- Email: [email protected]
Cost reduction in abdominal CT by weight-adjusted dose
European Journal of Radiology 70 (2009) 507–511
Cost reduction in abdominal CT by weight-adjusted dose夽 Estanislao Arana ∗ , Luis Mart´ı-Bonmat´ı, Eva Tobarra, Consuelo Sierra Department of Radiology, Hospital Quir´on, E-4610 Valencia, Spain Received 25 November 2007; accepted 28 January 2008
Abstract Aim: To analyze the influence of contrast dose adjusted by weight vs. fixed contrast dose in the attenuation and cost of abdominal computed tomography (CT). Materials and methods: A randomised, consecutive, parallel group study was conducted in 151 patients (74 men and 77 women, age range 22–67 years), studied with the same CT helical protocol. A dose at 1.75 ml/kg was administered in 101 patients while 50 patients had a fixed dose of 120 ml of same non-ionic contrast material (320 mg/ml). Mean enhancements were measured at right hepatic lobe, superior abdominal aorta and inferior cava vein. Statistical analysis was weight-stratified (<60, 61–70, 71–80 and >81 kg). Results: Aortic attenuation was significantly superior (p < 0.05) in the dose adjusted by weight group than in the fixed dose group. Patients who weighed >61 kg in dose-adjusted group, presented higher hepatic attenuation, being statistically significant in those >81 kg (p < 0.01). In doseadjusted group, there was a savings of D 4.1 per patient in patients weighing <80 kg. In patients weighing >80 kg, there was an over cost of D 10.7 per patient. Conclusions: An injection volume of 1.75 ml/kg offers an optimal diagnostic quality with a global savings of D 1.34 per patient. © 2008 Elsevier Ireland Ltd. All rights reserved. Keywords: Computed tomography (CT), contrast enhancement; Computed tomography (CT), helical technology; Abdomen, CT; Health Care Economics and Organizations; Cost savings
1. Introduction Non-ionic contrast material is the agent of choice for computed tomography (CT) because it is better tolerated by patients compared with ionic contrast material [1]. This product is more expensive, although its price has decreased. A dose reduction is also advantageous in patients with renal insufficiency and/or patients who require additional contrast studies in close temporal solution. Recently, flushing with saline solutions has been shown as a method to reduce the cost and contrast material dose in CT [2]. Although higher concentration doses are gaining acceptance in several CT applications as angiography and liver imaging [3–5], contrast material concentration decrease can be achieved with saline flush [6]. Contrast-enhanced CT in the portal venous phase is the most commonly used phase acquisition in the study of the abdomen. At this time point, most parenchymal structures are at the peak
夽 This work was presented as Scientific Exhibit at the European Congress of Radiology, Vienna, Austria, 5–9 March 2006. ∗ Corresponding author. Tel.: +34 963390405; fax: +34 963391147. E-mail address: [email protected] (E. Arana).
0720-048X/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejrad.2008.01.048
of their enhancement and most lesions will be seen as hypovascular masses. The technique-related factors, which have great influence on the degree of organ enhancement in dynamic contrast-enhanced CT, include the contrast material volume, rate, and type of injection. On the other hand, patient-related factors are mainly weight, gender and age. Patient weight is an important issue regarding quality of CT studies [7]. Several authors have reported that dose of intravenous contrast material should be adjusted to the patients’ weight to achieve adequate contrast enhancement [8]. Contrast media definition in iodine grams may be a more expedient way of improving risk assessment of contrast-medium-induced nephropathy [9]. Our objective was to assess whether a contrast dose of 1.75 ml/kg allows an iodine dose reduction without impairing parenchymal and vascular attenuation in abdominal CT compared to a standard monophasic fixed 120 ml volume, both with saline flush. 2. Materials and methods A randomised, parallel group study was designed and conducted in 171 consecutive patients. Patients who did not
508
E. Arana et al. / European Journal of Radiology 70 (2009) 507–511
consent, had renal failure, congestive heart failure, and contraindication for iodine contrast material or were unconscious or non-cooperative were excluded (four patients). In addition, 16 patients with solid liver and kidney tumours or ongoing chemotherapy were excluded to avoid errors in measurement of contrast material enhancement due to hemodynamic alterations or fatty infiltration of the liver. Patients were free to withdraw from trial at any moment, in accordance with the Helsinki declaration. Finally, there were 151 adult patients included in the study, referred for an abdominal contrast-enhanced helical CT examination. All patients signed an informed consent. There were 74 males and 77 females, with a mean age of 55.5 ± 16 years (range 27–80 years). Patients’ weight was 70.4 ± 13.9 kg. Patients were consequently studied with two monophasic injector protocol groups: (A) 50 patients, 120 ml at 2.5 ml/s; (B) 101 patients, weight-dosage