American Journal of Transplantation 2015; 15: 1632–1643
Wiley Periodicals Inc.


C

Copyright 2015 The American Society of Transplantation
and the American Society of Transplant Surgeons
doi: 10.1111/ajt.13162

Early Graft Loss After Kidney Transplantation:
Risk Factors and Consequences
M. O. Hamed1, Y. Chen2, L. Pasea3,
C. J. Watson1, N. Torpey4, J. A. Bradley1,
G. Pettigrew1 and K. Saeb-Parsy1,*
1

Department of Surgery, University of Cambridge, and
NIHR Cambridge Biomedical Research Centre,
Cambridge, UK
2
Department of Pure Mathematics and Mathematical
Statistics, Cambridge, UK
3
Centre for Applied Medical Statistics, University of
Cambridge, Cambridge, UK
4
Department of Renal Medicine, Addenbrooke’s Hospital,
Cambridge, UK


Corresponding author: Kourosh Saeb-Parsy,

Early graft loss (EGL) after kidney transplantation is
a catastrophic outcome that is assumed to be more
likely after the use of kidneys from suboptimal
donors. We therefore examined its incidence, risk
factors and consequences in our center in relation to
different donor types. Of 801 recipients who received
a kidney-only transplant from deceased donors, 50
(6.2%) suffered EGL within 30 days of transplantation. Significant risks factors for EGL were donation
after circulatory death (DCD) (odds ratio [OR] 2.88;
p ¼ 0.006), expanded criteria donor (ECD) transplantation (OR 4.22; p ¼ 0.010), donor age (OR 1.03;
p ¼ 0.044) and recipient past history of thrombosis
(OR 4.91; p ¼ 0.001). Recipients with EGL had 12.28
times increased risk of death within the first year,
but long-term survival was worse for patients
remaining on the waiting list. In comparison with
patients on the waiting list but not transplanted, and
with all patients on the waiting list, the risk of death
after EGL decreased to baseline 4 and 23 months
after transplantation, respectively. Our findings
suggest that DCD and ECD transplantation are
significant risk factors for EGL, which is a major
risk factor for recipient death. However, long-term
mortality is even greater for those remaining on the
waiting list.

Abbreviations: ATG, antithymocyte globulin; ATN,
acute tubular necrosis; BMI, body mass index; CIT,
cold ischaemia time; DBD, donation after brain death;

DCD, donation after circulatory death; DGF, delayed
graft function; ECD, expanded criteria donor; EGL, early
graft loss; ESRD, end-stage renal disease; HLA, human
1632

leukocyte antigen; PNF, primary non-function; SCD,
standard criteria donor
Received 26 September 2014, revised 12 December
2014 and accepted for publication 14 December 2014

Introduction
For selected patients with end-stage renal disease (ESRD),
kidney transplantation is associated with improved quality
of life, longer patient survival and reduced costs compared
to remaining on dialysis (1,2). Kidney transplant outcomes
have improved incrementally over the last two decades (3),
attributed to better surgical and medical care, improved
immunosuppressive regimens and Human Leucocyte
Antigen (HLA) testing and matching of recipients, HLAantibody screening, and better prophylaxis and management of infections and other complications (4). However,
long waiting lists and the shortage of optimal donors have
necessitated the use of kidneys from increasingly more
sub-optimal donors (5). The use of kidneys from suboptimal
donors has been shown to be associated with worse longterm graft outcomes in some studies (6,7) and there is a
perception that the use of suboptimal kidneys may also
increase the risk of early graft loss (EGL). However, the
assertion that the use of suboptimal kidneys increases the
risk of EGL after kidney transplantation has not been
definitively examined.
EGL, defined as graft loss occurring within 30 days after
kidney transplantation, is relatively uncommon and

occurs in approximately 5% of kidney transplants (3).
However, it is a physically and emotionally devastating
outcome for both the recipient and the transplant team:
The recipients are exposed to the risks of medical and
surgical postoperative complications, compounded by a
period of immunosuppression and with remaining on, or
returning rapidly to, dialysis. EGL may also lead to
sensitisation to HLA, reducing the likelihood and/or
success of re-transplantation (8,9). Although hyperacute
or accelerated acute rejections are occasionally responsible for EGL, the most common causes are nonimmunological; vascular thrombosis accounts for up to
one third of early kidney transplant loss (10,11). Primary
non-function (PNF) is also an important and potentially
avoidable cause of EGL that may reflect the quality of
the donor organ.


Early Kidney Graft Loss

Clinicians and patients offered a kidney from a suboptimal
donor, therefore, face a difficult and poorly understood
choice: should the patient be transplanted with that kidney
and face an increased risk of EGL, or remain on the waiting
list for a better offer while continuing to endure the risks
associated with dialysis? Informed judgment about this
difficult choice requires a detailed analysis of the incidence,
risk factors and consequences of EGL. It is important to
perform this analysis in the context of the increasing use of
kidneys from donation after circulatory death (DCD) and
expanded criteria (ECD) donors compared to donation after
brain death (DBD) donors and standard criteria (SCD)

donors.
In this study, we report our single center experience of EGL
following renal transplantation and examine the associated
risk factors and consequences. Because a high proportion
of renal transplants undertaken at our centre are from DCD
and ECD donors, we were able to systematically examine
the risk factors for EGL amongst the different deceased
donor types.

