**2. IV versus IA CM administration and CIN**

The alleged lower risk of CIN following CM-enhanced CT compared with PCA/PCI has lead to conclusions such as


Such statements and conclusions may jeopardize patient safety, since they were not based on any studies comparing the risk of CIN following CM-enhanced CT and coronary interventions in patients with matched risk factors and CM-doses. In addition, a recent study showed no difference in the incidence of CIN between CT-angiography and digital subtraction angiography (DSA) of the aortofemoral arteries in the same patients. The lack of difference occurred despite that the DSA-results may have been affected by the CM load from the CT performed 3-14 days prior to the DSA (Karlsberg et al., 2011).

It seems inexplicable that the same type of CM molecules passing through the coronary arteries via the coronary sinus to the right atrium should be more nephrotoxic than if the same molecules pass via the arm veins to the right atrium and then through the pulmonary circulation to finally reach the kidneys via the aorta. As a matter of fact in the vast majority of IA injections, the CM has to pass through the venous system before reaching the kidneys (IV relative to the kidneys), i.e. carotid, subclavian, celiac, mesenteric, distal aortic and iliaco-femoral. Left ventriculography or aortography in connection with PCA/PCI is an exception. However, in this case only a minor part will reach the kidneys directly through the aortic route, i.e. about 20% of cardiac output or e.g. 2-3 grams of iodine following a left ventriculography (6-8 mL of an injected volume of 30-40 mL of 320 to 370 mg I/mL) of a total mean dose commonly ranging between 50 to 100 grams of iodine during a coronary procedures. Spill-over into the aorta also occurs during selective coronary artery injections and through side-holes of guiding catheters during PCI. However, the amount during each injection is so small that it will hardly affect plasma osmolality to cause any hypertonic renal effects and will therefore only affect the kidneys with the same pathophysiological mechanisms as an IV injection will do.

In the relatively few published reports of CIN following CM-enhanced CT the incidence may vary between 0 and 42% depending on definitions, degree of renal impairment and number and degree of risk factors (Katzberg & Newhouse, 2010; Nguyen et al., 2008; Polena et al., 2005; Tepel et al., 2000; Thomsen et al., 2008b). In a recent prospective study of unselected emergency patients 11% (n=70/633) increased their serum creatinine ≥44 μmol/L or ≥25% of whom 9% (n=6) developed CM-induced severe renal failure, which contributed to death in 4 of the 6 patients (Mitchell et al., 2010). Another CIN study showed that IV CM injections were actually associated with a higher mortality risk than IA administration (From et al., 2008). One explanation may be that the entire CM dose in CT is injected within one minute and thus may strike the kidneys at a considerable higher dose rate compared with a coronary arterial procedure that may last for 15-30-60 minutes or even longer.

It should also be noted that in randomized studies comparing renal effects of various CM, high-risk patients (e.g. unstable renal function, heart failure, uncontrolled diabetes, recent CM examinations, etc.) are often excluded (Barrett et al., 2006; Kuhn et al., 2008; Nguyen et al., 2008; Thomsen et al., 2008b). This bias in patient selection compared with coronary studies, where high-risk patients can not be excluded from life-saving procedures, may in part explain the illusive opinion that an IV CM injection implies a lesser risk of CIN than an IA. Thus, it may seem premature to consider the risk of CIN less following IV injections than after IA administration.

## **3. Evaluation of renal function**

196 Coronary Interventions

offending agent itself, i.e. the contrast medium (Davidson et al., 2006; Kane et al., 2008; Sterner et al., 2001). Though low- and iso-osmolal CM should be substituted for highosmolal CM (Barrett & Carlisle, 1993; Rudnick et al., 1995), the benefit of iso- over lowosmolal CM is only suggestive but not statistically significant according to a recent meta-

• Using estimated glomerular filtration rate (eGFR) in absolute terms to evaluate renal

• Using gram iodine (g-I) to express CM-dose instead of simply volumes and promoting g-I/eGFR ratio to maximize CM doses as a predictor of CIN instead of the Cigarroa

• Potential means to reduce CM dose for CT coronary angiography (CTCA) in patients at

