The impact of N-acetylcysteine and ascorbic acid in contrast-induced nephropathy in critical care patients: an open-label randomized controlled study

The present study is a one-center, two-arm, randomized, open-label, controlled trial. The study took place in the Intensive Care Unit (ICU) of the University Hospital of Larissa (12 beds) between 2010 and 2013. Inclusion criteria were age ≥14 years and diagnostic need for contrast-enhanced CT. Exclusion criteria were history of intravascular administration of contrast agent during the 6-day period prior to randomization, the use of antioxidant agents during the last week before the examination, unstable renal function, use of RRT in the 3-day period before randomization and pregnancy. We defined the unstable renal function as a change in serum creatinine values greater than 20% between 2 consecutive days during the 3 days prior to randomization, independently from crude baseline renal function.

Patients were randomized to intravenously receive Nac (1200 mg) and Aa (2 g) dissolved separately in 100 ml of normal saline (N/S) 0.9%, 2 hours before and at 10 hours and 18 hours following the infusion of contrast agent (treatment group - NacA) or 200 ml of intravenous N/S 0.9% (control group - CG) at the same time points as NacA. The choice of Nac and Aa and their doses were based on previous studies that aimed to manage oxidative damage [9], whereas Nac presents vasodilatory properties [20] that could overcome vasoconstriction caused by the contrast agent [21].

All participants received the same amount of additional hydration with 1000 ml N/S 0.9% given intravenously before CT as protection against the intravenous constant medium, unless this was contraindicated based on concurrent hemodynamic assessment; the latter was assessed either by ultrasonography or thermo-dilution [22]. Randomization was performed using tables of random numbers.

Serum urea, creatinine concentrations were assessed before the infusion of the contrast agent and once daily until the 5^th day following radiocontrast infusion; serum cystatin-C assessed before and at 24 and 48 hours following radiocontrast infusion and 8-isoprostane was assessed before and at 2, 24 and 48 hours following radiocontrast infusion. Timing of renal function assessment was based on creatinine and cystatin-C expected peak serum levels following the infusion of radiocontrast agent; serum creatinine peaks at 3–5 days [23] while serum cystatin-C peaks at 1–2 days after the infusion of contrast agent [24].

CIN was defined as relative increase by 25% of serum creatinine from the baseline value within 5 days [24]. CT scans were performed according to the institute’s standard protocol with the use of the same agent, iopamidol, a low osmolarity, non ionic, iodinated contrast medium (Iopamiro 370, Bracco). The quantity of infused contrast agent was determined by the radiologists depending on the type of CT imaging, and patients related characteristics and was recorded in a dedicated chart. The dose of contrast medium for contrast-enhanced CT was generally 0.5–2 ml/Kg of body weight. The minimum used dose was 100 ml and 150 ml was the maximum dose that was given. Measurement of serum cystatin-C and 8-isoprostane were performed with commercial enzyme-linked immunoassay kits (Cayman CC, USA).

The main outcome was the incidence of CIN. In addition, we assessed serial changes in serum creatinine within 5 days, serum cystatin-C within 48 hours and 8-isoprostane serum levels within 48 hours. Secondary indices of outcome were the need for RRT for a 10-day period following the infusion of the contrast agent, ICU stay and mortality.

Assuming that the incidence of CIN is 50% [25] we estimated that a sample size of 58 patients per treatment group would be required to detect 50% relative reduction in the incidence of CIN in the NacA with 80% power and 95% confidence level.

All analysis was performed on an intention-to-treat basis. The results are expressed as means ± standard error (SE) unless otherwise stated. Data were compared between groups using Fisher’s exact test for categorical variables and the t test or Mann–Whitney test as appropriate for continuous variables. Differences in changes in serum creatinine, cystatin-C and 8-isoprostane concentrations (dependent variables) during time were analyzed by linear mixed model analysis. The kinetic of serum creatinine, cystatin-C and 8-isoprostane are indicated by mean regression lines. The slope of the regression line is the rate at which the examined parameter’s value changes day after day. The intercept of the regression line represents the value of the y (dependant variable) axis where the mean regression line crosses the y axis at theoretical day 0. Receiver operating characteristic (ROC) analysis was performed to illustrate the performance of clinical or laboratory variables in identifying patients with CIN. Only variables that were associated with CIN in univarate analysis were used in ROC analysis. P values <0.05 were considered to be statistically significant. Statistical analysis and preparation of graphs were performed using the statistical package SPSS 21.0 (SPSS Inc., Chicago, IL, USA), and GraphPad Prism (version 5.01).

The incidence of CIN in the CG and NacA was 15.6% and 18.33%, respectively (p = 0.81). There was no significant difference between the two groups in serum creatinine during the examination period. Table 3 shows characteristics of participants according to the presence of CIN or not. Additional file 1: Table S4 shows a classification of patients with CIN based on serum creatinine or cystatin-C changes or the Risk, Injury, Failure, Loss, End-Stage Renal Failure (RIFLE) score. Compared to patients without CIN, patients with CIN had significantly increased baseline values of urea/creatinine ratio (p = 0.01) and had more often received colimycin or at least one of the following drugs that have a known adverse impact in renal function: aminoglycosides, amphotericin, colimycin, vancomycin, teicoplanin or non-steroid anti-inflammatory drugs (Additional file 1: Table S5). Additional file 1: Table S6 shows fluid balance in patients with or without CIN. Multivariate analysis revealed that CIN was associated with a higher baseline ratio of serum urea/creatinine (odds ratio, 1.02; 95 CI%, 1.00–1.05) and with the concomitant use of nephrotoxic medications (odds ratio, 0.24; 95 CI%, 0.06–0.94), or with a higher baseline ratio of serum urea/creatinine (odds ratio, 1.03; 95 CI%, 1.00–1.05) and with the concomitant use of colimycin (odds ratio, 0.25; 95 CI%, 0.08–0.78) as independent risk factors. CIN did not have significant impact on ICU mortality (28.57% versus 19.42%) or on ICU stay (29 versus 25 days) (Table 3).

Serum cystatin-C (sCysC) concentration did not present significant differences between groups or during the examination period (p = 0.658). Patients in the CG who had CIN had significantly increased cystatin-C levels compared to patients with NacA who had CIN (p = 0.03) (Fig. 2). The percentage change (median ± interquartile range (IR)) in sCysC (mg/L) from baseline in patients who had CIN was 37.23% ± 12.54 and 93.20% ± 33.54 in NacA and in CG, respectively (p = 0.03). Data were also analyzed according to 25% increase of cystatin-C levels from the baseline value (25% DeltaCystC) - similarly to creatinine-based definition of CIN, but there was no significant difference between the NacA and CG (p = 0.26).

Serum levels of 8-isoprostane are presented in Fig. 3. Although the mean regression lines of serum levels of 8-isoprostane did not differ significantly between the two groups over time, significant differences between the NacA and CG were detected at 2 and 24 hours after the infusion of the contrast agent and in the CG between baseline and 48 hours.

ROC curve analysis showed that the best cutoff baseline value of the urea/creatinine ratio to predict CIN was 63.79 (area under the curve (AUC) (95% CI) 0.712 ± 0.076 (0.56–0.86), p = 0.002) (Fig. 4). ROC curve analysis also showed that the best cutoff baseline value of the serum urea/creatinine ratio to predict CIN in a patient who receives nephrotoxic medication was 65.70 (sensitivity 68.75%, specificity 69.81%, AUC (95%CI) 0.74 ± 0.078 (0.58–0.89), p = 0.003) (Fig. 5).