of 1.75 ml/kg. All patients were injected, through an ante brachial route the same non-ionic contrast media (Iodixanol 320, Visipaque® 320 mg/ml, Amersham Health, Norway). There were no statistically significant differences in age or body weight between the study groups. The power injector (Tomojet CT, Bruker, Switzerland) was a double power with two separated interconnected injectors for contrast material and saline solution, allowing the injection of saline to push the contrast agent (40 ml, 2.5 ml/s). Patients fasted for 6 h before the CT examination. Oral water was administered to every patient at a volume around 1500 ml to ensure bowel analysis. Scanning was performed with a single detector CT helical protocol (craniocaudal acquisition rotation time: 1 s; collimation: 7 mm; nominal slice thickness: 5 mm; pitch 1.5, 100 mAs, 120 kV) in the same scanner (CT Aura, Philips Medical Systems, The Netherlands). Scans were obtained with the patient in suspended respiration. The average abdominal helical CT examination lasted 21 ± 4 s. For all patients, a portal phase was obtained at a fixed delay time of 55 s after the starting of the contrast delivery. At this time, there is a high enhancement of the liver and portal vein. Contrast enhancement efficacy was evaluated by measuring attenuation coefficients. Right hepatic lobe, superior abdominal aorta and extra hepatic inferior vena cava just caudal to the entrance of the renal veins attenuation coefficients were obtained by operator-defined manually placed regions-of-interest (ROIs), and expressed in Hounsfield units (HU). A registered nurse and radiographer with experience in abdominal CT located the ROIs, with a size between 350 and 450 mm2 blinded to the injection protocol. Within the liver, ROIs were situated excluding vascular and biliary structures. Statistical analysis was performed for the two groups with the Pearson χ2 -test for contingency tables (gender) and, after testing the normal distribution with the Kolmogorov–Smirnov, an ANOVA with the Student–Newman–Keuls test (age, weight and attenuation measurements). Patient’s weight was divided in four groups (<60 kg, 42 patients; 61–70 kg, 48 patients; 71–80 kg, 32 patients and >80 kg, 30 patients). Results are expressed as mean ± standard error of the mean. A statistical significance was established with a p < 0.05. Patient-level cost data often have skewed distributions, with a result from standard tests
based on normality assumptions that may not be appropriate for small group comparisons [10], and for this reason, bootstrap method was chosen. The nonparametric bootstrap method is a resampling approach that successively recreates the data sample (repeated samples) by random selection (with replacement) from the original sample, from which values of interest such as average differences are computed [11]. Bootstrap estimation of cost with the 95% confidence limits were calculated using the 2.5 and 97.5 percentiles of the computed bootstrap average difference in 1000 resamples [11,12]. All exchange rates were calculated on 26 September 2006, inflation adjusted for previous years. 3. Results There were no statistically significant differences of gender or weight between the two groups. Aortic attenuation was significantly superior (p < 0.05) in the dose adjusted by weight group than in the fixed contrast dose group with mean values equal or greater than 200 HU, except in less than 60 kg (Fig. 1). Liver attenuation was higher in all patients with dose adjusted by weight who weighted more than 60 kg, being statistically significant in those with more than 81 kg (Fig. 2). However, in the fixed contrast dose group, attenuation was significantly higher in patients with less than 60 kg. Vena cava attenuation was higher in the fixed contrast dose group, with no statistical significance, except in patients weighing less than 60 kg (p < 0.01). Fig. 3 shows the iodine grams administered according to the injection group. Above 71 kg, there was more iodine grams in
Fig. 1. Graph illustrates comparison of aortic enhancement with patient weight divided into 10-kg intervals. Continuous bold line: fixed contrast dose group.
E. Arana et al. / European Journal of Radiology 70 (2009) 507–511
509
Fig. 4. Bootstrap estimation after 1000 resamples of the savings and 95% confidence interval per patient according to weight group (bars) compared to the administration of 120 ml.
Fig. 2. Graph illustrates comparison of liver enhancement with patient weight divided into 10-kg intervals. Continuous bold line: fixed contrast dose group.
the dose adjusted by weight group than in the fixed contrast dose, exceeding the 38 g, marked as Refs. [1,3]. 3.1. Cost analysis On the basis of the list price of iodixanol 320 (500 ml costs or D 0.47 ml/ml), the total cost of contrast if the 101 patients (who were actually scanned after receiving 1.75 ml/kg) had received a standard 120 ml dose would have been D 5417.6. D 223.5,
Fig. 3. Graph shows iodine grams according to patient’s weight per injection group. Straight parallel lines mark the 26 g (patient less than 83 kg) and 38 g (for more than 83 kg) references considered adequate for liver enhancement [15].
