Acute renal failure (ARF) is a serious complication following cardiac operations performed with cardiopulmonary bypass (CPB) and carries a high mortality rate . Prior studies have attempted to identify predictors of ARF or to develop risk stratification algorithms. Valve operations are an independent predictor of ARF [2,3]. To our knowledge, no study has attempted yet to find an independent association between ARF and a specific type of valve operation (i.e. repair or replacement).
The aim of this study is to define the incidence, predictors and mortality related to ARF after different types of mitral valve (MV) surgery.
From January 1998 to January 2003, we studied 1276 consecutive adult patients who underwent isolated MV surgery with CPB. We excluded from our population patients with dialysis dependence and combined procedures. MV repair was mostly performed with the edge-to-edge technique with the positioning of a rigid ring while patients who had mitral replacement mostly had a mechanical valve implanted.
All patients received a standard premedication and monitoring. Anaesthesia was induced in all patients with fentanyl–propofol–pancuronium, maintained with propofol, isoflurane, and additional doses of fentanyl. CPB was conducted with an institutional custom pack including a coated membrane oxygenator, with mild hypothermia (32–33°C). Non-pulsatile perfusion was used throughout the study, with perfusion maintained between 2 and 2.8L min−1 m−2. The pumps were primed with crystalloid solution, mannitol 18% 0.5g kg−1 formulated to achieve a haematocrit of 18% or more during CPB. Packed red blood cells were added to achieve the desired haematocrit and as required by the clinical circumstance. Intermittent cold (4°C) blood cardioplegia was infused by means of heat exchanger and two roller pumps.
Perioperative ventricular dysfunction occurring after cardiac surgery and CPB was managed with heart rate and rhythm control, preload and afterload optimization and, when these manoeuvres were ineffective, with inotropic drugs. Dopamine was the first sympathomimetic drug used and when it was unable to resolve the low-output syndrome it was coupled with epinephrine and/or enoximone and/or an intra-aortic balloon pump. Prophylactic use of a balloon pump was performed in patients with functional MV disease and with ejection fraction <30%.
Prophylactic strategies such as hydration, dopamine, fenoldopam and mannitol were used according to anaesthesiologist's preference. Loop diuretics have been administered early in the course of ARF to convert an oliguric to a nonoliguric state. ARF was defined as a postoperative 100% increase in serum creatinine (doubling from baseline values). Renal replacement therapy was initiated by the attending nephrologist and intensivist based on the clinical situation. Renal support was provided in all cases by continuous veno-venous haemofiltration (CVVH, Prisma CFM, Hospal Lyon, France) using high flux AN69 membranes with a membrane surface of 0.60 m2.
Statistical analysis. Data were analysed using the SAS statistics package. Dichotomous variables were compared using χ2-test with Yates correction. Continuous measures are expressed as mean + SD unless otherwise indicated and were compared with a t-test for paired or unpaired data, as appropriate. A multivariate stepwise logistic regression was used to assess the independent correlates of ARF.
Of 1276 patients included in the study, 32 (2.5%) developed postoperative ARF. The incidence of ARF for MV replacement and MV repair was 8% (25/312) and 0.7% (7/964) respectively (P < 0.001). The overall population was 57 ± 12.8-yr-old, 41% (528) female, 6.3% (81) with ejection fraction <40%. All perioperative clinical and patient characteristics are depicted in Table 1 together with a univariate analysis to study their association to ARF: patients who developed ARF had a higher incidence of preoperative comorbidities and perioperative complications.
At a multivariate analysis that included all factors in Table 1 with entry and exit values P < 0.05, MV replacement was an independent risk factor for the development of postoperative ARF (odd ratio (OR): 4.0, 95% confidence interval (CI): 1.49–10.59, P < 0.01) together with low-output syndrome (OR: 13.7, 5.4–34.9, P < 0.01), emergency sur-gery (OR: 8.5, 1.4–52.2, P = 0.02, creatinine > 124 μmol L−1 (OR: 7.9, 2.8–22, P < 0.01), reopening for bleeding (OR: 4.5, 1.4–14.3, P = 0.01), diabetes (OR: 4.4, 1.08–18.2, P = 0.04) and age (OR: 1.048 per year, 1.002–1.096, P = 0.04). Hospital death occurred in 23/1276 (1.8%) patients: those who developed ARF had a 46.9% (15/32) incidence of death vs. 0.6% (8/1244) in the patients who did not develop ARF. ARF requiring renal replacement therapy occurred in 20 patients (1.6%): 3/964 (0.3%) in the MV repair group and 17/312 (5.4%) in the MV replacement group (P < 0.001). Patients who developed ARF requiring renal replacement therapy had a 65% incidence of death. Death occurred in 19/312 (6.1%) MV replacement and in 4/964 (0.4%) MV repair patients (P < 0.0001).
The main result of our study is that finding the MV replacement is an independent risk factor for the development of ARF after MV surgery (OR: 4.0; 95% CI: 1.5–10.6). The other risk factors for this complication in our study population were perioperative low-output syndrome, emergency operation, preoperative renal impairment, reoperation for bleeding, diabetes and age. The aetiology of ARF after cardiac surgery is multifactorial. Factors include occult renal ischaemia, renal injury from endo- and exotoxins, and decreased renal reserve; MV surgery can be associated with an ischaemic injury to the kidneys due to a low-output syndrome. An alternative mechanism of renal ischaemia during MV surgery is embolic.
In conclusion, our study identifies risk factors for the development of ARF in MV surgery and, for the first time, shows that MV replacement is an independent risk factor for this complication.
We are indebted to Costantini Marco RN, Fichera Mariano RN and Tolja Marina for the care provided to these patients and for the support in data collection and data entry.
1. Chertow GM, Levy EM, Hammermeister KE et al
. Independent association between acute renal failure and mortality following cardiac surgery. Am J Med
1998; 104: 343–348.
2. Chertow GM, Lazarus JM, Christiansen CL et al
. Preoperative renal risk stratification. Circulation
1997; 97: 878–884.
3. Bove T, Calabro MG, Landoni G et al
. The incidence and risk of acute renal failure after cardiac surgery. J Cardiothor Vasc Anesth
2004; 18: 442–445.