Post-contrast acute kidney injury. Part 2: risk stratification, role of hydration and other prophylactic measures, patients taking metformin and chronic dialysis patients

1The Contrast Media Safety Committee (CMSC) of the European Society of Urogenital Radiology (ESUR) produced their most recent guidelines on what was then termed contrast-induced nephropathy (CIN) in 2011. Guidelines on the use of contrast media (CM) in patients on dialysis and on the use of CM in diabetic patients using metformin were published in 2002 and 2014. This review provides recommendations for updating the CMSC guidelines which were obtained using a structured literature review based on clinical questions and Patient–Intervention–Comparator–Outcome (PICO) formatting. Since the literature related to the topics considered is so large, the results of the review have been split into two papers. The review only considers post-contrast kidney injury (PC-AKI) after iodine-based CM because acute kidney injury is not associated with gadolinium-based contrast agents in doses approved for clinical magnetic resonance imaging.

2In this second paper on PC-AKI, the following topics related to patient management are considered:The role of questionnaires and risk scores to identify at-risk patients with reduced renal functionThe need to stop nephrotoxic medication before giving CMThe optimal hydration protocols for protecting against PC-AKIThe possible role of prophylactic drug treatment in preventing PC-AKIThe need to adapt metformin administration when giving CMThe need to alter schedules for dialysis in the period before and after CM administration

3The role of questionnaires and risk scores to identify at-risk patients with reduced renal function

4The need to stop nephrotoxic medication before giving CM

5The optimal hydration protocols for protecting against PC-AKI

6The possible role of prophylactic drug treatment in preventing PC-AKI

7The need to adapt metformin administration when giving CM

8The need to alter schedules for dialysis in the period before and after CM administration

9Recommendations are made for items 1–6. The recommendations have been incorporated into version 10 of the ESUR CMSC guidelines, at the end of this paper (Table 4).

10The recommendations were prepared using the Appraisal of Guidelines for Research and Evaluation (AGREE) II document. A guideline Writing Group (WG) prepared ten clinical questions in PICO format. Systematic search strings were developed with a professional librarian for four different biomedical literature databases (PubMed, Web of Science, Embase and the Cochrane Library). The titles and abstracts were screened for relevance and selected on predefined inclusion and exclusion criteria. Emphasis was put on comparative studies with strong scientific evidence, such as meta-analyses and systematic reviews, and prospective randomised controlled trials (RCTs). The six systematic searches in this manuscript yielded 3402 references of which 445 were selected on the basis of title and abstract. After review of the full text of these 445 publications, 145 were selected for inclusion in this paper. The quality of the evidence from the selected articles was evaluated according to the Oxford Centre for Evidence Based Medicine levels of evidence: grade A, established scientific evidence; grade B, scientific presumption; grade C, low level of evidence. When there was no scientific evidence, recommendations were based on WG consensus and were graded as expert opinion (grade D).

11The full description of the materials and methods appears in part 1.

12The term intra-arterial injection with first pass renal exposure indicates that contrast medium reaches the kidneys in a relatively undiluted form, e.g. injection into the left heart, thoracic and suprarenal abdominal aorta or the renal arteries. The term intra-arterial injection with second pass renal exposure indicates that contrast medium reaches the renal arteries after dilution either in the pulmonary or peripheral circulation, e.g. injection into the right heart, pulmonary artery, carotid, subclavian, coronary, mesenteric or infrarenal arteries.

13Questionnaires, such as that proposed by Choyke, use information about a history of renal disease or renal surgery, heart failure, diabetes, proteinuria, hypertension and gout to stratify patients for their PC-AKI risk so that sCr measurements need only be done selectively. This may work well and can save resources. Observational studies have shown that these questionnaires can identify patients with eGFR < 45 ml/min/1.73 m2 with adequate sensitivity, especially if they are aged less than 70. Since eGFR measurement can detect more patients with renal dysfunction than questionnaires, with easier patient logistics and similar cost-effectiveness, many hospitals have adopted a policy of sCr measurements in all patients scheduled for intravenous (IV) CM and no longer use questionnaires for risk stratification or selection for eGFR measurement (Table 1).