• The potential of using iodine concentrations and doses iso-attenuating with gadolinium

The alleged lower risk of CIN following CM-enhanced CT compared with PCA/PCI has

• "In clinical settings such as CM-enhanced multidetector CT makes it defensible to consider using CM even in patients with greater levels of background risk factors (e.g. greater degree of preexisting chronic renal insufficiency) than one would be

• "International radiologic professional organizations should revisit the basis of their practice guidelines to reduce their implications about the danger of CIN with CM-

Such statements and conclusions may jeopardize patient safety, since they were not based on any studies comparing the risk of CIN following CM-enhanced CT and coronary interventions in patients with matched risk factors and CM-doses. In addition, a recent study showed no difference in the incidence of CIN between CT-angiography and digital subtraction angiography (DSA) of the aortofemoral arteries in the same patients. The lack of difference occurred despite that the DSA-results may have been affected by the CM load

It seems inexplicable that the same type of CM molecules passing through the coronary arteries via the coronary sinus to the right atrium should be more nephrotoxic than if the same molecules pass via the arm veins to the right atrium and then through the pulmonary circulation to finally reach the kidneys via the aorta. As a matter of fact in the vast majority of IA injections, the CM has to pass through the venous system before reaching the kidneys (IV relative to the kidneys), i.e. carotid, subclavian, celiac, mesenteric, distal aortic and iliaco-femoral. Left ventriculography or aortography in connection with PCA/PCI is an exception. However, in this case only a minor part will reach the kidneys directly through the aortic route, i.e. about 20% of cardiac output or e.g. 2-3 grams of iodine following a left

(Gd) CM and other means to decrease CM-doses in patients at risk of CIN.

comfortable with in the IA setting" (Katzberg & Barrett, 2007) and

from the CT performed 3-14 days prior to the DSA (Karlsberg et al., 2011).

analysis (From et al., 2010).

function.

risk of CIN.

lead to conclusions such as

The present chapter will focus on:

formula (Cigarroa et al., 1989).

**2. IV versus IA CM administration and CIN** 

enhanced CT"s (Katzberg & Newhouse, 2010).

• The risk of CIN in IV versus IA CM administration.

It is well recognized that serum creatinine is a poor predictor of renal function (Perrone et al., 1992), especially in elderly patients with decreasing muscle mass, the major source of creatinine. In one study 50% of patients ≥70 years with a normal serum creatinine had a GFR ≤50 mL/min (Duncan et al., 2001).

Measurement of GFR based on exogenous markers such as inulin and I-CM is regarded the best indices of the level of renal function in health and disease (Stevens et al., 2006), but is work-intensive, relatively expensive, time-consuming and therefore unsuitable in clinical practice prior to CM administration. Instead, GFR should be estimated (eGFR) taking into account not only serum creatinine but also anthropometric (weight and height) and/or demographic (gender and age) data as a measure of muscle mass by using dedicated GFR prediction equations (Stevens et al., 2006) such as the MDRD (Modification of Diet in Renal Disease) (Levey et al., 2007), CKD-EPI (Levey et al., 2009) and Lund-Malmö equations (Nyman et al., 2006). Consequently, newly developed CIN risk scores include eGFR using prediction equations (Bartholomew et al., 2004; Mehran et al., 2004). Before adapting a GFR prediction equation the following should be considered:

Contrast Medium-Induced Nephropathy (CIN)

which is lacking in the Cigarroa formula.

Laskey, 2007 3179 Unselected

Worasuwannarak, 2010 248 Elective

Total 5026

2. Anticipated mean concentration.

in the remaining.

authors).