However, the actual cost of contrast medium (found by summing the individual weight-based cost incurred by the 101 patients) was D 5281.5, with an aggregate savings of D 136.1, or an average of D 1.34 per patient. In this weight-adjusted group patients weighing less than 80 kg, we saved a mean of D 4.1 per patient. In patients >80 kg, there was and over cost of D 10.7 per patient. Kruskal–Wallis test revealed statistical differences among the groups (χ2 87.4, p < 0.001) (Fig. 4). 4. Discussion Present benefits in using low osmolality non-ionic contrast media include improved organ enhancement and a decrease in the expected number of contrast material-related reactions. Previous investigations have provided information about the effect of different doses of intravenous contrast material in helical CT of the abdomen and about the minimum volume of intravenous contrast material required for abdominal CT [1,3,6,14]. However, controversy persists regarding the effect of the injection rate of contrast material on hepatic enhancement [1,3]. Previous studies with weight-adjusted dose found no clinically significant difference at doses greater than 1.5 ml/kg with 300 mg I/ml contrast [14]. These different studies stated that minimum dose of contrast material required for abdominal CT is approximately 26–30 g iodine [1]. Yamashita et al. [3] believe that 34–42 g iodine is required for adequate hepatic enhancement. There are few studies that have investigated saline flushing in abdominal helical CT [1,2,6] and neither using weight-adjusted contrast dose. In the present study, we wanted to evaluate the potential of a weightadjusted dose of 1.75 ml/kg of contrast with saline flush on contrast material dose reduction in abdominal CT. We used an iodine dose of 560 mg/kg patient body weight, a compromise between recommended doses of 521 mg I/kg [8] and 600 mg I/kg [3]. Our entire liver enhancement was above the 50 HU recommended by Brink et al. [15]. This enhancement was achieved with an iodine dose of 38.4 g in the fixed dose and above 45.3 g in heavy patients. The effect of patient weight has been noted in previous studies. Brink et al. [15] found that
510
E. Arana et al. / European Journal of Radiology 70 (2009) 507–511
in heavy patients, a dose of 38 g of iodine produced adequate enhancement whereas, doses as small as 26 g may be sufficient in thin patients. In the present study, subjects within more than 80 kg had more than 44.8 g of iodine than their controls at 120 ml, but showed higher liver attenuation. Likewise, Megibow et al. [14] showed a higher acceptable rate of liver scans in their heaviest patients using 1.5 ml/kg compared to 150 ml. Higher contrast doses did not increase statistically the acceptability of liver scans. In previous reports, contrast material costs can be saved by dose reduction, as in incremental CT by increasing the concentration and decreasing the volume [16]. Awai et al. selected an enhancement value of 200 HU to indicate adequate enhancement of the aorta [8]. In other work, the same authors concluded that the minimal vascular enhancement needed to discriminate lymph nodes from vessels was 44 HU [17]. In the dose adjusted by weight group, mean enhancement of the abdominal aorta was above 200 HU. This level of vascular enhancement is considered sufficient for diagnosis, because the purpose of examination is not performing CT angiography but to obtain a valid study of the whole abdominal cavity. Our aortic enhancement was consistently higher in patients weighing more than 61 kg in the weight-adjusted dose group. Similar results have been obtained in weight-adjusted doses for aortic imaging [8]. The use of contrast material dose based on patients’ weight with a fixed duration injection (as in our work) is equivalent to injecting a certain volume of contrast material into the central blood compartment in a given amount of time. When the rate of contrast material was fixed, the aortic enhancement value decreased as patient weight increased [8]. Awai et al. found similar aortic results in smaller subjects [8]. Our saving per patient is lower than Megibow et al. [14], who found no clinically significant differences in acceptability of scans at doses greater than 1.5 ml/kg, although their type of CT scanners were not detailed. They compared 1.5 ml/kg and 150 ml, with a wider range in the total amount of contrast administered and without saline flush with an average of D 27.27 per patient, inflation adjusted. As it has been shown, cost can also be lowered using saline flush. Schoellnast et al. [2] compared 120–100 ml plus 40 ml of saline flush, and found savings of D 5.89. However, all their patients who weighted more than 80 kg had less enhancement than in the present work. We think that our over cost of D 10.71 per patient with more than 81 kg is effective, because they presented statistically significant enhancement values at liver and aortic locations. However, we have not studied thoroughly their clinical implication. In abdominal CT, a parenchymal phase obtained 50–70 s after the start of the contrast administration will ensure a high parenchymal enhancement [13]. A high level of parenchymal and vascular enhancement is desirable to increase the radiologist’s performance. Our randomised study design was conducted to assess the differences in vascular and parenchymal enhancement after different contrast media concentrations observed during a single CT acquisition of the abdomen in the portal phase. Although patients were homogeneously distributed among groups, it does not imply equality. In obese patients, the well-perfused extracellular compartment accounts for a lower
