Barbara Ulibarri^{1}, Carlos MartinezDuncker^{2}, Claudia Camarillo^{2}, Enrique Omar Guadarrama^{1}, Pilar Garcia Roca^{3}, Ricardo Munoz Arizpe^{3}, Mara Medeiros^{3}. ^{1}Departamento de Nefrologia, Hospital Infantil de Mexico Federico Gomez, Mexico, ^{2}Departamento Medicina Nuclear, Hospital Infantil de Mexico Federico Gomez, Mexico, ^{3}Laboratorio de Investigacion en Nefrologia., Hospital Infantil de Mexico Federico Gomez, Mexico.CITE THIS ARTICLE
Ulibarri B, MartinezDuncker C, Camarillo C, Guadarrama E O, Roca P G, Arizpe R M, Medeiros M. Comparison of 99m Tc DTPA vs Five Methods to Estimate Glomerular Filtration Rate in Children with Renal Transplantation. Pediatr Oncall J. 2009;6: 6770.


 ADDRESS FOR CORRESPONDANCE Mara Medeiros, Laboratorio de Investigacion en Nefrologia, Hospital Infantil de Mexico Federico Gomez, Dr. Marquez 162 Col. Doctores., Mexico D.F., Mexico 06720. Email: maramedeiros@hotmail.com
 Abstract  The purpose of the study was to compare the GFR obtained with ^{99m}Tc DTPA against five other methods: the GFR measured by 24 hours creatinine clearance (Ccr), and four creatininebased equations (Schwartz, CounahanBarrat, Morris and Leger). A descriptive, crosssectional study was performed in 42 renal transplant children; a 24 hours urine collection was obtained for creatinine clearance. A ^{99m}TcDTPA renal gammagram was performed the same day. The mean GFR (mL/min1.73m2BSA) obtained with 99mTcDTPA was 66.9 ± 15, whereas by Ccr was 68.7 ± 24.4, by Schwartz formula 81.8 ± 16, by the CounahanBarrat equation was 57.6 ± 13.6, by Morris equation 60.7 ± 14.5 and by Leger equation 73.2 ± 18. The predictive performance was evaluated calculating bias as mean prediction error (MPE) and precision (Mean Squared Error, MSE). The Schwartz formula had a MPE of 14.8 and MSE of 9223. The Ccr had a MPE of 1.7 and MSE of 132, CounahanBarrat MPE 9.4, MSE 3727, Morris 6.28, MSE 1656, Leger 6.2 MSE 2000. Ccr had the lowest bias but is not precise to estimate GFR in renal transplant children. The Morris equation had better predictive performance in renal transplant children than the Schwartz, CounahanBarrat and Leger equations.   Keywords  Glomerular filtration rate, renal transplant, children, Schwartz, GFR measurement   Introduction  According to the NKFK/DOQI guidelines to evaluation, classification and stratification of chronic kidney disease in children and adolescents, the estimates of glomerular filtration rate (GFR) are the best indices of the level of kidney function, and the creatinine clearance using timed urine collections do not improve the estimated GFR over that provided by prediction equations [1]. The GFR estimation in children is made by exogenous markers such as inulin, cystatine C, iothalamate or the radiolabeled ^{99m}Tecnetium diethylenetriamine pentaacetic acid (^{99m}TcDTPA), accepted as the best methods. Nevertheless, these procedures are invasive, timeconsuming and expensive, and not easy to perform in the every day practice [2]. For these reasons children are more frequently monitored with the GFR prediction equations using serumcreatinine based formulas. The GFR prediction equations are more accurate when GFR is normal (>90mL/min per 1.73 m²), otherwise they tend to overestimate the GFR [3]. The majority of the equations used have been initially developed for native kidneys, but a caveat appears when the equation are used to estimate the GFR in transplanted kidneys, since the nephron mass of the graft is never taken into account [4]. There are several studies in adults comparing equations to estimate GFR in renal transplant recipients. In general the equation proposed by the Modification of Diet in Renal Diseases Study, also known as MDRD formula has shown better predictive performance than the CockcroftGault or the Nankivell equations [57], but it underestimate the number of patients with declining function [8]. These formulas are not recommended for children.   