transfusion, as compared with the group without a movements over the different clinical cut-offs likely has more documented alloimmunization event (P,0.002; Figure 6). effect on organ allocation strategies than negative movements,

JASN 30: ccc–ccc,2019 Pretransplant Anti-HLA Antibody Kinetics 7 CLINICAL RESEARCH www.jasn.org

A B Time between measurements <98 days <98 days 98 days - 175 days 176 days - 225 days 226 days - 361 days 98 days - 175 days > 361 days

176 days - 225 days

226 days - 361 days Time between measurements > 361 days

0 5 101520253035 0 5 101520253035 Number of threshold crossings Number of threshold crossings C D Time between measurements <97 days <97 days 97 days - 175 days 176 days - 225 days 226 days - 362 days 97 days - 175 days > 362 days

176 days - 225 days

226 days - 362 days Time between measurements > 362 days

0 5 101520253035 0 5 101520253035 Number of threshold crossings Number of threshold crossings

Figure 5. The IBM does not significantly affect the number of SAB specificities that cross the clinical cut-off of MFI 5000 between consecutive measurements. (A and C) Swarm plots of the number of SAB specificities, in individual patients, with MFI values that cross the 5000 MFI threshold between consecutive measurements for five different IBM categories (in days), for (A) class 1 and (C) class 2 HLA antigens. The IBM category is denoted by differently colored dots. (B and D) Individual swarm plots of the number of 5000 MFI threshold crossings for the five IBM categories. Each group contains the same number of measurements: (B) n=253 for class 1 and (D) n=242 for class 2. After correcting for multiple hypothesis testing, no pair of time differences showed a significantly different mean number of threshold crossings. we also investigated this in our subgroups. Previously trans- within our dataset where MFI measurements had been per- planted patients show the largest percentage of patients with formed before and after (1–200 days) the transfusion event. anti-HLA antibodies that increase from ,1000 to .5000 MFI Total mean MFI change after transfusions for class 1 and class between measurements during the study period (9.5% for 2 antibodies for this subgroup is displayed in Supplemental class 1 and 6.7% for class 2), as compared with the other Figure 4. The effect of transfusions was greater for antibodies groups (previously pregnant 1.9% and 1.1%, previously trans- against class 1 than for class 2, with 16% of the evaluated fused 1.5% and 1.6%, and without a documented immuniza- patients showing a mean total MFI increase .500 MFI after tion event 0.3% and 0.3%, for class 1 and 2, respectively) transfusion for class 1 as compared with 6% for class 2. In sum- (Supplemental Figure 3, Supplemental Table 12). To investi- mary, our data show that the amount of alloimmunization gate how the alloimmune status changes after immunizing status change over time is strongly related to alloimmuniza- events in individual patients, we investigated the total MFI tion history. This suggests that alloimmunity-monitoring change associated with transfusions occurring during the schemes could be further individualized, and that immuniza- waiting time. We identified 37 transfusion events in 24 patients tion history is a relevant factor to include.

8 JASN JASN 30: ccc–ccc,2019 www.jasn.org CLINICAL RESEARCH

A Previously transplanted Previously pregnant Previously transfused No documented immunizing events

<98 days

98 - 175 days

176 - 225 days

226 - 361 days

> 361 days Time between measurements

0102030 40 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 Number of threshold crossings Number of threshold crossings Number of threshold crossings Number of threshold crossings

B Previously transplanted Previously pregnant Previously transfused No documented immunizing events

<97 days

97 - 175 days

176 - 225 days

226 - 362 days

> 362 days Time between measurements

0102030 40 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 Number of threshold crossings Number of threshold crossings Number of threshold crossings Number of threshold crossings

Figure 6. The alloimmunization history affects pretransplant MFI variability. Individual swarm plots of the number of MFI crossings over the 1000 MFI threshold for five IBM groups and across four immunization categories: patients who had a previous transplant (n=115 overall), patients who did not have a previous transplant but who had been pregnant (n=98 overall), patients who did not have a previous transplant nor were pregnant but received blood transfusions (n=81 overall), and patients without a documented immunizing event (n=173 overall). The number of threshold crossing for (A) class 1 and (B) class 2 anti-HLA antibodies are reported. Patients without 2 a documented immunizing event show significantly less threshold crossings as compared with the rest of the patients (P,3310 11).

