American Journal of Transplantation 2009; 9: 217–221
Wiley Periodicals Inc.
Brief Communication
Histidine-Tryptophan-Ketoglutarate (HTK) Is
Associated with Reduced Graft Survival in Pancreas
Transplantation
Z. A. Stewart, A. M. Cameron, A. L. Singer,
N. N. Dagher, R. A. Montgomery
and D. L. Segev∗
Department of Surgery, The Johns Hopkins University
School of Medicine, Baltimore, MD
∗ Corresponding author: Dorry L. Segev, dorry@jhmi.edu
Prior single-center studies have reported that pan-
creas allograft survival is not affected by preserva-
tion in histidine-tryptophan-ketoglutarate (HTK) ver-
sus University of Wisconsin (UW) solution. To expand
on these studies, we analyzed the United Network
for Organ Sharing (UNOS) database of pancreas trans-
plants from July 2004, through February 2008, to deter-
mine if preservation with HTK (N = 1081) versus UW
(N = 3311) impacted graft survival. HTK preservation of
pancreas allografts increased significantly in this time
frame, from 15.4% in 2004 to 25.4% in 2008. After ad-
justing for other recipient, donor, graft and transplant
center factors that impact graft survival, HTK preserva-
tion was independently associated with an increased
risk of pancreas graft loss (hazard ratio [HR] 1.30, p =
0.014), especially in pancreas allografts with cold is-
chemia time (CIT) ≥12 h (HR 1.42, p = 0.017). This
reduced survival with HTK preservation as compared
to UW preservation was seen in both simultaneous
pancreas-kidney (SPK) transplants and pancreas alone
(PA) transplants. Furthermore, HTK preservation was
also associated with a 1.54-fold higher odds of early
(<30 days) pancreas graft loss as compared to UW
(OR 1.54, p = 0.008). These results suggest that the in-
creasing use of HTK for abdominal organ preservation
should be re-examined.
Key words: HTK solution, pancreas transplant, pan-
creas transplantation alone, simultaneous pancreas-
kidney (SPK) transplantation, UW preservation solu-
tion
Received 29 July 2008, revised 24 August 2008 and
accepted for publication 05 September 2008
Introduction
Successful solid organ transplantation requires minimiz-
ing the damage from cold ischemia and reperfusion during
C 2009 The Authors
Journal compilation 
C 2009 The American Society of
Transplantation and the American Society of Transplant Surgeons
doi: 10.1111/j.1600-6143.2008.02449.x
organ recovery. For abdominal organs, the two most com-
monly used preservative solutions in the United States
are University of Wisconsin (UW) solution and histidine-
tryptophan-ketoglutarate (HTK) solution. UW solution has
been the standard for abdominal organ preservation fluid
since 1987 (1); however, HTK has replaced UW solution at
many centers since it was approved for use in kidney and
liver preservation. In addition to significantly lower cost (2),
other potential advantages of HTK include reduced risk of
reperfusion hyperkalemia, better tissue perfusion due to
lower viscosity and better cell preservation over a wider
range of temperatures (3).
The experience with HTK in pancreas allograft preservation
has been limited to small series that have reported equiva-
lent patient and graft survival for allografts preserved with
UW and HTK solutions (2,4–10). However, these studies
have been powered to detect only very large differences
in graft survival, and have lacked sufficient sample size to
appropriately adjust for confounders through multivariate
regression models.
To expand on previous reports, we analyzed a large national
registry of pancreas transplants from 2004 through 2008.
The primary outcome measured was graft survival, which
is very reliably measured through this registry. We then
performed multivariate regression analyses to adjust for
potential confounding recipient, donor, graft and transplant
center characteristics. Finally, we analyzed subgroups of
pancreas transplants by cold ischemia time (CIT), donor age
and transplant type (simultaneous pancreas-kidney (SPK)
or pancreas alone (PA)), as well as transplant center volume
to determine if HTK versus UW preservation impacted graft
survival differentially in these subgroups.
Methods
The study population consisted of 4392 patients undergoing pancreas trans-
plantation between July 1, 2004, and February 28, 2008, as reported to the
United Network for Organ Sharing (UNOS). Patients <18 years of age and
patients in whom a solution other than HTK or UW was used were excluded
from the study. Patients were divided into two groups: ‘UW’ = organ re-
covery with UW solution (n = 3311) and ‘HTK’ = organ recovery with HTK
solution (n = 1081). The primary outcome was graft survival, defined as
postoperative time alive with a functioning allograft, censored for loss to
217
Stewart et al.

follow-up or administrative end-of-study. Patient death information was sup-

plemented by linkage with the Social Security Master Death File.

