Introduction

Lung transplantation has increased in the last decade, and is now the therapeutic option of choice for patients in the terminal phase of their illness. However, the supply of grafts cannot meet current demand, and waiting list mortality is high [1]. Only 15 % of brain death donors are lung donors, making lung graft maintenance an important field of research [2, 3]. Various strategies have been proposed to increase the pool of donors based on pulmonary hemodynamic monitoring, ventilatory management with recruitment maneuvers, and hormone treatment [4]. The hemodynamic approach seeks to achieve normovolemia, maintain blood pressure and cardiac output, and ensure perfusion pressure and blood flow to different organs with less use of vasoactive drugs [5, 6]. This is accomplished using the Pulse Contour Cardiac Output (PiCCO) system (Pulsion Medical Systems SE, Munich, Germany), a hemodynamic monitoring tool that combines pulse contour techniques with transpulmonary thermodilution. In a single-center study, later validated in a multicenter project, our group has shown that these measures combined with a donor treatment protocol are effective in increasing lung donation [7, 8].

The multicenter study also showed that introduction of this treatment strategy for thoracic donors is within the budget of any hospital and does not affect donation of other organs: thus, the rate of graft procurement remained the same in all centers. This protocol includes close monitoring to achieve hemodynamic objectives and controlling fluid balance with diuretics when necessary, and may be contrary to conventional donor kidney management strategies. Fluid therapy seems to help maintain kidney function, although fluid balance and/or restriction have been associated with more successful lung grafts. It has been suggested that a high cardiac preload can improve kidney graft function, although this was not confirmed in prospective studies [9, 18]. After having validated a specific thoracic donor therapy to increase lung donor rates, we analyzed the effect of this strategy on kidney donor rates. We also analyzed the impact of the protocol on long- and short-term renal allograft survival. To the best of our knowledge, our group has pioneered this approach.

Methods

Patients

In January 2009, we introduced a protocolized donor treatment strategy in the University Hospital “Marqués de Valdecilla” in Santander (Spain). The cornerstones of our approach were: hemodynamic support, respiratory support, and hormone treatment. Table 1 shows the characteristics of the protocol [8].

Table 1 Donor management protocol [8]

All patients undergoing kidney transplantation from 2009 to 2013 using organs from donors who were also lung donors and had been treated in our intensive care unit (ICU) were included in the study. These patients constituted the protocol cohort. All kidney transplants including combined transplants (kidney-pancreas transplantation) and re-transplantation were considered. We also analyzed outcomes in patients receiving kidney grafts from 2003 to 2008 from donors that were also lung donors (pre-protocol cohort). Kidney acceptance criteria remained the same during both periods.

A kidney biopsy was performed in all expanded criteria donors, and a ~20 % glomerulosclerosis was considered as the criterion for discarding the organs [10, 11]. Our nephrologists established that percentage given that it has been shown in several series that ~20 % glomerulosclerosis is associated to the presence of delayed graft function, and also to a reduced kidney function [12]. Kidneys obtained from our lung donors but ultimately transplanted in other centers were excluded in order to avoid the “center effect.”

Transplants were performed by the same surgeons during both periods, and no changes were made to post-transplant care. The immunosuppression protocol used in both periods was similar and initially consisted of steroids, tacrolimus and mycophenolate. The target trough levels of tacrolimus in the first postoperative month was 10–15 ng/ml for standard donors and 8–10 ng/ml for patients receiving organs from expanded criteria donors. Recipients of organs from expanded criteria donors received induction therapy with basiliximab. In patients at high immunological risk (human leukocyte antigens peak anti-[anti-HLA] antibodies >50 % or those who had lost a previous kidney through acute rejection), thymoglobuline was added. The management of kidney transplant patients in both periods was similar: all patients received prophylaxis against Pneumocystis jirovecii for 6 months post-transplantation. For cytomegalovirus (CMV)-negative recipients whose donors were CMV-positive, valganciclovir was prescribed for 3 months. Patients who received thymoglobuline also received a 2-week course of intravenous ganciclovir. Monitoring and treatment of antibody-mediated rejection has remained the same since 2006, with C4d screening in all renal allograft biopsies and treatment with intravenous immunoglobulin and plasmapheresis.

The pre-protocol cohort data were extracted from the kidney transplantation and organ donor database compiled in our hospital. The information from both databases was crossed-referenced anonymously using the identification number of the donor.

Outcome variables

The main outcome variables were rate of kidney donation and kidney graft survival at 3 and 5 years. The development of delayed graft function and evolution of creatinine levels in both periods were also evaluated. Delayed graft function was considered as constituting the need for dialysis in the first week. The follow-up continued until December 31, 2013.