proportion of the weight. Aortic CT studies have showed that small amounts of contrast material flows from the central blood compartment to the well-perfused extracellular compartment. Adipose tissue is considered to represent the poorly perfused extracellular compartment, being the highest proportion of the weight. Therefore, we agree with Awai et al. [8], that determination of contrast material doses according to patient weight should use body weight ratio in the future [8]. Notwithstanding, the ranges and mean values of body weight of the patients examined in our study are according to the European Statistics [18]. One of the inherent limitations in devising a universal scanning protocol is the inability to consider biological variability. Although we did not analyze the circulation time, for a contrast medium volume of 120 ml, a scanning delay of approximately 65 s ensures the depiction of the portal hepatic and pancreatic phases. A similar scanning delay (60 s) has shown to be the best compromise between the short gap between pancreatic and hepatic peaks using a single acquisition [13]. Because of individual variation in circulation time, slower injection (2–3 ml/s) of a large amount of contrast medium (longer injection) and an optimised CT protocol may be an effective method of obtaining good quality abdominal CT images [19]. Another bias is our velocity regarding single helical CT. With the increased availability of multidetector CT, these results should be transferred to their improved temporal resolution, although 300 mg I/ml and similar injection rates are currently used in abdominal MDCT studies [1,17]. In conclusion, when dose was tailored to patient weight, the use of 1.75 ml/kg of intravenous contrast material with saline flushing in abdominal SDCT allowed a reduction of contrast material dose. The savings are of approximately 0.96 g, or 2.51%, without significantly decreasing mean parenchymal or aortic attenuation. Consequently, a cost reduction of about D 1.34 per patient could be achieved. References [1] Schoellnast H, Tillich M, Deutschmann HA, et al. Improvement of parenchymal and vascular enhancement using saline flush and power injection for multiple-detector-row abdominal CT. Eur Radiol 2004;14:659– 64. [2] Schoellnast H, Tillich M, Deutschmann HA, et al. Abdominal multidetector row computed tomography: reduction of cost and contrast material dose using saline flush. J Comput Assist Tomogr 2003;27:847– 53. [3] Yamashita Y, Komohara Y, Takahashi M, et al. Abdominal helical CT: evaluation of optimal doses of intravenous contrast material—a prospective randomized study. Radiology 2000;216:718–23. [4] Awai K, Inoue M, Yagyu Y, et al. Moderate versus high concentration of contrast material for aortic and hepatic enhancement and tumor-to-liver contrast at multi-detector row CT. Radiology 2004;233:682–8. [5] Herman S. Computed tomography contrast enhancement principles and the use of high-concentration contrast media. J Comput Assist Tomogr 2004;28(Suppl. 1):S7–11. [6] Tatsugami F, Matsuki M, Kani H, et al. Effect of saline pushing after contrast material injection in abdominal multidetector computed tomography with the use of different iodine concentrations. Acta Radiol 2006;47:192–7. [7] Kalra MK, Maher MM, Prasad SR, et al. Correlation of patient weight and cross-sectional dimensions with subjective image quality at standard dose abdominal CT. Korean J Radiol 2003;4:234–8.
E. Arana et al. / European Journal of Radiology 70 (2009) 507–511 [8] Awai K, Hiraishi K, Hori S. Effect of contrast material injection duration and rate on aortic peak time and peak enhancement at dynamic CT involving injection protocol with dose tailored to patient weight. Radiology 2004;230:142–50. [9] Nyman U, Almen T, Aspelin P, et al. Contrast-medium-induced nephropathy correlated to the ratio between dose in gram iodine and estimated GFR in ml/min. Acta Radiol 2005;46:830–42. [10] Briggs AH, Gray AM. Handling uncertainty when performing economic evaluation of healthcare interventions. Health Technol Assess 1999;3:1–134. [11] Wood M. Statistical inference using bootstrap confidence intervals. Significance 2004;1:180–2. [12] Thompson SG, Barber JA. How should cost data in pragmatic randomised trials be analysed? Br Med J 2000;320:1197–200. [13] Marti-Bonmati L, Arana E, Tobarra E, et al. Low–high and high–low biphasic injection forms in computed tomography examinations of the upper abdomen. Acta Radiol 2006;47:10–4.
511
[14] Megibow AJ, Jacob G, Heiken JP, et al. Quantitative and qualitative evaluation of volume of low osmolality contrast medium needed for routine helical abdominal CT. AJR Am J Roentgenol 2001;176:583–9. [15] Brink JA, Heiken JP, Forman HP, et al. Hepatic spiral CT: reduction of dose of intravenous contrast material. Radiology 1995;197:83–8. [16] Baker ME, Beam C, Leder R, et al. Contrast material for combined abdominal and pelvic CT: can cost be reduced by increasing the concentration and decreasing the volume? AJR Am J Roentgenol 1993;160:637–41. [17] Awai K, Imuta M, Utsunomiya D, et al. Contrast enhancement for wholebody screening using multidetector row helical CT: comparison between uniphasic and biphasic injection protocols. Radiat Med 2004;22:303–9. [18] Eurostat. Health statistics-key data on health, 139–140. Luxembourg: Office for Official Publications of the European Communities; 2002. p. 197. [19] Han JK, Kim AY, Lee KY, et al. Factors influencing vascular and hepatic enhancement at CT: experimental study on injection protocol using a canine model. J Comput Assist Tomogr 2000;24:400–6.