Methods & Materials  Patients with renal transplant and stable renal function (defined as those in whom the serum creatinine levels had not changed by more than 0.2 mg per deciliter during the previous three months) were invited to participate in the study. Those patients with graft dysfunction or known recurrence of the original disease were excluded. The study was approved by the IRB, and parental and children informed consent/assent was obtained in all cases. The gold standard used for estimate GFR was ^{99m} TcDTPA, since it has a clearance similar to inulin [9]. All studies were performed according the Guidelines for Nuclear Medicine of the Health Laws in our country. Dynamic images were obtained in supine position after an intravenous administration of ^{99m} TcDTPA dose, using a gamma camera during 30 minutes and then processed to generate the timeactivity curves that allow to analyze the perfusion and filtration. A Gamma camera Siemens MultiSPECTII was used. Renograms were obtained and processed with a commercial nuclear medicine software (Siemens ICON). Serum and urine creatinine were measured by the kinetic Jaffe method using the Synchron CX3 automated analyzer.
24h Creatinine clearance (Ccr) was obtained with the formula:
Ccr = Ucr V/Scr ,
where: Ucr : urine creatinine concentration, V: flow rate of urine in millilitres per minute and Scr : serum creatinine. The creatinine clearance is normalized to body surface area x 1.73 m².
GFR was also estimated (eGFR) by the following equations:
1) Schwartz formula:
eGFR (ml/min/1.73 m²)= kL/ Scr (mg/dL), where:
L : length in centimeters, Scr : serum creatinine, k : constant value of 0.55 for children and adolescent girls, and 0.7 for adolescent boys [1012].
2) CounahanBarrat Formula [13]:
eGFR = 0.43 x Height (cm)/ Scr (mg/dL)
3) Morris equation [14]:
eGFR= 40 x height (cm) / Scr (μmol/L)
4) Leger formula [15]:
eGFR= 56.7 x body weight (kg) + 0.142 x Height (cm)2/ Scr (μmol/L)
The GFR value was corrected to body surface area x 1.73 m²
Statistical Analysis
We used the GraphPad Prism version 5.0 for Mac (GraphPad software, San Diego, CA) for data analysis. The GFR obtained by Ccr and the four equations were compared to ^{99m}TcDTPA by paired t test and Pearson correlation. Agreement between methods was tested by the Bland and Altman plot method, the difference between the two measurements is plotted as a function of the average of the two measurements of each sample. It also provides the standard deviation of the differences between the two assay methods to calculate the limits of agreement, computed as the mean bias 1.96 times its standard deviation. The predictive performance of the different methods to assess GFR were analyzed by measuring bias and precision according to the formula suggested by Sheiner and Beal [16].
Bias (Mean prediction Error) as: MPE = Σ(PREDOBS)/n where PRED is the predicted value, OBS is the GFR observed by ^{99m}TcDTPA, and n is the number of observations. An unbiased prediction results in a MPE value of zero. A negative MPE value indicates underestimation by the predictor and a positive MPE indicates overestimation. The precision of the predictions was estimated as Mean squared error (MSE) as: MSE= Σ(PREDOBS)2/n. The ideal prediction yields a MSE of zero, were PRED=OBS. A high MSE value indicates poor predictive performance.   Results  A total of 42 patients were included in the study. Demographic data is depicted in Table 1. The values of GFR (mL/min/1.73m^{2}BSA) by ^{99m}TcDTPA and the eGFR by 24hCcr as well as the values for the four equations in each patient are shown in Table 2. The mean GFR obtained with ^{99m} TcDTPA was 66.9 mL/min ± 15, whereas by Ccr was 68.7 ± 24.4, by Schwartz formula 81.8 ± 15.9, by the CounahanBarrat equation was 57.6 ± 13.6, by Morris equation 60.7 ± 14.5 and by Leger equation 73.2 ±18.3. GFR by all methods was normally distributed by D'Agostino & Pearson omnibus normality test. There was a significant correlation between GFR by ^{99m}TcDTPA vs. Ccr (Pearson r = 0.316, p= 0.040), ^{99m}TcDTPA vs. Schwartz (Pearson r = 0.500, p=0.0007), ^{99m}TcDTPA vs. CounahanBarrat formula ( r= 0.492, p=0.0009), ^{99m}TcDTPA vs. Morris (r= 0.492, p=0.0009) and ^{99m}TcDTPA vs. Leger (r= 0.494, p= 0.0009) (Figure 1). Nevertheless, when comparing by paired t test no difference was found between ^{99m}TcDTPA vs. Ccr (p=0.64) but there was statistically significant difference between ^{99m}TcDTPA and all the equations (Table 2).
Table 1. Patient's characteristics
Characteristic 
Value 
Age (years mean + SD) 
14.7 + 2.5 
Gender
Male
Female

