TVRB Diagnostic Yield
Characteristics and diagnostic yield of the 80 patients’ TVRBs are reported in Table 2. For the TVRB, the puncture route was transjugular (right internal jugular vein) for 75 patients (93.8%) and transfemoral (femoral vein) for five (6.2%). The median ICU-admission-to-TVRB interval was 8.5 days (2–24 d). Renal cortex was obtained from 78 patients (97.5%) with a median of 14 (7–20) glomeruli. A histologic diagnosis was obtained for 77 patients (96.3%) with ATN being the most frequent: n = 23 (29.9%). Histologic diagnoses were distributed as follows: 29.9% ATN, 14.3% thrombotic microangiopathy, 13% lupus nephritis, 7.8% drug toxicity, 7.8% acute interstitial nephritis, 3.9% membranoproliferative glomerulonephritis, 3.9% antineutrophil cytoplasm antibody-associated vasculitis, 2.6% Goodpasture syndrome, 2.6% acute renal graft rejection, 2.6% diabetes, 2.6% hemosiderosis, and 1.3% hypertension. A potentially treatable cause was found for 47 patients (58.8%).
Patients’ TVRB Day Characteristics
Patient’s clinical and biologic characteristics and bleeding risk factors are reported in Table 3. TVRB day organ-failure treatments were: 75% renal replacement therapy, 28.7% mechanical ventilation, 13.8% vasopressor support, 7.5% ECMO, 3.8% left ventricular assist device, and 1.3% total artificial heart. Preventive and curative anticoagulation on TVRB day were administered, respectively, to 42.5% and 31.3% of the patients and were transiently interrupted for 64.7% and 60% of the patients. Antiplatelet therapy (mainly aspirin) was given to 27.5% of the patients at the time of TVRB and interrupted for none. At the time of TVRB, the numbers of patients with platelet count less than 150 G/L, less than 100 Giga (G)/L, and less than 50 G/L, respectively, were 35 (43.8%), 17 (21.3%), and five (6.3%). Considering antiplatelet therapy, thrombopenia (< 150 G/L) and preventive or curative anticoagulation as bleeding risk factors, seven (8.8%), 37 (46.2%), 31 (38.7%), and five (6.3%) patients, respectively, had 0, 1, 2, or 3 bleeding risk factors at the time of TVRB.
Complications and Factors Associated With Them
Complications, transfusions, and SOFA score or hemoglobin changes post-TVRB are reported in Table 4. Four (5%) and two (2.5%) patients, respectively, developed renal hematoma and macroscopic hematuria; none required any specific treatment. During days 1–3 post-TVRB, 28 patients (35%) required transfusions, mostly RBCs (median: two [2–2] packs). TVRB day and day-3 post-TVRB SOFA scores were, respectively, 6 (4–7.7) and 6 (4–7.2). The day-3 SOFA score varied for 56.1% of 66 patients: increasing for 12 (18.2%). No deaths were attributed to TVRB, and median TVRB-to-death interval was 38 days (19.7–86 d). Thirty-six patients (45%) had a composite complication endpoint. Univariable and multivariable analyses of factors associated with a composite complication endpoint post-TVRB are reported in Supplemental Tables 1 and2 (Supplemental Digital Content 1, http://links.lww.com/CCM/E333). Multivariable logistic-regression analyses retained vasopressor use at the time of TVRB (odds ratio [OR], 6.8; 95% CI, 1.2–37.4; p = 0.026) and platelet count on TVRB day less than 175 G/L (OR, 3.1; 95% CI, 1.1–8.7; p = 0.026) as independent predictors of a composite complication endpoint. There were no differences in term of diagnostic yield and complications when comparing TVRB performed less than or equal to 2012 (n = 38) or greater than 2012 (n = 42) (Supplemental Table 3, Supplemental Digital Content 1, http://links.lww.com/CCM/E333).
