Contrast-medium–induced nephrotoxicity (CMIN) can lead to hospital-acquired renal failure. With the increasing number of contrast-enhanced diagnostic and interventional procedures, it is important that radiologists prevent this condition, and the associated morbidity and mortality. The author details the features and risk factors for CMIN, as well as current data on methods of preventing its occurrence.
is a Professor in the Department of Diagnostic Radiology,
Copenhagen University Hospital at Herlev, Herlev, Denmark.
is a Senior Lecturer at the University of Sheffield and a
Consultant Radiologist with North General Hospital, Sheffield
Teaching Hospitals NHS Trust, Sheffield, UK.
(CMIN) is widely used to refer to the reduction in renal function
induced by contrast media. It implies impairment in renal function
(an increase in serum creatinine by >25% or >0.5 mg/dL) that
occurs within 3 days following the intravascular administration of
contrast media and the absence of an alternative etiology.
Contrast-medium-induced nephrotoxicity is considered an important
cause of hospital-acquired renal failure.
This is not surprising because diagnostic and interventional
procedures requiring the use of contrast media are performed with
increasing frequency. In addition, the patient population subjected
to these procedures is progressively older with more comorbid
Prevention of this condition is important in order to avoid the
substantial morbidity and even mortality that can sometimes be
associated with CMIN. Even a small decrease in renal function may
greatly exacerbate morbidity and mortality caused by coexisting
Sepsis, bleeding, coma, and respiratory failure are frequently
observed in patients with acute renal failure.
Serum creatinine often peaks within 3 to 4 days after the
administration of contrast media.
Mild proteinuria and oliguria may also be observed. The majority of
patients with CMIN tend to be nonoliguric except those with
pre-existing advanced chronic renal failure. Heavy proteinuria is
an unusual feature of CMIN. Fortunately, most episodes of CMIN are
self-limited and resolve within 1 to 2 weeks. Several nonanuric
episodes are probably undetected because the serum creatinine is
rarely measured after administration of contrast media and these
patients have no symptoms, especially patients who are examined on
an outpatient basis. Permanent renal damage is rare and occurs only
in a very few instances.
A persistent nephrogram on plain radiography or computed
tomography (CT) of the abdomen for 24 to 48 hours after contrast
media injection has been described as a feature of CMIN.
However, its presence is not always associated with a reduction in
Also, gallbladder opacification is not correlated with the
occurrence of CMIN.
However, presence of these signs should be followed by assessment
of renal function and should discourage the administration of
further doses of contrast media if the results are abnormal.
The development of CMIN is low in people with normal renal
function varying from 0% to 2%.
Pre-existing renal impairment increases the frequency of this
complication. The incidence of CMIN has been reported to range from
12% to 27% in several prospective controlled studies.
In one study, an incidence as high as 50% was reported in patients
with diabetic nephropathy undergoing coronary angiography in spite
of the use of low-osmolar contrast media and adequate hydration.
Dialysis was necessary in 15% of these patients.
In another study, the incidence was significantly higher in
patients with diabetic nephropathy (19.7%) and in patients with
nephropathy (5.7%) due to other causes, whereas it was <1% in
patients with a serum creatinine level below 1.5 mg/dL.
Long-term renal effects of contrast media
High-osmolar contrast media can enhance the progression of
glomerulosclerosis and renal failure in old, spontaneously
hypertensive male rats.
However, the long-term effects of contrast media on renal function
in humans are not known.
Patients at highest risk for developing contrast-media-induced
acute renal failure are those with pre-existing renal impairment
(>1.5 mg/dL), particularly when the reduction in renal function
is secondary to diabetic nephropathy.
Diabetes mellitus per se, without renal impairment, is not a risk
The degree of renal insufficiency present before the administration
of contrast media determines, to a great extent, the severity of
CMIN. Large doses of contrast media and multiple injections within
72 hours increase the risk of developing CMIN. The route of
administration is also important and contrast media are less
nephrotoxic when administered intravenously than when given
intra-arterially in the renal arteries or in the aorta proximal to
the origin of the renal blood vessels. The acute intrarenal
concentration of contrast media is much higher after
intra-arte-rial injection than after intravenous administration.
Dehydration and congestive cardiac failure are risk factors, as
they are associated with a reduction in renal perfusion, which
enhances the ischemic insult of contrast media. In the past,
multiple myeloma was considered a risk factor for CMIN. However, if
dehydration is avoided, administration of contrast media rarely
leads to acute renal failure in patients with myeloma.
The concurrent use of nephrotoxic drugs such as nonsteriodal
anti-infiammatory drugs (NSAIDs) and aminoglycosides potentiate the
nephrotoxic effects of contrast media. Nephropathy is found more
frequently in patients with hypertension, hyperuricemia, or
proteinuria than in patients without these factors.