24 18 
Time postrenal transplant
(months + SD)

26.17 + 6 
Graft source
Deceased donor
Living donor

27
15

Immunosuppression regime
PDN/MMF/Tac PDN/AZA/Tac
PDN/MMF/CsA PDN/SRL/CsA PDN/MMF PDN/Tac PDN/CsA

28 (66.7%) 2 (4.8%) 1 (2.3%) 6 (14.2%) 2 (4.8%) 2 (4.8 %) 1 (2.3 %)

Acute rejection history n (%)

25 (59%) 
Serum creatinine (mg/dL mean + SD) 
1.18 + 0.3 
PDN: prednisone, MMF: Mycophenolate mofetil, CsA: Cyclosporin, Tac: Tacrolimus, SRL: Sirolimus
Table 2. GFR in 42 renal transplant children
Patient 
^{99}TcDTPA 
Ccr 
Schwartz 
CounahanBarrat 
Morris 
Leger 
1. 
53.1 
59.68 
93.30 
57.00 
60.61 
81.32 
2. 
50.6 
31.11 
59.60 
37.00 
38.76 
44.77 
3. 
75.2 
78.22 
83.46 
51.00 
54.20 
66.25 
4. 
88.7 
60.82 
78.00 
48.00 
50.65 
58.53 
5. 
74.7 
82 
91.90 
72.00 
76.01 
87.57 
6. 
71.75 
40.9 
92.90 
57.00 
60.33 
70.67 
7. 
75.7 
93.58 
99.10 
61.00 
64.39 
76.60 
8. 
54.2 
64.34 
50.00 
31.00 
32.47 
37.49 
9. 
56.7 
101.90 
100.30 
78.00 
82.95 
96.98 
10. 
75 
56.75 
65.50 
51.00 
54.20 
62.74 
11. 
63.6 
57.52 
81.50 
50.00 
52.92 
64.35 
12. 
73.8 
52.48 
103.00 
80.00 
85.06 
103.80 
13. 
75.5 
121 
103.90 
63.00 
66.94 
77.60 
14. 
85 
104.9 
89.20 
55.00 
57.95 
68.79 
15. 
50.3 
48.81 
67.94 
42.00 
44.12 
52.23 
16. 
58.9 
64.78 
81.50 
50.00 
52.92 
60.80 
17. 
59.4 
49.2 
57.30 
45.00 
47.08 
54.94 
18. 
37.5 
34.34 
55.41 
34.00 
35.98 
42.40 
19. 
50.6 
26.94 
72.40 
45.00 
47.06 
56.31 
20. 
58.1 
87.37 
88.00 
69.00 
72.73 
90.93 
21. 
64 
46.49 
78.00 
48.00 
47.88 
55.29 
22. 
47.6 
74.02 
73.00 
45.00 
47.40 
58.99 
23. 
63 
105.18 
101.00 
79.00 
83.52 
98.69 
24. 
88.7 
87.92 
99.27 
61.00 
64.46 
74.85 
25. 
46.6 
66.18 
67.37 
53.00 
55.68 
66.34 
26. 
78.3 
56.06 
78.31 
61.00 
64.77 
79.49 
27. 
56.7 
71.21 
98.46 
60.00 
63.94 
75.43 
28. 
78.4 
80.82 
88.70 
55.00 
57.64 
71.98 
29. 
101.2 
41.76 
114.70 
90.00 
94.81 
111.89 
30. 
78.4 
92.86 
104.50 
82.00 
86.36 
101.95 
31. 
76.6 
43.25 
66.00 
51.00 
54.20 
60.46 
32. 
37.9 
24.41 
63.80 
50.00 
52.60 
62.94 
33. 
66.2 
50.50 
68.29 
53.00 
56.44 
68.71 
34. 
69.4 
77.81 
71.75 
56.00 
59.30 
69.26 
35. 
78 
65.57 
68.50 
54.00 
56.64 
66.06 
36. 
67.1 
71.98 
108.80 
85.00 
89.94 
111.00 
37. 
62.2 
116.84 
76.00 
62.00 
63.13 
81.51 
38. 
74.2 
60.63 
87.30 
68.00 
72.16 
92.97 
39. 
102 
92.28 
88.00 
69.00 
72.73 
99.83 
40. 
76.2 
89.12 
72.80 
57.00 
60.23 
86.42 
41. 
55.3 
54.90 
74.00 
58.00 
60.98 
71.80 
42. 
56.8 
99.64 
72.80 
45.00 
47.27 
55.71 
Mean 
66.98 
68.7 
81.8 
57.6 
60.7 
73.2 
SD 
15 
24.4 
15.9 
13.6 
14.5 
18.3 
Paired t test p value 