Therapeutic Consequences of TVRB
TVRB and its histologic diagnoses (Supplemental Table 4, Supplemental Digital Content 1, http://links.lww.com/CCM/E333) led to treatment modifications: introduction for 18 patients (22.5%), discontinuation for 11 (13.8%), and continuation for 22 (27.5%). Immunosuppressants initiated during the ICU stay were: corticosteroids for 34 patients (42.5%), corticosteroid pulses for 23 (28.8%), plasmapheresis for 21 (26.3%), cyclophosphamide for 12 (15%), and rituximab for two (2.5%).
Univariable analyses identified the main following factors, associated with an in-ICU TVRB-based ATN diagnosis versus other diagnoses, respectively: age (53.8 ± 17.2 vs 44.7 ± 18 yr; p = 0.04), admission-to-TVRB interval (17 d [8–62 d] vs 5 d [2–17 d]; p = 0.003), admission-to-TVRB interval/in-ICU days (0.56 [0.37–0.83] vs 0.37 [0.17–0.6]; p = 0.028), ICU-admission cardiovascular SOFA score (13 [56.5%] vs 16 [28.1%]; p = 0.02), and any in-ICU vasopressor use (17 [73.9%] vs 27 [47.4%]; p = 0.046). There were no differences in terms of bleeding risk factors, complications, and outcomes between both groups (Supplemental Table 5, Supplemental Digital Content 1, http://links.lww.com/CCM/E333).
Comparison Between Transplant and Nontransplant Patients
When comparing transplant versus nontransplant patients, univariable analyses identified the main following differences: admission-to-TVRB interval (32 d [12.5–62 d] vs 4 d [2–10 d]; p < 0.0001), admission-to-TVRB interval/in-ICU days (0.58 [0.43–0.7] vs 0.37 [0.18–0.6]; p = 0.02), any in-ICU vasopressor use (19 [90.5%] vs 25 [42.4%]; p < 0.0001), mechanical ventilation (19 [90.5%] vs 30 [50.8%]; p = 0.001) or ECMO support (14 [66.7%] vs 12 [20.3%]; p < 0.0001), the presence of two bleeding risk factors on TVRB day (13 [61.9%] vs 18 [30.5]; p = 0.01), and tendency toward more frequent composite complication endpoint (13 [61.9%] vs 23 [39%]; p = 0.08) (Supplemental Table 6, Supplemental Digital Content 1, http://links.lww.com/CCM/E333).
AKI is a severe condition that dramatically impacts the prognosis of critically ill patients. ATN is the leading cause of kidney failure in ICU patients and usually does not require renal biopsy for its management. However, for some patients with another AKI etiology, obtaining a renal biopsy for histologic examination can be crucial for entity diagnosis and treatment. PCRB carries a nonnegligible risk for ICU patients because many of them have contraindication(s) to it (hemodynamic instability, coagulation disorders, anticoagulation, antiplatelet therapy). However, TVRB, as an alternative to conventional renal biopsy, has been poorly investigated in critically ill patients.
The authors of two earlier studies specifically reported on renal biopsy in 77 and 56 ICU patients, respectively, from 10 and five French ICUs over 10 years (12 , 13). All but one biopsy in each study had been obtained percutaneously. Our patients were similar to theirs in terms of initial disease severity, defined by SAPS II and SOFA scores, but differed by being younger (mean age 47.3 vs 62.3 and 60 yr), having more frequent preexisting CKD (30% vs 17% and 16%), being mixed medical and surgical patients requiring several mechanical cardiopulmonary supports and more having received organ transplants. Our patients’ in-ICU mortality (7.5% vs 22% and 23%) and dialysis-dependence at discharge (25.6% vs 55% and 43%) were less frequent, probably reflecting our center’s specificities (younger patients, high access to cardiopulmonary supports, and organ transplantation).