Cost reduction in abdominal CT by weight-adjusted dose夽 Estanislao Arana ∗ , Luis Mart´ı-Bonmat´ı, Eva Tobarra, Consuelo Sierra Department of Radiology, Hospital Quir´on, E-4610 Valencia, Spain Received 25 November 2007; accepted 28 January 2008
Abstract Aim: To analyze the influence of contrast dose adjusted by weight vs. fixed contrast dose in the attenuation and cost of abdominal computed tomography (CT). Materials and methods: A randomised, consecutive, parallel group study was conducted in 151 patients (74 men and 77 women, age range 22–67 years), studied with the same CT helical protocol. A dose at 1.75 ml/kg was administered in 101 patients while 50 patients had a fixed dose of 120 ml of same non-ionic contrast material (320 mg/ml). Mean enhancements were measured at right hepatic lobe, superior abdominal aorta and inferior cava vein. Statistical analysis was weight-stratified (<60, 61–70, 71–80 and >81 kg). Results: Aortic attenuation was significantly superior (p < 0.05) in the dose adjusted by weight group than in the fixed dose group. Patients who weighed >61 kg in dose-adjusted group, presented higher hepatic attenuation, being statistically significant in those >81 kg (p < 0.01). In doseadjusted group, there was a savings of D 4.1 per patient in patients weighing <80 kg. In patients weighing >80 kg, there was an over cost of D 10.7 per patient. Conclusions: An injection volume of 1.75 ml/kg offers an optimal diagnostic quality with a global savings of D 1.34 per patient. © 2008 Elsevier Ireland Ltd. All rights reserved. Keywords: Computed tomography (CT), contrast enhancement; Computed tomography (CT), helical technology; Abdomen, CT; Health Care Economics and Organizations; Cost savings
1. Introduction Non-ionic contrast material is the agent of choice for computed tomography (CT) because it is better tolerated by patients compared with ionic contrast material [1]. This product is more expensive, although its price has decreased. A dose reduction is also advantageous in patients with renal insufficiency and/or patients who require additional contrast studies in close temporal solution. Recently, flushing with saline solutions has been shown as a method to reduce the cost and contrast material dose in CT [2]. Although higher concentration doses are gaining acceptance in several CT applications as angiography and liver imaging [3–5], contrast material concentration decrease can be achieved with saline flush [6]. Contrast-enhanced CT in the portal venous phase is the most commonly used phase acquisition in the study of the abdomen. At this time point, most parenchymal structures are at the peak
夽 This work was presented as Scientific Exhibit at the European Congress of Radiology, Vienna, Austria, 5–9 March 2006. ∗ Corresponding author. Tel.: +34 963390405; fax: +34 963391147. E-mail address: [email protected] (E. Arana).
0720-048X/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejrad.2008.01.048
of their enhancement and most lesions will be seen as hypovascular masses. The technique-related factors, which have great influence on the degree of organ enhancement in dynamic contrast-enhanced CT, include the contrast material volume, rate, and type of injection. On the other hand, patient-related factors are mainly weight, gender and age. Patient weight is an important issue regarding quality of CT studies [7]. Several authors have reported that dose of intravenous contrast material should be adjusted to the patients’ weight to achieve adequate contrast enhancement [8]. Contrast media definition in iodine grams may be a more expedient way of improving risk assessment of contrast-medium-induced nephropathy [9]. Our objective was to assess whether a contrast dose of 1.75 ml/kg allows an iodine dose reduction without impairing parenchymal and vascular attenuation in abdominal CT compared to a standard monophasic fixed 120 ml volume, both with saline flush. 2. Materials and methods A randomised, parallel group study was designed and conducted in 171 consecutive patients. Patients who did not
508
E. Arana et al. / European Journal of Radiology 70 (2009) 507–511
consent, had renal failure, congestive heart failure, and contraindication for iodine contrast material or were unconscious or non-cooperative were excluded (four patients). In addition, 16 patients with solid liver and kidney tumours or ongoing chemotherapy were excluded to avoid errors in measurement of contrast material enhancement due to hemodynamic alterations or fatty infiltration of the liver. Patients were free to withdraw from trial at any moment, in accordance with the Helsinki declaration. Finally, there were 151 adult patients included in the study, referred for an abdominal contrast-enhanced helical CT examination. All patients signed an informed consent. There were 74 males and 77 females, with a mean age of 55.5 ± 16 years (range 27–80 years). Patients’ weight was 70.4 ± 13.9 kg. Patients were consequently studied with two monophasic injector protocol groups: (A) 50 patients, 120 ml at 2.5 ml/s; (B) 101 patients, weight-dosage of 1.75 ml/kg. All patients were injected, through an