0.64 
<0.0001 
0.0001 
0.009 
0.021 
The Bland Altman plots for the estimated GFR and the obtained by ^{99m}TcDTPA are shown in Figure 2. Ccr showed the best and Schwartz the poorest accuracy, respectively. The predictive performance was evaluated by Sheiner and Beal method, calculating bias as mean prediction error (MPE) and precision (Mean Squared Error, MSE), depicted in Table 3, Ccr had low bias but is unprecise (MPE 1.7, MSE 132) and Schwartz formula the more biased and unprecise (MPE 14.8, MSE 9223), from the equations, the Morris formula was the more unbiased and precise (MPE , MSE 1656).
Figure 1. Correlation Between ^{99m}TC DTPA and A) Ccr, B) Schwartz Formula, C) Counahan Barrat, D) Morris and E) Leger equations.
Figure 2.BlandAltman plots for the GFR estimated by Ccr and four equations (n=42). A) Ccr, B) Schwartz, C) Morris and D) Leger equation. The center line corresponds to mean level of bias, the dotted lines depict the upper and lower 95% limits of agreement.
Table 3. Predictive performance of Ccr and four equations to estimate GFR in 42 renal transplant children

Ccr 
Schwartz 
CounahanBarrat 
Morris 
Leger 
MPE 
1.73 
14.81 
9.40 
6.28 
6.27 
MSE 
132 
9223 
3727 
1656 
2000 
MPE: Mean prediction error, MSE: Mean squared error.
  Discussion  The equations used to estimate GFR in pediatric patients have an imprecision as high as ± 3040%, higher than the ± 20% reported in adults [17]. There are few studies evaluating this issue in the context of renal transplant. In adults, the MDRD formula is preferred than the CockcroftGault equation to be used in transplant recipients, nevertheless it is not precise [4, 18] and it can underestimate the patients with declining graft function [8]. The Schwartz formula is the most extensively used to estimate GFR in large pediatric studies such as NAPRTCS and has proven to be a valid outcome variable to measure posttransplant graft function [19, 20]. Zapitelli and coworkers evaluated the Schwartz and the Leger equations to predict iothalamate GFR in 195 children, they found that Schwartz was most unbiased and sensitive [21]. However, in the subset of transplant recipients the Leger had a better performance; the authors attribute this findings to the fact that 42% of the Leger study patients were kidney recipients, they also propose as Van Rossum et al [22] to use locally derived values of k.
More often than not, children receive adultsized kidneys, Naesens et al [23] demonstrated that when those large kidneys are transplanted into the smallest recipients show a high degree of interstitial fibrosis, tubular atrophy and tubular microcalcifications very early after transplantation. Beside, small children receiving an adult sized kidney have absolute GFR early after transplantation that is similar to the normal GFR that would be expected if they had 2 native kidneys. However, the absolute GFR would not increase as expected, with the growth of the child. This may be due to the nonimmunological injury caused by the renal ischemia occurring immediately after transplant, conditioned by the donorrecipient size mismatch. Therefore, the GFR estimated by any method, if normal, may not reflect the real fitness of the graft.