Statistical analysis

In the descriptive study, categorical variables are expressed as absolute values and percentages, and quantitative variables as mean ± standard deviation or median and interquartile range in the case of asymmetric variables. The t test or Mann–Whitney U test were used to compare quantitative variables between treatment groups. For categorical variables, the Chi-squared or Fisher’s exact test were used. The general linear model for repeated measures was used to evaluate creatinine measurements over time. Logistic regression models were used to identify risk factors associated with the development of delayed graft dysfunction. Risk factors associated with graft survival were identified using Cox regression. Survival was evaluated using Kaplan–Meier survival analysis, considering functional status recorded at the end of the observation period. Since survival of same-donor grafts may be related, robust variance estimators were used, using the donor as a cluster. Statistical significance was set at p < 0.05. Statistical analysis was performed using SPSS for Windows version 11.0.

Results

In the pre-protocol period (6 years), 29 lung donors (4.8 donor lungs/year) were procured, while in the protocol period (5 years) 72 lung donors (14.4 donor lungs/year) were obtained. Since the introduction of the multi-organ donor treatment protocol, 50 % of donors in our hospital have been lung donors. Table 2 shows the characteristics and ICU management of donors during both periods. It is important to note that all donors were Caucasian, brain-dead organ donors.

Table 2 Donor characteristics

During the pre-protocol period, kidney grafts were obtained in 86.2 % of donors (50 kidney grafts from 29 donors). The donation rate in the protocol period was also 86.2 % (100 kidney grafts from 58 donors) with no significant differences in the pre-protocol period (p > 0.05). A total of 44 grafts from the pre-protocol cohort and 82 grafts from the protocol cohort were transplanted in our hospital. The remaining kidney grafts were transplanted in other hospitals. In the protocol period, kidney donation was ruled out for a number of reasons, none of which was related to renal failure associated with patient management. During the protocol period, kidney grafts were discarded for a number of reasons, but never due to acute renal failure caused by our aggressive donor management protocol. During this period, both kidneys were unsuitable for transplantation in only 5 donors: 3 died from cardiac arrest during surgery before removal of organs; liver cancer was detected in 1 donor, so no grafts were obtained; finally, in a fifth donor (24 years) kidney transplant was ruled out due to medical reasons (neurogenic bladder due to spina bifida, urinary tract infection and chronic renal failure). A further 8 patients donated only 1 kidney due to structural causes or glomerular sclerosis diagnosed by intraoperative biopsy. A similar situation was found in the pre-protocol cohort, where both grafts were discarded in 3 donors due to cardiac arrest, while in another donor, 1 kidney only was obtained and transplanted and the other one was unsuitable for transplantation (glomerulosclerosis). The kidney discard rate was similar in both periods (protocol cohort vs. pre-protocol cohort): 18/144 versus 7/58.

With regard to kidney recipients, we observed that the protocol cohort was significantly older and with more comorbidities than the pre-protocol cohort. Table 3 shows the characteristics of kidney recipients. It is important to note that in the protocol cohort the number of expanded criteria donors was significantly higher than in the pre-protocol cohort: 26 (31.7 %) versus 5 (11.4 %); p = 0.011. Similarly, the kidney donor risk index (KDPI) was significantly higher in the protocol cohort: 55.6 ± 24.6 versus 30.6 ± 25; p < 0.001. Donor creatinine levels and the need for inotropic support were similar in both groups. Serial serum creatinine levels were significantly different in recipient groups (p = 0.001). At all time points, serum creatinine was statistically higher in the protocol cohort (Fig. 1). Multiple linear regression showed that expanded criteria donors explained the differences detected between creatinine levels at each time point, given that this difference did not remain significant after adjustment for that variable (p > 0.05).

Table 3 Kidney recipient characteristics
Fig. 1
figure 1

Sequential mean serum creatinine levels in kidney recipients in the two periods. A general linear model for a repeated measures test was used. Error bars represent the standard deviation (SD) of the mean for each time point. At all time points p < 0.05. D days, yr years

Although no significant differences in cases of delayed graft function were observed, it is important to note that there were more cases in the protocol cohort (28 % vs. 15.9 %; p = 0.127). However, neither univariate [odds ratio (OR) 2.1 (95 % confidence interval 0.8–5.3); p = 0.132] nor multivariate [OR 1.8 (0.7–4.9); p = 0.223] analysis, adjusted for recipient age and donor characteristics (expanded criteria donors), suggests that the protocol cohort had a greater risk of delayed graft dysfunction than the pre-protocol group. Asymmetric distribution of expanded criteria donors between groups of recipients was a confounding factor.

In the protocol period, some kidneys were transplanted from donors in whom this protocol was not applied (no donated lung). Of these, 23 % presented DPI, 28 % presented acute rejection, and 85 % had a functioning graft at the time of study closure. Similar outcomes were observed in kidney transplants during the same period from donors managed using the thoracic protocol (no significant differences between groups).

At the end of the study, the graft was lost in 16 patients from the pre-protocol period, and in 11 from the protocol period. Four patients with a functioning graft (1 from pre-protocol cohort and 3 from the protocol cohort) died of non-kidney-related causes. Risk of kidney graft failure was similar in both the protocol cohort and pre-protocol controls [hazard ratio (HR) 0.77 (0.33–0.8); p = 0.570]. The trend did not alter after adjusting for expanded criteria donors [HR 0.7 (0.3–1.8); p = 0.554].

