This retrospective multicentre observational study was conducted at 6 French adult renal transplantation centres: Gabriel Montpied Hospital, Clermont-Ferrand University Hospital; Michallon Hospital, Grenoble University Hospital; Necker-Enfants Malades Hospital, AP-HP; Edouard Herriot Hospital, Lyon University Hospital; Rangueil Hospital, Toulouse University Hospital; and Nouvel Hospital Civil, Strasbourg University Hospital.

The inclusion criteria were patients aged more than 18 years who had undergone renal transplantation between 01/01/1997 and 31/12/2020 complicated by venous thrombosis of the graft during the initial hospitalization (<14 days). To avoid selection bias, we submitted a request to the Biomedicine Agency database with the following terms: “vascular complications” and/or “no primary function.” We then checked all medical records and included only patients with early vein thrombosis of the allograft.

We collected the following demographic and clinical characteristics of the donor from the Biomedicine Agency’s prospective database CRISTAL and possibly from the patient’s file: type of donor (living or deceased), age, so-called “marginal” donor with extensive selection criteria (17), presence of thromboembolic risk factors, year of transplantation, laterality of the kidney, possible anatomical abnormalities, and conditions of retrieval. We also collected the following demographic, clinical and biological data of the recipient: age, sex, body mass index, thromboembolic history, haematological pathologies, history of miscarriage, smoking, diabetes, initial renal disease, presence of pretransplant anticoagulant or antiaggregant treatments, and induction immunosuppressive treatments.

Intraoperative graft data were collected from operative reports. We distinguished the type of graft (bitransplant, multiorgan transplant, or renal transplant alone) and the duration of cold and warm ischaemia; from the operative reports, we identified any difficulties that occurred during vascular anastomoses and the flushing or non-flushing of the vessels intraoperatively. We also noted signs suggestive of renal graft vein thrombosis, whether clinical and/or biological, and imaging studies allowing us to confirm this, as well as the management, the functional results in the long term, and the complications secondary to this management.

Of the 14,319 renal transplants (RTs) performed at these 6 centres in the period from 01/01/1997 to 12/31/2020, 182 (1.3%) patients presented with renal graft vein thrombosis during the initial transplant hospitalization. We excluded 2 patients with partial thromboses, 2 patients for whom the diagnosis was uncertain, and 2 patients because of a lack of data (Figure 1).

Patients were divided into 2 groups. Group 1 (G1; N = 16) included patients with intraoperative renal vein thrombosis. Group 2 (G2; N = 160) included patients with a postoperative diagnosis of eRVTPT.

Statistical analysis was essentially exploratory to describe management strategies in the event of eRVTPT. Data are presented as the mean and standard deviation or the median and interquartile range. The assumption of normality distribution was studied with the Shapiro–Wilk test. The chi-square test or, if appropriate, Fisher’s exact test was used to compare independent groups concerning graft outcomes, such as the proportion of patients with a functional graft rate at hospital discharge or the time to thrombosis. Statistical analysis was performed using the Stata software (version 15, StataCorp, College Station, Texas, US). The statistical tests were two-sided, with type I error set at 0.05. We performed a study of the factors associated with functional grafts at discharge using generalized linear mixed models with a logit link function to model between- and within-centre variability (as random effects). For multivariate analyses, we performed a multiple mixed logistic regression that considered covariables in terms of their significant results in univariate analysis (p < 0.10) (Table 1) as well as their clinical relevance (2,18–20). The results are expressed in terms of odds ratios (ORs) and 95% confidence intervals (95% CIs).

The characteristics of the patients are given in Table 1. The 176 patients included in our study were predominantly male (56.2%), aged 56.5 ± 10.0 years at transplantation, and had a mean body mass index of 26.2 ± 3 kg/m². In this cohort, thirty-one patients were diabetic (17.2%), and 16 (9.5%) were active smokers. A history of thrombosis was reported in 34 patients (19.0%), 12 of whom had multiple thromboses (6.8%). Haematological, haemostasis or immunological pathology posing a risk of thrombosis was found in 9.1% of cases (N = 16). Thirty-eight patients (21.6%) were receiving antiaggregation therapy, and 13 (7.3%) were receiving anticoagulant therapy (Table 1).

In the vast majority of cases, patients received a first kidney transplant (N = 144, 84.2%) from a deceased brain-dead donor (N = 166, 94.3%). The median donor age was 60 [46–71] years; 51.2% of the grafts (85/166) were considered marginal (Table 2).

Among the patients who received a kidney transplant from a living donor, 1 received a right kidney, 3 others received a left kidney with a short and thin vein and 1 with 2 veins that had been ligated. eRVTPT was more frequent in patients who received a right kidney than in those who received a left kidney during the interest period (101/6421 (1.57%) patients vs. 74/7797 (0.95%) patients, p = 0.002).

The median cold ischaemia time was 918 [714–1259] minutes. The median warm ischaemia time to perform vascular anastomosis was 45 [34–55] minutes. In 23 recipients (13.4%), the venous anatomy of the graft was abnormal (11 multiple veins, 10 venous wounds at harvesting, and 2 dysplastic veins). In addition, seven operators had to redo the venous anastomosis during the transplantation procedure (4.4%). During the initial operative procedure, 37 patients received a flush solution with heparinized saline or normal saline (27.4%), and 13 patients (9.4%) received heparin therapy (Table 2).