Pattern of Individual Anti-HLA Antibody Kinetics not fixed. The main difficulty of comparing patients this way Although our stratified analysis using previous alloimmuniza- comes from two aspects: (1) the irregularly sampled nature of tion events was able to somewhat predict the likelihood the data and (2) the fact that patients do not have the same of alloimmune status change over time, there was a lot of number of time steps at which their alloimmune status was heterogeneity within the individual alloimmunized groups. probed. As there are no standard methods to cluster multi- Previously transplanted patients for instance showed the full channel, irregularly sampled time-series, we performed an spectrum of anti-HLA antibody status variability within all of analysis that considered summary statistics of the time-series our investigated temporal groups (Figure 6). This suggests that of the patients, such as mean, SD, mean rate of change in MFI, the kinetics of anti-HLA antibody change is individual and and SD of the rate of change in MFI for each antibody. This does not follow a typical pattern over time. We therefore stud- resulted in a list of 320 features for class 1 and 174 features for ied anti-HLA antibody kinetics plots of individual patients class 2, for each patient. These extracted features were then with multiple measurements. As can be seen by the examples used with three clustering algorithms: k-means, hierarchical in Figure 7, the pattern of change over time is indeed highly clustering, and DBSCAN.11 Clustering success was measured individual. Not only are the individual patterns of alloimmune using the silhouette coefficient12 and evaluated via visual in- status change different ,but it also appears that the trajectories spection of t-SNE plots, a dimensionality reduction technique of antibodies directed at different HLA antigens vary consid- used to efficiently visualize high-dimensional data.13 No clear erably in individual patients (Figure 7). This suggests that clusters could be obtained via these approaches, indicating other methods are needed to efficiently cluster patients on that summary information on individual anti-HLA antibody the basis of the kinetics of alloimmune status change. kinetics are not detailed enough to provide valid ways of split- ting the patients. Clustering We attempted to cluster patients according to the detailed temporal pattern of their MFI signal across time. We consid- DISCUSSION ered each patient as a multichannel, irregularly sampled time- series, meaning that we had N antibodies measurements at Monitoring of alloimmune status by use of modern SAB every time point and that time between measurements was technologies, in combination with improved HLA typing of

JASN 30: ccc–ccc,2019 Pretransplant Anti-HLA Antibody Kinetics 9 CLINICAL RESEARCH www.jasn.org

20000

15000

10000

5000

0

20000

15000

10000

5000

0

MFI 20000 A 15000 B Cw 10000 DR DQ 5000 DP

0

20000

15000

10000

5000

0 2008 2009 2010 2011 2012 2013 2014 2015 2016

Figure 7. Individual variability in anti-HLA antibody kinetics likely explains the absence of clear global trends. Selected timelines for four individual patients with .10 SAB measurements; only values above the detection cut-off (MFI .500) are plotted. Trajectories for anti-HLA antibodies to individual HLA antigens are plotted and colored according to their respective HLA locus specificity. The gray line indicates 5000 MFI and the black line indicates 1000 MFI. donors, has vastly improved our ability to predict the immu- significant differences in the alloimmune status change, as nologic risk associated with kidney transplantation for indi- measured by anti-HLA antibody total MFI change, between vidual donor recipient combinations.2,14,15 SAB analyses are, measurements performed at different IBM. To investigate however, both time consuming and costly, and it is therefore this further, we also looked at the difference using an alter- crucially important to optimize alloimmune status monitor- native set of time intervals, in patients without a documen- ing. Optimally, monitoring should be performed so that im- ted immunizing event and in MFI movement over clinically portant changes in alloimmune status are not missed before important MFI cut-offs. None of these approaches were able transplantation, while at the same time limiting the analysis to identify an IBM with more change in alloimmune status, of patient samples were no alloimmune status changes are as compared with all other investigated IBM. A likely expla- occurring. To design new strategies we first need to under- nation for this is that alloimmune responses are highly in- stand how allo-responses, as measured by the presence of anti- dividual and the pattern of change over time evolves in an HLA antibodies, are changing over time in patients waiting individual fashion. The example plots of anti-HLA antibody for kidney transplantation. Our study aimed to investigate kinetics in individual patients in our study are in line with these kinetics and, interestingly, we were unable to find any this assessment.