Gram-Iodine/GFR Ratio to Predict CIN and Strategies to Reduce Contrast Medium Doses 199

uses serum creatinine instead of GFR; i.e. maximum CM volume = 5 mL × body weight/serum creatinine (mg/dL). From a female perspective, with a possible increased CIN-risk compared with males (Brown et al., 2008), the g-I/eGFR ratio is preferable since creatinine-based GFR prediction equation also contains coefficients for female gender,

Mounting evidence from coronary interventions indicate that a g-I/eGFR ratio roughly >1.0 represent a significant and independent predictor of CIN (Table 1). At a g-I/GFR ratio <1.0 the reported CIN frequency was <3% (Gurm et al, 2011; Laskey et al, 2007; Nyman et al, 2008).

population

Nyman, 2008 391 STEMI 2.91 350 1.00 Nozue, 2009 60 Stable angina 5.1 370 1.895

diabetics

Mager, 2010 871 STEMI 3.7 370 1.375 Liu et al, 2011 277 STEMI 2.39 3704 0.885

Weighted mean value 3.50 1.24

Table 1. Gram-iodine/eGFR ratio and CIN in coronary interventions. Studies defining CMvolume/eGFR ratio or gram-iodine/eGFR ratio as a significant and independent predictor of CIN (serum creatinine rise ≥25% or ≥44 μmol/L above baseline). Weighted mean value with individual study sizes as weights were finally calculated based on log-transformation of volume/eGFR and g-I/eGFR ratio. Absolute GFR was estimated in 3 reports (Laskey et al., 2007; Nyman et al., 2008; Worasuwannarak & Pornratanarangsi, 2010) and relative GFR

3. 96% 370 mg I/mL and 4% 320 mg I/mL (e-mail communication with the authors). 4. 271 patients 370 mg I/mL and 6 patients 320 mg I/mL (e-mail communication with the

A most recently published registry study involving about 50,000 patients recommended a planned gram-iodine dose restricted to 0.7 x eGFR value and not to exceed 1.0 x eGFR if a

Using a g-I/eGFR <1.0 implies a safer maximum dose compared with the Cigarroa formula. A 60-year old female with a height of 160 cm, weight 70 kg and serum creatinine of 150

eGFR ratio

Iodine Concentration (mg I/mL)

3.7 3502 1.305

2.60 3703 0.98

g-I/eGFR ratio

First author, year Number Indication Volume/

1. Calculated from the g-I/eGFR ratio and iodine concentration.

5. Calculated from the volume/eGFR ratio and iodine concentration.

CM concentration of 350 mg I/mL for PCI is anticipated (Gurm et al., 2011).


Body surface area (m2) = weight0.425 x (height0.725) x 0.007184

with weight expressed in kg and height in cm.

• Estimated GFR is only within 30% of measured GFR in 80-85% of the patients (Levey et al., 2009; Nyman et al., 2006). Thus, a patient with eGFR of 50 mL/min may actually only have a real GFR of 35 mL/min.

#### **4. Systemic drug exposure, gram-iodine/eGFR ratio and CIN**

#### **4.1 Area under the plasma concentration-time curve (AUC)**

Following injection of CM, blood samples may be used to calculate AUC. It is directly proportional to CM dose and inversely correlated with GFR (Frennby & Sterner, 2002). AUC is a fundamental pharmacokinetic parameter used to estimate *systemic exposure* of drugs that are distributed and eliminated according to linear kinetics, like contrast media (Chen et al., 2001; Sherwin et al., 2005). The systemic exposure of such a drug is often well correlated with its *toxicity* and hence is generally held as an index for dose optimization (Chen et al., 2001). The clinical value of AUC as a predictor of nephrotoxicity has been shown for a variety of drugs and CM dose/GFR ratio was first proposed as a potential indicator for the risk of CIN in 1997 (Altmann et al., 1997) and later in 2005 (Nyman et al., 2005; Sherwin et al., 2005).