Materials and Methods
Recipients of deceased donor kidney-only transplants between January 2002 and April 2012 in our center who suffered graft loss within 30 days
after transplantation were identified from a prospectively maintained
database. Patient demographics and medical history were obtained by
review of the database, pre- and post-transplant communication letters
(including with referral centers), discharge summaries relating to hospital
admissions, and by review of case notes where necessary. Waiting list
patients included active and suspended patients, as well as highly
sensitized patients, in our center only. Factors associated with EGL
were determined by univariate and multivariate analysis in comparison to
the cohort whose grafts survived beyond 30 days. EGL was defined as graft
nephrectomy or loss of kidney transplant function resulting in the recipient
becoming dialysis dependent within 30 days of kidney transplantation (and
never achieving graft function thereafter) or death with a non-functioning
graft within 30 days. Death with a functioning graft (i.e. no requirement for
dialysis) was not included as EGL. Recipients of double kidney transplants
who lost only one graft and subsequently did not require dialysis were not
included in the EGL group.
Primary non-function (PNF), defined as permanent lack of graft function
from the time of transplantation, was diagnosed when a kidney graft was
well perfused (confirmed by ultrasound examination) but never functioned, necessitating dialysis after kidney transplantation; it was

confirmed by at least one transplant biopsy to exclude other causes of
non-function. The diagnosis of acute vascular occlusion was suggested by
Duplex ultrasound examination and confirmed intra-operatively and/or by
histological analysis of the graft following transplant nephrectomy. Graft
loss secondary to acute rejection was defined as acute irreversible
deterioration of graft function not responding to immunosuppressive
treatment requiring renal replacement therapy within 30 days after
transplant and confirmed by biopsy. Further details on DCD and ECD
donors, the transplant operation and immunosuppressive regimens are
included in the supplementary materials.
Mann–Whitney rank sum tests and Pearson’s Chi-squared tests were used
to compare continuous and categorical variables respectively between the

American Journal of Transplantation 2015; 15: 1632–1643

two groups (graft survival 30 days and >30 days). Multivariate
logistic regression models were used to calculate adjusted estimate
effects of the variables on risk of EGL, and to assess the impact of EGL
on the short-term patient survival (defined as death within the first year
after kidney transplantation). A Cox proportional hazards model was
used to estimate an adjusted hazard ratio for long-term patient survival
(defined as survival for at least 1 year after kidney transplantation).
Kaplan–Meier plots and Greenwood’s formula were used to analyze
graft and patient survival between groups until all-cause mortality or
the end of follow-up (1st April 2013). The one-year risk of death
(defined as the probability of death within the next 365 days) was
calculated by Kaplan–Meier plot and the relative risk was taken as the
ratio of the risk of death among different groups. Missing observations
were assumed to be missing at random and partial deletion was used
to handle missing observations. All tests were performed at the 5%

significance level.
Except for univariate analysis, data were analyzed twice with deceased
donors categorized as DBD or DCD (donor type classification I) or as
SCD or ECD (donor-type classification II). This was because ECD and
DCD donor types were highly correlated and not independent: of the
288 ECD donors, 174 (60.4%) where also DCD donors. Including both
DCD and ECD donor types in the same analysis, therefore, would have
required a much higher number of observations to reach definitive
conclusions. Moreover, it was important to be able to determine the
impact of both DCD and ECD donor types separately, as one or the
other classification is often used by transplant centers and reported in
the literature.

Results
Incidence of EGL
During the 10-year study period, 801 adult patients received
a deceased donor kidney-only transplant, of which 435
(54.3%) were from DCD donors and 366 (45.7%) were from
DBD donors. Of the 801 transplants, 288 (36.0%) were
performed using ECD grafts and 465 (58.0%) using SCD
grafts and 48 donors (6.0%) could not be classified as ECD
or SCD due to missing data.
EGL occurred in 50 (6.2%) recipients. Causes of EGL are
summarised in Table 1 and included PNF (20; 2.5%),
arterial or venous thrombosis (18; 2.2%), haemorrhage (6;
0.7%) and acute rejection (3; 0.4%). Further details on
causes of EGL are provided in the supplementary
materials. In comparison, only 2 (of 288; 0.7%) recipients
of living donor kidneys suffered EGL during the study
period.

The incidence of EGL was higher among recipients of DCD
grafts compared to DBD grafts (8.3% vs. 3.8%; p ¼ 0.018) .
The increased EGL among DCD transplants was due to a
numerically higher incidence of PNF (3.2% vs. 1.6%;
p ¼ 0.180) and acute vascular occlusions (3.0% vs. 1.4%;
p ¼ 0.155), although these differences did not reach
statistical significance (Table 1a). The incidence of EGL
was also higher among recipients of ECD grafts compared
to SCD grafts (10.1% vs. 4.1%; p ¼ 0.003). The increased
EGL among ECD transplants was due to a higher number of
1633


Hamed et al
Table 1: Causes of EGL after kidney transplantation. a) The overall incidence of EGL was higher among DCD recipients compared to DBD
recipients (p < 0.001), consisting of a higher number of grafts lost due to PNF (3.2% vs. 1.6%; p ¼ 0.18) and acute vascular occlusions (3.0%
vs. 1.4%; p ¼ 0.16). b) The overall incidence of EGL was higher among ECD recipients compared to SCD recipients (p < 0.001), consisting of
a higher number of grafts lost due to PNF (4.2% vs. 1.5%; p ¼ 0.03) and acute vascular occlusions (4.2% vs. 1.3%; p ¼ 0.03)
Deceased donor type classification I (%)
Causes of early graft loss