10 JASN JASN 30: ccc–ccc,2019 www.jasn.org CLINICAL RESEARCH

Another explanation for our inability to find significant organ transplantation was, as predicted, associated with an MFI differences between different IBM could be that our increased amount of detectable anti-HLA antibodies, but dataset includes short-interval measurements performed be- was also associated with a highly significant increase in al- cause of new immunizing events, with significant changes in loimmune status change over time. This was evident both in MFI resulting from these events. These short interval mea- comparison with other immunizing events and with patients surements would then counteract a trend toward more without previous documented immunization events. The changes in MFI over long IBM. However, comparison of amount of difference between these categories of patients measurements between later time points as well as analysis suggests that individualized alloimmune monitoring could, on other time intervals did not support a time dependency in part, be structured so that patients with previous trans- for MFI changes. Furthermore, when all patients with docu- plantations are monitored more frequently than patients in mented immunizing events or with measurements per- the other categories. Factors underlying increased MFI formed ,30 days apart were excluded, there was still no change in previously transplanted patients could be changes significant effect of the IBM on MFI change. This somewhat in ongoing immunosuppression, transplantectomies, or argues against a general bias in our dataset that could coun- other proinflammatory events.16 Because of the lack of de- teract an underlying true difference in MFI change over time. tailed information on these events, we were unfortunately However, we cannot exclude that our individualized SAB not able to analyze this further in the current cohort. We monitoring introduced confounding that might affect our were, however, able to investigate the effect of transfusions analyses and results. Another important issue for longitudi- on alloimmune status change in a subgroup of patients with a nal SAB analyses is assay variability, which could potentially documented transfusion during the waiting time. Our data both introduce additional MFI change between measure- underlines the importance of accurate monitoring of immu- mentsaswellascloudourabilitytodetectrealdifferences nization events, as 22% of the transfusions were associated between different IBMs. Our measured internal coefficients with marked changes in alloimmune status post-transfusion. of assay variability are, however, unlikely to substantially Our dataset primarily comprises patients with detectable alter our results, even for highly immunized patients (where anti-HLA antibodies (92.3% .1000 MFI peak) that were the effect on absolute DMFI measurements would be most followed with multiple SAB measurements throughout their significant), because measured DMFIs in our dataset mark- waiting time. This means that our findings cannot be ex- edly exceeded predicted effects of assay variability for a large tended to nonalloimmunized patients without detectable proportion of the patients with detected MFI change. It is anti-HLA antibodies, where clinical experience suggests important to note that although the effect of variability likely that very little change in alloimmunization status occur hadalimitedeffectonourresults, SAB assay variability still over time, in the absence of new immunizing events. affects our clinical ability to evaluate true alloimmune status Despite the large difference in alloimmune status change change over time, especially over low MFI clinical cut-offs, in between the different categories, there are also huge differences an individual patient. Our data does not include single-bead within the immunized subgroups in the pattern of MFI change resolution MFI values, but instead uses values from the high- over time. We explored several different clustering approaches, est-ranked bead carrying the designated HLA antigen, and on the basis of summary statistics, to generate groups of pa- this could potentially influence our ability to evaluate MFI tients with similar alloimmune status change kinetics, but change over time. To improve upon these problems, addi- were unable to efficiently cluster the patients in a meaningful tional studies using single-bead resolution data and a fixed way. This suggests that the individual nature of alloimmune short IBM, preferably in a large group of immunized pa- kinetics is extremely complex and cannot be easily categorized. tients, should be performed. The differentially spaced time points for alloimmune moni- One way to optimize alloimmune status monitoring would toring in the current dataset that allowed for the analysis of be to divide patients on the waiting list into groups with dif- interaction between IBM and MFI change, unfortunately also ferent monitoring schedules on the basis of clinical parameters make it difficult to design clustering algorithms. Further stud- that could predict future alloimmune status change. Indeed, ies to improve on these analyses are underway. the investigated dataset already contains numerous data points In conclusion, we were unable to find a significant corre- that were the result of attempts at individualized monitoring, lation between the change in alloimmune status, as measured on the basis of waiting time, immunizing events, and the by the amount of MFI change, and the time interval of SAB previous presence of anti-HLA antibodies. Despite this fact, measurements. We were, however, able to find large differ- many measurements were made that did not detect a signifi- ences in the magnitude of alloimmune status change between cant change in alloimmune status, and it is highly likely that subgroups, stratified according to alloimmunization history. many relevant alloimmune status changes were also missed Our study has implications for the standardized monitoring by the current approach. of patients on the transplant waiting list and suggests that a To investigate potential individualized monitoring strate- more individualized monitoring schedule, partially on the gies, we tried several ways of stratifying the patients on the basis of alloimmunization history, is preferable to current basis of different clinical parameters. A history of previous recommendations.