In our study Ccr showed the lowest bias and the higher precision to estimate GFR in renal transplant children. This method is inconvenient as it is cumbersome to patients and their families, and patients without bladder control will require bladder catheterization. Nevertheless, it was the only method found to be not significantly different when evaluated by paired t test, and certainly the cost is lower than the gammagram. Other studies have found a large margin of error in the Ccr since it overestimates the GFR in patients with reduced glomerular function [24], but this was not the case in our study, perhaps due to the fact that our patients were older than 9 years (mean 14.7 years) thus, meaning a better compliance in urine collection. The Morris equation had better predictive performance in renal transplant children than the Schwartz, CounahanBarrat and Leger equations.
It is necessary to perform further long term comparative studies to determine which one of this methods to estimate GFR is the best to identify children with declining function of the graft.
  Acknowledgement  This work was supported by the Fondo Sectorial de Investigacion en Salud 2004 C01193.
  Compliance with Ethical Standards   Funding None   Conflict of Interest None  
References:   Hogg RJ, Furth S, Lemley KV, Portman R, Schwartz GJ, Coresh J, et al. National Kidney Foundation's Kidney Disease Outcomes Quality Initiative clinical practice guidelines for chronic kidney disease in children and adolescents: evaluation, classification, and stratification. Pediatrics 2003; 111: 14161421. [CrossRef]
 Filler G, Lepage N. Should the Schwartz formula for estimation of GFR be replaced by cystatin C formula? Pediatr Nephrol 2003; 18: 981985. [CrossRef]
 Seikaly MG, Browne R, Bajaj G, Arant BS, Jr. Limitations to body length/serum creatinine ratio as an estimate of glomerular filtration in children. Pediatr Nephrol 1996; 10: 709711. [CrossRef]
 Mariat C, Alamartine E, Afiani A, Thibaudin L, Laurent B, Berthoux P, et al. Predicting glomerular filtration rate in kidney transplantation: are the K/DOQI guidelines applicable? Am J Transplant 2005; 5: 26982703. [CrossRef]
 Poge U, Gerhardt T, Palmedo H, Klehr HU, Sauerbruch T, Woitas RP. MDRD equations for estimation of GFR in renal transplant recipients. Am J Transplant 2005; 5: 13061311. [CrossRef]
 Poge U, Gerhardt T, Woitas RP. Calculation of glomerular filtration rate using serum cystatin C in kidney transplant recipients. Kidney Int 2006; 70: 1878; author reply 18781879.
 Perico N, Gaspari F, Remuzzi G. Assessing renal function by GFR prediction equations in kidney transplantation. Am J Transplant 2005; 5: 11751176. [CrossRef]
 Gera M, Slezak JM, Rule AD, Larson TS, Stegall MD, Cosio FG. Assessment of changes in kidney allograft function using creatininebased estimates of glomerular filtration rate. Am J Transplant 2007; 7: 880887. [CrossRef]