#### **4.2 Gram-iodine/eGFR ratio**

CM doses in CIN risk scores and recommendations to minimize the risk of CIN have for obscure reasons often been based only on volumes (Bartholomew et al., 2004; Davidson et al., 2006; Mehran et al., 2004). It should rather be expressed in terms of gram iodine (g-I) since concentrations of commercially available CM varies from 140-400 mg I/mL and it will also reflect the attenuating capacity. This also makes it easier to compare CM doses and expand the experience of CIN made from one examination or department to another if different concentrations are used. Furthermore, common g-I doses for radiography-based procedures, i.e. 10-120 g-I, are in the same numerical range as patients' GFR, i.e. 10-120 mL/min. Thus, forming a g-I/eGFR ratio combines CM volume and concentration, serum creatinine, age and body size into a single continuous risk variable, and provides the examiner with a simple numerical relationship and an expedient way to predict the risk of CIN. This implies also a more sophisticated relationship between CM dose and renal function than the Cigarroa formula (Cigarroa et al., 1989) that lacks CM concentration and uses serum creatinine instead of GFR; i.e. maximum CM volume = 5 mL × body weight/serum creatinine (mg/dL). From a female perspective, with a possible increased CIN-risk compared with males (Brown et al., 2008), the g-I/eGFR ratio is preferable since creatinine-based GFR prediction equation also contains coefficients for female gender, which is lacking in the Cigarroa formula.

Mounting evidence from coronary interventions indicate that a g-I/eGFR ratio roughly >1.0 represent a significant and independent predictor of CIN (Table 1). At a g-I/GFR ratio <1.0 the reported CIN frequency was <3% (Gurm et al, 2011; Laskey et al, 2007; Nyman et al, 2008).


Table 1. Gram-iodine/eGFR ratio and CIN in coronary interventions. Studies defining CMvolume/eGFR ratio or gram-iodine/eGFR ratio as a significant and independent predictor of CIN (serum creatinine rise ≥25% or ≥44 μmol/L above baseline). Weighted mean value with individual study sizes as weights were finally calculated based on log-transformation of volume/eGFR and g-I/eGFR ratio. Absolute GFR was estimated in 3 reports (Laskey et al., 2007; Nyman et al., 2008; Worasuwannarak & Pornratanarangsi, 2010) and relative GFR in the remaining.


198 Coronary Interventions

• The creatinine assay in the local laboratory must be calibrated according to the specific method used when the equation was developed, in practice isotope dilution mass

• Dosing of drugs excreted by glomerular filtration should be based on GFR not adjusted for body surface area, i.e. absolute GFR in mL/min (Stevens et al., 2009). GFR adjusted to body surface area, i.e. relative GFR in mL/min/1.73 m2, will overestimate actual GFR in small subjects, especially children, and underestimate it in large individuals. The MDRD and CKD-EPI equations primarily gives relative GFR, which can be converted to absolute GFR using a body surface area equation such as the commonly used Dubois

Body surface area (m2) = weight0.425 x (height0.725) x 0.007184

• Estimated GFR is only within 30% of measured GFR in 80-85% of the patients (Levey et al., 2009; Nyman et al., 2006). Thus, a patient with eGFR of 50 mL/min may actually

Following injection of CM, blood samples may be used to calculate AUC. It is directly proportional to CM dose and inversely correlated with GFR (Frennby & Sterner, 2002). AUC is a fundamental pharmacokinetic parameter used to estimate *systemic exposure* of drugs that are distributed and eliminated according to linear kinetics, like contrast media (Chen et al., 2001; Sherwin et al., 2005). The systemic exposure of such a drug is often well correlated with its *toxicity* and hence is generally held as an index for dose optimization (Chen et al., 2001). The clinical value of AUC as a predictor of nephrotoxicity has been shown for a variety of drugs and CM dose/GFR ratio was first proposed as a potential indicator for the risk of CIN in 1997 (Altmann et al., 1997) and later in 2005 (Nyman et al., 2005; Sherwin et

CM doses in CIN risk scores and recommendations to minimize the risk of CIN have for obscure reasons often been based only on volumes (Bartholomew et al., 2004; Davidson et al., 2006; Mehran et al., 2004). It should rather be expressed in terms of gram iodine (g-I) since concentrations of commercially available CM varies from 140-400 mg I/mL and it will also reflect the attenuating capacity. This also makes it easier to compare CM doses and expand the experience of CIN made from one examination or department to another if different concentrations are used. Furthermore, common g-I doses for radiography-based procedures, i.e. 10-120 g-I, are in the same numerical range as patients' GFR, i.e. 10-120 mL/min. Thus, forming a g-I/eGFR ratio combines CM volume and concentration, serum creatinine, age and body size into a single continuous risk variable, and provides the examiner with a simple numerical relationship and an expedient way to predict the risk of CIN. This implies also a more sophisticated relationship between CM dose and renal function than the Cigarroa formula (Cigarroa et al., 1989) that lacks CM concentration and

spectrometry (IDMS) with modern equations (Myers et al., 2006).

formula (Dubois & Dubois, 1916 (DuBois & DuBois, 1916):

**4. Systemic drug exposure, gram-iodine/eGFR ratio and CIN** 

**4.1 Area under the plasma concentration-time curve (AUC)** 

with weight expressed in kg and height in cm.

only have a real GFR of 35 mL/min.

al., 2005).