JASN 30: ccc–ccc,2019 Pretransplant Anti-HLA Antibody Kinetics 11 CLINICAL RESEARCH www.jasn.org

ACKNOWLEDGMENTS Supplemental Table 11. Percentage of patients with more than N crossings of the 1000, 5000, and 10,000 MFI threshold (class 2) Mr. Togninalli, Prof. Borgwardt, and Dr. Nilsson designed the pretransplant, stratified on documented previous immunizing study, Mr. Togninalli, Dr. Yoneoka, and Dr. Nilsson analyzed the events. data. Dr. Kolios provided critically important intellectual content. Supplemental Table 12. Percentage of patients with at least one Mr. Togninalli, Dr. Nilsson, Dr. Kolios, and Prof. Borgwardt drafted SAB specificity showing increase in MFI from ,1000 to .5000 the manuscript. All authors approved the final version of the sub- MFI pretransplant, stratified on documented previous immunizing mitted manuscript and are accountable for all aspects of the work. events. Supplemental Figure 1. MFI variability is, on average, not affected by the interval between measurements (IBM), in patients without DISCLOSURES documented immunizing events. When removing measurements made less than 30 days apart and patients with a documented im- Prof. Borgwardt reports personal fees from mentoring and consulting for munizing events, no significant correlation between MFI variability Roche Pharmaceutical Research and Early Development (pRED), Basel, out- and time between measurements can be seen for class I (A,B) and side the submitted work. Dr. Nilsson reports personal fees from One Lambda class II (C,D). Each IBM contains the same number of measure- Inc., outside the submitted work. All of the remaining authors have nothing to disclose. ments (n=37 for class I, 36 for class II). Supplemental Figure 2. Gender or age does not affect pre-transplant MFI variability. Individual swarm plots of the number of MFI SUPPLEMENTAL MATERIAL crossings over the 1000 MFI threshold for five IBM groups stratified on sex (above) and age (below). For the analyses on the impact of This article contains the following supplemental material online at patient age on pre-transplant MFI variability, the data set was split http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2019060594/-/ into three age groups (below 30 years, 30–60 years, and above 60 DCSupplemental. years). There were no significant differences in the number of Supplemental Table 1. Detailed P-values for percentile pairwise crossings across either of these stratifications. Only class I mea- tests, class 1. surements are shown but similar observations were also made for Supplemental Table 2. Detailed P-values for percentile pairwise class II measurements (data not shown). tests, class 2. Supplemental Figure 3. The alloimmunization history affects pre- Supplemental Table 3. P-values for Welch t-test statistics between transplant MFI variability. Individual swarm plots of the number of groups of measurements across important clinical time intervals (3, MFI crossings over the 5000 MFI threshold for five IBM groups and 4, and 6 months). The analysis was performed both on the entire across four immunization categories. Patients who had a previous filtered dataset (marked with immunized) and on a subgroup of transplant (n=115 overall), patients who did not have a previous patients without a documented immunizing event (marked without transplant but who had been pregnant (n=98 overall), patients immunized) (as denoted in Supplemental Figure 1). who did not have a previous transplant nor were pregnant but re- Supplemental Table 4. Detailed values for percentile pairwise tests ceived blood transfusions (n=81 overall) and patients without a without immunized patients and measurements made closer than 30 documented immunizing event (n=173 overall). The number of days apart, class 1. threshold crossing for class I (A) and class II (B) anti HLA-antibodies Supplemental Table 5. Detailed values for percentile pairwise tests are reported. without immunized patients and measurements made closer than 30 Supplemental Figure 4. Mean total MFI change after transfu- days apart, class 2. sions occurring during the waiting time. Individual swarm plots Supplemental Table 6. Detailed P-values for percentile pairwise showing the mean total MFI change after transfusion for class I tests over the number of crossings over the clinical cut-off of (blue), and class II (green) measurements. Only measurements MFI=1000, class 1. performed . 1to, 200 days from the transfusion were included Supplemental Table 7. Detailed P-values for percentile pairwise in the analyses. tests over the number of crossings over the clinical cut-off of MFI=1000, class 2. REFERENCES Supplemental Table 8. Detailed P-values for percentile pairwise tests over the number of crossings over the clinical cut-off of 1. Lefaucheur C, Loupy A, Hill GS, Andrade J, Nochy D, Antoine C, et al.: MFI=5000, class 1. Preexisting donor-specific HLA antibodies predict outcome in kidney P Supplemental Table 9. Detailed -values for percentile pairwise transplantation. J Am Soc Nephrol 21: 1398–1406, 2010 tests over the number of crossings over the clinical cut-off of 2. Kamburova EG, Wisse BW, Joosten I, Allebes WA, van der Meer A, MFI=5000, class 2. Hilbrands LB, et al.: Differential effects of donor-specific HLA anti- Supplemental Table 10. Percentage of patients with more than bodies in living versus deceased donor transplant. Am J Transplant 18: – N 2274 2284, 2018 crossings of the 1000, 5000, and 10,000 MFI threshold (class 3. Young JS, McIntosh C, Alegre ML, Chong AS: Evolving approaches in fi 1) pretransplant, strati ed on documented previous immunizing the identification of allograft-reactive T and B cells in mice and humans. events. Transplantation 101: 2671–2681, 2017