Total, n ¼ 801

DCD, n ¼ 435

DBD, n ¼ 366

20 (2.5%)
15 (1.9%)
3 (0.4%)

6 (0.7%)
3 (0.4%)
3 (0.4%)
50 (6.2%)

14 (3.2%)
10 (2.3%)
3 (0.7%)
4 (0.9%)
2 (0.5%)
3 (0.7%)
36 (8.3%)

6 (1.6%)
5 (1.4%)
0
2 (0.5%)
1 (0.3%)
0
14 (3.8%)

a)
Primary non-function (PNF)
Acute venous thrombosis
Acute arterial thrombosis
Haemorrhage
Acute rejection
Other
Total


Deceased donor type classification II (%)
Causes of early graft loss

Total, n ¼ 801

SCD, n ¼ 465

ECD, n ¼ 288

Unknown*, n ¼ 48

12 (4.2%)
10 (3.5%)
2 (0.7%)
3 (1.0%)
2 (0.7%)
0
29 (10.1%)

1 (2.1%)
0
0
0
1 (2.1%)
0
2 (4.2%)

b)
Primary non-function (PNF)
Acute venous thrombosis

Acute arterial thrombosis
Haemorrhage
Acute rejection
Other
Total

20 (2.5%)
15 (1.9%)
3 (0.4%)
6 (0.7%)
3 (0.4%)
3 (0.4%)
50 (6.2%)

7
5
1
3

(1.5%)
(1.1%)
(0.2%)
(0.6%)
0
3 (0.7%)
19 (4.1%)

*

Unknown; unclassified deceased donors due to missing data.


grafts lost due to PNF (4.2% vs. 1.5%; p ¼ 0.033) and acute
vascular occlusions (4.2% vs. 1.3%; p ¼ 0.025, Table 1b).
The risk of EGL was numerically higher in recipients of
kidneys from donors who were both ECD and DCD (23 of
174; 13.2%) compared to recipients of DCD or ECD grafts
only (8.3% and 10.1%, respectively) but the difference was
not statistically significant (p ¼ 0.177).
Excluding kidneys that were transplanted in pairs or
when one kidney from the pair was discarded, the
relative risk of EGL of the second kidney in the pair was
higher when the first kidney suffered EGL (12.5 vs.
6.2%), but the difference was not statistically significant
(p ¼ 0.176).

Risk factors for EGL
Donor and recipient demographics are shown in Table 2.
Univariate analysis revealed that higher donor age (55.2 vs.
49.0 y, p ¼ 0.005), recipient past history of venous thrombosis (p ¼ 0.004), DCD donor type (p ¼ 0.014) and ECD
donor type (p < 0.001) were significantly associated with
EGL (Table 2). Of note, cold ischaemia time (CIT), recipient
age, recipient past history of hypertension or diabetes
mellitus, peritoneal dialysis, surgeon grade, sensitisation
status and HLA mismatch status were not associated
with EGL.
1634

In multivariate logistic regression analysis using deceased
donor type classification I (DBD/DCD), recipient past history
of thrombosis (odds ratio [OR] 4.91; p ¼ 0.001), donor age

(OR 1.03; p ¼ 0.044) and DCD donor type (OR 2.88 vs. DBD
donors; p ¼ 0.006) were significant risk factors for EGL
(Table 3a). Donor age was >50 years in 27 (of 36; 75.0%)
DCD recipients with EGL compared to 233 (of 399;
58.4%) DCD recipients in the control (No EGL) group
(p ¼ 0.344). Seven (of 50; 14.0%) recipients with EGL had
a past history of venous thrombosis; four recipients had a
past history of pulmonary embolism (PE), two recipients
had a past history of deep venous thrombosis (DVT) and
PE, two recipients with a past history of DVT only, and
one EGL recipient lost two previous renal grafts due to
venous thrombosis. When using deceased donor type
classification II (SCD/ECD), ECD donor type (OR 4.22 vs.
SCD donors; p ¼ 0.010) and history of thrombosis (OR
4.79; p ¼ 0.001) were significant risk factors for EGL
(Table 3b). Of note, donor age was not a significant risk
factor for EGL when including ECD donor type in the
analysis; this was not unexpected since age is included in
the definition of ECD donors and the two variables are
therefore highly correlated.
Although DCD donor type was significantly associated
with EGL, there was no significant difference in long-term
graft survival between recipients of DCD and DBD
American Journal of Transplantation 2015; 15: 1632–1643


Early Kidney Graft Loss
Table 2: Univariate analysis of risk factors for EGL
Graft survival
Variable

Recipient age (years)
Donor age (years)
Recipient gender male
Recipient comorbidities
Diabetes
Hypertension
History of thrombosis
HLA-A mismatch
0
1
2
Unknown
HLA-B mismatch
0
1
2
Unknown
HLA-DR mismatch
0
1
2
Unknown
Cold ischaemia time (hours)
Anastomosis time (hours)
Recipient BMI
Donor BMI
Donor type classification I
DBD
DCD
Donor type classification II

SCD
ECD
Unknown
Sensitization
Not sensitized
Sensitized
Highly sensitized
Unknown
1st transplant
2nd transplant
3rd transplant
Peritoneal dialysis
Surgeon grade
Consultant* not present
Consultant present – not scrubbed
Consultant present – scrubbed
Unknown
Gender mismatch
None
Male recipient/female donor
Female recipient/male donor
Side mismatch
None
Right recipient/left donor
Left recipient/right donor
Unknown