ante brachial route the same non-ionic contrast media (Iodixanol 320, Visipaque® 320 mg/ml, Amersham Health, Norway). There were no statistically significant differences in age or body weight between the study groups. The power injector (Tomojet CT, Bruker, Switzerland) was a double power with two separated interconnected injectors for contrast material and saline solution, allowing the injection of saline to push the contrast agent (40 ml, 2.5 ml/s). Patients fasted for 6 h before the CT examination. Oral water was administered to every patient at a volume around 1500 ml to ensure bowel analysis. Scanning was performed with a single detector CT helical protocol (craniocaudal acquisition rotation time: 1 s; collimation: 7 mm; nominal slice thickness: 5 mm; pitch 1.5, 100 mAs, 120 kV) in the same scanner (CT Aura, Philips Medical Systems, The Netherlands). Scans were obtained with the patient in suspended respiration. The average abdominal helical CT examination lasted 21 ± 4 s. For all patients, a portal phase was obtained at a fixed delay time of 55 s after the starting of the contrast delivery. At this time, there is a high enhancement of the liver and portal vein. Contrast enhancement efficacy was evaluated by measuring attenuation coefficients. Right hepatic lobe, superior abdominal aorta and extra hepatic inferior vena cava just caudal to the entrance of the renal veins attenuation coefficients were obtained by operator-defined manually placed regions-of-interest (ROIs), and expressed in Hounsfield units (HU). A registered nurse and radiographer with experience in abdominal CT located the ROIs, with a size between 350 and 450 mm2 blinded to the injection protocol. Within the liver, ROIs were situated excluding vascular and biliary structures. Statistical analysis was performed for the two groups with the Pearson χ2 -test for contingency tables (gender) and, after testing the normal distribution with the Kolmogorov–Smirnov, an ANOVA with the Student–Newman–Keuls test (age, weight and attenuation measurements). Patient’s weight was divided in four groups (<60 kg, 42 patients; 61–70 kg, 48 patients; 71–80 kg, 32 patients and >80 kg, 30 patients). Results are expressed as mean ± standard error of the mean. A statistical significance was established with a p < 0.05. Patient-level cost data often have skewed distributions, with a result from standard tests
based on normality assumptions that may not be appropriate for small group comparisons [10], and for this reason, bootstrap method was chosen. The nonparametric bootstrap method is a resampling approach that successively recreates the data sample (repeated samples) by random selection (with replacement) from the original sample, from which values of interest such as average differences are computed [11]. Bootstrap estimation of cost with the 95% confidence limits were calculated using the 2.5 and 97.5 percentiles of the computed bootstrap average difference in 1000 resamples [11,12]. All exchange rates were calculated on 26 September 2006, inflation adjusted for previous years. 3. Results There were no statistically significant differences of gender or weight between the two groups. Aortic attenuation was significantly superior (p < 0.05) in the dose adjusted by weight group than in the fixed contrast dose group with mean values equal or greater than 200 HU, except in less than 60 kg (Fig. 1). Liver attenuation was higher in all patients with dose adjusted by weight who weighted more than 60 kg, being statistically significant in those with more than 81 kg (Fig. 2). However, in the fixed contrast dose group, attenuation was significantly higher in patients with less than 60 kg. Vena cava attenuation was higher in the fixed contrast dose group, with no statistical significance, except in patients weighing less than 60 kg (p < 0.01). Fig. 3 shows the iodine grams administered according to the injection group. Above 71 kg, there was more iodine grams in
Fig. 1. Graph illustrates comparison of aortic enhancement with patient weight divided into 10-kg intervals. Continuous bold line: fixed contrast dose group.
E. Arana et al. / European Journal of Radiology 70 (2009) 507–511
509
Fig. 4. Bootstrap estimation after 1000 resamples of the savings and 95% confidence interval per patient according to weight group (bars) compared to the administration of 120 ml.
Fig. 2. Graph illustrates comparison of liver enhancement with patient weight divided into 10-kg intervals. Continuous bold line: fixed contrast dose group.
the dose adjusted by weight group than in the fixed contrast dose, exceeding the 38 g, marked as Refs. [1,3]. 3.1. Cost analysis On the basis of the list price of iodixanol 320 (500 ml costs or D 0.47 ml/ml), the total cost of contrast if the 101 patients (who were actually scanned after receiving 1.75 ml/kg) had received a standard 120 ml dose would have been D 5417.6. D 223.5,
Fig. 3. Graph shows iodine grams according to patient’s weight per injection group. Straight parallel lines mark the 26 g (patient less than 83 kg) and 38 g (for more than 83 kg) references considered adequate for liver enhancement [15].