12 JASN JASN 30: ccc–ccc,2019 www.jasn.org CLINICAL RESEARCH

4. Tait BD: Detection of HLA antibodies in organ transplant recipients - 2017 working group meeting report. Am J Transplant 18: 1604– triumphs and challenges of the solid phase bead assay. Front Immunol 1614, 2018 7: 570, 2016 11. Ester M, Kriegel HP, Sander J, Xu X: A density-based algorithm for 5. Reed EF, Rao P, Zhang Z, Gebel H, Bray RA, Guleria I, et al.: Compre- discovering clusters in large spatial databases with noise. Presented at hensive assessment and standardization of solid phase multiplex-bead the 2nd International Conference on Knowledge Discovery and Data arrays for the detection of antibodies to HLA. Am J Transplant 13: Mining (KDD ’96), Portland, OR, 1996 1859–1870, 2013 12. Rousseeuw PJ: Silhouettes: A graphical aid to the interpretation and 6. Akgul SU, Ciftci HS, Temurhan S, Caliskan Y, Bayraktar A, TefikT, validation of cluster analysis. J Comput Appl Math 20: 53–65, 1987 et al.: Association between HLA antibodies and different sensitiza- 13. Maaten Lvd, Hinton GE: Visualizing data using t-SNE. JMachLearnRes tion events in renal transplant candidates. Transplant Proc 49: 425– 9: 2579–2605, 2008 429, 2017 14. Wehmeier C, Hönger G, Cun H, Amico P, Hirt‐Minkowski P, Georgalis 7. Resse M, Paolillo R, Pellegrino Minucci B, Costa D, Fiorito C, A, et al.: Donor specificity but not broadness of sensitization is associ- Santangelo M, et al.: Effect of single sensitization event on human ated with antibody-mediated rejection and graft loss in renal allograft leukocyte antigen alloimmunization in kidney transplant candidates: recipients. Am J Transplant 17: 2092–2102, 2017 A single-center experience. Exp Clin Transplant 16: 44–49, 2018 15. Amico P, Hirt-Minkowski P, Hönger G, Gürke L, Mihatsch MJ, Steiger J, 8. Katerinis I, Hadaya K, Duquesnoy R, Ferrari-Lacraz S, Meier S, van et al.: Risk stratification by the virtual crossmatch: A prospective study in Delden C, et al.: De novo anti-HLA antibody after pandemic H1N1 and 233 renal transplantations. 24: 560–569, 2011 seasonal influenza immunization in kidney transplant recipients. Am J 16. Locke JE, Zachary AA, Warren DS, Segev DL, Houp JA, Montgomery Transplant 11: 1727–1733, 2011 RA, et al.: Proinflammatory events are associated with significant in- 9. Hricik DE, Formica RN, Nickerson P, Rush D, Fairchild RL, Poggio ED, creases in breadth and strength of HLA-specific antibody. Am J et al.: Clinical Trials in Organ Transplantation-09 Consortium: Adverse Transplant 9: 2136–2139, 2009 outcomes of tacrolimus withdrawal in immune-quiescent kidney transplant recipients. JAmSocNephrol26: 3114–3122, 2015 10. Tambur AR, Campbell P, Claas FH, Feng S, Gebel HM, Jackson AM, See related editorial, “Tracking HLA Antibody Changes among Kidney Waitlist et al.: Sensitization in transplantation: Assessment of risk (STAR) Candidates: One Protocol May Not Fit All,” on pages XXX–XXX.