Herein, the number of diagnosis-contributive TVRBs and the median number of glomeruli were high and comparable to those of the PCRB studies. A specific diagnosis was obtained for 96% of the patients, most frequently ATN (29.9%), with a potentially treatable cause for ~ 60% of them. The frequency of ATN in our patients was concordant with the data from literature. Indeed, ATN was identified in 15 of 78 (19%) and 26 of 56 (46%) patients in the two previous study on kidney biopsy in ICU. Furthermore, in a postmortem renal biopsies study of critically ill children, ATN was the most frequent lesion (n = 19/62 [30.6%]) (14). One of the identified ATN-associated factors was the long admission-to-TVRB interval, independently of time spent in the ICU. Conversely, for patients with a treatable cause of AKI, admission-to-TVRB interval (4 d [2–17 d] vs 13 d [4–62 d]; p = 0.001) was significantly shorter and admission-to-biopsy interval/in-ICU days ratio was significantly lower (0.31 [0.16–0.61] vs 0.56 [0.41–0.81]; p = 0.002) (data not reported). Those findings suggest that TVRB is of particular interest early during the ICU stay, whereas, later during that stay, TVRBs from patients without recovery from AKI more frequently contained ATN lesions. Pertinently, even a diagnosis of ATN or chronic noninflammatory lesions can be determinant for the management of an ICU patient, as it can: prevent administration of deleterious inappropriate immunosuppressants, accelerate the start of chronic dialysis, help decide whether or not a combined transplantation is required.
As reported above, numerous patients’ TVRB histologic findings led to treatment changes. Even though the progress of noninvasive diagnostic tests (e.g., antineutrophil cytoplasm, anti-glomerular basement membrane, antinuclear, anti-DNA, a disintegrin and metalloprotease with thrombospondin type-I repeats-13 antibodies) allows treatment of systemic rheumatic disease without histologic proof, renal biopsy retains an important place in evaluating prognosis and determining appropriate treatment intensity. Therefore, assessment of the clinical impact of TVRB on the management of an ICU patient is difficult and should not be limited to the introduction, continuation or interruption of immunosuppressant(s).
The two previous studies provided little information about their patients’ coagulation disorders, anticoagulation or antiplatelet agent(s), which are contraindications for PCRB, and probably led to exclusion of patients receiving such treatments. Most of our patients had at least one bleeding risk factor(s) (91.3%), and 45% cumulated two or three. Anticoagulation was temporarily interrupted for a few hours to minimize the bleeding risk but many patients were biopsied under their full anticoagulation doses, sometimes combined with aspirin, because anticoagulation could not be safely stopped (left ventricular assist device, antiphospholipid syndrome). Despite those multiple bleeding risk factors, only six of our patients (7.5%) experienced a clinically significant TVRB complication, none requiring specific management or causing death. Conversely, 17 (22%) (12) and 7 (12.5%) (13) ICU patients who underwent PCRB experienced a complication, with two in each study requiring arterial embolization and one was fatal. Nevertheless, a significant number of our patient (35%) required a blood transfusion during the 3 days after TVRB. Even though, all transfusions were not necessarily linked to the procedure itself and could be associated with other factors (as being under ECMO).
Owing the retrospective nature of this work, to maximize the identification of TVRB-related complications, we constructed a composite complication endpoint comprising, during days 1–3 post-TVRB, any transfusion, or SOFA or hemoglobin changes. The objective of this composite score was to be in a “maximum bias hypothesis” in case some biopsy-related adverse events have gone unnoticed during the ICU stay. Although those endpoint items are frequent in unstable, critically ill patients, herein they were not necessarily linked to TVRB itself. Vasopressor use and low platelet counts, but not the number of bleeding risk factors, were independently associated with a composite complication endpoint. Pertinently, those factors should not, in our opinion, preclude TVRB if it is needed to improve patient management. Notably, TVRB complications in critically ill patients were infrequent and rarely serious, even in the presence of multiple bleeding risk factors.
When comparing transplant patients versus nontransplant patients, we found no differences in term of TVRB diagnostic yield. Although transplant patients had more frequent and severe circulatory failure, ATN frequency was similar between both groups. As transplant patients had more frequently two bleeding risk factor on TVRB day, there was a tendency toward more frequent composite complication endpoint, through a higher number of RBCs transfusions. Altogether, those results suggest that TRVB diagnostic yield is good in transplant patients but may be at the cost of higher RBCs transfusions.