The type of contrast media is also an important predisposing
factor. High-osmolar contrast media are more nephrotoxic in
comparison to low-osmolar and isosmolar contrast media.
Identifying patients at risk of CMIN
Patients with pre-existing renal impairment are at particularly
high risk of CMIN. In spite of the limitations of this measurement,
serum creatinine is often used to determine the state of renal
function and to identify high-risk patients. This section will
address the value of serum creatinine in the assessment of renal
function and when this measurement should be performed before
Validation of serum creatinine measurements
Serum creatinine is not an ideal marker of renal function. The
serum creatinine level depends on muscle mass and is not usually
raised until the glomerular filtration rate has fallen by at least
50%. Endogenous serum creatinine clearance as a measure of
glomerular filtration rate is also inaccurate (especially when
renal function is low, because of a compensatory increase in
tubular secretion), which limits its validity as a glomerular
filtration marker. Radionuclide techniques are preferable,
but each of these tests is labor-intensive and impossible to
perform in all patients undergoing contrast-enhanced imaging.
Alternatively, renal function can be estimated by using specially
derived predictive equations. The most accurate results are
obtained with the Cockroft-Gault equation, whereas the most precise
formula is the Modification of Diet in Renal Disease (MDRD) study
Unfortunately, the predictive capabilities of these formulae are
suboptimal for ideal patient care.
However, compared with a simple serum creatinine measurement, these
methods are far superior for assessing renal function. Another
alternative is to use cut-off values for serum creatinine as an
indicator of several levels of renal impairment. However, the use
of cut-off values (especially the low levels) will include several
patients with normal renal function, and the use of high cut-off
values will exclude patients with renal impairment.
Despite the inaccuracies of serum creatinine measurements, it is an
adequate measure for identifying those patients at risk for CMIN,
because patients with normal serum creatinine (<1.5 mg/dL) have
almost no risk.
When should serum creatinine be measured?
A questionnaire designed to elicit a history of renal disorders
as well as additional risk factors for CMIN may identify patients
with normal serum creatinine in whom blood testing would be
The majority of patients (85%) in this study
had normal serum creatinine values (<1.3 mg/dL for women, 1.4
mg/dL for men). All except 2 patients (99%) who gave negative
answers to the questionnaire had serum creatinine levels
<1.7 mg/dL. A strong association was found between raised
serum creatinine values and a history of renal disease,
proteinuria, prior kidney surgery, hypertension, gout, and
diabetes. Only 6% of patients with negative answers to the 6
questions had abnormal serum creatinine levels. In a study of 2034
consecutive outpatients referred for CT examinations, only 3.2% (66
patients) had elevated serum creatinine levels (>2.0 mg/dL) and
the majority (97%) of these patients had risk factors for CMIN.
Two of the 66 patients (0.1% of the total number of patients) with
a raised creatinine level had no identifiable risk factors. Serum
creatinine was measured in a prospective study involving 640
consecutive adult patients presenting to the emergency department
with a clinical indication for intravenous administration of
iodinated contrast medium.
Thirty-five (5.5%) patients had abnormal serum creatinine (>1.6
mg/dL). Of this group, 77% (27 patients) of them were considered to
have risk factors for renal insufficiency. The remaining 8 patients
(1.3% of the total number) had no identifiable risk factors for
renal insufficiency. Thus, the majority of patients at risk of CMIN
can be identified by appropriate questions, but a questionnaire
does not completely exclude the presence of renal
Measures to reduce the incidence of CMIN
Several measures have been recommended to reduce the incidence
including: volume expansion, hydration with intravenous
administration of normal saline solution (NaCl 0.9%) or
half-strength saline solution (NaCl 0.45%), infusion of mannitol,
pharmacological manipulation (administration of atrial natriuretic
peptide, loop diuretics, calcium antagonists, theophylline,
dopamine, dopamine-1 receptor antagonist [fenoldopam],
acetylcysteine), use of low-osmolar nonionic contrast media instead
of high-osmolar ionic contrast media, use of isosmolar contrast
media instead of low-osmolar contrast media, use of
gadolinium-based contrast media instead of iodine-based contrast
media for radiography and CT, hemodialysis quicky after contrast
administration, hemofiltration during and after contrast
administration, an injection of a small volume of contrast medium,
and avoiding short intervals (<48 hours) between procedures
requiring intravascular administration of contrast media.
Extracellular volume expansion
Of all the measures mentioned above, extracellular volume
expansion and use of low-osmolar contrast media have been found to
be most effec-
This can be achieved with the intravenous injection of 100 mL/hour
of 0.9% saline solution, starting 4 hours before contrast
administration and continuing for 24 hours afterward.