JASN 30: ccc–ccc,2019 Pretransplant Anti-HLA Antibody Kinetics 13 Supplemental information table of contents

Table S1. Detailed p-values for percentile pairwise tests. Class I.

Table S2. Detailed p-values for percentile pairwise tests. Class II.

Table S3. P-values for Welch’s t-test statistics between groups of measurements across important clinical time intervals (3, 4 and 6 months).

Table S4. Detailed values for percentile pairwise tests without immunized patients and measurements made closer than 30 days apart. Class I.

Table S5. Detailed values for percentile pairwise tests without immunized patients and measurements made closer than 30 days apart. Class II.

Table S6. Detailed p-values for percentile pairwise tests over the number of crossings over the clinical cut-off MFI=1000. Class I.

Table S7. Detailed p-values for percentile pairwise tests over the number of crossings over the clinical cut-off MFI=1000. Class II.

Table S8. Detailed p-values for percentile pairwise tests over the number of crossings over the clinical cut-off MFI=5000. Class I.

Table S9. Detailed p-values for percentile pairwise tests over the number of crossings over the clinical cut-off MFI=5000. Class II.

Table S10. Percentage of patients with more than N crossings of the 1000, 5000 and 10 000 MFI threshold (class I) pre-transplant, stratified on documented previous immunizing events.

Table S11. Percentage of patients with more than N crossings of the 1000, 5000 and 10 000 MFI threshold (class II) pre-transplant, stratified on documented previous immunizing events.

Table S12. Percentage of patients with at least one SAB specificity showing increase in MFI from <1000 MFI to >5000 MFI pre-transplant, stratified on documented previous immunizing events. Figure S1. MFI variability is, on average, not affected by the interval between measurements (IBM) in patients without documented immunizing events.

A B

C D Figure S2. Gender or age does not affect pre-transplant MFI variability. Figure S3. The alloimmunization history affects pre-transplant MFI variability.

A

B Figure S4. Mean total MFI change after transfusions occuring during the waiting time. Table S1. Detailed p-values for percentile pairwise tests. Class I.

<98 days 98 days - 175 176 days - 225 226 days - > 361 days days: days 361 days <98 days 0.00954 0.16152 0.07960 0.91378 98 days - 175 0.26924 0.46471 0.01884 days: 176 days - 225 0.72639 0.21631 days 226 days - 0.11580 361 days > 361 days

Table S2. Detailed p-values for percentile pairwise tests. Class II.

<97 days 97 days - 175 176 days - 225 226 days - > 362 days days: days 362 days <97 days 0.71681 0.31121 0.89956 0.27325 97 days - 175 0.50188 0.67425 0.16082 days: 176 days - 225 0.34266 0.05541 days 226 days - 0.40267 362 days > 362 days

Table S3. P-values for Welch’s t-test statistics between groups of measurements across important clinical time intervals (3, 4 and 6 months). Class I Class II