Risk factors for graft loss were determined using Cox proportional hazards
models, and risk factors for early graft loss (<30 days) were determined us-
ing logistic regression. Multivariate models were adjusted for the following
confounders that were felt to be (or shown in prior studies to be) clinically
relevant: recipient age, gender, ethnicity, body mass index (BMI), hyperten-
sion, hospitalized and prior transplant; donor age, gender, ethnicity, BMI
and cause of death and transplant type (SPK or PA), donor type (donation
after cardiac death or donation after brain death) and CIT.

To adjust for center-specific effects, models incorporating all the covariates

listed above in addition to center volume were performed, with clustered
variance estimates to account for correlation among patients and practice
patterns by transplant center. We also evaluated the proportion of HTK use
by transplant center volume divided into deciles.

Unless otherwise specified, all tests were 2-sided with statistical signifi-
cance set at a = 0.05. Comparisons of donor and transplant characteristics
between UW and HTK subgroups are reported using unpaired 2-tailed t -
tests for continuous covariates and chi-squared tests of independence for
categorical variables. For Cox models, proportional hazards assumptions
were confirmed by inspection of complementary log-log plots. Data were
missing for <5% of all covariates and handled by casewise deletion, with
the exception of CIT that was imputed as previously described (11,12). All
analyses were performed using Stata 10.0/MP for Linux (StataCorp, College
Station, TX).

Results
The proportion of pancreas allografts recovered with HTK
preservation increased from 15.4% in 2004 to 25.4%
in 2008. Of note, during this time frame, 50% of pan-
creas transplants were performed by 24 transplant centers
(Figure 1A). However, HTK utilization was not associated
Figure 1: Distribution of pancreas transplants and HTK use by
with transplant center volume (Figure 1B). HTK-preserved transplant center. (A) Cumulative proportion of pancreas trans-
allografts were less likely to be utilized in SPK transplants plants performed by transplant centers with decreasing volumes.
and had a statistically significantly shorter CIT (Table 1). Fur- Note that 50% of pancreas transplants were performed by 24 cen-
ther, donors of HTK-preserved allografts were more likely ters. (B) Average proportion of HTK use among centers by deciles
to be Caucasian as compared to UW-preserved allograft of pancreas transplant volume.
(Table 1).

Despite the fact that HTK preservation was used in recip-
ients with shorter CIT, HTK preservation was associated
with a significantly increased risk of graft loss on univariate
analysis (hazard ratio [HR] 1.27, CI 1.08–1.48, p = 0.003).
After adjusting for donor, recipient, graft and transplant cen-
ter factors impacting graft survival, HTK preservation was
still independently associated with a 30% increased risk
of graft loss on multivariate analyses when compared with
UW (HR 1.30, p = 0.014) (Table 2). In subgroup analy-
ses, HTK-preserved SPK allografts had an increased risk
of pancreas graft loss as compared to UW-preserved SPK
allografts (HR 1.32, p = 0.023). The effect of HTK preser-
vation on reduced graft survival of both SPK and PA allo-
grafts was observed immediately after transplant, consis-
tent with the effect resulting from a difference in the effi-
cacy of the preservative solution (Figure 2). Furthermore,
HTK-preserved allografts with CIT ≥12 h (HR 1.42, p =
0.017) or a history of a prior transplant (HR 2.10, p = 0.016)
had an even greater risk of graft loss when compared to
UW-preserved allografts in adjusted models (Table 2). Fi-
nally, center volume did not seem to modify the effect
of HTK preservation on graft survival, as higher risk was
observed at both larger volume centers (≥50 pancreas
transplants during the study period; adjusted HR 1.37, p =
0.008) and smaller volume centers (<50 pancreas trans-
plants during the study period; adjusted HR 1.33, p = 0.03).
A closer examination of the allograft survival curves in
Figure 2 suggested a dominant early effect of preservation
solution; this was consistent with a recent single-center
report (2). To differentiate between early and late effects
in the national population, we developed a multivariate