In areas with a hot climate, more fiuid should be given. This
regime is suitable for patients who are not in congestive heart
failure and are not allowed to drink or eat before undergoing an
interventional or surgical procedure. If there is no
contraindication to oral administration, free fiuid intake should
be encouraged. At least 500 mL of water or soft drinks before and
2500 mL during the following 24-hour period should be recommended
orally. In addition, concurrent administration of nephrotoxic
drugs, such as gentamicin and nonsteroid anti-infiammatory drugs,
should be avoided. Mannitol and furosemide enhance the risk.
The main factors in the pathophysiology are considered to be a
reduction in renal perfusion caused by a direct effect of contrast
media on the kidney and toxic effects on the tubular cells,
although the latter effect is contentious.
Calcium channel blockers prevent the infiux of calcium ions
through voltage-operated channels. This will cause a vasorelaxant
effect in all vascular beds, including the kidney. In one study, a
3-day treatment with 20 mg of nitrendipine prevented the
development of CMIN in patients with moderate renal impairment,
whereas in another study, a single dose (20 or 10 mg) given 1 hour
before contrast medium administration failed to prevent CMIN.
Thus, the role of calcium channel blockers remains uncertain, and
their protective effect in patients with advanced renal impairment
has not been proven. Use of the vasodilators dopamine and arterial
natriuretic peptide may even be harmful in patients with diabetic
Selective dopamine-1 receptor agonist (fenoldopam), in contrast
to dopamine, increases both cortical and medullary blood fiow. One
study has shown that fenoldopam offers some protection against
whereas 2 studies showed that it offers no protection.
Furthermore, fenoldopam has several disadvantages: it has to be
given intravenously, and it induces hypotension and requires
regular monitoring of blood pressure.
Experimental studies have indicated that the endogenous
vasoactive peptide endothelin may play an important role in
mediating these events. Therefore, it was expected that endothelin
would reduce the incidence of CMIN in man, but a clinical study has
shown the opposite effect
; a nonselective endothelin receptor antagonist and intravenous
hydration exacerbated CMIN when compared with hydration alone.
However, there were several fiaws in this study.
The choice of the endothelin receptor antagonist was not
appropriate. It was given only 12 hours after contrast medium
injection, avoiding sustained drug cover.
The nonselective adenosine receptor antagonist theophylline has
also been advocated to reduce the risk of CMIN. However, the
results from published studies are confiicting. In one study,
administration of 200 mg theophylline intravenously was shown to
have a preventive effect,
whereas after 819 mg theophylline orally and daily for 3 days did
not offer additional protective effect when compared with hydration
Thus, the effectiveness of theophylline in preventing CMIN remains
uncertain. In addition, theophylline may induce cardiac arrhythmia
in patients with ischemic heart disease.
Administration of acetylcysteine, which is an antioxidant and a
scavenger of oxygen-free radicals, has also been shown to be both
effective in preventing CMIN in some studies
and to be without any effect in others.
One meta-analysis based on 7 studies (a total of 805 patients)
concluded that the use of acetylcysteine reduced the risk of CMIN
whereas a subsequent analysis based on 12 studies (1307 patients)
concluded that conclusive data indicating that it offers effective
protection are lacking.
In 7 of the 12 studies, no effect was found. In a meta-analy-sis,
inconsistency of study results reduces the confidence about
Thus, pharmacological manipulation with renal vasodilators,
receptor antagonists of endogenous vasoactive mediators, or
cytoprotective drugs does not seem to offer consistent protection
Nonionic contrast media
The use of low-osmolar nonionic monomeric and, more recently,
isosmolar nonionic dimeric contrast media has been recommended in
high-risk patients to reduce the risk of contrast media
nephrotoxicity. A multicenter study of 129 patients with renal
impairment and diabetes mellitus (serum creatinine between 1.5 and
3.5 mg/dL) undergoing angiography with either the isosmolar dimer
iodixanol or the nonionic monomer iohexol showed that CMIN
developed only in 3% of the patients after the dimer and in 26%
after the monomer.
The study concluded that the isosmolar dimer iodixanol is
significantly less nephrotoxic in comparison with the nonionic
monomers. In another study of patients with nephropathy due to
various causes, CMIN developed in 4% after the isosmolar nonionic
dimer and in 10% after a low-osmolar nonionic monomer.
In a further study, CMIN developed in 17% of patients with severe
renal impairment who received intravenous injection of either a
nonionic monomer or an isosmolar dimer for body or cranial CT. No
difference in the incidence of CMIN between the monomer and dimer
was found in this study.
Further studies are required to elucidate whether there is a
significant difference in the nephrotoxic effects of the nonionic
dimer in comparison with nonionic monomers.
Prophylactic hemodialysis or hemofiltration
Dialysis has been used in the prevention of CMIN. Hemodialysis
and peritoneal dialysis safely remove both iodinated and
gadolinium-based contrast media from the body.