30 days n ¼ 50

>30 days n ¼ 751


52.9  12.8
55.2  15.7
34 (68%)

49.4  13.3
49.0  16.6
493 (66%)

0.063
0.005
0.853

5 (10%)
24 (48%)
7 (14%)

60 (8%)
293 (39%)
30 (4%)

0.813
0.268
0.004

9 (18%)
26 (52%)
15 (30%)
0


172 (23%)
386 (51%)
189 (25%)
4 (1%)

0.753

4 (8%)
33 (66%)
13 (26%)
0

141 (19%)
493 (66%)
114 (15%)
3 (0.4%)

0.083

24 (48%)
25 (50%)
1 (2%)
0
15.1  4.0
0.79  0.22
26.0  4.3
27.2  6.9

405 (54%)
307 (41%)

33 (4%)
6 (1%)
14.6  4.6
0.78  0.44
25.9  4.4
26.5  5.8

0.516

14 (28%)
36 (72%)

352 (47%)
399 (53%)

18 (36%)
30 (60%)
2 (4%)

447 (60%)
258 (34%)
46 (6%)

28 (56%)
17 (34%)
4 (8%)
1 (2%)
43 (86%)
7 (14%)
0

15 (30%)

477 (63%)
201 (27%)
65 (9%)
4 (1%)
658 (88%)
78 (10%)
14 (2%)
251 (34%)

0.486

13 (26%)
7 (14%)
29 (58%)
1 (2%)

257 (34%)
98 (13%)
379 (51%)
17(2%)

0.680

29 (58%)
17 (34%)
4 (8%)

388 (52%)

218 (29%)
145 (19%)

0.137

32 (64%)
7 (14%)
6 (12%)
5 (10%)

424
115
107
105

0.746

(56%)
(15%)
(14%)
(14%)

p-value

0.178
0.509
0.785
0.833
0.014


<0.001

0.467

0.732

*

The equivalent surgeon grade in the United States is Attending Surgeon.
Continuous variables are reported as mean ( SD) and categorical variables are reported as n (%). p values are derived from Mann–Whitney
rank sum tests for continuous data and Chi-squared tests for categorical data. Sensitization level was grouped according to the percentage
Calculated Reactive Frequency at the time of transplant (not sensitized 0%, sensitized 1–85%, highly sensitized >85%)

American Journal of Transplantation 2015; 15: 1632–1643

1635


Hamed et al

whose graft survived more than 30 days (76.0% vs.
98.1%; p ¼ 0.005). A multivariate logistic regression
analysis using donor type classification I (DBD/DCD)
revealed that patients with EGL had 12.28 times
increased short-term (within 1 year after kidney transplantation) odds ratio death compared to the control
cohort whose grafts survived beyond 30 days (p < 0.001;
Table 5a). The increased odds ratio was 12.17 times
(p < 0.001; Table 5b) when using donor type classification
II (SCD/ECD). No other variable, including donor type, was
a significant risk factor for patient death at 1 year in the

presence of the EGL indicator (Table 5a and b).
To assess the effect of EGL on long-term patient survival,
a Cox proportional hazard analysis was performed on a
subgroup of patients who survived at least 1 year after the
kidney transplantation (Table 6). This revealed that EGL
had a significant negative impact on long-term patient
survival using both donor type classification I (HR of 5.36;
p ¼ 0.003; Table 6a) and donor type classification II (HR of
5.33; p ¼ 0.004; Table 6b). EGL, therefore, had a
detrimental impact on both short- and long-term patient
survival (Figure 2A). Importantly, in the presence of EGL,
donor type itself was not a risk factor for long term patient
death using donor type classification I (DCD 96.7% vs.
DBD 97.0%; p ¼ 0.732; Figure 2B) or donor type
classification II (ECD 95.1% vs. SCD 97.8%; p ¼ 0.071;
Figure 2C).

Figure 1: Long-term graft survival. (A) There was no significant
difference in long-term graft survival between recipients of DCD
and DBD transplants (5-year graft survival; DCD 86.4% vs. SCD
90.0%; p ¼ 0.257). (B) Long-term graft survival was inferior in
recipients of ECD compared to SCD grafts (5-year graft survival;
ECD 83.7% vs. SCD 91.4%; p ¼ 0.017).

kidneys (5-year graft survival; DCD 86.4% vs. SCD 90.0%;
p ¼ 0.257; Figure 1A). However, 5-year graft survival in
recipients of ECD kidneys was inferior compared to SCD
kidneys (ECD 83.7% vs. SCD 91.4%; p ¼ 0.017;
Figure 1B).


Consequences of EGL
Sixteen (of 50; 32.0%) recipients with EGL died within the
study period. Causes of patient mortality in the EGL group
are shown in Table 4. One-year patient survival was
markedly inferior in the EGL group compared to those
1636