218 American Journal of Transplantation 2009; 9: 217–221

 Reduced Pancreas Graft Survival with HTK

Table 1: Donor and transplant characteristics lized in allografts with shorter CIT, which are less likely to
UW HTK fail.
(n = 3311) (n = 1081) p
Age (y) 25.5 25.8 0.5 One strength of this study is the large sample size identi-
Female (%) 32.3 31.7 0.7 fied in the national registry. Prior studies trying to assess
Ethnicity, Caucasian (%) 64.0 77.3 <0.001 the impact of HTK versus UW preservation on pancreas
BMI 24.1 24.1 0.9 allograft survival have been severely underpowered to de-
Cause of death (%) 0.1 tect meaningful differences in graft survival (2,4–9). As a
Anoxia 10.5 13.1 conservative estimate in a previous study (4), it would re-
CVA 22.8 20.5 quire 157 patients in each arm to have an 80% chance
Head trauma 66.4 66.1 of showing a 10% difference in graft survival. Thus, it is
Donor after cardiac death (%) 2.4 1.6 0.1
not surprising that prior reports with series of 10 to 78
SPK (%) 68.0 63.6 0.009
patients (matched to historical controls) have not been
CIT (h) 13.5 11.5 <0.001
able to identify a statistically significant difference with
By 2-sided t -test for continuous variables and by chi-squared test HTK versus UW preservation, and have further lacked the
for categorical variables. Bold values represent statistically signif- ability to perform stable multivariate regression analyses
icant differences between HTK and UW.
to control for confounding recipient, donor and transplant
variables.
logistic regression model to examine the effect of HTK
preservation on early pancreas graft loss within the first 30 Trends from previously reported series of HTK preserva-
postoperative days. Indeed, a dominant early effect was tion of pancreas allografts are consistent with several of
seen for most subgroups. For all pancreas transplants, HTK our statistically significant findings. A review of the Michi-
preservation resulted in a 1.54-fold higher odds of early gan Organ Procurement Organization experience with HTK
graft loss (Odds Ratio, OR 1.54, p = 0.008) (Table 3). SPK in pancreas transplantation identified 36 HTK-preserved al-
transplants (OR 1.50, p = 0.021) and grafts with CIT ≥12 h lografts and matched them to historical UW controls (2).
(OR 1.58, p = 0.006) also had a higher odds of early graft While not achieving statistical significance, there was a
loss with HTK preservation. clear trend toward poorer graft survival with HTK preser-
vation at 30 days (HTK = 87% and UW = 94%) (2). This
Discussion difference was mostly due to notably poorer PA graft sur-
vival with HTK preservation, as the 30-day PA graft lost was
In this national study of pancreas transplants between 14.3% for HTK as compared to 5.9% for UW (2). Agarwal
2004 and 2008, we found that the use of HTK solution et al. reported on 78 HTK-preserved pancreas allografts
is an independent risk factor for decreased graft survival compared to historical UW controls and found shorter CIT
and early graft loss. This effect was seen in both SPK and for HTK-preserved allografts (9 h) versus UW-preserved
PA transplants. The increased risk of graft loss with HTK allografts (11 h) with similar 1-year graft survival for HTK
preservation was seen despite the fact that HTK was uti- (93%) and UW (92%) (4). Certainly, one explanation for

Table 2: Subgroup analysis of the effect of HTK preservation on pancreas allograft survival
Multivariate patient
Multivariate patient level∗ and center level∗∗
HR (95% CI) p-Value HR (95% CI) p-Value
All pancreas transplants 1.27 (1.07–1.50) 0.006 1.30 (1.05–1.60) 0.014
(n = 1081 HTK and 3311 UW)
SPK 1.29 (1.03–1.63) 0.029 1.32 (1.04–1.68) 0.023
(n = 688 HTK and 2250 UW)
CIT ≥12 h 1.42 (1.10–1.85) 0.008 1.42 (1.06–1.90) 0.017
(n = 391 HTK and 1499 UW)
Prior transplant 2.10 (1.27–3.49) 0.004 2.10 (1.15–3.84) 0.016
(n = 86 HTK and 266 UW)
Donors age > 40 years 1.29 (0.88–1.88) 0.20 1.34 (0.85–2.11) 0.20
(n = 151 HTK and 424 UW)
∗ Adjusted for recipient age, gender, ethnicity, BMI, hospitalized and prior transplant; donor age, gender, ethnicity, BMI and cause of

death; transplant type (SPK or PA), donor type (donation after cardiac death or donation after brain death) and CIT.
∗∗ Adjusted for same covariates as above and also adjusted for center volume, with clustered variance estimates to account for correlation

among patients and practice patterns by transplant center.

American Journal of Transplantation 2009; 9: 217–221 219

Stewart et al.

these results are the short CIT (<12 h) in this (4) and prior
case series (2, 5–9) of HTK preserved pancreas allografts,
as the UNOS data would suggest that HTK more strongly
impacts graft survival when CIT ≥12 h.