On analyzing graft survival probability, we found that that this was similar in both periods (Fig. 2), with no statistically significant difference (χ² = 0.325, df = 1; 0.569), giving a probability of graft survival in the protocol cohort of 0.85 (0.76–0.94) at 3 years post-transplant. The trend in the pre-protocol group was similar, with a probability of surviving kidney graft of 0.84 (0.80–0.93) at 3 years post-transplant. Considering survival at 1 year, the results were also similar between groups. The probability of surviving kidney graft at 1 year post-transplant was 0.91 (0.87–0.96) in the historical group and 0.93 (0.88–0.99) in the protocol cohort. Lung transplant outcomes were similar between groups (data not shown).

Fig. 2
figure 2

Survival of kidney grafts

Discussion

Much can be done to improve specific maintenance protocols for lung donors, as they account for only 16.1 % of all organ donors [3]. Several scientific societies and consensus guidelines recommend potential lung donor treatment based on 3 pillars: hemodynamic, respiratory and hormonal [4, 6]. In both a single- and multicenter-center validation study, our group has shown that simultaneous use of these 3 strategies increases lung donation rates [7, 8] and that survival of transplanted lungs was comparable to rates observed prior to the introduction of the protocol [15].

One of the cornerstones of this protocol is hemodynamic management based on PiCCO system monitoring to achieve fluid balance using restrictive fluid therapy and/or diuretics. This approach may be contrary to conventional recommendations for renal donor maintenance based on fluid intake to achieve high renal perfusion pressure and abundant diuresis. However, the results of this study show that kidney donation rates were identical in both groups of patients, showing that a more restrictive blood volume did not affect donation. In the protocol period, some renal grafts for transplantation were ruled out, although this was not related to renal failure due to the treatment donor protocol. Moreover, recent good clinical practice guidelines on the assessment of renal donors and guidelines to implement strategies to increase the pool of kidney donors make no mention of high volemia donor management, a fact that supports our conclusions [16, 17].

Donor maintenance variables such as ischemia or use of fluids have been associated with delayed graft function (DGF) [1821]. However, our DGF series was similar to findings reported by other authors [22, 23]. There were no significant differences between groups of patients, but the rate of DGF was higher in the protocol cohort even though donors and recipients from the protocol period were not comparable in terms of age or comorbidities to donors and recipients from the pre-protocol cohort (Tables 2, 3). We consider that the worse characteristics of donors and recipients in the protocol cohort are the main cause of this finding. In fact, these comorbidities in expanded criteria donors would explain serum creatinine level trends (which were higher in the protocol cohort than in the pre-protocol group), since this variable has a bearing on linear regression. Sequential creatinine levels were as expected, given the characteristics of both groups of patients (the protocol cohort had the highest proportion of expanded criteria donors). Despite that, the long-term survival of these kidney grafts was similar to that observed in the pre-protocol series with better quality donors. These survival rates are comparable to those reported by other groups [24]. Our group was the first to evaluate short-term kidney grafts from donors treated with fluid restriction therapy, and the results of this study are consistent with the findings of a previous pilot study conducted by our group, which showed that negative or equalized fluid balance with a central venous pressure (CVP) <6 mm Hg did not affect short-term kidney graft survival or incidence of DGF [25]. Recipient and donor conditions remained unchanged, but management differed in the two periods. This would explain why, even though creatinine levels were higher in the transplant protocol cohort, the survival of these long-term kidney grafts was similar to that observed in the pre-protocol series with better quality donors. The more proactive attitude needed to implement this protocol coupled with extensive monitoring strategies could contribute to the results observed in our study. Our study shows that careful application of this protocol on poorer quality donors optimizes oxygenation and preserves the viability of the kidney, without affecting long-term graft survival.

Our study has some limitations. It is a single-center study with a small sample size; however, we consider that our approach should increase the number of harvested lungs without affecting other grafts such as the kidney. Nevertheless, we believe our results should be validated by other groups, and a prospective randomized trial to investigate donor management using this protocol is needed.

We consider that the strict hemodynamic monitoring involved in this donor management strategy requires donors to be monitored as carefully as critical patients, and should therefore be performed only by qualified intensive care personnel. Addressing the clinical complexity and managing potential donor brain death requires knowledge, experience and training. Studies have shown that donor management by experienced physicians is associated with an increase in valid grafts for transplantation [26]. In this context, invasive beat-to-beat hemodynamic monitoring is mandatory (PiCCO systems) [2]. This, along with other indicators of pulmonary edema, permits real-time bedside management to maintain euvolemia, reduce fluid intake, and minimize risk of pulmonary edema, thereby increasing lung donation and preserving other grafts such as the kidney [7, 27, 28].