We next examined how patient survival after EGL
compared to remaining on the kidney transplant waiting
list. As expected, patients remaining on the transplant
waiting list (i.e. listed but not transplanted) had a higher
mortality rate compared to all patients on the waiting list
(i.e. including patients who received a transplant; p < 0.001;
Figure 3A). This analysis confirmed the survival benefit to
those receiving a transplant compared to all patients on the
waiting list. Importantly, it also enabled us to quantify, at a
population level, the ‘‘best-case scenario’’ and ‘‘averagecase scenario’’ impact of EGL on patient survival compared
to remaining on the waiting list. This analysis can potentially
help the recipient and the transplant team when faced with
the choice to accept or decline a kidney that is associated
with a high risk of EGL, because the recipient and the
transplant team are generally unable to predict if, and when,
the recipient will be offered another organ if the transplant
is declined.
As shown in Figure 3B, compared to all patients on the
waiting list, EGL resulted in a very high (>8 fold) initial risk of
recipient death. This increased risk of patient death
attributable to EGL persisted for 23 months (time to equal
risk) after transplantation; patients experiencing EGL,
therefore, must survive for a mean of 23 months after

transplantation before their risk of death equals that of other
patients on the waiting list. The corresponding time to equal
survival was 76 months compared to all patients on the
waiting list, defined as the time needed before the
American Journal of Transplantation 2015; 15: 1632–1643


Early Kidney Graft Loss
Table 3: Multivariate analysis of risk factors of EGL. a) DCD donor type, donor age, and recipient past history of thrombosis were significant
risk factors for EGL. b) ECD donor type and recipient past history of thrombosis were significant risk factors for EGL
Variable

Odds ratio

95% CI

p value

CIT

1.00
–
1.01
–
4.91
1.03
–
2.88
1.04


0.97–1.03
–
0.52–1.98
–
1.89–12.73
1.00–1.05
–
1.36–6.13
0.96–1.12

0.864
–
0.975
–
0.001
0.044
–
0.006
0.223

Variable

Odds ratio

95% CI

p value

1.00
–

0.96
–
4.79
0.99
–
4.22
1.02

0.97–1.03
–
0.49–1.88
–
1.87–12.27
0.96–1.03
–
1.41–12.59
0.95–1.10

0.928
–
0.899
–
0.001
0.625
–
0.010
0.541

a)
Recipient age

Recipient gender
Recipient past history of thrombosis
Donor age
Donor type classification I

Female
Male
No
Yes
DBD
DCD

b)
Recipient age
Recipient gender
Recipient past history of thrombosis
Donor age
Donor type classification II

Female
Male
No
Yes
SCD
ECD

CIT

transplant offers a survival advantage compared to
remaining on the transplant waiting list.

When compared to patients on the waiting list that do not
receive a transplant (Figure 3C), however, EGL was
associated with a much smaller (<2 fold) increased risk
of death. The increased risk of death attributable to EGL
persisted for only 4 months (time to equal risk) after
transplantation and the corresponding time to equal survival
was 37 months. Despite the high mortality rate among the
EGL group, their long-term survival was still better than
either control group, especially compared to patients on the
waiting list that do not receive a transplant (Figures 3B
and C). The age of recipients with EGL (52.9  12.8 years)
was comparable to patients on the waiting list that did not
receive a transplant (53.4  12.5 years; p ¼ 0.892). The
worse outcome of non-transplanted patients was therefore
not attributable to a difference in age.

Outcomes for patients who suffered EGL are shown in
Figure 4. Of the 16 recipients who died after EGL, only
one was re-listed for kidney transplantation before death
and died while active on the waiting list. Of the 34
patients that remained alive during the study period, 24
were re-listed for transplantation and 14 were re-transplanted at a median of 24.5 months after their graft loss
(range 8–89 months). One-year re-transplant graft
survival was 85.7% (two graft losses within 1 year) in
patients re-transplanted after EGL; one graft failed due to
rejection 9 months after transplantation and one patient
had another EGL secondary to venous thrombosis (the
same aetiology as for the first graft loss). In comparison,
during the study period, 98 patients received a second
kidney transplant without having had early loss of the

first graft. One-year re-transplant graft survival in this
group was similar to patients re-transplanted after EGL
(90.8%; p ¼ 0.635).

Table 4: Causes of death in patients with early kidney graft loss
Cause of death

n

Time of death after transplant (days)

Pneumonia
Haemorrhage related to transplant
Cardiac arrest secondary to hyperkalaemia
Perforated bowel
Non-lymphoid malignant disease
Others: uraemia, subdural haemorrhage, ESRD
Total

4
3
2
2
2
3
16

28, 60, 720, 1275
14, 14, 28
1, 1

30, 150
180, 720
150, 210, 450

American Journal of Transplantation 2015; 15: 1632–1643

1637


Hamed et al
Table 5: Multivariate logistic regression analysis of risk of recipient death within 1 year after kidney transplantation
Variable

Odds ratio

95% CI

p value

1.03
–
0.96
–
2.71
1.03
–
0.97
1.06
–
12.28


0.98–1.08
–
0.36–2.59
–
0.73–10.16
0.99–1.07
–
0.35–2.70
0.96–1.17
–
4.60–32.91

0.243
–
0.932
–
0.138
0.205
–
0.958
0.252
–
<0.001

Odds ratio

95% CI

p value


1.02
–
1.11
–
2.66
1.05
–
0.50
1.07
–
12.17

0.97–1.07
–
0.39–3.21
–
0.68–10.40
0.99–1.12
–
0.1–2.47
0.96–1.18
–
4.30–34.44

0.428
–
0.846
–
0.160

0.109
–
0.393
0.232
–
<0.001

a)
Recipient age
Recipient gender
Recipient past history of thrombosis
Donor age
Donor type classification I
CIT
EGL

Female
Male
No
Yes
DBD
DCD
No
Yes

Variable
b)
Recipient age
Recipient gender
Recipient past history of thrombosis

Donor age
Donor type classification II
CIT
EGL

Female
Male
No
Yes
SCD
ECD
No
Yes

EGL was the only significant risk factor for short-term patient mortality, using both donor type classification I (a) and II (b)