While the precise mechanism by which HTK preservation

may impact human pancreas allograft survival remains un-
clear, there have been extensive studies of HTK and UW
preservation of porcine pancreas allografts that suggest
porcine pancreas allografts stored in HTK are more ede-
matous than those stored in UW (13–17). These gross
findings also correlated with histological evidence of in-
creased interstitial edema and cellular swelling (15–17).
At least one group noted clinical correlation with HTK-
preserved human pancreas allografts, as they reported that
all HTK-flushed pancreas allografts appeared more edema-
tous as compared to UW-flushed pancreas allografts (7).
In vitro models of both human hepatocytes (18) and hu-
man liver endothelial cells (19) found that UW was signif-
icantly superior to HTK at reduction of cellular necrosis,
mitochondrial dysfunction and intracellular ATP depletion
during ischemia-reperfusion. Clearly, further work remains
to elucidate the critical cellular signaling pathways to target
during the ischemia-reperfusion that occurs during organ
recovery and transplantation.

It is important to acknowledge several limitations to our

Figure 2: Kaplan–Meier graft survival curves for SPK and PA
transplant based on preservative solution. (A) SPK graft survival
by HTK versus UW. (B) PA graft survival by HTK versus UW.
study. As with any analysis utilizing the UNOS database,
our conclusions rely on the assumption that there is no sys-
tematic bias generated by reporting error or missing data.
However, the primary endpoint for this analysis is graft
survival, which is reliably captured in the UNOS database.
Residual or unmeasured confounders that could impact
organ preservation and survival include: surgeon tech-
nique, volume of initial flush used at the time of organ
recovery, differences in immunosuppression protocols,
the fat content/quality of the allograft and center-specific

Table 3: Subgroup analysis of the effect of HTK preservation on early (< 30 days) pancreas allograft loss
Multivariate patient
Multivariate patient level∗ and center level∗∗
OR (95% CI) p-Value OR (95% CI) p-Value
All pancreas transplants 1.44 (1.13–1.84) 0.003 1.54 (1.12–2.13) 0.008
(n = 1081 HTK and 3311 UW)
SPK 1.45 (1.06–1.98) 0.019 1.50 (1.06–2.12) 0.021
(n = 688 HTK and 2250 UW)
CIT ≥12 h 1.59 (1.08–2.33) 0.019 1.58 (0.99–2.53) 0.006
(n = 391 HTK and 1499 UW)
Prior transplant 1.84 (0.77–4.38) 0.20 1.93 (0.72–5.19) 0.20
(n = 86 HTK and 266 UW)
Donors age > 40 years 1.50 (0.88–2.55) 0.10 1.61 (0.92–2.82) 0.09
(n = 151 HTK and 424 UW)
∗ Adjusted for recipient age, gender, ethnicity, BMI, hospitalized and prior transplant; donor age, gender, ethnicity, BMI and cause of
death; transplant type (SPK or PA), donor type (donation after cardiac death or donation after brain death) and CIT.
∗∗ Adjusted for same covariates as above and also adjusted for center volume, with clustered variance estimates to account for correlation

among patients and practice patterns by transplant center.