14No risk models have been produced yet for IV or IA CM administration with second pass renal exposure. For patients having coronary angiography (CA) or percutaneous coronary intervention (PCI), many different risk scores have been proposed to stratify the patient’s PC-AKI risk, and most include pre-procedural and procedural data. A model with only pre-procedural data would be more practical for selecting suitable preventive measures. Risk scores should be verified in relation to improvements in clinical outcome. In clinical practice, a prediction rule would require a high discriminatory value, i.e. a C-statistic greater than 0.80.

15The best-known risk model is the eight-variable Mehran score, which has been studied in more than 15,000 patients and has been externally validated in multiple studies, but with variable C-statistic values of 0.57–0.85. The Mehran score correlates relatively well with clinical outcomes. Newer risk scores with good discriminatory value, available in user-friendly calculators or smartphone applications, still need external validation. A recent systematic review of 16 risk models concluded that they had only modest predictive value and a review and meta-analysis of 74 risk models noted their heterogeneity and concluded that further research was needed to evaluate the effect of such models on clinical care (Table 1).

16Optimal nephrologic care involves minimizing the use of nephrotoxic drugs where clinically possible. Many frequently prescribed medications, such as nonselective NSAIDs, selective Cox-2 inhibitors, several classes of antimicrobial agents and chemotherapeutic agents have nephrotoxic potential and can induce AKI.

17There is little good quality data about the relationship between these drugs and PC-AKI. A retrospective cohort study showed that concurrent use of four or more nephrotoxic agents was significantly predictive for PC-AKI in patients given IV CM. A meta-analysis of PC-AKI incidence following CM-enhanced CT found that concurrent administration of NSAIDs was an independent risk factor for PC-AKI.

18The effect of withholding angiotensin converting enzyme inhibitors (ACEI) and angiotensin-2 receptor blockers (ARB) in chronic users has been extensively evaluated. Multiple RCTs and observational studies gave conflicting results and are limited by small sample sizes and significant heterogeneity. However, meta-analyses of RCTs found no lower risk. Withholding ACEI/ARB may be associated with a slightly lower risk of PC-AKI but the evidence is not sufficiently strong to recommend this (Table 1).

19Evidence for prevention of PC-AKI with IV saline hydration (volume expansion) comes from RCTs in patients who received intra-arterial (IA) CM during percutaneous intervention, and in patients who received bicarbonate hydration before IV enhanced emergency CTPA. One RCT evaluated the evidence for IA CM administration during CA. These studies found that, for both IA and IV CM administration, the incidence of PC-AKI was significantly lower in patients who received IV hydration compared to placebo, and that hydration prevented emergency dialysis. Significant differences for mortality or other adverse events were not found. There were few patients with severe renal impairment (eGFR < 30 ml/min/1.73 m2) in almost all studies. The recent AMACING trial showed that for patients with eGFR > 30 ml/min/1.73 m2 receiving IV CM there was no difference between no hydration and hydration in preventing PC-AKI.

20Oral intake of clear fluids by patients as an alternative to IV saline to prevent PC-AKI is difficult to monitor or control. Nine studies evaluated oral hydration, but were limited by small patient numbers and by the absence of patients with severe renal impairment. Three meta-analyses concluded that there is no evidence that oral hydration is associated with more risk of PC-AKI compared to IV hydration, but the studies were limited by heterogeneity and lack of hard clinical outcomes. The CMSC does not recommend the use of oral hydration as the sole preventive strategy for PC-AKI, but unrestricted intake of clear oral fluids in addition to IV volume expansion is supported (Table 2).

21Normal saline (NaCl 0.9%) and sodium bicarbonate solution (1.4% or 154 mmol NaHCO3 in D5W) are the two most commonly studied crystalloid solutions. The rationale for using bicarbonate is that alkalinisation can reduce the formation of free reactive oxygen species. Initial studies favoured bicarbonate, but this was not replicated in later studies, so IV hydration with bicarbonate can be considered equivalent to normal saline.