**4.2 Gram-iodine/eGFR ratio** 


A most recently published registry study involving about 50,000 patients recommended a planned gram-iodine dose restricted to 0.7 x eGFR value and not to exceed 1.0 x eGFR if a CM concentration of 350 mg I/mL for PCI is anticipated (Gurm et al., 2011).

Using a g-I/eGFR <1.0 implies a safer maximum dose compared with the Cigarroa formula. A 60-year old female with a height of 160 cm, weight 70 kg and serum creatinine of 150

Contrast Medium-Induced Nephropathy (CIN)

• CIN risk score ≥16 (Table 3) or ≥three risk factors OR

Hypotension (<80 mm Hg for at least 1 h requiring inotropic

Intra-aortic balloon pump 5 Congestive heart failure (New York Heart Association III/IV) 5 Age >75 years 4 Anemia (hematocrit value <39% for men and <36% for women) 3 Diabetes mellitus 3

Serum creatinine >133 μmol/L (1.5 mg/dL) 4

**5. Reducing CM doses in CT-angiography of azotemic patients** 

• Congestive heart failure (NYHA III/IV) OR • Multiple CM exposures within 72 hours

support or intra-aortic balloon pump within 24 h

GFR <60 mL/min/1.73 m2 40-60 20-40 <20

Kristiansson et al., 2010).

Table 3. Mehran CIN risk score (Mehran et al., 2004).

**5.1 CM distribution volume and injected dose rate** 

of the ten studies in Table 2.

• GFR <40 mL/min OR

periprocedurally)

include:

Gram-Iodine/GFR Ratio to Predict CIN and Strategies to Reduce Contrast Medium Doses 201

significant independent predictor of CIN may give erroneous results. Half of the reports in Table 1 used relative eGFR (Liu et al., 2011; Mager et al., 2010; Nozue et al., 2009) and three

If a CM-based examination is deemed necessary in high risk patients, the author's strategy is to keep the g-I/GFR ratio as low as reasonably achievable, preferably below 0.5. Features classifying a patient at high risk of CIN (Kakkar et al., 2008; Mehran et al., 2004) may

Risk factors Integer score

Contrast medium volume 1 for each 100 mL

During the past decade, CTCA has become a clinical reality as a consequence of major advances in CT technology. Vascular enhancement in CT-angiography is dependent on a number of factors such as CM dose, injection rate, plasma volume, cardiac output (CO) and x-ray tube potential (Bae & Heiken, 2005; Fleischmann, 2003; Kormano et al., 1983;

The distribution volume of CM includes the plasma volume and the extravascular extracellular space, both related to body weight. By dosing CM in relation to body weight

5

2 4 6

μmol/mL (1.7 mg/dL) results in an eGFR of 31 mL/min if the IDMS-traceable MDRD equation is used (Levey et al., 2007). At a CM concentration of 350 mg I/mL, 31 grams of iodine will give a maximum CM volume of 88 mL (31,000/350). The corresponding figures in a male will be 41 grams of iodine and 118 mL. According to the Cigarroa formula the maximum volume will be 206 mL (5 × 70/1.7) for both females and males.

Individual patient data from CT studies are lacking, but weighted mean data from CTstudies shows an 8% incidence of CIN at a g-I/eGFR ratio of 0.9 (Table 2), indicating that the ratio should also be kept <1.0 also at CT.