The effectiveness of hemodialysis depends on many factors,
including blood and dialysate fiow rate, permeability of dialysis
membrane, duration of hemodialysis, and molecular size, protein
binding, hydrophilicity, and electrical charge of the contrast
medium. Generally, several hemodialysis sessions are needed to
remove all contrast medium, whereas it takes 3 weeks for continuous
ambulatory dialysis to remove the agent completely. There is no
need to schedule the dialysis in relation to the injection of
iodinated or MR-contrast media or the injection of a contrast agent
in relation to the dialysis program. Hemodialysis does not protect
poorly functioning kidneys against CMIN.
In addition, hemodialysis may cause deterioration of renal function
through activation of infiammatory reactions with the release of
vasoactive substances that may induce acute hypotension.
Hemofiltration is a continuous form of renal replacement therapy
and requires the intravenous infusion of a large volume of isotonic
replacement fiuid (1000 mL/hour). This is exactly matched with the
rate of ultrafiltrate production, so that no net fiuid loss or
overload occurs. A recent study has shown that, in patients with
chronic renal failure who are undergoing coronary interventions,
periprocedural hemofiltration given in the intensive care unit
(ICU) setting appears to be effective in preventing CMIN and
reduces the rate of in-hospital events and in-hospital mortality.
However, it is very costly and requires an intensive treatment
Gadolinium-based contrast agents for radiography in
patients with renal impairment
Gadolinium-based contrast agents are believed to be safe and not
nephrotoxic in the usual MRI doses up to 0.3 mmol/kg body weight.
Therefore, it has been suggested that gadolinium-based contrast
media could be used in place of iodinated agents for radiological
examinations in patients with significant renal impairment.
However, the dose requirement for a satisfactory diagnostic study
differs between MRI and radiography because different properties of
the gadolinium are being used in each of the 2 techniques. The
commonly used dose for body CT is 150 mL of a 300 mg I/mL (2.38
mmol I/mL) solution. The standard dose for contrast-enhanced MRI is
0.2 mL/kg body weight of a 0.5 mmol/mL gadolini-um-based contrast
agent. For body CT, a patient weighing 70 kg would receive 120 mmol
of the iodinated agent molecule and 360 mmol of iodine. For MRI,
this same 70 kg patient would receive 7 mmol of the
gadolinium-based agent molecule and 7 mmol of gadolinium.
Thus, the number of iodinated contrast agent molecules administered
would be almost 17 times that of gadolinium-containing molecules,
and the number of iodine atoms administered is 51 times that of
The results of a study involving 64 patients undergoing MRI with
a gadolinium-based agent and a radiographic examination with an
iodinated contrast medium indicated that gadolinium chelates are
significantly less nephrotoxic than are iodinated agents.
Eleven of the 64 patients had a significant increase in serum
creatinine after intravenous or intra-arterial administration of
iodine-based contrast media, whereas none had increased serum
creatinine levels after intravenous administration of a
gadolinium-based contrast agent. However, the molar doses and
concentrations of the iodine and gadolinium atoms were not
comparable. In 3.5% of 195 patients with abnormal pre-examination
creatinine clearance levels, acute renal failure (anuria) developed
after gadolinium-based contrast medium administration; for
MR-angiography the incidence was 1.9% and for digital subtraction
angiography it was 9.5%.
Dialysis was required in 3 of the 7 patients who developed acute
renal failure. The doses of gadolinium-DTPA ranged from 0.31 to
0.41 mmol/kg for MR angiography and 0.27 to 0.42 mmol/kg for
digital subtraction angiography. Other reports have shown the
nephrotoxic potential of gadolinium-based contrast media in high
doses (>0.3 mmol/kg).
An experimental study in pigs has indicated that gadolin-ium-based
contrast media are more nephrotoxic than are iodinated contrast
Thus, the use of gadolinium-based contrast media for radiographic
examinations cannot be recommended to avoid nephrotoxicity in
patients with renal impairment.
In patients with risk factors for CMIN, one should not: 1) give
high osmolar contrast media; 2) administer large doses of contrast
media; 3) administer mannitol and diuretics, particularly
loop-diuretics; or 4) perform multiple studies with contrast media
within a short period of time. It is very important that one: 1)
makes sure that the patient is well hydrated (give ≥100 mL of
liquid orally [eg, soft drinks] or intravenously [normal saline],
depending on the clinical situation) per hour starting 4 hours
before and continuing for 24 hours after contrast administration,
in warm areas increase that fiuid volume); 2) uses low-osmolar or
isosmolar contrast media; 3) stops administration of nephrotoxic
drugs for at least 24 hours prior to contrast administration; and
4) considers alternative imaging techniques that do not require the
administration of intravascular radiographic contrast media.