22There is no consensus on the optimal hydration regime. Most studies have compared bicarbonate given pre- and post-CM for less than 6 h to longer duration saline pre- and post-CM protocols (12–24 h). In all studies, there are few patients with eGFR < 30 ml/min/1.73 m2, and evidence about whether short duration bicarbonate is better than long duration saline is conflicting. There is limited evidence on whether pre-hydration only is inferior to pre- and post-hydration, and only one very short duration bicarbonate protocol has been evaluated.

23Most studies have been performed in cardiac patients admitted for CA or PCI. Three studies evaluated hydration protocols in patients having contrast-enhanced CT, and did not favour bicarbonate over saline. No studies were identified assessing the beneficial effect of other crystalloids. However, balanced crystalloid solutions, such as Ringer’s lactate, may be preferable in critical care populations, because they avoid the harmful effects of hyperchloraemic acidosis.

24The CMSC considers that for IV and IA CM injection with second pass renal exposure either a short bicarbonate hydration regime before CM or a conventional protocol with saline given before and after CM may be used. For IA CM injection with first pass renal exposure conventional protocols with either bicarbonate or saline given before and after CM should be used (Table 2).

25Newer approaches for patients with impaired left ventricular function combine controlled saline hydration with a forced high urinary flow rate to maintain euvolemia and avoid overhydration and several RCTs showed better results than conventional hydration protocols. Other catheter-based strategies used left ventricular end-diastolic pressure or central venous pressure to guide hydration. In these RCTs the incidence of PC-AKI was lower than with standard IV hydration. Since the forced diuresis studies have heterogeneous populations, interventions and control hydration protocols, their findings cannot be pooled. The CMSC considers that there is not sufficient evidence to recommend forced diuresis.

26There is no data to suggest that patients with severe renal impairment (CKD grade V) or severe heart failure (NYHA grade 3–4) should receive different hydration protocols. However, IV hydration with large volumes may exacerbate acute heart failure and induce pulmonary oedema. The opinion of the CMSC is that hydration protocols in these patients should be individualized for type, volume and duration.

27Most recent RCTs or meta-analyses do not show a protective effect of NAc against PC-AKI following coronary or peripheral angiography. NAc also failed to affect clinical outcome in coronary or peripheral angiography or to have a protective effect in CT or in patients with diabetes mellitus undergoing coronary or peripheral angiography. Comparative studies with NAc combined with saline or sodium bicarbonate protocols did not show any additional effect of protective effect of NAc. However, more recent meta-analyses showed a benefit of NAc, with or without high-dose statins, when added to hydration for preventing PC-AKI.

28Several meta-analyses showed lower overall PC-AKI rates with the use of high-dose, short-term statin treatment compared to controls. Lower PC-AKI rates were also found in subgroups, such as older patients, patients with acute coronary syndromes and for high-dose statin regimes. Some of these meta-analyses showed a reduced need for RRT after statins, but no reduction in all-cause mortality. However, the US Agency for Healthcare Research and Quality (AHRQ) meta-analysis showed that the risk of PC-AKI was only significantly reduced when statins were added to hydration and NAc. A reduction in PC-AKI risk could not be shown when statins plus hydration were compared to hydration alone in patients not taking statins. The standard of evidence grade was low in both analyses.

29Despite the many positive results, it is difficult to make a general recommendation for statins because the patients studied were invariably cardiac, and a variety of statin and hydration protocols were used. Patients with CKD grade 3B–5 (eGFR < 45 ml/min/1.73 m2) are under-represented in the studies and results in these patients remain inconclusive. Most patients undergoing CA/PCI are already taking long-term statins, and results in these patients are unclear.

30While the CMSC recognises the potential preventive effects of short-term statins, it does not advise the use of short-term, high-dose statins as a single strategy for preventing PC-AKI (Table 2).

31Administration of renin–angiotensin–aldosterone system (RAAS) blockade as a preventive measure for patients not taking these drugs did not show a significant effect on the incidence of PC-AKI in recent meta-analyses (Table 2).