However, the actual cost of contrast medium (found by summing the individual weight-based cost incurred by the 101 patients) was D 5281.5, with an aggregate savings of D 136.1, or an average of D 1.34 per patient. In this weight-adjusted group patients weighing less than 80 kg, we saved a mean of D 4.1 per patient. In patients >80 kg, there was and over cost of D 10.7 per patient. Kruskal–Wallis test revealed statistical differences among the groups (χ2 87.4, p < 0.001) (Fig. 4). 4. Discussion Present benefits in using low osmolality non-ionic contrast media include improved organ enhancement and a decrease in the expected number of contrast material-related reactions. Previous investigations have provided information about the effect of different doses of intravenous contrast material in helical CT of the abdomen and about the minimum volume of intravenous contrast material required for abdominal CT [1,3,6,14]. However, controversy persists regarding the effect of the injection rate of contrast material on hepatic enhancement [1,3]. Previous studies with weight-adjusted dose found no clinically significant difference at doses greater than 1.5 ml/kg with 300 mg I/ml contrast [14]. These different studies stated that minimum dose of contrast material required for abdominal CT is approximately 26–30 g iodine [1]. Yamashita et al. [3] believe that 34–42 g iodine is required for adequate hepatic enhancement. There are few studies that have investigated saline flushing in abdominal helical CT [1,2,6] and neither using weight-adjusted contrast dose. In the present study, we wanted to evaluate the potential of a weightadjusted dose of 1.75 ml/kg of contrast with saline flush on contrast material dose reduction in abdominal CT. We used an iodine dose of 560 mg/kg patient body weight, a compromise between recommended doses of 521 mg I/kg [8] and 600 mg I/kg [3]. Our entire liver enhancement was above the 50 HU recommended by Brink et al. [15]. This enhancement was achieved with an iodine dose of 38.4 g in the fixed dose and above 45.3 g in heavy patients. The effect of patient weight has been noted in previous studies. Brink et al. [15] found that
510
E. Arana et al. / European Journal of Radiology 70 (2009) 507–511
in heavy patients, a dose of 38 g of iodine produced adequate enhancement whereas, doses as small as 26 g may be sufficient in thin patients. In the present study, subjects within more than 80 kg had more than 44.8 g of iodine than their controls at 120 ml, but showed higher liver attenuation. Likewise, Megibow et al. [14] showed a higher acceptable rate of liver scans in their heaviest patients using 1.5 ml/kg compared to 150 ml. Higher contrast doses did not increase statistically the acceptability of liver scans. In previous reports, contrast material costs can be saved by dose reduction, as in incremental CT by increasing the concentration and decreasing the volume [16]. Awai et al. selected an enhancement value of 200 HU to indicate adequate enhancement of the aorta [8]. In other work, the same authors concluded that the minimal vascular enhancement needed to discriminate lymph nodes from vessels was 44 HU [17]. In the dose adjusted by weight group, mean enhancement of the abdominal aorta was above 200 HU. This level of vascular enhancement is considered sufficient for diagnosis, because the purpose of examination is not performing CT angiography but to obtain a valid study of the whole abdominal cavity. Our aortic enhancement was consistently higher in patients weighing more than 61 kg in the weight-adjusted dose group. Similar results have been obtained in weight-adjusted doses for aortic imaging [8]. The use of contrast material dose based on patients’ weight with a fixed duration injection (as in our work) is equivalent to injecting a certain volume of contrast material into the central blood compartment in a given amount of time. When the rate of contrast material was fixed, the aortic enhancement value decreased as patient weight increased [8]. Awai et al. found similar aortic results in smaller subjects [8]. Our saving per patient is lower than Megibow et al. [14], who found no clinically significant differences in acceptability of scans at doses greater than 1.5 ml/kg, although their type of CT scanners were not detailed. They compared 1.5 ml/kg and 150 ml, with a wider range in the total amount of contrast administered and without saline flush with an average of D 27.27 per patient, inflation adjusted. As it has been shown, cost can also be lowered using saline flush. Schoellnast et al. [2] compared 120–100 ml plus 40 ml of saline flush, and found savings of D 5.89. However, all their patients who weighted more than 80 kg had less enhancement than in the present work. We think that our over cost of D 10.71 per patient with more than 81 kg is effective, because they presented statistically significant enhancement values at liver and aortic locations. However, we have not studied thoroughly their clinical implication. In abdominal CT, a parenchymal phase obtained 50–70 s after the start of the contrast administration will ensure a high parenchymal enhancement [13]. A high level of parenchymal and vascular enhancement is desirable to increase the radiologist’s performance. Our randomised study design was conducted to assess the differences in vascular and parenchymal enhancement after different contrast media concentrations observed during a single CT acquisition of the abdomen in the portal phase. Although patients were homogeneously distributed among groups, it does not imply equality. In obese patients, the well-perfused extracellular compartment accounts for a lower