Augusto et al obtained PCRBs from 8.4 of 10,000 patients admitted to the ICU during their study period, whereas our TVRB frequency was 73 of 10,000. This more extensive experience underscores TVRB safety in critically ill patients and highlights some unmet needs of ICU patients’ diagnoses based on PCRB histology.
Our study has several limitations. First, it is a single-center study with a retrospective design. Second, TVRB was the only biopsy technique used, as part of our local protocol, regardless of the presence or absence of contraindication(s). Third, the study population is very particular with young patients, high numbers of transplantees, connective tissue diseases, or vasculitides and patients receiving mechanical cardiopulmonary support. Fourth, there were no criteria for systematic TVRB in our ICU during the time of the study. The indication for TVRB was retained on a patient-basis after weighting the benefice-risk ratio during a multidisciplinary discussion. The population of our study is therefore highly selected, and the findings of our study may not be relevant for any ICU patient. Last, TVRB safety and efficacy in our center relies strongly on the remarkable dexterity and experience of our interventional radiologists, and those results may not be immediately reproducible in every center.
TVRB use to investigate AKI in selected ICU patients was shown to be safe, with high diagnostic yields for in critically ill patients, even in the presence of multiple bleeding risk factors.
1. Hoste EAJ, Kellum JA, Selby NM, et al. Global epidemiology and outcomes of acute kidney injury
. Nat Rev Nephrol 2018; 14:607–625.
2. Nisula S, Kaukonen KM, Vaara ST, et al; FINNAKI Study Group: Incidence, risk factors and 90-day mortality of patients with acute kidney injury
in Finnish intensive care units: The FINNAKI study. Intensive Care Med 2013; 39:420–428.
3. Srisawat N, Sileanu FE, Murugan R, et al; Acute Kidney Injury
-6 Study Group: Variation in risk and mortality of acute kidney injury
in critically ill
patients: A multicenter study. Am J Nephrol 2015; 41:81–88.
4. Hoste EA, Bagshaw SM, Bellomo R, et al. Epidemiology of acute kidney injury
in critically ill
patients: The multinational AKI-EPI study. Intensive Care Med 2015; 41:1411–1423.
5. Cohen AH, Nast CC, Adler SG, et al. Clinical utility of kidney biopsies in the diagnosis and management of renal disease. Am J Nephrol 1989; 9:309–315.
6. Madaio MP. Renal biopsy. Kidney Int 1990; 38:529–543.
7. Striker GE. Controversy: The role of renal biopsy in modern medicine. Am J Kidney Dis 1982; 1:241–243.
8. Mal F, Meyrier A, Callard P, et al. Transjugular renal biopsy
. Lancet 1990; 335:1512–1513.
9. Mal F, Meyrier A, Callard P, et al. The diagnostic yield of transjugular renal biopsy
. Experience in 200 cases. Kidney Int 1992; 41:445–449.
10. Sofocleous CT, Bahramipour P, Mele C, et al. Transvenous transjugular renal core biopsy with a redesigned biopsy set including a blunt-tipped needle. Cardiovasc Intervent Radiol 2002; 25:155–157.
11. Cluzel P, Martinez F, Bellin MF, et al. Transjugular versus percutaneous renal biopsy for the diagnosis of parenchymal disease: Comparison of sampling effectiveness and complications. Radiology 2000; 215:689–693.
12. Augusto JF, Lassalle V, Fillatre P, et al. Safety and diagnostic yield of renal biopsy in the intensive care unit
. Intensive Care Med 2012; 38:1826–1833.
13. Philipponnet C, Guérin C, Canet E, et al. Kidney biopsy in the critically ill
patient, results of a multicentre retrospective case series. Minerva Anestesiol 2013; 79:53–61.
14. Rameshkumar R, Krishnamurthy S, Ganesh RN, et al. Histopathological changes in septic acute kidney injury
in critically ill
children: A cohort of post-mortem renal biopsies. Clin Exp Nephrol 2017; 21:1075–1082.
acute kidney injury; acute tubular necrosis; critically ill; intensive care unit; transjugular renal biopsy; transvenous renal biopsy
Supplemental Digital Content
Copyright © by 2019 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.