Half of the diagnoses of intraoperative RT vein thrombosis were made at one centre. All patients with intraoperative eRVTPT at this centre had a functional graft at discharge. This team frequently used intraoperative ultrasound during kidney transplant procedures to assess graft vascular anastomoses and flow. Recently, another centre introduced this technique and diagnosed one case of intraoperative eRVTPT with a favourable outcome.

In the majority of cases, venous thrombosis was symptomatic (86.2%). It manifested as oligoanuria (63.1%), abnormal pain in the renal pelvis (26.9%) and frank haematuria (17.5%). Venous thrombosis was revealed by haemodynamic disorders, such as the use of vasopressor amines in 14 patients (8.6%) or haemorrhagic shock in 7 other recipients (4.3%). The main biological criterion leading to the diagnosis was increased creatinine levels (81/160; 50.6%). The vast majority of patients underwent imaging to confirm graft vein thrombosis: 68/160 (42.5%) by graft Doppler ultrasound, 22 (13.8%) by abdominopelvic computed tomography angiography, 61 (38.1%) by both aforementioned modalities, and 6 (3.7%) by ultrasound combined with MRI (Table 3).

In group 1, 10/16 (62.5%) intraoperative venous thromboses were treated surgically by repair of the venous anastomosis (Figure 1), and 6/16 (37.5%) were treated by venotomy for thrombectomy. Four (25.0%) patients underwent intraoperative transplantectomy, and 1 (6.3%) underwent secondary transplantectomy (Figure 1).

In group 2, the time to onset of venous thrombosis was 48 [24–120] hours (Table 3) after transplant surgery. Management was almost exclusively surgical (159/160, 99.4%), with the exception of one patient who received heparin therapy alone (Figure 1). In 94/159 patients (59.1%), transplantectomy was performed immediately because of a necrotic renal graft. For the other 65 patients (40.9%), the 2 main surgical techniques used were venotomy and thrombectomy (N = 37) or thrombectomy added to venous anastomosis repair (N = 22). For one patient, the type of surgical revision was not specified (Figure 1). Surgical revision was accompanied by primary procedure failure in 23/159 (14.5%) cases with intraoperative transplantectomy. For 42 patients (42/159; 26.4%), surgical revision was said to be satisfactory because of macroscopically satisfactory revascularization of the graft (Figure 1). However, 22/42 (52.4%) patients required a secondary transplantectomy during the initial stay, and only 13/42 (31.0%) had a functional graft vs. 10/11 (90.9%) in G1 (p = 0.001). Among the 7 patients discharged with a non-functional graft in place, 2 died in the weeks that followed, 3 benefited from a transplantectomy several months after the transplant for graft intolerance syndrome, 1 benefited from graft embolization, and the last patient kept the graft in place.

In 96 patients (62.8%), a pathology report of the allograft nephrectomies was available for analysis and confirmed the renal infarction due to renal thrombosis. None of the patients had signs of acute rejection.

The majority of patients in our cohort had significant blood loss defined as the need for at least one red blood cell transfusion (N = 131; 74.4%). Among them, 5 patients (2.8%) presented haemorrhagic shock, and 29 developed large haematomas (29/176; 16.5%). Thirty-one patients presented concomitant deep vein thrombosis and/or pulmonary embolism (Table 3). All received curative anticoagulant treatment that may contribute to significant blood loss. Thirty-eight patients presented a postoperative infection (Table 4).

Five patients (5/176; 2.8%) died during initial management, including 4/176 (2.3%) within the first 15 days of transplantation. In G1, 1 patient died on day 4 from haemorrhagic shock. Four patients died in G2: 1 in the operating room from haemorrhagic shock during transplantectomy, 1 on day 6 from anaemia and hyperkalaemia, and the other 2 on day 11 and day 61 because of multivisceral failure, preceded by multiple repeat operations (Table 4).

Overall, the proportion of patients with a functional graft at discharge was 13.1% (N = 23). The proportion of patients with a functional graft at discharge from the hospital in the G1 group compared with the G2 group was 10/16 (62.5%) and 13/160 (8.1%), respectively (p < 0.001). A sensitivity analysis excluding patients who underwent dual kidney transplantation (N = 8) or multiorgan transplantation (N = 5) exhibited a similar rate of remission (data not shown). In multivariate analysis, patients who had a postoperative diagnosis of eRVTPT had a lower probability of having a functional graft at discharge (OR = 0.016, 95% CI [0.002; 0.119], p < 001).

All these grafts were also functional at 1 year. The median serum creatinine at 1 year was 155 [130–207] µmol/L, with similar values in the 2 groups (data not shown).

At 5 years, 16 patients had a functional graft, 2 patients were dialysed, 1 patient died, and 4 were lost to follow-up. In addition, 44/153 (29.1%) of the patients who lost their graft were able to receive a new transplant. None of the patients had thrombosis of their new graft (Table 4).