Table 2. Gram-iodine/eGFR ratio and CIN in CT studies. Literature review of nonrandomized and randomized CT-studies reporting mean gram-iodine dose (or volume and concentration), mean eGFR (A = absolute GFR, R = relative GFR), g-I/eGFR ratio (calculated by the author) and incidence of CIN (serum creatinine rise ≥25% or ≥44 μmol/L above baseline). Only results for low-osmolal contrast media (LOCM) included unless there was no significant difference between LOCM and IOCM (iso-osmolal contrast media). Weighted mean value with individual study sizes as weights were finally calculated. The weighted mean of the g-I/eGFR ratio was based on log-transformation.


Note that if GFR adjusted to body surface area is used to form the g-I/GFR ratio, a higher maximum dose may be permitted in small individuals while large individuals may tolerate a larger dose certain ratio would indicate. In addition analyzing g-I/GFR ratio as a significant independent predictor of CIN may give erroneous results. Half of the reports in Table 1 used relative eGFR (Liu et al., 2011; Mager et al., 2010; Nozue et al., 2009) and three of the ten studies in Table 2.

If a CM-based examination is deemed necessary in high risk patients, the author's strategy is to keep the g-I/GFR ratio as low as reasonably achievable, preferably below 0.5. Features classifying a patient at high risk of CIN (Kakkar et al., 2008; Mehran et al., 2004) may include:

• GFR <40 mL/min OR

200 Coronary Interventions

μmol/mL (1.7 mg/dL) results in an eGFR of 31 mL/min if the IDMS-traceable MDRD equation is used (Levey et al., 2007). At a CM concentration of 350 mg I/mL, 31 grams of iodine will give a maximum CM volume of 88 mL (31,000/350). The corresponding figures in a male will be 41 grams of iodine and 118 mL. According to the Cigarroa formula the

Individual patient data from CT studies are lacking, but weighted mean data from CTstudies shows an 8% incidence of CIN at a g-I/eGFR ratio of 0.9 (Table 2), indicating that the

Tepel, 20001 LOCM 42 23 A34 0.7 21 Lufft, 2002 LOCM 33 49 A63 0.8 9.1 Kolehemainen, 2003 LOCM/IOCM 50 35 ?29 1.2 16 Garcia-Ruiz, 2004 LOCM 50 48 A30 1.6 4.0 Becker, 2005 LOCM 100 27 R41 0.7 9.0 Barrett, 2006 LOCM/IOCM 150 40 A45 1.02 3.9 Thomsen, 2008b3 LOCM/IOCM 148 40 A42 1.0 6.1 Nguyen, 2008 LOCM 56 37 A53 0.7 28 Kuhn, 2008 LOCM/IOCM 248 36 R49 0.7 5.2 Weisbord, 2008 LOCM 421 48 R53 0.9 6.5

Weighted mean data 40 47 0.9 7.8

randomized and randomized CT-studies reporting mean gram-iodine dose (or volume and concentration), mean eGFR (A = absolute GFR, R = relative GFR), g-I/eGFR ratio (calculated by the author) and incidence of CIN (serum creatinine rise ≥25% or ≥44 μmol/L above baseline). Only results for low-osmolal contrast media (LOCM) included unless there was no significant difference between LOCM and IOCM (iso-osmolal contrast media). Weighted mean value with individual study sizes as weights were finally calculated. The weighted

Note that if GFR adjusted to body surface area is used to form the g-I/GFR ratio, a higher maximum dose may be permitted in small individuals while large individuals may tolerate a larger dose certain ratio would indicate. In addition analyzing g-I/GFR ratio as a

Table 2. Gram-iodine/eGFR ratio and CIN in CT studies. Literature review of non-

(gram iodine)

eGFR (AmL/min or RmL/min/1.73 m2)

g-I/eGFR ratio

CIN (%)

maximum volume will be 206 mL (5 × 70/1.7) for both females and males.

ratio should also be kept <1.0 also at CT.

First author, year Type of CM N CM dose

Total 1301

mean of the g-I/eGFR ratio was based on log-transformation. 1. Only control group not receiving acetylcysteine included

3. Based on the CIN definition ≥25% serum creatinine increase

2. Based on individual data in the report



Table 3. Mehran CIN risk score (Mehran et al., 2004).
