Statins for the Prevention of Contrast-Induced Nephropathy: A Systematic Review and Meta-Analysis

Contrast-induced nephropathy (CIN) is an important potential complication of percutaneous coronary intervention. It occurs in approximately 7% of patients undergoing cardiac procedures currently [1,2,3]. Although the incidence of CIN has decreased due to improved contrast media and better risk prevention measures in the past decade, there are continued substantial increases in CIN cases because of the increasing numbers of patients undergoing percutaneous coronary intervention [4,5,6]. CIN generally resolves spontaneously in most instances, but patients may have prolonged hospital stay, increased risk of in-hospital death and higher long-term mortality rates [7,8,9]. There has been considerable interest in searching for effective strategies to prevent CIN.

Besides periprocedural hydration, pharmacological prophylactic strategies have received considerable attention in recent years [10,11,12]. However, the pathophysiology of CIN is not well known. Some studies have suggested that oxidative stress, inflammation, reduction in renal blood flow and direct tubular cell damage by contrast media might play important roles in the organ injury process [13,14,15]. Statins are widely used in patients with coronary heart disease due to their cholesterol-lowering effect and cholesterol-independent effects, such as improving endothelial function as well as decreasing oxidative stress and inflammation [16].

Some retrospective and prospective studies have demonstrated that statins have a protective effect in the prevention of CIN, but there are currently no established guidelines on when patients should be given statins or how to use them. In this article, we performed a systematic review and meta-analysis of published human cohort studies and randomized controlled trials (RCTs) to determine whether the administration of statins is protective against CIN and to assess the magnitude of their effects.

We performed a systematic literature search of MEDLINE (from 1966 to July 2010), EMBASE (from 1974 to July 2010), Cochrane Library (July 2010), CNKI (from 1994 to July 2010) and ISI Proceedings for the last 5 years. We derived two comprehensive search themes that were then combined using the Boolean operator ‘AND’. For the theme ‘contrast-induced nephropathy’, we used combinations of MeSH, entry terms and text words: contrast media, contrast induced nephropathy, iodinated contrast medium, contrast agents, contrast medium, contrast dye, radiographic contrast, radiocontrast media, radiocontrast medium, Iohexol, Iobitridol, Ioxaglate, Ioversol, Iopromide, Iopamidol, Iodixanol, contrast-associated nephropathy, contrast nephropathy, contrast-induced acute renal failure, contrast-induced acute kidney injury, radiocontrast-induced acute kidney injury, radiocontrast-induced acute renal failure and nephrotoxicity. For the theme ‘statins’, we used: statins, atorvastatin, pravastatin, lovastatin, simvastatin, cerivastatin, fluvastatin, hydroxymethylglutaryl-CoA reductase inhibitors and HMG-CoA inhibitor. To include cohort studies, we did not use the Cochrane highly sensitive search strategy for randomized controlled studies in our search strategy.

We included cohort studies comparing the incidence of CIN in a chronic statin pretreatment group with a statin-naïve group and RCTs comparing short-term high-dose statin treatment with short-term low-dose statin treatment or placebo, and reported the incidence of CIN in both arms. There was no restriction on language or journal type. The studies were reviewed by 2 independent investigators (T.Z. and L.-H.S.) to determine whether they met the eligibility criteria for inclusion. Discrepancies were resolved by consensus. Using a standardized form, we recorded baseline demographics and clinical and procedural characteristics of each study, including mean age, baseline creatinine, diabetes, average contrast volume, periprocedural hydration, type of radio procedure, type of contrast media, type of statin, statin dose, other prophylactic agents used and the definition of CIN. We evaluated the quality characteristics of the included RCTs using the previously published PRISMA statement [17].

The endpoint was the incidence of CIN, which was defined as an increase in serum creatinine level >25% or 0.5 mg/dl (44.2 µmol/l). We performed qualitative analysis of the cohort studies and quantitative analysis of the RCTs to estimate the pooled risk ratios (RRs) for preventive effect of statins. Quantitative analysis was performed with review manager software (Rev-Man Analyses Version 5.0.4 Copenhagen; The Nordic Cochrane Center, The Cochrane Collaboration, 2008). Heterogeneity between studies was analyzed by the Q and I² statistics. A p value of Q statistic <0.1 was defined as an indicator of heterogeneity, and I² <40% indicated magnitude of the heterogeneity might not be important.