proportion of the weight. Aortic CT studies have showed that small amounts of contrast material flows from the central blood compartment to the well-perfused extracellular compartment. Adipose tissue is considered to represent the poorly perfused extracellular compartment, being the highest proportion of the weight. Therefore, we agree with Awai et al. [8], that determination of contrast material doses according to patient weight should use body weight ratio in the future [8]. Notwithstanding, the ranges and mean values of body weight of the patients examined in our study are according to the European Statistics [18]. One of the inherent limitations in devising a universal scanning protocol is the inability to consider biological variability. Although we did not analyze the circulation time, for a contrast medium volume of 120 ml, a scanning delay of approximately 65 s ensures the depiction of the portal hepatic and pancreatic phases. A similar scanning delay (60 s) has shown to be the best compromise between the short gap between pancreatic and hepatic peaks using a single acquisition [13]. Because of individual variation in circulation time, slower injection (2–3 ml/s) of a large amount of contrast medium (longer injection) and an optimised CT protocol may be an effective method of obtaining good quality abdominal CT images [19]. Another bias is our velocity regarding single helical CT. With the increased availability of multidetector CT, these results should be transferred to their improved temporal resolution, although 300 mg I/ml and similar injection rates are currently used in abdominal MDCT studies [1,17]. In conclusion, when dose was tailored to patient weight, the use of 1.75 ml/kg of intravenous contrast material with saline flushing in abdominal SDCT allowed a reduction of contrast material dose. The savings are of approximately 0.96 g, or 2.51%, without significantly decreasing mean parenchymal or aortic attenuation. Consequently, a cost reduction of about D 1.34 per patient could be achieved. References [1] Schoellnast H, Tillich M, Deutschmann HA, et al. Improvement of parenchymal and vascular enhancement using saline flush and power injection for multiple-detector-row abdominal CT. Eur Radiol 2004;14:659– 64. [2] Schoellnast H, Tillich M, Deutschmann HA, et al. Abdominal multidetector row computed tomography: reduction of cost and contrast material dose using saline flush. J Comput Assist Tomogr 2003;27:847– 53. [3] Yamashita Y, Komohara Y, Takahashi M, et al. Abdominal helical CT: evaluation of optimal doses of intravenous contrast material—a prospective randomized study. Radiology 2000;216:718–23. [4] Awai K, Inoue M, Yagyu Y, et al. Moderate versus high concentration of contrast material for aortic and hepatic enhancement and tumor-to-liver contrast at multi-detector row CT. Radiology 2004;233:682–8. [5] Herman S. Computed tomography contrast enhancement principles and the use of high-concentration contrast media. J Comput Assist Tomogr 2004;28(Suppl. 1):S7–11. [6] Tatsugami F, Matsuki M, Kani H, et al. Effect of saline pushing after contrast material injection in abdominal multidetector computed tomography with the use of different iodine concentrations. Acta Radiol 2006;47:192–7. [7] Kalra MK, Maher MM, Prasad SR, et al. Correlation of patient weight and cross-sectional dimensions with subjective image quality at standard dose abdominal CT. Korean J Radiol 2003;4:234–8.
E. Arana et al. / European Journal of Radiology 70 (2009) 507–511 [8] Awai K, Hiraishi K, Hori S. Effect of contrast material injection duration and rate on aortic peak time and peak enhancement at dynamic CT involving injection protocol with dose tailored to patient weight. Radiology 2004;230:142–50. [9] Nyman U, Almen T, Aspelin P, et al. Contrast-medium-induced nephropathy correlated to the ratio between dose in gram iodine and estimated GFR in ml/min. Acta Radiol 2005;46:830–42. [10] Briggs AH, Gray AM. Handling uncertainty when performing economic evaluation of healthcare interventions. Health Technol Assess 1999;3:1–134. [11] Wood M. Statistical inference using bootstrap confidence intervals. Significance 2004;1:180–2. [12] Thompson SG, Barber JA. How should cost data in pragmatic randomised trials be analysed? Br Med J 2000;320:1197–200. [13] Marti-Bonmati L, Arana E, Tobarra E, et al. Low–high and high–low biphasic injection forms in computed tomography examinations of the upper abdomen. Acta Radiol 2006;47:10–4.
511
[14] Megibow AJ, Jacob G, Heiken JP, et al. Quantitative and qualitative evaluation of volume of low osmolality contrast medium needed for routine helical abdominal CT. AJR Am J Roentgenol 2001;176:583–9. [15] Brink JA, Heiken JP, Forman HP, et al. Hepatic spiral CT: reduction of dose of intravenous contrast material. Radiology 1995;197:83–8. [16] Baker ME, Beam C, Leder R, et al. Contrast material for combined abdominal and pelvic CT: can cost be reduced by increasing the concentration and decreasing the volume? AJR Am J Roentgenol 1993;160:637–41. [17] Awai K, Imuta M, Utsunomiya D, et al. Contrast enhancement for wholebody screening using multidetector row helical CT: comparison between uniphasic and biphasic injection protocols. Radiat Med 2004;22:303–9. [18] Eurostat. Health statistics-key data on health, 139–140. Luxembourg: Office for Official Publications of the European Communities; 2002. p. 197. [19] Han JK, Kim AY, Lee KY, et al. Factors influencing vascular and hepatic enhancement at CT: experimental study on injection protocol using a canine model. J Comput Assist Tomogr 2000;24:400–6.