With immunized Without immunized With immunized Without immunized

3 months 0.06373 0.94191 0.77056 0.53905

4 months 0.27305 0.88502 0.59098 0.46024

6 months 0.71886 0.89899 0.49940 0.71302

Table S4. Detailed values for percentile pairwise tests without immunized patients and measurements made closer than 30 days apart. Class I. <140 days 140 days - 187 188 days - 255 256 days - 365 > 365 days days: days days <140 days 0.67765 0.08365 0.47584 0.62537 140 days - 187 0.23153 0.80525 0.93078 days: 188 days - 255 0.24689 0.31032 days 256 days - 365 0.88932 days > 365 days Table S5. Detailed values for percentile pairwise tests without immunized patients and measurements made closer than 30 days apart. Class II. <133 days 133 days - 187 188 days - 265 266 days - 365 > 365 days days: days days <133 days 0.44897 0.53719 0.92756 0.98960 133 days - 187 days: 0.91364 0.53248 0.61052 188 days - 265 days 0.59027 0.66522 266 days - 365 0.95064 days > 365 days

Table S6. Detailed p-values for percentile pairwise tests over the number of crossings over the clinical cut-off MFI=1000. Class I. <98 days 98 days - 175 176 days - 225 226 days - > 361 days days: days 361 days <98 days 0.12826 0.21375 0.47844 0.13927 98 days - 175 0.86042 0.50212 0.00704 days: 176 days - 225 0.63728 0.01453 days 226 days - 0.04659 361 days > 361 days Table S7. Detailed p-values for percentile pairwise tests over the number of crossings over the clinical cut-off MFI=1000. Class II. <97 days 97 days - 175 176 days - 225 226 days - > 362 days days: days 362 days <97 days 0.73094 0.95973 0.82772 0.02806 97 days - 175 0.74687 0.91505 0.00871 days: 176 days - 225 0.85289 0.01742 days 226 days - 0.01747 362 days > 362 days

Table S8. Detailed p-values for percentile pairwise tests over the number of crossings over the clinical cut-off MFI=5000. Class I. <98 days 98 days - 175 176 days - 225 226 days - > 361 days days: days 361 days <98 days 0.04489 0.14690 0.27112 0.87147 98 days - 175 0.59026 0.43918 0.04146 days: 176 days - 225 0.78560 0.12664 days 226 days - 0.23028 361 days > 361 days

Table S9. Detailed p-values for percentile pairwise tests over the number of crossings over the clinical cut-off MFI=5000. Class II. <97 days 97 days - 175 176 days - 225 226 days - > 362 days days: days 362 days <97 days 0.71673 0.11055 0.97677 0.65514 97 days - 175 0.22686 0.72822 0.44599 days: 176 days - 225 0.15655 0.06545 days 226 days - 0.70667 362 days > 362 days Table S10. Percentage of patients with more than N crossings of the 1000, 5000 and 10 000 MFI threshold (class I) pre-transplant, stratified on documented previous immunizing events. 1000 MFI 5000 MFI 10 000 MFI

N crossings N crossings N crossings N crossings N crossings N crossings > 0 > 1 > 0 > 1 > 0 > 1

None of the 41.79 % 19.40 % 10.15 % 2.69 % 1.19 % 0.30 % previous

Previously 59.85 % 34.47 % 28.79 % 18.18 % 13.64 % 8.33 % pregnant

Previously 51.44 % 27.88 % 13.46 % 7.69 % 4.33 % 2.40 % transfused

Previously 74.11 % 59.29 % 50.66 % 39.38 % 34.73 % 26.33 % transplanted

Table S11. Percentage of patients with more than N crossings of the 1000, 5000 and 10 000 MFI threshold (class II) pre-transplant, stratified on documented previous immunizing events. 1000 MFI 5000 MFI 10 000 MFI

N crossings N crossings N crossings N crossings N crossings N crossings > 0 > 1 > 0 > 1 > 0 > 1

None of the 38.80 % 20.07 % 4.35 % 1.00 % 0.33 % 0.33 % previous

Previously 47.81 % 27.89 % 12.65 % 6.77 % 7.57 % 3.59 % pregnant

Previously 42.11 % 21.58 % 8.95 % 5.79 % 1.05 % 1.05 % transfused

Previously 63.42 % 41.12 % 39.61 % 22.73 % 32.25 % 18.40 % transplanted

Table S12. Percentage of patients with at least one SAB specificity showing increase in MFI from <1000 MFI to >5000 MFI pre-transplant, stratified on documented previous immunizing events. Class I Class II

None of the previous 0.30 % 0.33 %

Previously pregnant 1.52 % 1.59 %

Previously transfused 1.92 % 1.05 %

Previously transplanted 9.51 % 6.71 %