32The majority of RCTs or meta-analyses do not demonstrate a protective effect of vitamin C against PC-AKI in patients with CKD predominantly undergoing coronary angiography or any benefit of the use of vitamin C, NAc or a combination of both over the standard hydration regimen in preventing PC-AKI (Table 2). Combining vitamin C with pentoxifylline also failed to show an advantage. Only two publications have shown a protective effect of vitamin C in patients with CKD undergoing CA.

33There is no convincing evidence in favour of preventive haemodialysis or RRT alone or combined with hydration in patients with CKD, predominantly undergoing CA (Table 2). There is no evidence of an increased risk of permanent anuria in patients on peritoneal dialysis undergoing CA. There is a single study showing better late-stage (day 5–30) renal protection against PC-AKI with simultaneous haemodialysis.

34The data on the protective effects of several agents, such as trimatizidine, theophylline, alprostadil, nebivolol, fenoldopam and iloprost, is not conclusive and does not support recommending their use to reduce the risk of PC-AKI.

35Metformin is the standard drug for monotherapy of type 2 diabetes mellitus. The effect of CM on the risk of metformin-associated lactic acidosis is indirect, since an episode of AKI following intravascular CM administration may lead to metformin accumulation. The use of metformin in patients with eGFR 30–59 ml/min/1.73 m2 is considered safe if doses are reduced appropriately. Limiting the metformin dose to a maximum of 2000 mg/day for eGFR 45–60 ml/min/1.73 m2 and to a maximum of 1000 mg/day for eGFR 30–44 ml/min/1.73 m2 has been recommended. In patients with eGFR 30–59 ml/min/1.73 m2 metformin drug levels remain within therapeutic ranges. For patients with eGFR < 30 ml/min/1.73 m2 metformin administration is not approved.

36Multiple studies and meta-analyses have shown that the risk of lactic acidosis is very low and linked more to the underlying disease and possible co-morbidities rather than the use of metformin. Because of the lack of published evidence on metformin and CM, early guidelines about the need to stop metformin before intravascular CM were based on consensus, and were strict. As the low risk of lactic acidosis became apparent, guidelines have become less restrictive.

37Since no new published evidence is available, the CMSC has updated its recommendations based on recent recommendations from the FDA, and on guidelines from the ACR and RSTN (Table 3).

38Iodine-based CM can be safely removed by haemodialysis (HD) or haemofiltration (HF). Many factors influence the effectiveness of HD, such as flow rate of blood and dialysate, dialysis membrane permeability, HD duration, and CM characteristics such as molecular size, protein binding, hydrophilicity and electrical charge.

39Although HF concomitant with radiological procedures has been shown to be feasible and well tolerated, the fractional removal of iodine-based CM contrast agents is modest and several HF or HD sessions are needed to remove 95% of the administered CM. Also, there is no evidence for the necessity of emergency HD after administration of iodine-based CM in patients on chronic HD. However, to avoid volume overload, CM administration may be synchronised with scheduled HF or HD (Table 3).

40Assessment of the risk of PC-AKI before intravascular CM is administered is best done by measuring eGFR but the alternative of a questionnaire for patients detects most patients with eGFR less than 45 ml/min/1.73 m2. Volume expansion with normal saline or sodium bicarbonate remains the mainstay of PC-AKI prevention, but there is still uncertainty about the optimal protocol. The additional benefit of a number of drugs, such as N-acetylcysteine, statins, ACE inhibitors and angiotensin-II receptor blockers, and vitamin C in preventing PC-AKI has not been proved conclusively. Stopping nephrotoxic medications appears to be of limited value in preventing PC-AKI. Recommendations for discontinuing metformin when CM is given have been relaxed and now only apply to patients with eGFR < 30 ml/min/1.73 m2 receiving IV CM or IA CM with second pass renal exposure, and to all patients receiving IA CM with first pass renal exposure or who have AKI. There is no need to adapt dialysis schedules in patients being given intravascular CM.

41The recommendations made in this paper have been incorporated into the ESUR CMSC guidelines version 10 (Table 4).