220 American Journal of Transplantation 2009; 9: 217–221


practices. However, presumably these differences would
only bias our results if there was a substantial interac-
tion between preservative solution and these other vari-
ables, such that these variables not only impacted graft
survival but also modified the impact of preservative solu-
tion on graft survival. Further, since causes of graft failure
are poorly captured in the UNOS database (over 50% of
the data are missing), a meaningful analysis of specific
causes of graft loss in HTK versus UW-preserved pancreas
allografts could not be performed.
In conclusion, our analysis of the national experience with
HTK and UW pancreas allograft preservation demonstrates
that HTK preservation is an independent risk factor for re-
duced graft survival for pancreas allografts and in particular
those with prolonged CIT. These findings may be utilized
to help guide clinical decision making in the selection of
organ preservation solution and suggest that despite its
lower cost, enthusiasm about HTK as a universal preserva-
tive solution should be re-examined.
Acknowledgments
The UNOS National Data Registry is supported in part by Health Resources
and Services Administration contract 231–00-0115. The analyses described
here are the responsibility of the authors alone and do not necessarily reflect
the views or policies of the Department of Health and Human Services, nor
does mention of trade names, commercial products or organizations imply
endorsement by the U.S. Government.
References
1. Belzer FO, Southard JH. Principles of solid-organ preservation by
cold storage. Transplantation 1988; 45(4): 673–676.
2. Englesbe MJ, Moyer A, Kim DY et al. Early pancreas transplant
outcomes with histidine-tryptophan-ketoglutarate preservation: A
multicenter study. Transplantation 2006; 82(1): 136–139.
3. Muhlbacher F, Langer F, Mittermayer C. Preservation solutions for
transplantation. Transplant Proc 1999; 31(5): 2069–2070.
4. Agarwal A, Murdock P, Pescovitz MD, Goggins WC, Milgrom ML,
Fridell JA. Follow-up experience using histidine-tryptophan ketog-
lutarate solution in clinical pancreas transplantation. Transplant
Proc 2005; 37(8): 3523–3526.
5. Fridell JA, Agarwal A, Milgrom ML, Goggins WC, Murdock P,
Pescovitz MD. Comparison of histidine-tryptophan-ketoglutarate
solution and University of Wisconsin solution for organ preser-
vation in clinical pancreas transplantation. Transplantation 2004;
77(8): 1304–1306.
6. Becker T, Ringe B, Nyibata M et al. Pancreas transplantation with
histidine-tryptophan-ketoglutarate (HTK) solution and University of
American Journal of Transplantation 2009; 9: 217–221
Reduced Pancreas Graft Survival with HTK
Wisconsin (UW) solution: Is there a difference? Jop 2007; 8(3):
304–311.
7. Potdar S, Malek S, Eghtesad B et al. Initial experience us-
ing histidine-tryptophan-ketoglutarate solution in clinical pancreas
transplantation. Clin Transplant 2004; 18(6): 661–665.
8. Mangus RS, Tector AJ, Fridell JA, Kazimi M, Hollinger E, Vianna
RM. Comparison of histidine-tryptophan-ketoglutarate solution
and University of Wisconsin solution in intestinal and multivisceral
transplantation. Transplantation 2008; 86(2): 298–302.
9. Agarwal A, Powelson JA, Goggins WC, Milgrom ML, Fridell JA.
Organ preservation with histidine-tryptophan ketogluatarate solu-
tion in clinical pancreas transplantation: An update of the Indiana
university experience. Transplant Proc 2008; 40(2): 498–501.
10. Schraibman V, Goldenberg A, Koh IH et al. Evaluation of se-
quential perfusion with Euro-Collins and Belzer solutions for pan-
creas preservation in rats. Transplant Proc 2006; 38(9): 2780–
2783.
11. Lee KW, Simpkins CE, Montgomery RA, Locke JE, Segev DL,
Maley WR. Factors affecting graft survival after liver transplantation
from donation after cardiac death donors. Transplantation 2006;
82(12): 1683–1688.
12. Lee KW, Cameron AM, Maley WR, Segev DL, Montgomery RA.
Factors affecting graft survival after adult/child split-liver transplan-
tation: Analysis of the UNOS/OPTN data base. Am J Transplant
2008; 8(6): 1186–1196.
13. Hesse UJ, Troisi R, Jacobs B, Berrevoet F, De Laere S, de
Hemptinne B. Pancreas preservation with HTK solution in the pig.
Transplant Proc 1997; 29(8): 3522–3523.
14. Hesse UJ, Troisi R, Jacobs B et al. Cold preservation of the porcine
pancreas with histidine-tryptophan-ketoglutarate solution. Trans-
plantation 1998; 66(9): 1137–1141.
15. Troisi R, Meester D, Regaert B et al. Physiologic and metabolic
results of pancreatic cold storage with histidine-tryptophan-
ketoglutarate-HTK solution (Custodiol) in the porcine autotrans-
plantation model. Transpl Int 2000; 13(2): 98–105.
16. Troisi R, Meester D, Van Den Broecke C et al. Functional and struc-
tural integrity of porcine pancreatic grafts subjected to a period
of warm ischemia and cold preservation with histidine-tryptophan-
ketoglutarate (custodiol) or University of Wisconsin solution. Trans-
plantation 2003; 75(11): 1793–1799.
17. Troisi R, Meester D, Regaert B, van den Broecke C, Cuvelier
C, Hesse UJ. Tolerance of the porcine pancreas to warm and
cold ischemia: comparison between University of Wisconsin and
histidine-tryptophan-ketoglutarate solution. Transplant Proc 2002;
34(3): 820–822.
18. Janssen H, Janssen PH, Broelsch CE. Celsior solution com-
pared with University of Wisconsin solution (UW) and histidine-
tryptophan-ketoglutarate solution (HTK) in the protection of human
hepatocytes against ischemia-reperfusion injury. Transpl Int 2003;
16(7): 515–522.
19. Janssen H, Janssen PH, Broelsch CE. UW is superior to Cel-
sior and HTK in the protection of human liver endothelial cells
against preservation injury. Liver Transpl 2004; 10(12): 1514–
1523.
221