As shown in figure 1, 6 cohort studies and 6 RCTs were identified under our search criteria.

The main characteristics of the cohort studies are listed in table 1[18,19,20,21,22,23]. The definition of CIN as an increase in serum creatinine >25% or 0.5 mg/dl was used in 4 studies, and the other 2 studies used serum creatinine >0.5 mg/dl as a definition. Demographic and baseline characteristics of these studies are listed in table 2. Most risk factors associated with CIN were comparable in both arms. Demographic and baseline characteristics of the two groups were not available in the study by Ivanova et al. [18], but multivariate analysis was performed with these risk factors to adjust the odds ratio (OR) of statin pretreatment.

Results of the cohort studies are listed in table 2. The incidence of CIN was available in all the studies. Four studies reported adjusted ORs of statin pretreatment. According to these results, 4 studies supported that chronic statin pretreatment significantly reduced the incidence of CIN. In the study by Yoshida et al. [21], even high-risk patients combined with renal insufficiency defined as baseline serum creatinine >1.1 mg/dl also benefited from statin pretreatment. Although there was no significant difference in the incidence of CIN between the two groups in the study by Ivanova et al. [18], multivariate analysis indicated that only statin treatment was nephroprotective and was a significant variable in predicting the rise in serum creatinine. In the study by Kandula et al. [19], a higher incidence of CIN was observed in patients pretreated with statins, but there was no significant difference between the groups after adjusting for the propensity of receiving statins.

In addition, Khanal et al. [20] reported a significant reduction in the need for renal dialysis in the statin group (0.32 vs. 0.49%, p = 0.03). In the study by Patti et al. [23], 430 of 434 patients (99%) were followed up for 4 years, and no patient required dialysis. Moreover, the incidence of late major adverse cardiac events were significantly lower in the statin group (6 vs. 36%, p < 0.05). Actuarial life-table analysis showed that 4-year event-free survival was similar between statin-pretreated patients with CIN and statin-naïve patients without CIN (72 vs. 71%).

The main characteristics of RCTs are listed in table 3, and the demographic and baseline characteristics of the RCTs are listed in table 4. Overall, 1,194 patients were included: 591 patients who received short-term high-dose statin treatment and 603 patients who received short-term low-dose statin treatment or placebo [24,25,26,27,28,29]. Two statins were used, and the duration of statin treatment ranged from 2 to >7 days. The total statin dose ranged from 140 to >460 mg. All of the patients received low-osmolar or iso-osmolar contrast media and periprocedural hydration. Patients in 2 studies received oral N-acetylcysteine as a prophylactic agent. Two studies included high-risk patients combined with chronic kidney disease, which means baseline serum creatinine was higher than the other studies. The quality characteristics of the studies are shown in table 5.

Figure 2 shows the results from all RCTs with corresponding RRs. The fixed-effects model was chosen as a test for heterogeneity, and showed p = 0.16 and I² = 38%. The overall effect was a nonsignificant protective trend toward decreased incidence of CIN with periprocedural short-term high-dose statin treatment (RR: 0.70; 95% CI: 0.48–1.02). Further, similar results were shown under a random-effects model (RR: 0.63; 95% CI: 0.36–1.33).

Two studies followed up long-term effects. In the study by Toso et al. [25], there was 1 death in the statin group and 1 patient undergoing temporary hemofiltration in the control group within 1 month. No differences were observed between the 2 groups at the 6-month follow-up in the study by Jo et al. [26].

To the best of our knowledge, this is the first study to pool the current evidence of the preventive effect of statins in CIN.

Statins, which are mainly used as cholesterol-lowering agents, are also known to have pleiotropic effects, such as antioxidative and anti-inflammatory properties [16,30,31,32]. Given the potential role of oxidative stress in the pathophysiology of CIN, statins might reduce contrast media nephrotoxicity by removing free radicals. In our study, we included cohort studies comparing the incidence of CIN in a chronic statin pretreatment group to a statin-naïve group.