Table 6: Cox proportional hazards analysis of risks factors for long-term (>1 year after transplantation) patient mortality
Variable

Hazard ratio

95% CI

p value

1.03
–
0.71
–
0.78

1.00

0.99–1.06
–
0.31–1.62
–
0.10–5.88
0.97–1.03

0.169
–
0.419
–
0.808
0.975

0.52
0.99
–
5.36

0.22–1.24
0.91–1.08
–
1.74–16.47

0.142
0.834
–
0.003


Hazard ratio

95% CI

p value

1.01
–
0.74
–
0.97
1.00

0.97–1.05
–
0.30 - 1.83
–
0.12–7.47
0.97–1.04

0.517
–
0.523
–
0.977
0.835

0.82
1.01


0.21–3.16
0.92–1.10

0.772
0.851

5.33

1.69–16.83

0.004

a)
Recipient age
Recipient gender
Recipient past history of thrombosis
Donor age
Donor type classification I
CIT
EGL

Female
Male
No
Yes
DBD
DCD
No
Yes


Variable
b)
Recipient age
Recipient gender
Recipient past history of thrombosis
Donor age
Donor type classification II
CIT
EGL

Female
Male
No
Yes
SCD
ECD
No
Yes

EGL was the only significant risk factor for long-term patient mortality, using both donor type classification I (a) and II (b)

1638

American Journal of Transplantation 2015; 15: 1632–1643


Early Kidney Graft Loss

Figure 2: Long-term patient survival. (A) Long-term patient survival was significantly lower after EGL: 1-year patient survival was 76.0%

compared to 98.1% in patients whose grafts survived >30 days (p ¼ 0.005). (B) Long-term patient survival was similar in recipients of DBD
and DCD grafts: 1 year patient survival was 97.0% and 96.7% in recipients of DBD and DCD grafts, respectively (p ¼ 0.732). (C) Long-term
patient survival was also similar in recipients of SCD and ECD grafts: 1 year patient survival was 97.8% in recipients of SCD grafts and 95.1%
in recipients of ECD grafts (p ¼ 0.071).

Discussion
The disparity between organ supply and demand justifies
the increasing usage of DCD and ECD kidneys (12). A
specific aim of this study was to examine the possible
association between deceased donor type and EGL in a
cohort which consisted of a high proportion of transplants
from DCD and ECD donors; previous studies made no
distinction between DCD and DBD donors (3,13). Our
results show that while EGL was more frequent among
DCD and ECD kidney recipients, the long-term outcomes
were similar between DCD versus DBD kidney recipients,
which is consistent with the findings of previous studies (6,14,15). ECD donor type kidney recipients are known
American Journal of Transplantation 2015; 15: 1632–1643

to have worse long-term graft survival compared to SCD
kidney recipients (16,17), as confirmed by our results. The
striking finding of our study is the significant negative
impact of EGL, irrespective of the donor type, on patient
survival. This confirms the view of transplant clinicians that
all efforts should be made to minimize the risk of the
disastrous outcome of EGL. However, long-term patient
outcomes were even worse for those remaining on the
transplant waiting list. At a population level, therefore, the
detrimental consequences of EGL should be considered in
the context of the mortality rates on the waiting list.

The association of higher donor age and recipient past
history of venous thrombosis and DCD/ECD donor type in
1639


Hamed et al

Figure 3: Mortality risks of patients listed on the kidney transplant waiting list. (A) Patient survival from the time of listing on the
kidney transplant waiting list. Patient survival was significantly lower in patients remaining on the kidney transplant waiting list (No
Transplant group) compared to all patients listed on the kidney transplant (log-rank test p < 0.001). (B) Adjusted Relative Risk of Death within
1 year among the early graft loss (EGL) group, control (No EGL) group, and all kidney transplant recipients group (Transplant: All Patients). The
reference group (relative risk of 1.0) was all patients on the kidney transplant waiting list (Waiting List: All Patients), including those who were
transplanted (‘‘All Patients’’ from Figure 2A). The increased relative risk of patient death, attributable to EGL, persisted for 23 months (time to
equal risk) after transplantation. Time to equal survival was 76 months. (C) Adjusted Relative Risk of Death within 1 year among early graft
loss (EGL) group, control (No EGL) group, and all kidney transplant recipients group (Transplant: All Patients). The reference group (Waiting
List: No Transplant; relative risk of 1.0) was patients on the kidney transplant waiting list that did not receive a transplant (‘‘No Transplant’’
from Figure 2A). The increased relative risk of patient death, attributable to EGL, persisted for 4 months (time to equal risk) after
transplantation. Time to equal survival was 37 months. The first point of calculation of relative risk of death within 1 year was at 1 week
posttransplantation.

our univariate analysis with EGL is not surprising as these
factors have been associated with graft thrombosis and
worse long-term outcome (10,18,19). The increased EGL
among DCD and ECD transplants in our series was due to
a higher incidence of PNF and acute vascular occlusions.
The increased incidence of PNF in DCD kidneys may be
attributable to warm ischemic injury in the donor;
preliminary results in animal models suggest that DCD
kidneys are more vulnerable to ischemia reperfusion
injury (19,20).