 Hellerstein S, Berenbom M, Erwin P, Wilson N, DiMaggio S. Timedurine collections for renal clearance studies. Pediatr Nephrol 2006; 21: 96101. [CrossRef]
 Schwartz GJ, Brion LP, Spitzer A. The use of plasma creatinine concentration for estimating glomerular filtration rate in infants, children, and adolescents. Pediatr Clin North Am 1987; 34: 571590. [CrossRef]
 Schwartz GJ, Furth SL. Glomerular filtration rate measurement and estimation in chronic kidney disease. Pediatr Nephrol 2007; 22: 18391848. [CrossRef]
 Schwartz GJ, Haycock GB, Spitzer A. Plasma creatinine and urea concentration in children: normal values for age and sex. J Pediatr 1976; 88: 828830. [CrossRef]
 Counahan R, Chantler C, Ghazali S, Kirkwood B, Rose F, Barratt TM. Estimation of glomerular filtration rate from plasma creatinine concentration in children. Arch Dis Child 1976; 51: 875878. [CrossRef]
 Morris MC, Allanby CW, Toseland P, Haycock GB, Chantler C. Evaluation of a height/plasma creatinine formula in the measurement of glomerular filtration rate. Arch Dis Child 1982; 57: 611615. [CrossRef]
 Leger F, Bouissou F, Coulais Y, Tafani M, Chatelut E. Estimation of glomerular filtration rate in children. Pediatr Nephrol 2002; 17: 903907. [CrossRef]
 Sheiner LB, Beal SL. Some suggestions for measuring predictive performance. J Pharmacokinet Biopharm 1981; 9: 503512. [CrossRef]
 Gretz N, Schock D, Sadick M, Pill J. Bias and precision of estimated glomerular filtration rate in children. Pediatr Nephrol 2007; 22: 167169. [CrossRef]
 Gaspari F, Ferrari S, Stucchi N, Centemeri E, Carrara F, Pellegrino M, et al. Performance of different prediction equations for estimating renal function in kidney transplantation. Am J Transplant 2004; 4: 18261835. [CrossRef]
 Sorof JM, Goldstein SL, Brewer ED, Steiger HM, Portman RJ. Serial estimation of glomerular filtration rate in children after renal transplant. Pediatr Nephrol 1999; 13: 737741. [CrossRef]
 Ellis EN, Martz K, Talley L, Ilyas M, Pennington KL, Blaszak RT. Factors related to longterm renal transplant function in children. Pediatr Nephrol 2008;23 : 11491155. [CrossRef]
 Zappitelli M, Joseph L, Gupta IR, Bell L, Paradis G. Validation of child serum creatininebased prediction equations for glomerular filtration rate. Pediatr Nephrol 2007; 22: 272281. [CrossRef]
 Van Rossum LK, Mathot RA, Cransberg K, Zietse R, Vulto AG. Estimation of the glomerular filtration rate in children: which algorithm should be used? Pediatr Nephrol 2005; 20: 17691775. [CrossRef]
 Naesens M, Kambham N, Concepcion W, Salvatierra O, Jr., Sarwal M. The evolution of nonimmune histological injury and its clinical relevance in adultsized kidney grafts in pediatric recipients. Am J Transplant 2007; 7: 25042514. [CrossRef]

Cite this article as:  Ulibarri B, MartinezDuncker C, Camarillo C, Guadarrama E O, Roca P G, Arizpe R M, Medeiros M. Comparison of 99m Tc DTPA vs Five Methods to Estimate Glomerular Filtration Rate in Children with Renal Transplantation. Pediatr Oncall J. 2009;6: 6770. 