Most studies have suggested that chronic use of statins have a preventive effect against CIN, and a beneficial effect in reducing the incidence of dialysis and long-term mortality was also reported. Only 1 study has reported a null effect for statin use [19] and an even higher incidence of CIN in the statin group. Higher contrast volume administered in the statin group might have contributed to this result, but when adjusted for the amount of contrast used, no differences in the incidence of CIN were noted. We also performed quantitative analysis of RCTs comparing periprocedural short-term high-dose statins with low-dose statins or placebo. Pooled RR indicated that short-term high-dose statin treatment was associated with a reduction in the incidence of CIN, but there was no statistical significance.

Considering the inherent limitations of cohort studies and meta-analysis, these results must be considered with several caveats. Among the cohort studies included, most risk factors associated with CIN were well balanced in the two arms, and adjusted ORs were also available in 4 studies. But there were still some confounding variables. For instance, there was a higher percentage of patients with diabetes mellitus or chronic heart failure in the statin group than in the statin-naïve group in the study by Khanal et al. [20]. In the study by Kandula [19], the mean total contrast dose per procedure in the statin group was more than that in the statin-naïve group. These characteristics could have introduced clinical heterogeneity and bias against the beneficial effects of chronic statin pretreatment. Consequently, we did not perform quantitative analyses on these studies.

There is also clinical heterogeneity in RCTs associated with the varied study settings. One of the factors is the timing of statin pretreatment. For example, simvastatin was used both in the studies by Jo et al. [26] and Xinwei et al. [28]. Patients in Jo’s study received simvastatin 80 mg before and after the procedure (a total of 160 mg), while patients in Xinwei’s study received simvastatin pretreatment for more than 5 days (>460 mg in total). The former, however, showed a null effect for this drug and the latter showed a positive effect. The duration of statin use might have been too short to exert an effect in Jo’s study. On the other hand, the timing of drug administration might also be important in cohort studies. Chronic statin administration could contribute to the positive effect of statins.

Secondly, other preventive strategies applied in these studies are also related to the effect of statins. In the studies by Toso et al. [25] and Ozhan et al. [24], the patients received oral N-acetylcysteine at the same time. There was no difference between the two arms in CIN frequency. To date, N-acetylcysteine, an anti-oxidant agent, is considered the most effective drug at decreasing the incidence of CIN [33,34,35]. Statins and N-acetylcysteine might share the same mechanism for renal protection. The preventive effect of statins might be obscured by N-acetylcysteine.

Finally, CIN is associated with poor long-term outcome, but current studies focus primarily on short-term rather than long-term outcomes of harder endpoints, such as incidence of dialysis requirement, morbidity and mortality. Although 2 RCTs followed up for 1 and 6 months, the sample sizes were too small to detect differences. Thus, a larger multicenter study will be necessary to detect differences in more meaningful outcomes.

The major limitation of this study is based on the combined data of cohort studies and heterogeneous RCTs. Another potential limitation is the inclusion of low-quality studies for meta-analysis. All the RCTs, except the one by Jo et al. [26], did not report loss to follow-up. We presumed that loss was not significant due to the short duration of observation in these studies. Some studies did not have blinding. However, considering that the endpoint of these studies was objective, we believe that including these studies did not bring any notable bias.

In conclusion, most cohort studies showed chronic using of statins might have a preventive effect against CIN. Short-term high-dose statin treatment was associated with a protective trend toward decreased CIN incidence in RCTs, but there was no statistical significance. Considering the limitations of the included studies, the current data are not conclusive as to whether statins are protective for CIN. Future large well-designed studies are needed to address the effect of these drugs and their longer-term clinical outcomes.

This work was supported by Grants No. 30971185 and 81070239 from the National Natural Science Foundation, and Grants No. 10410701200, 09JC1409400 and 10JC1409400 from the Shanghai Municipal Natural Science Foundation.

Tuo Zhang and Ling-Hong Shen contributed equally to this work.