Our findings are consistent with previous reports that have
identified renal vascular thrombosis as a major cause of
EGL, accounting for up to 30% of cases (10,11,21). The
1640

pathogenesis of renal graft thrombosis is poorly understood
and likely multifactorial, including donor, recipient and
technical (operative) factors (10,11,21). The risk of renal
graft thrombosis as related to different deceased donor
types (DCD vs. DBD) has, however, not been previously
reported. The increased incidence of EGL from acute
thrombosis and hemorrhage in recipients of DCD and ECD
kidneys may be attributable to technical complications
secondary to poor quality vessels and/or endothelial
activation in DCD and ECD donation. It is likely, however,
that DCD and ECD donor types were not solely responsible
for the higher incidence and that other predisposing factors,
such as donor age or co-morbidity, are also relevant. In our
analysis, recipient past history of major venous thrombosis
American Journal of Transplantation 2015; 15: 1632–1643


Early Kidney Graft Loss

Figure 4: Outcomes of patients with early kidney graft loss. One-year graft survival in patients re-transplanted after a previous EGL was
85.7% (12/14).

(DVT or PE) was a significant risk factor for EGL. Although it
is possible that a past history of venous thrombosis was
under-reported, the observed increased risk (OR >4) was

sufficiently high to suggest that this group may benefit from
intensive monitoring or modified management such as
detailed screening for thrombotic tendencies and/or modified posttransplant anti-thrombotic prophylaxis.
In our analysis, the detrimental impact of EGL on patient
survival was most pronounced within the first year after
transplantation. However, the increased mortality
amongst recipients with early kidney graft loss was
directly related to surgical complications of the transplant
in only five cases. It is likely, therefore, that in addition to
postoperative medical and surgical complications, the
negative impact of returning to dialysis, the risks and
complications associated with ESRD and exposure to
immunosuppression combine to dramatically increase
mortality risk after EGL. This is consistent with a previous
study which reported an increased risk of death in the
60 days post graft loss, irrespective of when graft loss
occurred (22).
Despite the association of DCD donor type with EGL, there
were no significant differences in the long-term graft
survival outcomes between DCD and DBD donors, as
previously reported by our center and others (6,12,15). This
finding suggests that the early impact of the DCD donation
American Journal of Transplantation 2015; 15: 1632–1643

on graft survival is ‘‘diluted’’ by other factors affecting graft
survival later in the first year.
An important finding of our study is that, despite their
increased mortality risk, recipients with EGL had better longterm outcomes than patients on the transplant waiting list.
Starting from an initial increased risk of 1 year mortality of
more than eightfold, patient survival in the EGL group

gradually reduced to equal the risk of death for all patients on
the transplant waiting list at 23 months (Figure 2B). The
increased mortality risk after EGL was much smaller (less
than twofold) when compared to patients on the waiting list
who were not transplanted, and reduced to baseline by
4 months (Figure 2C). In comparison, the relative risk of death
of recipients has previously been reported to be 2.8 times
higher than patients on the waiting list after transplantation,
decreasing to unity (time to equal risk) by 3.5 months (23).
Our results are consistent with the high mortality rate
reported for patients on dialysis (24) and suggest that, in the
face of lengthening kidney transplant waiting lists, increasing use of kidneys from suboptimal donors must be
considered despite their predisposition to the catastrophic
outcome of EGL. A caveat to this conclusion, and a
limitation of this study, is the selection bias that is inherent
in this comparison: it is likely that patients selected from a
waiting list to receive a kidney transplant are a more optimal
group (e.g. less sensitized) than those that remain on the
1641


Hamed et al

waiting list without being transplanted. These and other
factors may have resulted in a selection bias, not quantified in
our analysis, which affects the comparison between patients
on the waiting list who are transplanted and those who are
not. Nonetheless, our results suggest that the additional
mortality risk associated with EGL is, in the long-term, less
than the alternative of remaining on the transplant waiting

list. While this does not justify the injudicious use of
suboptimal organs, it serves to draw attention to the risks
that patients on the waiting list are exposed to every time a
kidney is declined for transplantation. Our findings are
consistent with a previous report that demonstrated DCD
kidney transplantation offers a survival benefit over those
waiting for DBD grafts (25). They go further, however, and
suggest that when the waiting list is considered in its
entirety, use of even very suboptimal kidneys may offer a
survival advantage for the population on the waiting list.
In this study, 1-year graft survival after kidney retransplantation following EGL was 85.7% (compared to
90.8% after re-transplantation following late graft loss).
This is consistent with previous reports on outcomes for
second kidney grafts (1-year graft survival 85–89%) (26,27).
Our findings, therefore, suggest that re-transplantation
after EGL should be considered for selected cases.
Our results suggest that EGL is more frequent after DCD
and ECD kidney transplantation and is a major risk factor for
patient mortality. However, long-term patient survival after
EGL is still better than those remaining on the kidney
transplant waiting list. Kidney re-transplantation after EGL is
associated with a good outcome and should be considered.

Acknowledgments
We are grateful to the NHS Blood and Transplant (NHSBT) for their help in
providing the relevant data for this paper from the UK transplant registry. The
study was approved by our institution’s review board as a service evaluation
audit. Yining Chen was supported by the Engineering and Physical Sciences
Research Council Grant EP/J017213/1.


Disclosure
The authors of this manuscript have no conflicts of interest
to disclose as described by the American Journal of
Transplantation.

References
1. Tonelli M, Wiebe N, Knoll G, et al. Systematic review: Kidney
transplantation compared with dialysis in clinically relevant outcomes. Am J Transplant 2011; 11: 2093–2109.
2. Yildirim A. The importance of patient satisfaction and healthrelated quality of life after renal transplantation. Transplant Proc
2006; 38: 2831–2834.

1642

3. Phelan PJ, O’Kelly P, Tarazi M, et al. Renal allograft loss in the first
post-operative month: Causes and consequences. Clin Transplant
2012; 26: 544–549.
4. Marcen R, Fernandez-Rodriguez A, Rodriguez-Mendiola N, et al.
Evolution of rejection rates and kidney graft survival: A historical
analysis. Transplant Proc 2009; 41: 2357–2359.
5. Perico N, Ruggenenti P, Scalamogna M, Remuzzi G. Tackling the
shortage of donor kidneys: How to use the best that we have. Am J
Nephrol 2003; 23: 245–259.
6. Summers DM, Johnson RJ, Allen J, et al. Analysis of factors that
affect outcome after transplantation of kidneys donated after
cardiac death in the UK: A cohort study. Lancet 2010; 376: 1303–
1311.
7. Port FK, Bragg-Gresham JL, Metzger RA, et al. Donor characteristics associated with reduced graft survival: An approach to
expanding the pool of kidney donors. Transplantation 2002; 74:
1281–1286.
8. Marfo K, Lu A, Ling M, Akalin E. Desensitization protocols their

outcome. Clin J Am Soc Nephrol 2011; 6: 922–936.
9. Pour-Reza-Gholi F, Nafar M, Saeedinia A, et al. Kidney retransplantation in comparison with first kidney transplantation. Transplant Proc 2005; 37: 2962–2964.
10. Keller AK, Jorgensen TM, Jespersen B. Identification of risk factors
for vascular thrombosis may reduce early renal graft loss: A review
of recent literature. J Transplant 2012; 2012: 793461.
11. Penny MJ, Nankivell BJ, Disney AP, Byth K, Chapman JR. Renal
graft thrombosis. A survey of 134 consecutive cases. Transplantation 1994; 58: 565–569.
12. Gagandeep S, Matsuoka L, Mateo R, et al. Expanding the donor
kidney pool: Utility of renal allografts procured in a setting of
uncontrolled cardiac death. Am J Transplant 2006; 6: 1682–
1688.
13. Zukowski M, Kotfis K, Biernawska J, et al. Donor-recipient gender
mismatch affects early graft loss after kidney transplantation.
Transplant Proc 2011; 43: 2914–2916.
14. Wijnen RM, Booster MH, Stubenitsky BM, de Boer J, Heineman E,
Kootstra G. Outcome of transplantation of non-heart-beating donor
kidneys. Lancet 1995; 345: 1067–1070.
15. Wells AC, Rushworth L, Thiru S, et al. Donor kidney disease and
transplant outcome for kidneys donated after cardiac death. Br J
Surg 2009; 96: 299–304.
16. Pascual J, Zamora J, Pirsch JD. A systematic review of kidney
transplantation from expanded criteria donors. Am J Kidney Dis
2008; 52: 553–586.
17. Metzger RA, Delmonico FL, Feng S, Port FK, Wynn JJ, Merion RM.
Expanded criteria donors for kidney transplantation. Am J
Transplant 2003; 3: 114–125.
18. van der Vliet JA, Warle MC, Cheung CL, Teerenstra S, Hoitsma AJ.
Influence of prolonged cold ischemia in renal transplantation. Clin
Transplant 2011; 25: E612–E616.
19. Phelan PJ, Magee C, O’Kelly P, et al. Immediate re-transplantation

following early kidney transplant thrombosis. Nephrology (Carlton)
2011; 16: 607–611.
20. Jani A, Zimmerman M, Martin J, et al. Perfusion storage reduces
apoptosis in a porcine kidney model of donation after cardiac death.
Transplantation 2011; 91: 169–175.
21. Terasaki PI, Cecka JM, Gjertson DW, Takemoto S, Cho YW, Yuge
J. Risk rate and long-term kidney transplant survival. Clin
Transplant 1996; 443–458.
22. McDonald SP, Russ GR. Survival of recipients of cadaveric kidney
transplants compared with those receiving dialysis treatment in
Australia and New Zealand, 1991–2001. Nephrol Dial Transplant
2002; 17: 2212–2219.

American Journal of Transplantation 2015; 15: 1632–1643


Early Kidney Graft Loss
23. Wolfe RA, Ashby VB, Milford EL, et al. Comparison of mortality in
all patients on dialysis, patients on dialysis awaiting transplantation, and recipients of a first cadaveric transplant. New Eng J Med
1999; 341: 1725–1730.
24. Medin C, Elinder CG, Hylander B, Blom B, Wilczek H.
Survival of patients who have been on a waiting list for
renal transplantation. Nephrol Dial Transplant 2000; 15:
701–704.
25. Evenson AR. Utilization of kidneys from donation after circulatory
determination of death. Curr Opin Organ Transplant 2011; 16: 385–
389
26. Ingsathit A, Kantachuvesiri S, Rattanasiri S, et al. Long-term
outcome of kidney retransplantation in comparison with first


American Journal of Transplantation 2015; 15: 1632–1643

kidney transplantation: A report from the Thai Transplantation
Registry. Transplant Proc 2013; 45: 1427–1430.
27. Barocci S, Valente U, Fontana I, et al. Long-term outcome on
kidney retransplantation: A review of 100 cases from a single
center. Transplant Proc 2009; 41: 1156–1158.

Supporting Information
Additional Supporting Information may be found in the
online version of this article.
Supplementary Materials and Methods

1643